Add files using upload-large-folder tool

00071c19a357d74c148b616e4f6f54f029c63bf7dc9c70b0f9dba908075f6ed8/preprint/preprint.md

@@ -0,0 +1,473 @@
+Switch-like Gene Expression Modulates Disease Susceptibility
+
+Omer Gokcumen
+gokcumen@gmail.com
+
+Article
+
+Keywords:
+
+Posted Date: September 13th, 2024
+
+DOI: https://doi.org/10.21203/rs.3.rs-4974188/v1
+
+License: This work is licensed under a Creative Commons Attribution 4.0 International License.
+Read Full License
+
+Additional Declarations: There is NO Competing Interest.
+
+Version of Record: A version of this preprint was published at Nature Communications on June 18th, 2025. See the published version at https://doi.org/10.1038/s41467-025-60513-x.
+Switch-like Gene Expression Modulates Disease Susceptibility
+
+Authors: Alber Aqil1,†, Yanyan Li2,†, Zhiliang Wang2, Saiful Islam3, Madison Russell2, Theodora Kunovac Kallak4, Marie Saitou5, Omer Gokcumen1*, Naoki Masuda2,3*
+
+Affiliations:
+1. Department of Biological Sciences, State University of New York at Buffalo, Buffalo, NY, USA.
+2. Department of Mathematics, State University of New York at Buffalo, Buffalo, NY, USA.
+3. Institute for Artificial Intelligence and Data Science, State University of New York at Buffalo, Buffalo, NY, USA.
+4. Department of Women’s and Children’s Health, Uppsala University, Uppsala, Sweden.
+5. Faculty of Biosciences, Norwegian University of Life Sciences, Aas, Norway
+
+Correspondence:
+Omer Gokcumen, gokcumen@gmail.com
+Naoki Masuda, naokimas@gmail.com
+
+Abstract
+A fundamental challenge in biomedicine is understanding the mechanisms predisposing individuals to disease. While previous research has suggested that switch-like gene expression is crucial in driving biological variation and disease susceptibility, a systematic analysis across multiple tissues is still lacking. By analyzing transcriptomes from 943 individuals across 27 tissues, we identified 1,013 switch-like genes. We found that only 31 (3.1%) of these genes exhibit switch-like behavior across all tissues. These universally switch-like genes appear to be genetically driven, with large exonic genomic structural variants explaining five (~18%) of them. The remaining switch-like genes exhibit tissue-specific expression patterns. Notably, tissue-specific switch-like genes tend to be switched on or off in unison within individuals, likely under the influence of tissue-specific master regulators, including hormonal signals. Among our most significant findings, we identified hundreds of concordantly switched-off genes in the stomach and vagina that are linked to gastric cancer (41-fold, \( p < 10^{-4} \)) and vaginal atrophy (44-fold, \( p < 10^{-4} \)), respectively. Experimental analysis of vaginal tissues revealed that low systemic levels of estrogen lead to a significant reduction in both the epithelial thickness and the expression of the switch-like gene *ALOX12*. We propose a model wherein the switching off of driver genes in basal and parabasal epithelium suppresses cell proliferation therein, leading to epithelial thinning and, therefore, vaginal atrophy. Our findings underscore the significant biomedical implications of switch-like gene expression and lay the groundwork for potential diagnostic and therapeutic applications.
+Introduction
+The study of gene expression began in earnest with the characterization of lactose-metabolizing switch-like genes in E. coli 1. The presence of lactose triggered the production of enzymes needed to metabolize it, while these enzymes were absent when lactose was not present. These genes acted like switches, toggling between “on” and “off” states based on the presence or absence of lactose, respectively. In subsequent decades, the discovery of enhancer elements 2–4, epigenetic modifications 5–8, and transcription factor dynamics 9 revealed that gene expression in humans is more nuanced, resembling a dimmer more often than a simple on-and-off mechanism. Consequently, the study of switch-like genes in humans was largely relegated to the narrow realm of Mendelian diseases 10–12.
+
+The recent availability of population-level RNA-sequencing data from humans has made it possible to systematically identify switch-like versus dimmer-like genes. For dimmer-like genes in a given tissue, we expect expression levels across individuals to be continuously distributed with a single mode, i.e., a unimodal distribution. In contrast, expression levels of switch-like genes in a given tissue are expected to exhibit a bimodal distribution, with one mode representing the "off" state and the other representing the "on" state. As we will detail, bimodal expression across individuals is a characteristic of a gene in a specific tissue, referred to as a gene-tissue pair. We define a gene as switch-like if it exhibits bimodal expression in at least one tissue. Most of the recent studies on bimodal gene expression are related to cancer biology, associating on and off states to different disease phenotypes and their prognoses 13–15. These cancer studies have already produced promising results for personalized medicine 16. However, to our knowledge, the only study focusing on switch-like genes in non-cancerous tissues across individuals restricted their analysis to muscle tissue 17. As a result, the dynamics of switch-like expression across the multi-tissue landscape remain unknown. We hypothesize that switch-like expression is ubiquitous but often tissue-specific. We further hypothesize that these tissue-specific expression trends underlie common disease states. Therefore, the analysis of switch-like genes across tissues and individuals may provide a means for early diagnosis and prediction of human disease.
+
+Here, we systematically identified switch-like genes across individuals in 27 tissues. Our results explain the regulatory bases of switch-like expression in humans, highlighting genomic structural variation as a major factor underlying correlated switch-like expression in multiple tissues. Furthermore, we identified groups of switch-like genes in the stomach and vagina for which the “off” state predisposes individuals to gastric cancer and vaginal atrophy, respectively. Overall, these findings improve our understanding of the regulation of switch-like genes in humans. They also suggest promising future paths for preventative biomedical interventions.
+Figure 1. Methodological framework. A. List of 27 tissues used in this study. B. Distribution of 19,132 genes by the number of tissues in which they are highly expressed. C. Bimodal expression is a property of a gene-tissue pair. We tested 516,564 gene-tissue pairs (19,132 genes x 27 tissues) for bimodal expression across individuals. When a gene-tissue pair exhibits switch-like (bimodal) expression, the individuals divide into two subpopulations: one with the gene switched off, and the other with the gene switched on.
+RESULTS
+
+Tissue-specificity of bimodal expression
+The misregulation of highly expressed genes often has consequences for health and fitness. To systematically identify biomedically relevant switch-like genes in humans, we focused on 19,132 genes that are highly expressed (mean TPM > 10) in at least one of the 27 tissues represented in the GTEx database (Figure 1A; Figure 1B; Table S1). For each of the 516,564 gene-tissue pairs (19,132 genes x 27 tissues), we applied the dip test of unimodality \(^{18}\) to the expression level distribution across individuals (Figure 1C). Employing the Bejamini-Hochberg procedure for multiple hypotheses correction, we identified 1,013 switch-like genes (Figure 1C; Methods; Table S2). The expression of these genes is bimodally distributed in at least one tissue, such that it is switched “off” for one subset of individuals and switched “on” for the rest of the individuals.
+
+Expression of different switch-like genes may be bimodally distributed in different numbers of tissues. We contend that genes that are bimodally expressed across all tissues are likely so due to a germline genetic polymorphism driving switch-like expression across tissues. If this is the case, the expression of these genes would be highly correlated across pairs of tissues. Given this insight, discovering universally bimodal genes is more tractable using tissue-to-tissue co-expression of each gene. Therefore, for each gene, we calculated the pairwise correlation of expression levels across pairs of tissues (Methods; Table S3). To visualize tissue-to-tissue co-expression patterns of genes, we performed principal component analysis (PCA) on the tissue-to-tissue gene co-expression data (Table S4). We emphasize that we are referring to the co-expression of the same gene across pairs of tissues instead of the co-expression of pairs of genes in the same tissue. In the space spanned by the first two principal components (explaining 35.3% and 3.47% of the variance, respectively), switch-like genes form two major clusters (cluster 1 and cluster 2; Methods), dividing along PC1 (Figure 2A). Applying PCA exclusively to switch-like genes reveals the further division of cluster 2 into two distinct subclusters – cluster 2A and cluster 2B – in the space spanned by the first two principal components (explaining 58.1% and 4.25% of the variance, respectively) (Figure 2B; Table S5).
+
+Manual inspection reveals that cluster 1, which contains 954 genes, represents genes, such as *KRT17*, with bimodal expression in a small subset of tissues (Figure 2C). Cluster 2A consists of 23 genes, such as *GPX1P1*, with bimodal expression in all tissues (Figure 2D). Lastly, cluster 2B represents eight genes, such as *EIF1AY*, with bimodal expression in all non-sex-specific tissues but not in sex-specific tissues (Figure 2E). We will refer to genes in cluster 1 as “tissue-specific switch-like genes.” Although some of them are bimodally expressed in more than one tissue, these genes tend to exhibit high tissue specificity in their bimodal expression. Genes in cluster 2 will be referred to as “universally switch-like genes.”
+Figure 2. Categorization of switch-like genes. A. PCA analysis of tissue-pair correlations of gene expression. Each point represents a gene. When we perform PCA on the tissue-to-tissue co-expression vectors for 19,132 genes, the switch-like genes divide into two clusters. Cluster 1 primarily represents genes that are bimodally expressed in a tissue-specific manner, while cluster 2 represents genes that are bimodally expressed in at least all non-sex-specific tissues. B. Performing PCA on the co-expression vectors of only switch-like genes further divides cluster 2 into two subclusters: cluster 2A, which contains genes that are bimodally expressed across all 27 tissues, and cluster 2B, which contains genes that are bimodally expressed in all 22 tissues common to both sexes, but not in the five sex-specific tissues. C-E. Violin plots display the expression levels in all 27 tissues for representative genes from cluster 1, cluster 2A, and cluster 2B, respectively.
+
+Genetic variation underlies universally switch-like genes
+We found that 3.1% of all switch-like genes (i.e., the proportion of switch-like genes that are in cluster 2) show clear bimodal expression, at least in all tissues common to both sexes. We contend that germline genetic variation across individuals likely underlies the universally switch-like gene expression, specifically due to four major types of genetic variants. Firstly, we expect genes on the Y chromosome to show bimodal expression in all tissues common to both sexes since these genes are present in males and absent in females (Figure 3A). Consistent with this reasoning, seven out of the eight genes in cluster 2B lie within the male-specific region of the Y-chromosome \(^{19}\); the remaining
+gene in cluster 2B is *XIST*, showing female-specific expression. Secondly, a homozygous gene deletion would result in the gene being switched off (**Figure 3B**). We found five such genes in cluster 2A for which genomic structural variants likely underlie the observed universally switch-like expression; four genes are affected by gene deletions, and the remaining one by an insertion into the gene. Thirdly, the homozygous deletion of a regulatory element can also switch off a gene (**Figure 3C**). While we did not find any examples of this scenario, it remains a theoretical possibility. Lastly, a loss-of-function single nucleotide variant (SNV) or short indel, which disrupts gene function, can switch off the gene (**Figure 3D**). We identified five genes in cluster 2A where such SNVs cause universal bimodality.
+
+Remarkably, we could genetically explain the expression of 10 out of 23 (43%) cases in cluster 2A despite the small number of genes fitting our conservative definition for universally switch-like genes. SNVs underlie five of these cases (**Figure 3B**), while structural variants underlie the remaining five cases (**Figure 3D**). Thus, out of the 10 cases where we can explain the genetic underpinnings of switch-like expression, 50% involve genomic structural variation, highlighting the importance of this type of genetic variation. Although we could not identify the genetic variation underlying the bimodal expression of the remaining 13 genes in cluster 2A, their consistent and highly correlated switch-like expression across all tissues strongly suggests a genetic basis. We anticipate that better resolution assemblies and detailed regulatory sequence annotations will help identify the genetic variants responsible for the remaining universally switch-like genes.
+
+![Diagram showing genetic mechanisms for switch-like gene expression, including gene deletions, insertions, and structural variants in regulatory regions](page_186_1097_1207_496.png)
+Figure 3. Genetic bases of universally switch-like gene expression (cluster 2). A. Genes on the Y chromosome are expressed only in males, leading to bimodal expression in non-sex-specific tissues. B. Common structural variants, such as deletions or insertions, may lead to increased, decreased, or no expression in all tissues relative to individuals who carry the alternative allele. C. Common structural variants affecting a genomic region regulating a gene may lead to increased, decreased, or no expression in all tissues, relative to individuals who carry the alternative allele. D. Common single nucleotide variants or short indels affecting a gene or its regulatory region may lead to increased, decreased, or no expression in all tissues relative to individuals who carry the alternative allele.
+
+We highlight a clear example of a common structural variant leading to universally switch-like expression (Figure 3B). *USP32P2* and *FAM106A* – both universally switch-like genes – are bimodally expressed in all 27 tissues. Both genes show high levels of tissue-to-tissue co-expression. A common 46 kb deletion (esv3640153), with a global allele frequency of ~25%, completely deletes both genes (**Figure 4A-B**). We propose that this deletion accounts for the universal switch-like expression of both *USP32P2* and *FAM106A* in all tissues. For illustration, we show the expression level distributions of *USP32P2* and *FAM106A* in the cerebellum (**Figures 4C-D**). Indeed, the haplotype harboring this deletion is strongly associated with the downregulation of both genes in all 27 tissues (*p*<10^{-5} for every single gene-tissue pair, **Methods**). We note that the under-expression of *USP32P2* in sperm is associated with male infertility^{20}, and plausibly, homozygous males for the deletion may be prone to infertility. Additionally, *FAM106A* interacts with SARS-CoV-2 and is downregulated after infection, at least in lung-epithelial cells^{21-23}. Individuals with *FAM106A* already switched off may develop more severe COVID-19 symptoms upon infection, though further investigation is needed. The case of *FAM106A* and *USP32P2* exemplifies the link between disease and bimodal gene expression, a theme we will explore further in the remainder of this text.
+
+We caution that we base our results regarding bimodality on expression at the RNA level. The bimodal expression of genes across individuals at the RNA level may not necessarily lead to bimodal expression at the protein level. For example, the universally switch-like expression of *RPS26* at the RNA level can be explained by a single nucleotide variant (rs1131017) in the gene's 5'-untranslated region (UTR). In particular, *RPS26* has three transcription states based on the SNV genotypes. The ancestral homozygote C/C corresponds to a high transcription state, the heterozygote C/G to a medium state, and the derived homozygote G/G to a low state (See **Supplement** for a discussion on why an expression distribution driven by three genotypes at a polymorphic site might still appear bimodal). Remarkably, this pattern is reversed at the translation level^{24}: Messenger RNA carrying the derived G allele produces significantly more protein. This reversal may be due to a SNV in the 5'-UTR that can abolish a translation-initiation codon^{25}. This finding demonstrates how the same SNV can regulate a gene’s expression level in opposite directions during transcription and translation. This multi-level regulation in opposite directions likely serves to dampen protein expression variability. It has been shown previously that RNA variability is greater than protein variability in primates^{26,27}; the presence of dampening variants discussed here may be one reason behind these findings. Such compensatory mechanisms for gene expression remain fascinating areas for future research.
+Figure 4. An example of a polymorphic gene deletion resulting in universally switch-like gene expression. A. FAM106A and USP32P2 (not drawn to scale) are overlapping genes on chromosome 17. Two alternative haplotype classes exist for these genes: one in which both genes are completely deleted and the other without the deletion. B. Frequency distribution of the deletion across diverse populations. Each pie chart represents one of the 26 populations from the 1000 Genomes Project. Purple indicates the frequency of the deletion, while gray indicates the frequency of the alternative haplotype. C-D. Expression level distribution in the cerebellum (as an example) across individuals for FAM106A and USP32P2,
+respectively. The gene deletion presumably leads to the switched-off expression state in both genes.
+
+Tissue-specific switch-like genes have a shared regulatory framework
+Tissue-specific expression patterns are crucial for tissue function. Thus, we now turn our attention to tissue-specific switch-like genes. We found that the stomach, vagina, breast, and colon show a higher number of tissue-specific switch-like genes compared to other tissues (Figure 5A), after controlling for confounding factors (Methods; Supplement; Table S6). Furthermore, within these tissues, the expression of switch-like genes is not independent; instead, they exhibit high pairwise co-expression between genes (Figure 5B-C; Table S7). Hence, tissue-specific switch-like genes tend to be either all switched off or switched on within an individual. This result suggests a shared regulatory mechanism for the expression of these genes in each tissue. Given that hormonal regulation plays a substantial role in shaping tissue-specific expression patterns \(^{28,29}\), we hypothesize that hormones may regulate genes that are bimodally expressed in specific tissues (cluster 1; Figure 2B).
+
+Sexual differences in hormonal activity are well documented \(^{30,31}\). To explore this further, we investigated whether hormone-mediated sex-biased expression underlies the co-expression of tissue-specific switch-like genes within tissues. Under this scenario, a gene would be largely switched on in one sex and off in the other in a given tissue. Among tissue-specific switch-like genes, we identified 186 gene-tissue pairs with sex-biased bimodal expression (Figure 6A; Table S8). These instances are biologically relevant; for example, we found switch-like immunoglobulin genes with female-biased expression in the thyroid, heart, tibial nerve, and subcutaneous adipose tissue. This observation may relate to previous findings \(^{32,33}\) of higher antibody responses to diverse antigens in females than in males.
+
+More dramatically, we found that 162 out of 164 tissue-specific switch-like genes (cluster 1) in the breast tissue are female-biased, explaining their correlated expression levels (Figure 6A). However, the sex-based disparity in the on-versus-off states of these genes is not absolute, but rather a statistical tendency. In other words, the gene is not switched off in all males and switched on in all females. Instead, the proportion of individuals with the gene switched on significantly differs between sexes. Notably, multiple sex-biased switch-like genes—including *SPINT1* and *SPINT2*\(^{34}\), multiple keratin genes \(^{35}\), and the oxytocin receptor gene \(^{36,37}\) (*OXTR*; Figure 6B)—in the breast tissue are differentially expressed in breast cancers relative to matched non-cancerous tissues. Future investigations could reveal whether the toggling of these genetic switches affects breast cancer risk in females. We caution that sex-biased switch-like expression in the breast may result from differences in cell-type abundance between females and males. Nevertheless, the differential expression of some genes between sexes might developmentally drive such differences in cell-type abundance. In summary, our results indicate that sex is a major contributor to bimodal gene expression, with breast tissue standing out as particularly sex-biased in this context.
+
+We note that the intra-tissue co-expression of tissue-specific switch-like genes in the
+stomach and colon cannot be explained by sex. By biological definition, the variation in vaginal expression levels in our sample is not sex-biased. Thus, the intra-tissue co-expression of tissue-specific switch-like genes in the stomach, colon, and vagina may be explained by one of two reasons: 1) Most of the tissue-specific switch-like genes in each tissue are directly regulated by the same hormone in that tissue, or 2) Most of the tissue-specific switch-like genes in each tissue are regulated by the same transcription factor which is, in turn, under regulation by a hormone or other cellular environmental factors. In the case of hormonally controlled gene expression, genes are likely switched off when the systemic hormone levels drop below a certain threshold. We will discuss this idea further, specifically for the vagina, later in the text.
+
+![Bar chart showing number of genuine tissue-specific switch-like genes per tissue, and scatterplots illustrating pairwise correlation coefficients for selected gene pairs in stomach, vagina, breast, and colon](page_184_670_1207_563.png)
+
+Figure 5. Characterization of genuine tissue-specific switch-like genes (cluster 1). The results shown here exclude genes that showed switch-like expression due to confounding factors like ischemic time. **A.** Number of tissue-specific switch-like genes showing bimodal expression in each of the 27 tissues. The stomach, vagina, breast, and colon show disproportionately more tissue-specific switch-like genes than other tissues. **B.** An illustration of how Pearson’s correlation coefficients were calculated for each pair of bimodally expressed tissue-specific switch-like genes within the stomach, vagina, breast, and colon. We show the scatterplots for two arbitrarily chosen gene pairs for each of the four tissues. The axes in each dot plot represent the log( TPM + 1 ) for the labeled gene in the relevant tissue. Panel C was generated using the pairwise correlation coefficients thus obtained. **C.** Tissue-specific switch-like genes within the four tissues shown are highly co-expressed. Tissue-specific master regulators, such as endocrinological signals, likely drive their concordant on and off states.
+Figure 6. Sex-biased expression of tissue-specific switch-like genes (cluster 1). A. Number of tissue-specific switch-like genes that show female- and male-biased expression. Only those tissues are shown that have at least one tissue-specific switch-like gene showing sex bias. The number in the central grid next to each tissue image represents the number of genuine tissue-specific switch-like genes in that tissue. In orange, the numbers to the left of the central grid indicate the count of female-biased genes in each of the 10 tissues shown. In blue, the numbers to the right of the grid indicate the count of male-biased genes. B. Violin plots showing the expression level distribution in the breast for five female-biased tissue-specific switch-like genes discussed in the main text.
+
+Concordantly switched-off genes in the stomach may indicate a predisposition to gastric cancer
+Gene expression levels have been studied as a diagnostic marker for disease states\(^{38}\).
+Therefore, we asked whether tissue-specific switch-like genes co-expressed with each other across individuals are linked to human disease, with each of the two expression states corresponding to different risks. To address this question, we investigated whether the identified switch-like genes in a given tissue are overrepresented among genes implicated in diseases of the same tissue.
+
+We overlapped the switch-like genes in the stomach with a previously published list \(^{39}\) of differentially expressed genes in gastric carcinomas. We found that switch-like genes in the stomach are significantly enriched (41-fold enrichment, \(p<10^{-4}\)) among genes that are downregulated in gastric carcinomas. Specifically, nine switch-like genes are downregulated in gastric carcinomas (\(ATP4A\), \(ATP4B\), \(CHIA\), \(CXCL17\), \(FBP2\), \(KCNE2\), \(MUC6\), \(TMEM184A\), and \(PGA3\)). Additionally, these nine genes are concordantly expressed in 92.5% (332/359) of the stomach samples, being either all switched off or on in a given individual (**Methods**). Our data suggest that individuals with these nine genes switched off in the stomach may be susceptible to developing cancers. This preliminary observation provides exciting avenues to investigate both the cause of the concordant toggling of these genes and their potential role in cancer development.
+
+**Concordantly switched-off genes result in vaginal atrophy**
+We found that switch-like genes in the vagina are significantly overrepresented (44-fold enrichment; \(p<10^{-4}\); see methods) among genes linked to vaginal atrophy in postmenopausal women. Vaginal atrophy, affecting nearly half of postmenopausal women, is triggered by sustained low levels of systemic estrogen and is marked by increased microbial diversity, higher pH, and thinning of the epithelial layer in the vagina \(^{40,41}\). It is also known as atrophic vaginitis, vulvovaginal atrophy, estrogen-deficient vaginitis, urogenital atrophy, or genitourinary syndrome of menopause, depending on the specialty of the researchers. Symptoms experienced by women include dryness, soreness, burning, decreased arousal, pain during intercourse, and incontinence \(^{42}\). Our analysis of switch-like genes in the vagina provides new insights into the development of vaginal atrophy.
+
+Specifically, we overlapped a previously published list \(^{43}\) of genes that are transcriptionally downregulated in vaginal atrophy with our list of bimodally expressed genes in the vagina. We found that the genes *SPINK7*, *ALOX12*, *DSG1*, *KRTDAP*, *KRT1*, and *CRISP3* are both bimodally expressed in the vagina and transcriptionally downregulated (presumably switched off) in women with vaginal atrophy (**Figure 7A**). We refer to these genes as “atrophy-linked switch-like genes.” Indeed, these six genes are either all switched on, or all switched off concordantly in 84% (131/156) of the vaginal samples we studied. The pairwise concordance rates (percentage of individuals with both genes switched on or both genes switched off) for these genes are shown in **Figure 7B**. Among postmenopausal women with this concordant gene expression, 50% are in the “off” state – a fraction that closely matches the prevalence of vaginal atrophy in postmenopausal women \(^{40,44}\). Therefore, our data suggest that estrogen-dependent transcription underlies concordant expression of atrophy-linked switch-like genes, with
+the “off” state of these genes associated with vaginal atrophy.
+
+For background, the vaginal epithelial layers are differentiated from the inside out. The basal and parabasal layers of the epithelium consist of mitotic progenitor cells with differentiation potential, while the outermost layer comprises the most differentiated cells \(^{45,46}\). When basal and parabasal cells stop proliferating, the death of mature cells leads to a thin epithelium, and the symptoms of vaginal atrophy appear. Given this background, atrophy-linked switch-like genes may either be a cause or a consequence of vaginal atrophy. In particular, if an atrophy-linked switch-like gene encodes a protein necessary for the continued proliferation and differentiation of basal and parabasal cells, we call it a “driver” gene. In the absence of the driver gene’s protein, cell differentiation ceases, and the outer layer gradually disappears, resulting in vaginal atrophy (**Figure 8A**). On the other hand, if the product of an atrophy-linked switch-like gene is not required for basal and parabasal cell proliferation, we refer to it as a “passenger” gene, borrowing the terminology from cancer literature \(^{47}\). In healthy vaginas with a thick epithelium, there are more cells in which passenger genes would be expressed. By contrast, in atrophic vaginas, the epithelium thins, resulting in fewer cells where these genes can be expressed. This contrast would lead to the bimodal expression of passenger genes across vagina samples in whole-tissue RNA-sequencing datasets. We hypothesize that at least some of the atrophy-linked switch-like genes are driver genes.
+
+Two key findings allowed us to construct this hypothesis. Firstly, switch-like genes in the vagina show a 26-fold ontological enrichment for the establishment of the skin barrier (FDR=1.26 x10\(^{-6}\)) and a 25-fold enrichment for keratinocyte proliferation (FDR=1.75 x 10\(^{-4}\)), both related to epithelial thickness and differentiation. Notably, two atrophy-linked switch-like genes in the vagina that we identified, *KRTDAP* and *KRT1*, are crucial for the differentiation of epithelial cells in the vagina \(^{48,49}\). Protein stainings available through Human Protein Atlas \(^{50}\) show that all six atrophy-linked switch-like genes are expressed at the protein level, predominantly in the vaginal epithelium. Secondly, administering 17β-estradiol (a type of estrogen) to postmenopausal women with vaginal atrophy leads to the upregulation of the same six genes, causing symptoms to subside \(^{51}\). According to our hypothesis, administering estrogen activates the expression of the driver switch-like genes in the vagina, resuming the proliferation of basal and parabasal cells in the epithelium. This process leads to the reformation of a thick and healthy vaginal mucosa, thereby alleviating the symptoms of vaginal atrophy.
+
+Thus, it is essential to distinguish driver genes from passenger genes to understand the etiology of vaginal atrophy. However, we expect driver and passenger genes to show the same expression patterns in healthy versus atrophic vaginas using bulk RNA-sequencing data. In order to make this distinction, we need comparative expression data, specifically from the basal and parabasal epithelium from healthy versus atrophic vaginas. We expect driver genes to be differentially expressed in the basal and parabasal layers of the epithelium. By contrast, we expect passenger genes to show no differential expression in the basal and parabasal layers between healthy and atrophic
+vaginas.
+
+To look at the expression levels in the basal and parabasal layers of the epithelium, we arbitrarily chose *ALOX12* from the six atrophy-linked switch-like genes for immunohistochemical staining of its protein product in the vaginal mucosa (which includes the epithelium and the underlying connective tissue). We found that the ALOX12 protein is present in the epithelial cells, and its abundance directly correlates with epithelial thickness, as expected from our RNA-sequencing results. However, we found no significant difference in the staining of the ALOX12 protein in the basal or parabasal epithelial layers between healthy and atrophic samples (**Figure 8B**). This suggests that the gene is not differentially expressed in the basal or parabasal layers of the vaginal epithelium between healthy and atrophic vaginas. Therefore, *ALOX12* is a passenger gene for vaginal atrophy. Comparative immunohistochemical staining of the protein product of the other five atrophy-linked switch-like genes may identify the driver gene in the future. Indeed, the KRT1 protein is recognized as a marker of basal cell differentiation in mouse vaginas \(^{52}\), a finding that may also be true for humans. Overall, our results open up several new paths for potential pre-menopausal risk assessment and intervention frameworks targeting cell differentiation pathways in the clinical setting.
+Figure 7. Atrophy-linked switch-like genes tend to be either all switched off, or all switched on within individuals. A. The distribution of expression levels in the vagina of the six switch-like genes implicated in vaginal atrophy. The x-axes represent log(TPM +1) values for each gene in the vagina, and the y-axes represent the probability density. We obtained the probability densities using kernel density estimation. In each case, the global minimum (excluding endpoints) is considered the switching threshold. A gene is deemed “on” in an individual if the expression level is above this threshold; otherwise, the gene is deemed “off.” B. Pairwise concordance rates (percentage of individuals in which the two genes are either both switched on or both switched off).
+A
+
+High systemic levels
+Low systemic levels
+Estradiol
+High expression
+Low expression
+Driver gene is switched on in the basal and parabasal layer
+Basal and parabasal cell proliferation and differentiation
+Healthy vagina
+Driver gene is switched off in the basal and parabasal layer
+Basal and parabasal cells do not proliferate
+Vaginal atrophy
+
+B
+ALOX12 Protein Staining
+Healthy vagina
+Atrophic vagina
+ALOX12 is not differentially expressed in the basal and parabasal layer
+ALOX12 does not match the driver gene expression pattern shown in A
+ALOX12 is a passenger gene
+
+Figure 8. ALOX12 is a passenger gene. A. Model for the etiology of vaginal atrophy. High levels of
+estrogen keep the driver genes switched on in basal and parabasal epithelium, impelling basal and parabasal cells to proliferate and mature, resulting in healthy vaginal mucosa. Conversely, low levels of estrogen switch off the driver genes. The lack of basal and parabasal cell proliferation leads to a thin vaginal epithelium, resulting in vaginal atrophy. **B.** Representative immunohistochemical staining of Arachidonate 12-Lipoxygenase (ALOX12) in vaginal tissue. We show healthy vaginal tissue from a woman with higher systemic estrogen levels and a thicker vaginal epithelial layer, along with atrophic vaginal tissue from a woman with low systemic estrogen levels and a thinner vaginal epithelial layer. There is no difference in *ALOX12* expression in the basal or parabasal cells between healthy and atrophic epithelium, implicating it as a passenger gene. Images taken with Axio Observer Z1 (Carl Zeiss AG) with a 40X objective.
+
+**Discussion**
+In this study, we investigated factors underlying switch-like gene expression and its functional consequences. Our systematic analysis revealed 1,013 switch-like genes across 943 individuals. Some of these genes show bimodal expression across individuals in all tissues, suggesting a genetic basis for their universally switch-like behavior. We found several single nucleotide and structural variants to explain the switch-like expression of these genes. Most of the switch-like genes, however, exhibit tissue-specific bimodal expression. These genes tend to be concordantly switched on or off in individuals within the breast, colon, stomach, and vagina. This concordant tissue-specific switch-like expression in individuals is likely due to tissue-specific master regulators, such as endocrinological signals. For example, in the vagina, switch-like genes tend to get concordantly switched off in a given individual when systemic estrogen levels fall below a certain threshold. On the biomedical front, our work linked switch-like expression to the susceptibility to gastric cancer and vaginal atrophy. Furthermore, this study has paved two major paths forward toward early medical interventions, as discussed below.
+
+First, we emphasize that bimodal expression that is correlated across all tissues is driven by genetic polymorphisms. However, the genetic bases for 13/23 universally switch-like genes remain elusive. We propose that the underlying genetic bases for these universally switch-like genes are structural variants, which are not easily captured by short-read DNA sequencing. These structural variants may be discovered in the future as population-level long-read sequencing becomes more common. The first biomedical path forward is to use long-read DNA sequencing to pinpoint the genetic polymorphisms responsible for the bimodal expression of disease-related genes. Of particular interest are the genes *CYP4F24P* and *GPX1P1*, both long non-coding RNAs, which are implicated in nasopharyngeal cancer. The genetic basis for their bimodal expression remains unknown. *CYP4F24P* is significantly downregulated in nasopharyngeal cancer tissues \(^{53}\), while *GPX1P1* is significantly upregulated in nasopharyngeal carcinomas treated with the potential anticancer drug THZ1 \(^{54}\). Investigating whether individuals with naturally switched-off *GPX1P1* and *CYP4F24P* are at a higher risk of nasopharyngeal cancer will enable genotyping to identify individuals at elevated risk for nasopharyngeal cancer, facilitating early interventions and improving patient outcomes.
+Secondly, switch-like genes present a promising avenue for exploring gene-environment interactions, an area of growing interest. Recent studies indicate that environmental factors can significantly modulate genetic associations \(^{55,56}\). Polymorphisms that result in switch-like gene expression have already been linked to several diseases within specific environmental contexts \(^{57}\). For instance, the deletion of *GSTM1* has been associated with an increased risk of childhood asthma, but only in cases where the mother smoked during pregnancy \(^{58}\). Even more critically, switch-like genes potentially create unique cellular environments that could modulate the impact of genetic variations. We hypothesize that switch-like expression can produce diverse cellular environments, whether in a single gene (as in genetically determined cases) or in multiple genes (as in tissue-specific, hormonally regulated cases). These environments may, in turn, influence the effect of genetic variations and their associations with disease. Thus, much like current gene-environment association studies that control for factors such as birthplace, geography, and behaviors like smoking, it is conceivable that controlling for switch-like gene expression states could enhance the power of such studies. By cataloging these switch-like genes and developing a framework to classify them as “on” or “off” in various samples, our work lays the groundwork for more robust association studies in future research.
+
+In summary, our study has significant implications for understanding the fundamental biology of gene expression regulation and the biomedical impact of switch-like genes. Specifically, it contributes to the growing repertoire of methods for determining individual susceptibility to diseases, facilitating early therapeutic interventions. By providing a new approach to studying gene expression states, our study will enhance the predictive accuracy of disease susceptibility and improve patient outcomes.
+
+Acknowledgment
+O.G. and N.M. acknowledge support from the National Institute of General Medical Sciences (under grant no.1R01GM148973-01). N.M. also acknowledges support from the Japan Science and Technology Agency (JST) Moonshot R&D (under grant no.JPMJMS2021), the National Science Foundation (under grant no.2052720), and JSPS KAKENHI (under grant no.JP 24K14840). O.G. acknowledges support from the National Science Foundation (under grant nos.2049947 and 2123284). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
+
+METHODS
+
+Data
+The Genotype-Tissue Expression (GTEx) project is an ongoing effort to build a comprehensive public resource to study tissue-specific gene expression and regulation. The data we use are transcript per million (TPM) obtained from human samples across
+54 tissues and 56,200 genes (as of December 1st, 2023). We excluded laboratory-grown cell lines from our analysis. Since we need a reasonable number of individuals from each tissue, we excluded tissues with less than 50 individuals for our calculations. Of the remaining tissues, there were instances of multiple tissues from the same organ. In such cases, we randomly chose one tissue per organ. We thus focus our analysis on 27 tissues (Figure 1). Additionally, we retained only those genes for which the mean TPM across individuals was greater than 10 in at least one of the 27 focal tissues. This filter was applied because the analysis of lowly expressed genes may lead to false positive calls for bimodal expression and, as a result, to assign biological significance to cases where there is none. After these filtering steps, we are left with TPM data from 19,132 genes in each of the 27 tissues. We note that each tissue contains data from a different number of samples (individuals), totaling 943 across tissues. We will refer to this set of 19,132 genes as \( G \) in our equations and the rest of the methods.
+
+Dip test
+
+There are many tests of bimodality of gene expressions \( ^{16,59} \). We use a dip test described as follows. We denote by \( S_i \) the number of samples (individuals) available for tissue \( i \). We also denote by \( x_{g,i,s} \) the TPM value for gene \( g \) in tissue \( i \), for sample \( s \in \{1, ..., S_i\} \) and \( g \in G \). According to convention, we log-transform the TPM, specifically by \( \log(x_{g,i,s}+1) \)^{60} to suppress the effect of outliers; TPM is extremely large for some samples. Note that \( \log(x_{g,i,s} + 1) \) conveniently maps \( x_{g,i,s} = 0 \) to 0. For each pair of gene \( g \) and tissue \( i \), we carried out a dip test, which is a statistical test for multimodality of distributions, on the distribution of \( \log(x_{g,i,s} + 1) \) across the samples \( S_i \). We performed the dip test using the dip.test() function within the “dipTest” package in R, with the number of bootstrap samples equal to 5000. We applied the Benjamini-Hochberg procedure for multiple hypothesis correction to the results with a false discovery rate of 5%. Additionally, to reduce false positive calls of bimodal expression, we only retained results where the dip statistic \( D > \max[0.05, 0.05/\log(\bar{x}_{g,i})] \), where
+
+\[
+\bar{x}_{g,i} = \frac{1}{S_i} \sum_{s=1}^{S_i} x_{g,i,s}
+\]
+
+We obtained this threshold of 0.05 by visual inspection of \( \log(x_{g,i,s} + 1) \) distributions in the stomach and adipose subcutaneous tissues, starting with those with the highest values of \( D \). For statistically significant results, the distribution was almost always bimodal if \( D \) exceeded 0.05. The only exceptions were genes with low \( \bar{x}_{g,i} \). Thus, we penalized gene-tissue pairs with low \( \bar{x}_{g,i} \) across samples by requiring a higher \( D \) in order to classify them as bimodally distributed. Genes identified as bimodally distributed in at least one tissue are referred to as “switch-like” genes.
+
+Tissue-to-tissue co-expression of genes
+We sought to identify switch-like genes whose expression exhibits bimodal expression in all tissues. One seemingly straightforward approach is to count the number of tissues
+showing bimodal distribution of expression levels for each gene. However, even if a gene genuinely exhibits bimodal expression across all tissues, our methodology may fail to recognize it as such if the mean expression levels (\( \bar{x}_{g,i} \)) of the gene are low in some tissues. This is because our effect size threshold penalizes gene-tissue pairs with low \( \bar{x}_{g,i} \). Moreover, if gene expression follows a bimodal distribution across all tissues, then it does so likely due to a genetic polymorphism affecting expression. Thus, the expression of such genes would be highly correlated between pairs of tissues. Given this insight, discovering universally bimodal genes is more tractable using tissue-to-tissue co-expression of each gene.
+
+For each gene, we construct the co-expression matrix among pairs of tissues as follows. To calculate the co-expression between a pair of tissues, we need to use the samples whose TPM is measured for both tissues \( ^{61} \). In general, even if the number of samples is large for both of the two tissues, it does not imply that there are sufficiently many common samples. Therefore, using the sample information described in GTEX_Analysis_v8_Annotations_SampleAttributesDD.xlsx in the GTEx data portal, we counted the number of samples shared by each tissue pair and excluded the 41 tissue pairs that share less than 40 samples. For each of the remaining 27 x 26/2 - 41 = 310 tissue pairs, we denote by \( S_{i,j} \) the number of samples shared by the two tissues \( i \) and \( j \). We also denote by \( x_{g,i,s} \) and \( x_{g,j,s} \) the TPM value for gene \( g \) in tissues \( i \) and \( j \), respectively, for sample \( s \in \{1, 2, \ldots, S_{i,j}\} \). Then, we calculated the Pearson correlation coefficient between \( \log(x_{g,i,s} + 1) \) and \( \log(x_{g,j,s} + 1) \) across the \( S_{i,j} \) samples and used it as the strength of the co-expression of gene \( g \) between tissues \( i \) and \( j \). Specifically, we calculate
+
+\[
+r_g(i,j) = \frac{\sum_{s=1}^{S_{i,j}} [\log(x_{g,i,s} + 1) - m_{g,i}] [\log(x_{g,j,s} + 1) - m_{g,j}]}{\sqrt{\sum_{s=1}^{S_{i,j}} [\log(x_{g,i,s} + 1) - m_{g,i}]^2 \sum_{s=1}^{S_{i,j}} [\log(x_{g,j,s} + 1) - m_{g,j}]^2}}
+\]
+
+where
+
+\[
+m_{g,i} = \frac{1}{S_{i,j}} \sum_{s=1}^{S_{i,j}} \log (x_{g,i,s} + 1),
+\]
+
+and
+
+\[
+m_{g,j} = \frac{1}{S_{i,j}} \sum_{s=1}^{S_{i,j}} \log (x_{g,j,s} + 1).
+\]
+
+For each gene \( g \), we then vectorize the correlation matrix, \( (r_g(i,j)) \), into a 310-dimensional vector. If, for a given gene, \( g \), \( \log(x_{g,i,s} + 1) \) or \( \log(x_{g,j,s} + 1) \) were 0 across all \( S_{i,j} \) samples for any of the 310 tissue pairs, the gene was removed. In this process, 28 out of 1,013 switch-like genes were removed. Note that the correlation matrix is symmetric, so we only vectorize the upper diagonal part of the matrix. We denote the
+generated vector by \( \vec{v}_g \). Vector \( \vec{v}_g \) characterizes the gene. We ran a principal component analysis (PCA), using the prcomp() function in R, on vectors, \( \vec{v}_g \) for all genes for which we could calculate \( r_g(i,j) \) for all 310 tissue pairs. In parallel, we also ran PCA on only the set of vectors (genes) characterizing only the 985 (1013 - 28) switch-like genes.
+
+In the space spanned by the first two principal components, we calculated the pairwise distance between genes using the dist() function in R with method = "euclidean". We then performed hierarchical clustering using the hclust() function with method = "complete". Finally, we used the cuttree() function with k=2 and k=3 to obtain two and three clusters, respectively.
+
+Identifying the genetic basis of universal bimodality
+In order to identify the genetic basis of bimodality for switch-like genes in cluster 2A, we obtained the coordinates of the genes for both hg19 and hg38 using their Ensembl IDs as keys through Ensembl BioMart. We obtained coordinates of common structural variants using both the 1000 genomes project (hg19) \( ^{62} \) and the HGSV2 dataset (hg38) \( ^{63} \). We performed an overlap analysis using BedTools \( ^{64} \) to identify polymorphic deletions of or insertions into these genes. We thus obtained five universally bimodal genes being affected by structural variants. These were USP32P2, FAM106A, GSTM1, RP11-356C4.5, and CYP4F24P. Additionally, we obtained the GTEx dataset for the expression quantitative trait loci (eQTL). We identified genes in cluster 2A that had at least one eQTL, which was consistently associated with either increased or decreased expression of a given gene across all 27 tissues analyzed. We thus obtained five genes from cluster 2A whose expression was associated with a short variant across tissues. These were NPIPA5, RPS26, PSPHP1, PKD1P2, and PKD1P5.
+
+Controlling for confounders
+A bimodal distribution of expression levels of universally switch-like genes is unlikely to be driven by confounding factors such as ischemic time, and time spent by the tissue in chemical fixatives (PAXgene fixative). For example, the expression of genes on the male-specific region of chromosome Y is bimodally distributed across tissues regardless of confounding factors because females do not possess these genes. Similarly, regardless of confounding factors, USP32P2 is bimodally distributed due to a polymorphic gene deletion. However, tissue-specific switch-like genes are particularly prone to being affected by confounding variables. Specifically, we investigated whether the switch-like expression of genes can be explained by ischemic time and PAXgene fixative using the following approach.
+
+Ischemic time for a sample s in a given tissue i, denoted by \( k_{i,s} \), is a continuous variable representing the time interval between death and tissue stabilization. Time spent by a tissue i from a sample s in PAXgene fixative, denoted by \( f_{i,s} \), is also a continuous variable. For each gene-tissue pair (\( g, i \)), we calculated, across the \( S_i \) samples, the Pearson correlation between 1) \( \log(1 + x_{g,i,s}) \) and \( k_{i,s} \) and 2) \( \log(1 + x_{g,i,s}) \) and \( f_{i,s} \). For
+each tissue \( i \) and confounder \( c \), where \( c \in \{k_{i,s}, f_{i,s}\} \), we denote the correlation coefficient between \( \log(1 + x_{g,i,s}) \) and \( c \) as \( r_{g,i,c} \).
+
+We partition the set of switch-like genes into two subsets: cluster 1 and cluster 2 (the union of clusters 2A and 2B). We treat cluster-2 genes as internal controls since their correlated bimodal expression across tissues is robust to the presence of confounding factors. Thus, we eliminated a cluster-1 gene \( g1 \) if, for any confounder \( c \), \( (r_{g1,i,c})^2 > \left( \max_{g2 \in \text{cluster 2}} r_{g2,i,c} \right)^2 \).
+
+**Gene-to-gene co-expression within tissues**
+We performed gene-to-gene co-expression analysis within the stomach, breast, vagina, and colon tissues. In a given tissue \( i \), we denote the set of genuine cluster-1 genes (excluding genes affected by confounding variables) by \( C_i \). Then, for \( i \in \{\text{stomach, breast, vagina, colon}\} \), we calculated the Pearson correlation, across the \( S_i \) samples, between \( \log(x_{g,i,s} + 1) \) and \( \log(x_{h,i,s} + 1) \) for every \( g, h \in C_i \) where \( g \neq h \).
+
+**Quantifying sex bias in cluster-1 gene expression**
+For every gene-tissue pair (\( g, i \)), where \( g \) is a switch-like gene, and \( i \) is a tissue common to both sexes, we tested the hypothesis that the distribution of \( \log(x_{g,i,s} + 1) \) across male samples differed from that across female samples using the Wilcoxon rank-sum test. We applied the Benjamini-Hochberg procedure of multiple hypotheses correction with FDR = 5%. We quantified the effect size of the sex bias using Cohen’s \( d \). Statistically significant results were considered to represent true sex bias only if \( |d| > 0.2 \) [65].
+
+**Enrichment of switch-like genes among disease-linked genes**
+We performed enrichment analysis for switch-like genes in the stomach and vagina that are downregulated in gastric cancer and vaginal atrophy, respectively. We denote the set of genes downregulated in disease \( y \) as \( Z_y \), where \( y \in \{\text{gastric cancer, vaginal atrophy}\} \). We calculated the fold enrichment of genuine cluster-1 genes in the stomach among genes downregulated in gastric cancer by:
+
+\[
+\frac{|C_{\text{stomach}} \cap Z_{\text{gastric cancer}}|}{|G \cap Z_{\text{gastric cancer}}|} / \frac{|C_{\text{stomach}}|}{|G|}
+\]
+
+We calculated the fold enrichment of genuine cluster-1 genes in the vagina among genes downregulated in vaginal atrophy by:
+
+\[
+\frac{|C_{\text{vagina}} \cap Z_{\text{vaginal atrophy}}|}{|G \cap Z_{\text{vaginal atrophy}}|} / \frac{|C_{\text{vagina}}|}{|G|}
+\]
+To calculate the \( p \)-values associated with these enrichments, we obtained 10,000 uniformly random samples (with replacement) of size \( |C_i| \) from \( G \). The \( p \)-value for the enrichment of switch-like genes in tissue \( i \) among genes linked to disease \( y \) is then given by the fraction of random samples among the 10,000 samples for which \( |q_j \cap Z_y| > |C_i \cap Z_y| \). Here, \( q_j \) is the set of genes in random sample \( j \) where \( j \in \{1, ..., 10000\} \).
+
+Discretizing expression levels
+We performed kernel density estimation using the density() function in R on the distributions of 1) \( \log(x_{g,\text{stomach},s} + 1) \) across the \( S_{\text{stomach}} \) samples for \( g \in C_{\text{stomach}} \cap Z_{\text{gastric cancer}} \); and 2) \( \log(x_{g,\text{vagina},s} + 1) \) across the \( S_{\text{vagina}} \) samples for \( g \in C_{\text{vagina}} \cap Z_{\text{vaginal atrophy}} \).
+
+We used the minimum of the estimated density as the switching threshold; if an individual had an expression level above the threshold in a given tissue, the gene was considered “on” in the individual in that tissue. The gene was considered “off” otherwise. We then calculate the concordance of expression among genes in any arbitrary set of switch-like genes \( G^A \) in a given tissue \( i \) as follows:
+
+\[
+\frac{1}{S_i} \sum_{s=1}^{S_i} \left[ \prod_{g \in G^A} \mathbf{1}_{(g \text{ is "on" in sample } s \text{ in tissue } i)} + \prod_{g \in G^A} \mathbf{1}_{(g \text{ is "off" in sample } s \text{ in tissue } i)} \right],
+\]
+
+where \( \mathbf{1}_{(\cdot)} \) is the indicator function.
+
+Gene ontology enrichment of tissue-specific switch-like genes in the vagina
+We performed Gene Ontology (GO) enrichment analysis for genes in \( C_{\text{vagina}} \) using the online database available at https://geneontology.org/.\(^{66}\)
+
+Immunohistochemistry
+Vaginal biopsies were taken by use of punch biopsies from postmenopausal women, fixed and stained as previously described by use of ALOX12 (HPA010691 polyclonal antirabbit, Sigma-Aldrich) \(^{67,68}\).
+Supplementary Information
+
+Principal component analysis on tissue-to-tissue co-expression vectors
+We applied a principal component analysis to the 19,132 vectors of tissue-to-tissue co-expression, one vector for each gene. We find that PC1 (Figure 2A), explaining 35.3% of the variation, is nearly perfectly correlated with mean tissue-to-tissue co-expression across tissue-tissue pairs (\( r^2 = 0.998, p\text{-value} < 2.2 \times 10^{-16} \); Figure S1). This result indicates that the 35.3% of the variation in the tissue-to-tissue co-expression of genes is primarily explained by the mean tissue-to-tissue co-expression of genes.
+
+![Scatter plot showing the correlation between PC1 and mean tissue-to-tissue co-expression, with \( r^2 = 0.998 \) and \( p < 2.2 \times 10^{-16} \)](page_340_670_670_420.png)
+
+Figure S1. The mean tissue-to-tissue co-expression of genes shows a near-perfect correlation with PC1.
+
+Universally switch-like genes and their biomedical implications
+In the main text, we discussed the *USP32P2* and *FAM106A*. Here, we discuss some other interesting examples of universally switch-like genes. The violin plots for the expression level distributions for all cluster-2A and cluster-2B switch-like genes not shown in the main text are present in Figure S2 and Figure S3, respectively.
+
+Firstly, a common ~20kb whole-gene deletion (esv3587154) of the *GSTM1* gene \(^{69,70}\) is associated with bladder cancer in humans \(^{71}\). *GSTM1* is bimodally expressed across individuals in all tissues (Figure S2D) that we analyzed, as well as across multiple tumor types \(^{15}\), with different expression peaks corresponding to differential prognoses among patients. These findings suggest a compelling hypothesis: the common deletion of *GSTM1*, maintained either by drift or balancing selection \(^{72}\), has no significant effect on the health of non-cancerous individuals; however, it could have significant implications for prognosis once certain types of tumors develop. Therefore, screening
+patients with certain tumor types for the *GSTM1* deletion could significantly advance our ability to predict the course of tumor progression in an individualized manner.
+
+Secondly, genes that are bimodally expressed across multiple tissues raise an evolutionary paradox. Typically, genes with a wide expression breadth (i.e., expression across a large number of tissues) affect fitness and are thus constrained at both the sequence and expression level \(^{26,73-75}\). However, universally switch-like genes, despite having a high expression breadth, are not conserved at the expression level. This could imply different health consequences for individuals with off versus on state of the genes. For example, the universally switch-like gene RP4-765C7.2 (ENSG00000213058; **Figure S2K**) is upregulated in the peripheral blood mononuclear cells of patients with ankylosing spondylitis \(^{76}\), eutopic endometrium in endometriosis patients \(^{77}\), and peripheral blood mononuclear cells of multiple sclerosis patients \(^{78}\). Conversely, it is downregulated in the peripheral blood mononuclear cells of Sjögren's syndrome patients \(^{79}\). These results suggest that this gene being switched on versus off may predispose individuals to certain diseases while protecting them against others. This balance between susceptibility and protection could explain why both high-expression and low-expression states are maintained in the population at comparable frequencies.
+
+Thirdly, the bimodality of *NPIPA5* (**Figure S2G**), too, can be explained by a single eQTL. The T allele of the SNV rs3198697 is associated with *NPIPA5* being switched on across tissues, while the C allele is associated with the gene being switched off. *NPIPA5* has been reported as one of the top differentially expressed genes among patients with multiple sclerosis in both blood and brain \(^{80}\). Moreover, this study \(^{80}\) showed that this gene is co-expressed in blood and brain. Here, we have shown that this gene is switch-like and that the co-expression of *NPIPA5* is not restricted to blood and brain but extends to all pairs of tissues.
+
+Lastly, a single eQTL can explain the bimodality of a member of the PKD1 gene family in cluster 2A, *PKD1P5* (**Figure S2I**). For *PKD1P5*, the C allele of the SNV rs201525245 is associated with the gene being switched on, while the G allele is associated with the gene being switched off.
+Figure S2. Violin plots for expression level distributions of switch-like genes in cluster 2A.
+Figure S3. Violin plots for expression level distributions of switch-like genes in cluster 2A.
+
+Conceptual issues regarding bimodal expression distributions driven by genetic polymorphisms
+In the main text, we claimed that genetic polymorphisms drive the bimodal expression of universally switch-like genes in cluster 2A. For a polymorphism with two alleles (A and a), there are three possible genotypes (\( aa, Aa, \) and \( AA \)). Since each of the three genotypes can lead to three different expression levels, we expect expression distributions of a cluster-2A gene to have three modes. This leads to the question: Why do we not see trimodal, as opposed to bimodal, expression distributions for genes in cluster 2A? To answer this question, we develop the following frameworks. Let us assume that a genetic polymorphism exists with two alleles, \( A \) and \( a \), with frequencies \( p_A \) and \( (1-p_A) \), respectively. The three genotypes for this polymorphism, \( aa, Aa, \) and \( AA \), lead to three different expression states (TPM levels) for the gene with averages
+\( \mu_{aa}, \mu_{Aa}, \) and \( \mu_{AA} \), respectively. Let us also assume that the Hardy-Weinberg equilibrium holds for this locus. Then, the frequency of \( aa = (1-p_A)^2 \), the frequency of \( Aa = 2p_A(1-p_A) \), and the frequency of \( AA = p_A^2 \). We assume that \( \mu_{aa} \leq \mu_{Aa} \leq \mu_{AA} \). Next, we define a dominance coefficient \( 0 \leq \alpha \leq 1 \) by,
+
+\[
+\mu_{Aa} = \mu_{aa} + (\mu_{AA} - \mu_{aa})\alpha.
+\]
+
+If we define the ratio \( R \) by
+
+\[
+R = \frac{\mu_{AA}}{\mu_{aa}},
+\]
+
+then, we obtain
+
+\[
+\mu_{Aa} = \mu_{aa} (1 - \alpha + R\alpha)
+\]
+
+and
+
+\[
+\mu_{AA} = R\mu_{aa}.
+\]
+
+We can then divide individuals into three groups depending on their genotypes. Let us assume that the coefficient of variation (CV) of expression is the same for each genotypic group. Then, we can model the TPM value of this gene in a given individual a normal random variable with:
+    1) mean = \( \mu_{aa} \) and standard deviation = CV x \( \mu_{aa} \) if the genotype is \( aa \);
+    2) mean = \( \mu_{Aa} \) and standard deviation = CV x \( \mu_{Aa} \) if the genotype is \( Aa \); and
+    3) mean = \( \mu_{AA} \) and standard deviation = CV x \( \mu_{AA} \) if the genotype is \( AA \).
+The value of \( \mu_{aa} \) is irrelevant for gauging the effect of polymorphisms on the shape of the expression level distributions. Therefore, we set \( \mu_{aa} = 1 \).
+
+Under these mathematical assumptions, we performed simulations using 36 distinct models. These models vary by four parameters: \( p_A \in \{0.05, 0.1, 0.5\} \), \( CV \in \{0.1, 0.3\} \), \( R \in \{10, 1000\} \), and \( \alpha \in \{0.2, 0.5, 0.8\} \). For each model, defined by a unique combination of the values of these four parameters, we performed a two-step sampling procedure. First, we obtained a random sample of 500 genotypes, based on \( p_A \) and the Hardy-Weinberg equilibrium. Next, for each of the 500 genotypes sampled, we sample a TPM value from the normal distribution corresponding to that genotype. Thus, for each of the 36 models, we simulated 500 TPM values. We present these values as histograms with and without log transformation. The results for \( p_A = 0.05 \), \( p_A = 0.1 \), and \( p_A = 0.5 \) are shown in Figure S4, Figure S5, and Figure S6, respectively. These simulations help us answer our question we first asked: Why do we not see a trimodal distribution if a genetic polymorphism drives expression-level variability in a gene?
+
+Firstly, even when the minor allele (\( A \)) frequency is not low (e.g., 10%), the frequency of the genotype \( AA \) is still quite low (e.g., 1%). Therefore, the third peak is not always conspicuously visible. We see this in all models with \( p_A = 0.05 \) and \( p_A = 0.1 \) (Figures S4 and S5), regardless of CV, \( R \), and \( \alpha \) values. At higher allele frequencies (e.g., 50%), the effect of the remaining parameters becomes more apparent. Figure S6 shows that a
+higher dominance coefficient \( \alpha \) makes the expression level distribution more bimodal. By contrast, a lower dominance coefficient \( \alpha \) makes the expression level distribution more trimodal. The lack of observed trimodality in the GTEx data may suggest that expression levels of switch-like genes tend to be more dominant than additive with regard to causal genetic polymorphisms. Secondly, greater variation (CV) in the data can also obscure the third peak. For example, by comparing Figure S6B to Figure S6H, we find that increasing the CV can change the distribution from being trimodal to bimodal when the other parameters are held constant. However, \( R \) does not seem to have much effect on whether the expression level distribution is bimodal or trimodal.
+
+![TPM simulations for a hypothetical gene whose expression is driven by a genetic polymorphism with an allele frequency of 5%.](page_184_672_1207_670.png)
+
+Figure S4. TPM simulations for a hypothetical gene whose expression is driven by a genetic polymorphism with an allele frequency of 5%.
+Figure S5. TPM simulations for a hypothetical gene whose expression is driven by a genetic polymorphism with an allele frequency of 10%.
+Figure S6. TPM simulations for a hypothetical gene whose expression is driven by a genetic polymorphism with an allele frequency of 50%.
+
+Tissue-specific switch-like genes
+We divided switch-like genes into three clusters in the space spanned by the first two principal components (Figure 2A). While we said that genes in cluster 1 (Figure 2A-B) are tissue-specific switch-like genes, manual inspection reveals this is not true for all genes in cluster 1. In particular, the transcript ENSG00000273906 coming from chr Y was labeled cluster 1 by hierarchical clustering even though it is universally switch-like in tissues common to both sexes. Indeed, we removed all chr-Y genes from our analyses of genuine cluster-1 genes. Other cluster-1 genes bimodally expressed in a large number of tissues lie on the autosomes. For example, CLPS, PRSS1, CELA3A, and CELA3B, despite having low overall tissue-to-tissue co-expression, are bimodally expressed across tissues. Indeed, we have shown previously that CELA3A and CELA3B have a shared regulatory architecture in the pancreas 81.
+
+Controlling for confounders
+We removed cluster-1 genes affected by confounders in each tissue using an approach outlined in Methods. Here, we present the number of genuine cluster-1 genes versus
+those affected by confounders in **Figure S7**. In particular, we show that the cluster-1 genes in the colon and the intestine are particularly prone to being affected by confounding factors. We also present in **Figure S8** examples of genes whose bimodal expression in specific tissues is correlated with variation in the sample ischemic time distribution.
+
+![Bar chart showing the number of total tissue-specific switch-like genes (cluster 1) for various tissues, with two bars per tissue: one for genuine tissue-specific switch-like genes and one for genes affected by confounding factors.](page_324_629_1017_563.png)
+
+**Figure S7.** Switch-like genes in cluster 1 that are genuine versus those affected by confounders.
+Figure S8. Examples of cluster-1 genes affected by confounders. Their bimodal distribution is caused by ischemic time (a confounding factor).
+
+The copy number variation at the PGA3 locus does not affect the gene’s expression levels
+PGA3 exhibits a high copy number variation among humans \(^{82}\), but the copy number seems to have no impact on \(PGA3\) expression, at least in cancer samples \(^{83}\). The bimodal expression of PGA3 in the stomach is likely not due to its copy number variation. This is because PGA3’s expression in the stomach is highly correlated with other tissue-specific genes in the stomach. The only way in which a copy number-driven bimodality of PGA3 could be correlated with other switch-like genes is if the product of PGA3 was regulating the correlated genes. Without this evidence, we surmise that the copy number variation at the PGA3 locus does not affect the gene’s expression levels, at least in the stomach.
+
+Table S1. A list of tissues used in this study along with the number of individuals for each tissue.
+Table S2. A list of 1,013 switch-like genes.
+Table S3. Tissue-to-tissue co-expression (Pearson’s correlation) for all genes across 310 tissue-tissue pairs.
+Table S4. Results from principal component analysis on tissue-to-tissue co-expression data for all genes.
+Table S5. Results from principal component analysis on tissue-to-tissue co-expression data for only switch-like genes.
+Table S6. Correlation between gene expression levels and confounding factors for switch-like
+genes.
+Table S7. Gene-to-gene co-expression of genuine tissue-specific switch-like genes in the stomach, vagina, breast, and colon.
+Table S8. Analysis of sex bias among genuine tissue-specific switch-like genes.
+
+1. Jacob, F. & Monod, J. Genetic regulatory mechanisms in the synthesis of proteins. J. Mol. Biol. **3**, 318–356 (1961).
+2. Banerji, J., Rusconi, S. & Schaffner, W. Expression of a beta-globin gene is enhanced by remote SV40 DNA sequences. *Cell* **27**, 299–308 (1981).
+3. Gillies, S. D., Morrison, S. L., Oi, V. T. & Tonegawa, S. A tissue-specific transcription enhancer element is located in the major intron of a rearranged immunoglobulin heavy chain gene. *Cell* **33**, 717–728 (1983).
+4. Serfling, E., Jasin, M. & Schaffner, W. Enhancers and eukaryotic gene transcription. *Trends Genet.* **1**, 224–230 (1985).
+5. Allfrey, V. G., Faulkner, R. & Mirsky, A. E. Acetylation and methylation of histones and their possible role in the regulation of RNA synthesis. *Proc. Natl. Acad. Sci. U. S. A.* **51**, 786–794 (1964).
+6. Riggs, A. D. & Jones, P. A. 5-methylcytosine, gene regulation, and cancer. *Adv. Cancer Res.* **40**, 1–30 (1983).
+7. Bestor, T., Laudano, A., Mattaliano, R. & Ingram, V. Cloning and sequencing of a cDNA encoding DNA methyltransferase of mouse cells. The carboxyl-terminal domain of the mammalian enzymes is related to bacterial restriction methyltransferases. *J. Mol. Biol.* **203**, 971–983 (1988).
+8. Jenuwein, T. & Allis, C. D. Translating the histone code. *Science* **293**, 1074–1080 (2001).
+9. Levine, M. & Tjian, R. Transcription regulation and animal diversity. *Nature* **424**, 147–151 (2003).
+10. Muntoni, F., Torelli, S. & Ferlini, A. Dystrophin and mutations: one gene, several proteins, multiple phenotypes. *Lancet Neurol.* **2**, 731–740 (2003).
+11. Sakai, T. *et al.* Allele-specific hypermethylation of the retinoblastoma tumor-suppressor gene. *Am. J. Hum. Genet.* **48**, 880–888 (1991).
+12. Cutting, G. R. Cystic fibrosis genetics: from molecular understanding to clinical application. *Nat. Rev.*
+Genet. **16**, 45–56 (2015).
+
+13. Ertel, A. Bimodal gene expression and biomarker discovery. *Cancer Inform.* **9**, 11–14 (2010).
+
+14. Bessarabova, M. *et al*. Bimodal gene expression patterns in breast cancer. *BMC Genomics* **11 Suppl 1**, S8 (2010).
+
+15. Justino, J. R., Reis, C. F. dos, Fonseca, A. L., Souza, S. J. de & Stransky, B. An integrated approach to identify bimodal genes associated with prognosis in cancer. *Genet. Mol. Biol.* **44**, e20210109 (2021).
+
+16. Moody, L., Mantha, S., Chen, H. & Pan, Y.-X. Computational methods to identify bimodal gene expression and facilitate personalized treatment in cancer patients. *J. Biomed. Inform.* **100S**, 100001 (2019).
+
+17. Mason, C. C. *et al*. Bimodal distribution of RNA expression levels in human skeletal muscle tissue. *BMC Genomics* **12**, 98 (2011).
+
+18. Hartigan, J. A. & Hartigan, P. M. The dip test of unimodality. *Ann. Stat.* **13**, 70–84 (1985).
+
+19. Jangravi, Z. *et al*. A fresh look at the male-specific region of the human Y chromosome. *J. Proteome Res.* **12**, 6–22 (2013).
+
+20. Cheung, S., Parrella, A., Rosenwaks, Z. & Palermo, G. D. Genetic and epigenetic profiling of the infertile male. *PLoS ONE* **14**, e0214275 (2019).
+
+21. Turjya, R. R., Khan, M. A.-A.-K. & Mir Md Khademul Islam, A. B. Perversely expressed long noncoding RNAs can alter host response and viral proliferation in SARS-CoV-2 infection. *Future Virol.* **15**, 577–593 (2020).
+
+22. Talotta, R., Bahrami, S. & Laska, M. J. Sequence complementarity between human noncoding RNAs and SARS-CoV-2 genes: What are the implications for human health? *Biochim. Biophys. Acta Mol. Basis Dis.* **1868**, 166291 (2022).
+
+23. Arman, K., Dalloul, Z. & Bozgeyik, E. Emerging role of microRNAs and long non-coding RNAs in COVID-19 with implications to therapeutics. *Gene* **861**, 147232 (2023).
+
+24. Li, Q. *et al*. Genome-wide search for exonic variants affecting translational efficiency. *Nat. Commun.* **4**, 2260 (2013).
+25. Liu, L. et al. Mutation of the CDKN2A 5' UTR creates an aberrant initiation codon and predisposes to melanoma. Nat. Genet. 21, 128–132 (1999).
+
+26. Khan, Z. et al. Primate transcript and protein expression levels evolve under compensatory selection pressures. Science 342, 1100–1104 (2013).
+
+27. Wang, S. H., Hsiao, C. J., Khan, Z. & Pritchard, J. K. Post-translational buffering leads to convergent protein expression levels between primates. Genome Biol. 19, 83 (2018).
+
+28. Chia, D. J. Minireview: mechanisms of growth hormone-mediated gene regulation. Mol. Endocrinol. 28, 1012–1025 (2014).
+
+29. Mayne, B. T. et al. Large scale gene expression meta-analysis reveals tissue-specific, sex-biased gene expression in humans. Front. Genet. 7, 183 (2016).
+
+30. McEwen, B. S. & Milner, T. A. Understanding the broad influence of sex hormones and sex differences in the brain. J. Neurosci. Res. 95, 24–39 (2017).
+
+31. Goel, N., Workman, J., -Y. Lee, T. T., Innala, L. & Viau, V. Sex differences in the HPA axis. Compr. Physiol. 4, 1121–1155 (2014).
+
+32. Fink, A. L. & Klein, S. L. The evolution of greater humoral immunity in females than males: implications for vaccine efficacy. Curr. Opin. Physiol. 6, 16–20 (2018).
+
+33. Laffont, S. & Guéry, J.-C. Deconstructing the sex bias in allergy and autoimmunity: From sex hormones and beyond. Adv. Immunol. 142, 35–64 (2019).
+
+34. Wu, Q. et al. Comprehensive analysis of the expression and prognostic value of spint1/2 in breast carcinoma. Front. Endocrinol. 12, 665666 (2021).
+
+35. Takan, I., Karakülah, G., Louka, A. & Pavlopoulou, A. ‘In the light of evolution:’ keratins as exceptional tumor biomarkers. PeerJ 11, e15099 (2023).
+
+36. Behtaji, S. et al. Identification of oxytocin-related lncRNAs and assessment of their expression in breast cancer. Sci. Rep. 11, 6471 (2021).
+
+37. Fiaz, T. et al. Peripheral mRNA expression and prognostic significance of emotional stress biomarkers in metastatic breast cancer patients. Int. J. Mol. Sci. 23, (2022).
+
+38. Emilsson, V. et al. Genetics of gene expression and its effect on disease. Nature 452, 423–428
+39. Li, H. et al. Characterization of differentially expressed genes involved in pathways associated with gastric cancer. PLoS ONE 10, e0125013 (2015).
+
+40. Goldstein, I., Dicks, B., Kim, N. N. & Hartzell, R. Multidisciplinary overview of vaginal atrophy and associated genitourinary symptoms in postmenopausal women. Sex. Med. Today 1, 44–53 (2013).
+
+41. Szymański, J. K., Słabuszewska-Jóźwiak, A. & Jakiel, G. Vaginal aging—what we know and what we do not know. Int. J. Environ. Res. Public Health 18, 4935 (2021).
+
+42. Kim, H.-K., Kang, S.-Y., Chung, Y.-J., Kim, J.-H. & Kim, M.-R. The recent review of the genitourinary syndrome of menopause. J. Menopausal Med. 21, 65–71 (2015).
+
+43. Hummelen, R. et al. Vaginal microbiome and epithelial gene array in post-menopausal women with moderate to severe dryness. PLoS ONE 6, e26602 (2011).
+
+44. Faubion, S. S., Sood, R. & Kapoor, E. Genitourinary syndrome of menopause: management strategies for the clinician. Mayo Clin. Proc. 92, 1842–1849 (2017).
+
+45. Buchanan, D. L. et al. Role of stromal and epithelial estrogen receptors in vaginal epithelial proliferation, stratification, and cornification. Endocrinology 139, 4345–4352 (1998).
+
+46. Anderson, D. J., Marathe, J. & Pudney, J. The structure of the human vaginal stratum corneum and its role in immune defense. Am. J. Reprod. Immunol. 71, 618–623 (2014).
+
+47. Greenman, C. et al. Patterns of somatic mutation in human cancer genomes. Nature 446, 153–158 (2007).
+
+48. Oomizu, S. et al. Kdap, a novel gene associated with the stratification of the epithelium. Gene 256, 19–27 (2000).
+
+49. Ho, M. et al. Update of the keratin gene family: evolution, tissue-specific expression patterns, and relevance to clinical disorders. Hum. Genomics 16, 1 (2022).
+
+50. Pontén, F., Jirström, K. & Uhlen, M. The Human Protein Atlas--a tool for pathology. J. Pathol. 216, 387–393 (2008).
+
+51. Cotreau, M. M. et al. A study of 17β-estradiol-regulated genes in the vagina of postmenopausal women with vaginal atrophy. Maturitas 58, 366–376 (2007).
+52. Miyagawa, S. & Iguchi, T. Epithelial estrogen receptor 1 intrinsically mediates squamous differentiation in the mouse vagina. Proc. Natl. Acad. Sci. U. S. A. **112**, 12986–12991 (2015).
+
+53. Zhang, X. *et al.* Identification of key pseudogenes in nasopharyngeal carcinoma based on RNA-Seq analysis. *BMC Cancer* **21**, 483 (2021).
+
+54. Gao, L. *et al.* Gene expression profile of THZ1-treated nasopharyngeal carcinoma cell lines indicates its involvement in the inhibition of the cell cycle. *Transl. Cancer Res.* **10**, 445–460 (2021).
+
+55. Dudbridge, F. & Fletcher, O. Gene-environment dependence creates spurious gene-environment interaction. *Am. J. Hum. Genet.* **95**, 301–307 (2014).
+
+56. Vetr, N. G., Gay, N. R., MoTrPAC Study Group & Montgomery, S. B. The impact of exercise on gene regulation in association with complex trait genetics. *Nat. Commun.* **15**, 3346 (2024).
+
+57. Thomas, D. Gene–environment-wide association studies: emerging approaches. *Nat. Rev. Genet.* **11**, 259–272 (2010).
+
+58. Gilliland, F. D. *et al.* Effects of glutathione S-transferase M1, maternal smoking during pregnancy, and environmental tobacco smoke on asthma and wheezing in children. *Am. J. Respir. Crit. Care Med.* **166**, 457–463 (2002).
+
+59. Hellwig, B. *et al.* Comparison of scores for bimodality of gene expression distributions and genome-wide evaluation of the prognostic relevance of high-scoring genes. *BMC Bioinformatics* **11**, 276 (2010).
+
+60. Shalek, A. K. *et al.* Single-cell transcriptomics reveals bimodality in expression and splicing in immune cells. *Nature* **498**, 236–240 (2013).
+
+61. Dobrin, R. *et al.* Multi-tissue coexpression networks reveal unexpected subnetworks associated with disease. *Genome Biol.* **10**, R55 (2009).
+
+62. 1000 Genomes Project Consortium *et al.* A global reference for human genetic variation. *Nature* **526**, 68–74 (2015).
+
+63. Ebert, P. *et al.* Haplotype-resolved diverse human genomes and integrated analysis of structural variation. *Science* **372**, (2021).
+
+64. Quinlan, A. R. & Hall, I. M. BEDTools: a flexible suite of utilities for comparing genomic features.
+Bioinformatics **26**, 841–842 (2010).
+
+65. Cohen, J. Statistical Power Analysis for the Behavioral Sciences. (Routledge, London, England, 2013). doi:10.4324/9780203771587.
+
+66. Thomas, P. D. et al. PANTHER: Making genome-scale phylogenetics accessible to all. Protein Sci. **31**, 8–22 (2022).
+
+67. Kallak, T. K. et al. Aromatase inhibitors affect vaginal proliferation and steroid hormone receptors. Menopause **21**, 383–390 (2014).
+
+68. Kallak, T. K. et al. Vaginal gene expression during treatment with aromatase inhibitors. Clin. Breast Cancer **15**, 527–535.e2 (2015).
+
+69. Saitou, M. & Gokcumen, O. An evolutionary perspective on the impact of genomic copy number variation on human health. J. Mol. Evol. **88**, 104–119 (2020).
+
+70. Saitou, M., Satta, Y., Gokcumen, O. & Ishida, T. Complex evolution of the GSTM gene family involves sharing of GSTM1 deletion polymorphism in humans and chimpanzees. BMC Genomics **19**, 293 (2018).
+
+71. Rothman, N. et al. A multi-stage genome-wide association study of bladder cancer identifies multiple susceptibility loci. Nat. Genet. **42**, 978–984 (2010).
+
+72. Aqil, A., Speidel, L., Pavlidis, P. & Gokcumen, O. Balancing selection on genomic deletion polymorphisms in humans. eLife https://doi.org/10.7554/eLife.79111 (2023).
+
+73. Duret, L. & Mouchiroud, D. Determinants of substitution rates in mammalian genes: expression pattern affects selection intensity but not mutation rate. Mol. Biol. Evol. **17**, 68–74 (2000).
+
+74. Zhang, L. & Li, W.-H. Mammalian housekeeping genes evolve more slowly than tissue-specific genes. Mol. Biol. Evol. **21**, 236–239 (2004).
+
+75. Park, J., Xu, K., Park, T. & Yi, S. V. What are the determinants of gene expression levels and breadths in the human genome? Hum. Mol. Genet. **21**, 46–56 (2011).
+
+76. Li, C., Qu, W. & Yang, X. Comprehensive IncRNA and mRNA profiles in peripheral blood mononuclear cells derived from ankylosing spondylitis patients by RNA-sequencing analysis. Medicine **101**, e27477 (2022).
+77. von Grothusen C. Endometrial receptivity and regeneration in health and disease: Molecular, cellular and clinical perspectives. Karolinska Institutet (Sweden, 2022).
+
+78. Almsned, F. M. Understanding the genetic nature of multiple sclerosis using next-generation sequencing genomic analysis methods. (Doctoral dissertation, George Mason University, 2020).
+
+79. Chery, G. Understanding Sjögren’s Syndrome as a Systemic Autoimmune Disorder. (State University of New York at Albany, 2022).
+
+80. Islam, T. et al. Detection of multiple sclerosis using blood and brain cells transcript profiles: Insights from comprehensive bioinformatics approach. Informatics in Medicine Unlocked **16**, 100201 (2019).
+
+81. Russell, M., Aqil, A., Saitou, M., Gokcumen, O. & Masuda, N. Gene communities in co-expression networks across different tissues. *PLoS Comput. Biol.* **19**, e1011616 (2023).
+
+82. Otto, M., Zheng, Y., Grablowitz, P. & Wiehe, T. Detecting adaptive changes in gene copy number distribution accompanying the human out-of-Africa expansion. *bioRxiv* 2023.08.14.553171 (2024) doi:10.1101/2023.08.14.553171.
+
+83. Shen, S., Li, H., Liu, J., Sun, L. & Yuan, Y. The panoramic picture of pepsinogen gene family with pan-cancer. *Cancer Med.* **9**, 9064–9080 (2020).

015d965c0cb0c74c387f9596e3f91c0197ff39a7333b1c0b5bef96993bcbbd18/metadata.json

@@ -0,0 +1,130 @@
+{
+    "title": "The emergence of three-dimensional chiral domain walls in polar vortices",
+    "pre_title": "The emergence of three-dimensional chiral domain walls in polar vortices",
+    "journal": "Nature Communications",
+    "published": "25 July 2023",
+    "supplementary_0": [
+        {
+            "label": "Supplementary Information",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-023-40009-2/MediaObjects/41467_2023_40009_MOESM1_ESM.docx"
+        },
+        {
+            "label": "Peer Review File",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-023-40009-2/MediaObjects/41467_2023_40009_MOESM2_ESM.pdf"
+        }
+    ],
+    "supplementary_1": NaN,
+    "supplementary_2": NaN,
+    "source_data": [],
+    "code": [],
+    "subject": [
+        "Ferroelectrics and multiferroics",
+        "Topological defects"
+    ],
+    "license": "http://creativecommons.org/licenses/by/4.0/",
+    "preprint_pdf": "https://www.researchsquare.com/article/rs-2551328/v1.pdf?c=1690369728000",
+    "research_square_link": "https://www.researchsquare.com//article/rs-2551328/v1",
+    "nature_pdf": "https://www.nature.com/articles/s41467-023-40009-2.pdf",
+    "preprint_posted": "07 Feb, 2023",
+    "research_square_content": [
+        {
+            "section_name": "Abstract",
+            "section_text": "Chirality or handedness of a material can be used as an order parameter to uncover emergent electronic properties for quantum information science. Conventionally, chirality is found in naturally occurring biomolecules and magnetic materials. Chirality can be engineered in a topological polar vortex ferroelectric/dielectric system via atomic-scale symmetry-breaking operations. We use four-dimensional scanning transmission electron microscopy (4D-STEM) to map out topology-driven three-dimensional domain walls, where the handedness of two neighbor topological domains change or remain the same. The nature of the domain walls is governed by the interplay of local perpendicular (lateral) and parallel (axial) polarization with respect to the tubular vortex structures. Unique symmetry-breaking operations and the finite nature of domain walls results in a triple point at the junction of chiral and achiral domain walls. The unconventional nature of the domain walls with triple point pairs may result in unique electrostatic and magnetic properties potentially useful for quantum sensing applications.Physical sciences/Materials science/Condensed-matter physics/Ferroelectrics and multiferroicsPhysical sciences/Materials science/Condensed-matter physics/Topological matter/Topological defects",
+            "section_image": []
+        },
+        {
+            "section_name": "Additional Declarations",
+            "section_text": "There is NO Competing Interest.",
+            "section_image": []
+        },
+        {
+            "section_name": "Supplementary Files",
+            "section_text": "SI.pdf",
+            "section_image": []
+        }
+    ],
+    "nature_content": [
+        {
+            "section_name": "Abstract",
+            "section_text": "Chirality or handedness of a material can be used as an order parameter to uncover\u00a0the\u00a0emergent electronic properties for quantum information science. Conventionally, chirality is found in naturally occurring biomolecules and magnetic materials. Chirality can be engineered in a topological polar vortex ferroelectric/dielectric system via atomic-scale symmetry-breaking operations. We use four-dimensional scanning transmission electron microscopy (4D-STEM) to map out the\u00a0topology-driven three-dimensional domain walls, where the handedness of two neighbor topological domains change or remain the same. The nature of the domain walls is governed by the interplay of\u00a0the local perpendicular (lateral) and parallel (axial) polarization with respect to the tubular vortex structures. Unique symmetry-breaking operations and the finite nature of domain walls result in a triple point\u00a0formation at the junction of chiral and achiral domain walls. The unconventional nature of the domain walls with triple point pairs may result in unique electrostatic and magnetic properties potentially useful for quantum sensing applications.",
+            "section_image": []
+        },
+        {
+            "section_name": "Introduction",
+            "section_text": "Chirality is a unique topological feature that drives many-body interactions in naturally occurring organic molecules and proteins1, subatomic particle physics2, and solid-state physics3. The symmetries in a chiral system are configured in such a way that its mirror image cannot be superimposed on itself, manifesting a handedness to the system as exemplified by screws and our own hands. Chirality also exists at the microscale/nanoscale level in inorganic and organic materials such as liquid crystals4, spin textures in ferromagnets5,6, amino acids, and D/L-glucose molecules1 with applications in spin selectivity-based quantum sensing7, non-linear optics8, and biosensing applications9. However, there are very a\u00a0few examples of chiral inorganic ferroelectric crystals which could have fundamentally different domain textures10,11,12,13. Over the past few years, novel polarization textures in ferroelectrics such as merons14, polar flux-closure domains15,16, vortices17, bubble domains18,19, antivortex20, three-fold polar vertices21, super crystals22,23 and skyrmions24 have been engineered in oxide superlattices, emerging from the careful interplay of elastic, electrostatic and gradient energies of electric dipoles. The electric dipole arrangement and complex orbital hybridization in these systems have been probed by the X-ray scattering techniques25, scanning transmission electron microscopy (STEM)17-electron energy loss spectroscopy (EELS)26, phase-field simulations12,14,15,17,22 and atomistic first- and second-principles calculations17,18,22,24,25. Surprisingly, the presence of chirality has been observed in one such topological texture i.e., polar vortices in PbTiO3/SrTiO3 (PTO/STO) superlattices from resonant soft X-ray scattering (RXS)25, second harmonic generation (SHG) and second principles calculations27. The presence of chirality in polar vortices is an emergent phenomenon because none of the parent compounds such as STO or PTO are known to be chiral. It has been recently shown experimentally and theoretically that the presence of chirality in this system might be due to different sources. First\u00a0factor, the strongest one, is the coexistence of vortices with an axial component of the polarization, perpendicular to the vortex plane28. The second factor is the buckling of the vortices (i.e., a staggered vortex configuration where the center of the clockwise and counterclockwise vortices are located at different heights) combined with different sizes of the up and down domains results in a chiral structure, although its strength is smaller in comparison to\u00a0the first scenario. This last source of chirality can be reversed by the\u00a0external electric fields. The first experimental demonstration was in refs. 27,29, where the\u00a0chirality switches in a reversible, deterministic, and non-destructive fashion over mesoscale regions27.\n\nWithout any prior knowledge, one would expect the as-grown sample as a racemic mixture, i.e., equal amounts of left-handed and right-handed domain enantiomers, where the\u00a0chirality within each domain comes from a combination of the two sources described above. To have a non-destructive switchable chirality, it is essential to understand the role of the domain walls separating the enantiomers. In other words, what local physical parameters play a role when the handedness in the\u00a0neighboring domains change? This includes the offset between the centers of the core, the axial component at the centers of the clockwise/counterclockwise vortex, the sense of rotation of the vortices when they merge at the domain wall, the presence of dislocations, or the combined effect of all of them. A proper understanding of how the left/right-handed domains evolve at the nanoscale is crucial to design\u00a0the new electrically switchable chiral devices that can be measured without scientifically sophisticated techniques. Indeed, a proper answer to this question will pave the way for the use of these chiral textures in next-generation technologies.\n\nFour-dimensional (4D)-STEM can precisely measure strain, and thus spontaneous polarization in ferroelectrics due to the violation of Friedel\u2019s Law30,31,32. This makes 4D-STEM a unique tool to probe polarization in three dimensions and understand emergent chirality in polarization vortices. In the current work, we have used 4D-STEM to probe three-dimensional domain walls in polar vortices in oxide superlattices and understand the nanoscale nature of chirality. We find that both achiral and chiral domain walls coexist in the same system. The chiral to achiral domain wall transition is driven by the change in the\u00a0local axial and lateral polarization direction across the domain wall. We have discovered a\u00a0new pair of triple points with the opposite/same sense of rotation at the junction of achiral and chiral domain walls. Finally, we unravel all the possible configurations of chiral and achiral domain walls in this system through different symmetry-breaking operations.",
+            "section_image": []
+        },
+        {
+            "section_name": "Results",
+            "section_text": "Trilayer (STO)16/(PTO)16/(STO)16 were grown on orthorhombic DyScO3 (DSO) [110]o substrates with SrRuO3 as a buffer layer using reflection high-energy electron diffraction (RHEED)-assisted pulsed-laser deposition (PLD) (Fig.\u00a01a). The polar vortex phase in this system is stabilized as a consequence of the interplay between depolarization energy at the PTO/STO interfaces, elastic energy from the tensile strain imposed by the DSO substrate, and the gradient energy in the ferroelectric17.\n\na Bright-Field STEM image showing the trilayer\u00a0stack of 16 STO/16 PTO/16 STO with SRO buffer layer on DSO substrate. The red regions near the center of the PTO layer indicate the position of the vortex core approximately. b (i) Vector mapping of the local displacements of A sites of highlighted region in (a) overlaid on HAADF-STEM image. The local red- and blue contrast at the center of the PTO layer indicates the local non-zero curl of displacement. b(i\u2013ii) The net lateral polarization resulting from vortex off-centering is indicated at the top. c Dark field TEM image along [110]o direction displaying a\u00a0long tubular vortex structure with domain walls shown as white dashed lines.\n\nA direct way to measure the polar textures in vortex topologies is through electron microscopy using atomic resolution images. We have used different types of STEM and TEM techniques to characterize the exact positions of the vortex centers. Figure\u00a01a shows a low-magnification bright-field STEM image of the superlattice trilayer with an SRO buffer layer along the \\({[1\\bar{1}0]}_{{{{{{\\rm{o}}}}}}}\\) zone axis. The diffraction contrast in BF-STEM allows us to directly locate the vortex center as dark contrast; marked using red circles in Fig.\u00a01a. To precisely understand the polarization or displacement texture around each vortex center, we obtained the A-site displacement vector maps at atomic resolution via gaussian fitting at the\u00a0A-sites (\u201cMethods\u201d, Supplementary Information). Figure\u00a01b and Fig.\u00a0S1 shows the High angle annular dark field (HAADF)-STEM image of trilayer STO/PTO/STO along \\({[1\\bar{1}0]}_{{{{{{\\rm{o}}}}}}}\\) zone axis where the\u00a0brighter regions are PTO and the\u00a0darker regions are STO. Overlaid yellow arrows show the\u00a0clockwise and counterclockwise rotating curls in the displacement of the A-cation in PTO/STO superlattices. The coexistence of the concomitant non-zero curl of polarization (red/blue contrast) with an alternating axial component of the polarization (perpendicular to the plane defined by the vortices) is the first symmetry-breaking operation that results in the\u00a0emergent chirality in an otherwise non-chiral system. The non-zero curl is larger in continuously rotating polarization textures such as polar vortices27,30, merons33, and skyrmions24 than in other polarization textures such as flux closure domains15,16 where the curl vanished in the central regions with 1800 domain walls. In addition, we observe that the cores of the polarization curls (indicated as blue/red contrast) are not located exactly at the center of the PTO layer, but follow a zig-zag type pattern, giving rise to a net in-plane polarization rotation along [001]o (lateral component) indicated as Px in Fig.\u00a01b(i\u2013ii) and Figure\u00a0S2. This buckling, combined with a small difference in the size of the up and down domains, is the second symmetry-breaking operation that results in\u00a0a net chirality; in agreement with previous observations27.\n\nThe various permutations and combinations of atomic scale symmetry-breaking operations such as, a non-zero curl of polarization together with the presence of an axial component, and the buckling of the vortices that yield a non-zero polarization component along [001]o, result in different types of domains walls at the mesoscale. We can visualize the mesoscale domain walls in these topological structures by imaging the vortices along the [110]o zone axis using weak beam dark field (WBDF) TEM (Fig.\u00a01c). We can observe the tubular nature of vortex textures by long bright and dark stripes regions in the image. In addition, we observe different domain wall features (indicated as white dashed lines) cutting across vortex tubes. Overall, if we combine our observations from WBDF-TEM and HR-STEM, we observe a three-dimensional vortex\u00a0structure in the PTO layer sandwiched between two\u00a0STO layers where the polarization curls follow a tubular pattern (Fig.\u00a02a). Unfortunately, the images formed by WBDF-TEM, HAADF-STEM, and BF-STEM are merely atomic projections and cannot give us an estimate of the\u00a0physical quantities such as polarization and chirality. On the other hand, 4D-STEM allows us to collect a diffraction pattern at each probe position, which can then be used to create precise maps of the\u00a0physical quantities such as strain and polarization. The diffraction pattern in 4D-STEM offers a unique advantage over HAADF-STEM in polar vortices because direction of polarization can be accurately measured. For the present experiment, we performed 4D-STEM imaging on a trilayer STO/PTO/STO along the [110]o zone axis (Fig.\u00a02a). We used a probe size of ~7\u2009\u00c5, larger than the STO/PTO unit cell dimensions (~4\u2009\u00c5) to remove the atomic-scale\u00a0signal and to estimate the polarization at unit cell length\u00a0scale. The 4D-STEM analysis was carried out using\u00a0an open source py4DSTEM analysis package34. We define the \\({[1\\bar{1}0]}_{{{{{{\\rm{o}}}}}}}\\) direction as axial and the\u00a0[001]o direction as lateral. The rotation calibration was performed between the real space and the\u00a0diffraction space\u00a0images to determine the lateral and the\u00a0axial zone\u00a0axis. Details are given in the supplementary information. For\u00a0the initial visualization of the polar textures, we created a virtual dark field image using the \\({[2\\bar{1}0]}_{{{{{{\\rm{o}}}}}}}\\) disk (disk 3) as shown in Fig.\u00a02b, c (More dark images in Figure\u00a0S4). The polarization from the PTO layer can be determined qualitatively by subtracting the intensity of the opposite Friedel pair disks due to the violation of Friedel\u2019s law30,31,32. This polarization mapping is slightly different from strain mapping. In the former, intensity of the opposite freidel pairs disks are used whereas in the latter the precise position of the diffracted disks is determined35. The method for subtracting the intensities of opposite Friedel\u2019s pair disks would also work for analogous PbxZr1-xTiO3 (PZT) where polarization is suppressed under large tetragonality36. The (hkl) Friedel pairs were chosen along [001]o and \\({[1\\bar{1}0]}_{{{{{{\\rm{o}}}}}}}\\) directions to determine the\u00a0pure lateral and the pure\u00a0axial polarization, respectively. We note that the other (hkl) directions may also break the\u00a0Friedel\u2019s law, but the associated polarization will be a combination of the\u00a0lateral and the\u00a0axial component. The polarization maps corresponding to regions delimited by the\u00a0rectangles in Fig.\u00a02b are shown in Fig.\u00a02d, e. We observe alternate longitudinal red and blue stripes representing the\u00a0positive and negative polarization in both the lateral ([001]o) and the\u00a0axial \\({[1\\bar{1}0]}_{{{{{{\\rm{o}}}}}}}\\) directions together\u00a0with the\u00a0domain walls as seen from the WBDF-TEM images marked with green and magenta lines in Fig.\u00a02d and with a black line in Fig.\u00a02e. We observe that the axial polarization magnitude is relatively smaller than the lateral polarization in agreement with the predictions from previous second principles calculations25,27.\n\na Schematic showing the e-beam scanned in 4D STEM mode across the vortex sample in the\u00a0in-plane geometry. Most of the signal is coming from the top half of the\u00a0PTO layer due to\u00a0the\u00a0strong scattering\u00a0by the\u00a0Pb atomic columns. b Virtual dark field image of the vortex region obtained via integrating the intensity of disk 3 from the summed diffraction pattern in (c). Zoomed-in images of the d dash-dot region and e solid line region showing the\u00a0lateral (P [001]o) and the\u00a0axial (P \\({[1\\bar{1}0]}_{{{{{{\\rm{o}}}}}}}\\)) polarization maps in vortices. The\u00a0\u03b1, \u03b2, and \u03b3 domain walls can be identified. The\u00a0\u03b1 domain wall (magenta curve) has an anti-parallel lateral (P [001]o) component, the\u00a0\u03b2 (green curve) domain wall has an anti-parallel axial component (P \\({[1\\bar{1}0]}_{{{{{{\\rm{o}}}}}}}\\)), the\u00a0\u03b3 domain wall\u00a0has both axial and lateral components anti-parallel across the domain wall.\n\nWe track the relative domain shift across the wall by the black dotted line as shown in Fig.\u00a02d, e. In Fig.\u00a02d, we observe an \u03b1-domain wall (magenta line) where the lateral polarization shifts whereas the axial polarization remains the same, as shown in Fig.\u00a03. We also detect a \u03b2-domain wall (green line), where the axial polarization shifts, but the lateral polarization remains the same. Finally, in Fig.\u00a02e there is a third domain wall configuration i.e., a \u03b3-domain wall (black line) as well where both the lateral and the\u00a0axial polarization shift. We verified this observation with the\u00a0average line profiles across the domain walls (Supplementary information, Figure\u00a0S3). To detect whether a change of chirality occurs at these domain walls, a way to quantify the chirality is required. The order parameter that best captures the breakdown of chiral symmetry is the helicity H of the chiral field. In our case, the chiral field is the polarization, and for the helicity, we borrow the definition from fluid dynamics:\n\nwhere \\(\\vec{{{{{{\\rm{p}}}}}}}\\) is the local value of polarization. Note that H changes sign upon a mirror symmetry reflection37,38. A non-zero helicity means chirality or lack of mirror symmetry of the polarization texture: right (left) handedness can be associated with positive (negative) values of H. Assuming a vortex structure where the polarization lines in the plane defined by the vortices are closed, and that we can measure the axial and the\u00a0lateral components of the polarization at the topmost PTO layer, then the previous equation can be estimated by\n\nwhere plateral and paxial are polarization along lateral and axial directions (Supplementary information, Figure\u00a0S5). Using this equation we can understand the nature of domain walls found in Fig.\u00a02d, e. For the \u03b1/\u03b2 domain wall, only one of the lateral/axial polarization sign changes across the domain wall. This causes a change in the overall sign of helicity, thus making them chiral domain walls. On the other hand, the \u03b3-domain wall has both the lateral and the\u00a0axial polarization switch, which doesn\u2019t change the overall helicity of the system, thus making it an achiral domain wall.\n\na The helicity map formed from Fig.\u00a02b shows left and right-handed domains separated by \u03b1 (magenta) and \u03b2 (green). In addition, achiral domain walls (black line) also coexist. The resultant triple point topologies formed due to the coexistence of chiral and achiral domain walls are shown in encircled regions. The sense of rotation of these triple point topologies is indicated in the encircled region. b, c Possible pairs of triple points. b Triple points with opposite sense of rotation with the point of inversion along \u03b2, \u03b3, and \u03b1, and c represents triple points with the same sense of rotation. The green ticks on the side shows what has been observed in the present\u00a0experiments.\n\nWe can use the qualitative lateral and axial polarization data from 4D-STEM and create a map of the helicity over a large scale using the equation\u00a02 as shown in Fig.\u00a03. The red and\u00a0the blue regions in the chirality maps indicate different signs of helicity in the system, making them left-handed and right-handed chiral domains separated by the \u03b1 or the\u00a0\u03b2 domain wall. We find that most of these chiral and achiral domain walls were not visible in the virtual dark field images. Further, we also observe a unique triple point topology at the mesoscale whenever the two types of chiral domain walls meet an achiral domain wall as seen from black-encircled areas, thus forming a quasi-1D defect in the network of chiral and achiral domain walls. These triple points tend to exist in pairs and exhibit a sense of rotation via the transition from \u03b1 to \u03b2 to \u03b3 domain wall and vice-versa (Fig.\u00a03, Figs.\u00a0S3\u2013S4), similar to what has been observed previously in the trimerized domain walls in hexagonal manganates39, vortex-antivortex phases in intercalated Vander-Waal ferromagnets6, and ferroelectric vortex cores in BiFeO340. The sense of rotation in a triple point can be the same or opposite depending on the arrangement of \u03b1, \u03b2, or \u03b3 domain walls. Figure\u00a03b,c illustrates this situation.\n\nWhenever a pair of domain walls (\u03b2, \u03b3 or \u03b1, \u03b3 or \u03b1, \u03b2) break the\u00a0inversion symmetry across \u03b1, \u03b2 or \u03b3 domains, we form a\u00a0triple point pair with the\u00a0opposite sense of rotation. If the inversion symmetry across \u03b1, \u03b2 or \u03b3 domains is not broken, we form triple point pairs with the same sense of rotation. The origin of triple point pairs can be understood by the following hypothesis. Consider an example of the first triple point pair in Fig.\u00a03b. If this particular type of triple point has to be isolated, then the\u00a0\u03b2 boundary would infinitely separate the positive and negative chirality regions. If \u03b2 boundary is not infinite, then it has to meet an \u03b1 boundary somewhere\u00a0to continue separating the positive and negative chirality regions. Now, if a \u03b2 boundary (change in axial polarization) meets an \u03b1 boundary (change in lateral polarization), a \u03b3 boundary appears (change in both lateral and axial polarization). In such a scenario, we get another triple point in the vicinity of the first triple point thus explaining the origins behind the existance of\u00a0pairs for the majority of cases. \u03b1/\u03b2 boundary could be isolated only in very special circumstances when they are born/die at the surface or they are infinitely long.\n\nFrom a theoretical perspective, there are three symmetry-breaking operations for the formation of domain walls, (1) Change in the lateral polarization direction, (2) Change in axial polarization direction, and (3) Change in the net lateral polarization due to the vortex core shifting away from the center. If we consider all three factors, then we expect to have six types of domain walls as shown in Fig.\u00a04. The first pair of chiral domain walls, \u03b1/\u03b1\u2032, results from a combination of the\u00a0factors 1 and 3. The second chiral domain wall pair, \u03b2/\u03b2\u2032, results from a combination of the\u00a0factors 2 and 3. The third achiral domain wall pair results from all the\u00a0three factors. Unfortunately, it is challenging to measure the quantitative net lateral component in the vortices due to the very low sensitivity of electron scattering to sample changes along the beam direction. Due to this, \u03b1/\u03b1\u2032, \u03b2/\u03b2\u2032, and \u03b3/\u03b3\u2032 are degenerate in this 4D-STEM\u00a0experiment and thus we observe only three types of domain walls. This has been consistent in multiple such 4D STEM images\u00a0as observed in Figs.\u00a0S6\u2013S7. Future studies, such as depth sectioning or sample tilting experiments, may be able to probe this variation41.\n\na Four possible combinations of alternating clockwise/counterclockwise vortices are displaced, Top and bottom cartoons differ by the direction of the axial component of the polarization (red dot and blue cross). Left and right cartons differ by the curl of the polarization. The possible domain walls between these configurations are marked as \u03b1/\u03b1\u2032, \u03b2/\u03b2\u2032, and \u03b3/\u03b3\u2032. \u03b1/\u03b1\u2032 and \u03b2/\u03b2\u2032 domain walls change the chirality at the domain wall. \u03b3/\u03b3\u2032 domains preserve the chirality. b\u2013d Three-dimensional representation of three types of chiral domain walls observed by 4D-STEM measurements.",
+            "section_image": [
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-023-40009-2/MediaObjects/41467_2023_40009_Fig1_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-023-40009-2/MediaObjects/41467_2023_40009_Fig2_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-023-40009-2/MediaObjects/41467_2023_40009_Fig3_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-023-40009-2/MediaObjects/41467_2023_40009_Fig4_HTML.png"
+            ]
+        },
+        {
+            "section_name": "Discussion",
+            "section_text": "We have unraveled the nanoscale three-dimensional domain wall network in topological polar vortices using quantitative 4D-STEM techniques. The polar vortex oxide superlattice has\u00a0an emergent chirality through different symmetry-breaking operations in the non-zero curl of polarization along with the alternate axial polarization component and vortex buckling-induced net-in plane lateral polarization component. The interplay of these symmetry-breaking operations result in the formation of two types of chiral and an achiral domain wall within\u00a0the tubular vortex topologies. Topology-driven domain wall existence in our work is unusual in comparison to other electrostatic conditions-driven domain walls in the ferroic materials16,39. The finite nature of the\u00a0chiral and the\u00a0achiral domain walls result in the formation of unique triple points whenever these domain walls intersect. The most probable existence of these points is in pairs with the same/different handedness, similar to multiferroic materials such as barium hexaferrite42 and BiFeO340. To our best understanding, such an unconventional scenario has not been seen yet in improper ferroics literature. We hope that our studies could inspire future experiments to understand the electronic and the\u00a0magnetic transport at these triple points within the network of chiral and achiral domain walls in polar vortices oxide superlattices.",
+            "section_image": []
+        },
+        {
+            "section_name": "Methods",
+            "section_text": "[(PTO)16/(STO)16] trilayer with SrRuO3 buffer layer was synthesized on single-crystalline DyScO3 (011) substrates via reflection high-energy electron diffraction (RHEED)-assisted pulsed-laser deposition (KrF laser). The PTO and the STO layers were grown at 610\u2009\u00b0C in 100 mTorr oxygen pressure.\n\nIn-plane [(PTO)16/(STO)16] trilayer grown on SrRuO3/DyScO3 substrate were mechanically polished using a 0.5\u00b0 wedge in Allied Multiprep. The samples were subsequently Ar ion milled in a Gatan Precision Ion Milling System, starting from 3.5\u2009keV at 4\u00b0 down to 1\u2009keV at 1\u00b0 for the final polish. The HAADF-STEM images were acquired using double aberration corrected TEAM I microscope operated at 300\u2009kV under non-monochromated mode.\n\nThe vector mapping was performed via Gaussian fitting of A site atomic positions on the drift-corrected HR-STEM images43. First, all the A-sites in the drift-corrected images were identified using \u201cAtomap\u201d atom finding tool44. Once the atoms were identified, the atomic planes were divided into different zone axis such as along [001]0 and [110]0. The deviation in local A-displacement was found by taking the difference between the local A site displacement and the corresponding average displacement in the local zone axis plane. The displacement vectors were further interpolated into a grid Cartesian grid and then differentiated to obtain strain tensor maps. The infinestimal rotation or the curl of the displacement of vortices was calculated using the following equation:\n\nThe color bar in the curl of displacement plot is plotted with respect to the mean intensities in the PTO layer.\n\nAll 4D STEM experiments were carried out on TEAM I microscope (aberration-corrected Thermo Fisher Scientific Titan 80-300) using a Gatan K3 direct detection camera located at the end of a Gatan Continuum imaging filter. The microscope was operated at 300\u2009kV with a probe current of 100 pA. The probe semi-angle used for the measurement was 2 mrad. Diffraction patterns were collected using a step size of 1\u2009nm with 514 by 399 scan positions. The K3 camera was used in full-frame electron counting mode with a binning of 4 pixels by 4 pixels and a camera length of 1.05\u2009mm. The exposure time for each diffraction pattern was 47\u2009ms. The 4D STEM analysis was carried out using the py4DSTEM modules. Briefly, rotation calibration was performed between the diffraction and image plane to identify the right orientation of the zone axis. For that process, the defocused image in the Ronchigram was compared to the focused scan image and the relative orientation of the two images was compared. Once the zone axis was identified, all the disks in the diffraction pattern at each probe position were fitted using the disk fitting function. The so-called polarization maps were generated by taking the normalized intensity difference between the opposite Friedel pair disks. Subsequently, the signal-to-noise in these polarization maps was improved by using a combination of band pass Gaussian filters. By using a high-pass Gaussian filter, we also minimized the dominating thickness contrast.",
+            "section_image": []
+        },
+        {
+            "section_name": "Data availability",
+            "section_text": "All data are available in the main text or the supplementary materials. The raw data can be made available upon request.",
+            "section_image": []
+        },
+        {
+            "section_name": "Code availability",
+            "section_text": "The code can be made available upon request.",
+            "section_image": []
+        },
+        {
+            "section_name": "References",
+            "section_text": "Sowa, W. Synthesis of L-glucurone. Conversion of D-glucose into L-glucose. Can. J. Chem. 47, 3931\u20133934 (1969).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nNambu, Y. Spontaneous symmetry breaking in particle physics: a case of cross fertilization. Int. J. Mod. Phys. A 24, 2371\u20132377 (2009).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSpaldin, N. A., Fiebig, M. & Mostovoy, M. The toroidal moment in condensed-matter physics and its relation to the magnetoelectric effect. J. Phys. Condens. Matter 20, 434203 (2008).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nNakata, M., Shao, R.-F., Maclennan, J. E., Weissflog, W. & Clark, N. A. Electric-field-induced chirality flipping in smectic liquid crystals: the role of anisotropic viscosity. Phys. Rev. Lett. 96, 067802 (2006).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nChen, G. et al. Unlocking Bloch-type chirality in ultrathin magnets through uniaxial strain. Nat. Commun. 6, 6598 (2015).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHoribe, Y. et al. Color theorems, chiral domain topology, and magnetic properties of FexTaS2. J. Am. Chem. Soc. 136, 8368\u20138373 (2014).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nQian, Q. et al. Chiral molecular intercalation superlattices. Nature 606, 902\u2013908 (2022).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nFischer, P. & Hache, F. Nonlinear optical spectroscopy of chiral molecules. Chirality 17, 421\u2013437 (2005).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMa, W., Xu, L., Wang, L., Xu, C. & Kuang, H. Chirality-based biosensors. Adv. Funct. Mater. 29, 1805512 (2019).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nCherifi-Hertel, S. et al. Non-Ising and chiral ferroelectric domain walls revealed by nonlinear optical microscopy. Nat. Commun. 8, 15768 (2017).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLee, S. et al. Single ferroelectric and chiral magnetic domain of single-crystalline BiFeO3 in an electric field. Phys. Rev. B Condens. Matter Mater. Phys. 78, 100101 (2008).\n\nLiu, D. et al. Phase-field simulations of vortex chirality manipulation in ferroelectric thin films. npj Quantum Mater. 7, 34 (2022).\n\nHu, Z.-B. et al. An effective strategy of introducing chirality to achieve multifunctionality in rare-earth double perovskite ferroelectrics. Small Methods 6, e2200421 (2022).\n\nArticle\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWang, Y. J. et al. Polar meron lattice in strained oxide ferroelectrics. Nat. Mater. 19, 881\u2013886 (2020).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLi, X. et al. Atomic-scale observations of electrical and mechanical manipulation of topological polar flux closure. Proc. Natl Acad. Sci. USA 117, 18954 (2020).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nTang, Y. L. et al. Ferroelectrics. Observation of a periodic array of flux-closure quadrants in strained ferroelectric PbTiO3 films. Science 348, 547\u2013551 (2015).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nYadav, A. K. et al. Observation of polar vortices in oxide superlattices. Nature 530, 198\u2013201 (2016).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nZhang, Q. et al. Nanoscale bubble domains and topological transitions in ultrathin ferroelectric films. Adv. Mater. 29, 1702375 (2017).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nZhang, Q. et al. Deterministic switching of ferroelectric bubble nanodomains. Adv. Funct. Mater. 29, 1808573 (2019).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nAbid, A. Y. et al. Creating polar antivortex in PbTiO3/SrTiO3 superlattice. Nat. Commun. 12, 2054 (2021).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLi, M. et al. Electric-field control of the nucleation and motion of isolated three-fold polar vertices. Nat. Commun. 13, 6340 (2022).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHadjimichael, M. et al. Metal-ferroelectric supercrystals with periodically curved metallic layers. Nat. Mater. 20, 495\u2013502 (2021).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nStoica, V. A. et al. Optical creation of a supercrystal with three-dimensional nanoscale periodicity. Nat. Mater. 18, 377\u2013383 (2019).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nDas, S. et al. Observation of room-temperature polar skyrmions. Nature 568, 368\u2013372 (2019).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nShafer, Padraic et al. Emergent chirality in the electric polarization texture of titanate superlattices. Proc. Natl Acad. Sci. USA 115, 915\u2013920 (2018).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSusarla, S. et al. Atomic scale crystal field mapping of polar vortices in oxide superlattices. Nat. Commun. 12, 6273 (2021).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBehera, P. et al. Electric field control of chirality. Sci. Adv. 8, eabj8030 (2022).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLouis, L., Kornev, I., Geneste, G., Dkhil, B. & Bellaiche, L. Novel complex phenomena in ferroelectric nanocomposites. J. Phys. Condens. Matter 24, 402201 (2012).\n\nArticle\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nChen, P. et al. Electrically driven motion, destruction, and chirality change of polar vortices in oxide superlattices. Sci. China Phys. Mech. Astron. 65, 237011 (2022).\n\nNguyen, K. X. et al. Transferring orbital angular momentum to an electron beam reveals toroidal and chiral order. Phys Rev. B. 107, 20519 (2023)\n\nOphus, C. Four-dimensional scanning transmission electron microscopy (4D-STEM): from scanning nanodiffraction to ptychography and beyond. Microsc. Microanal. 25, 563\u2013582 (2019).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nDeb, P. et al. Imaging polarity in two dimensional materials by breaking Friedel\u2019s law. Ultramicroscopy 215, 113019 (2020).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nShao, Y.-T. et al. Emergent chirality in a polar meron to skyrmion phase transition, Nat. Commun. 14, 1355 (2023).\n\nSavitzky, B. H. et al. Py4DSTEM: Open source software for 4D-STEM data analysis. Microsc. Microanal. 25, 124\u2013125 (2019).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nZeltmann, S. E. et al. Patterned probes for high precision 4D-STEM bragg measurements. Ultramicroscopy 209, 112890 (2020).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGao, P. et al. Atomic mechanism of polarization-controlled surface reconstruction in ferroelectric thin films. Nat. Commun. 7, 11318 (2016).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMoffatt, H. K. & Ricca, R. L. Helicity and the C\u0103lug\u0103reanu invariant. Proc., Math. Phys. Sci. 439, 411\u2013429 (1992).\n\nMATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMoffatt, H. K. Helicity and singular structures in fluid dynamics. Proc. Natl Acad. Sci. USA 111, 3663\u20133670 (2014).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHoltz, M. E. et al. Topological defects in hexagonal manganites: inner structure and emergent electrostatics. Nano Lett. 17, 5883\u20135890 (2017).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBalke, N. et al. Enhanced electric conductivity at ferroelectric vortex cores in BiFeO3. Nat. Phys. 8, 81\u201388 (2012).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nZeltmann, S. E. et al. Uncovering polar vortex structures by inversion of multiple scattering with a stacked Bloch wave model. Ultramicroscopy 250, 113732 (2023).\n\nKarpov, D. et al. Nanoscale topological defects and improper ferroelectric domains in multiferroic barium hexaferrite nanocrystals. Phys. Rev. B. 100, 054432 (2019).\n\nOphus, C., Ciston, J. & Nelson, C. T. Correcting nonlinear drift distortion of scanning probe and scanning transmission electron microscopies from image pairs with orthogonal scan directions. Ultramicroscopy 162, 1\u20139 (2016).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nNord, M., Erik Vullum, P., MacLaren, I., Tybell, T. & Holmestad, R. Atomap - automated analysis of atomic resolution STEM images. Microsc. Microanal. 23, 426\u2013427 (2017).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nDownload references",
+            "section_image": []
+        },
+        {
+            "section_name": "Acknowledgements",
+            "section_text": "All the electron microscopy experiments were carried out at the National Center for Electron Microscopy (NCEM) located in the Molecular Foundry user facility at Lawrence Berkeley National Laboratory. Work at the Molecular Foundry was supported by the Office of Science, Office of Basic Energy Sciences, of the U.S. Department of Energy under Contract No. DE-AC02-05CH11231. S.D. acknowledges Science and Engineering Research Board (SRG/2022/000058) and Indian Institute of Science start-up grant for financial support. S.S., P.B., S.L., and R.R. are supported by the DOE Office of Science, Basic Energy Sciences, Materials Sciences, and Engineering Division under contract DE-AC02-05-CH11231 within the Quantum Materials program (KC2202). C.O. acknowledges support from a DOE Early Career Research Award. F.G.-O., P.G.-F., and J.J. acknowledge financial support from Grant No. PGC2018-096955-B-C41 funded by MCIN/AEI/10.13039/501100011033 and by ERDF \u201cA way of making Europe,\u201d by the European Union. F.G.-O. acknowledges financial support from Grant No. FPU18/04661 funded by MCIN/AEI/10.13039/501100011033. B.H.S. was supported by the Toyota Research Institute.",
+            "section_image": []
+        },
+        {
+            "section_name": "Author information",
+            "section_text": "Sandhya Susarla\n\nPresent address: School for Engineering of Matter, Transport, and Energy, Arizona State University, Tempe, 85280, AZ, USA\n\nThese authors contributed equally: Sandhya Susarla, Shanglin Hsu.\n\nNational Center for Electron Microscopy, Molecular Foundry, Lawrence Berkeley National Laboratory, Berkeley, 94720, CA, USA\n\nSandhya Susarla,\u00a0Shanglin Hsu,\u00a0Benjamin H. Savitzky,\u00a0Peter Ercius,\u00a0Ramamoorthy Ramesh\u00a0&\u00a0Colin Ophus\n\nMaterials Sciences Division, Lawrence Berkeley Laboratory, Berkeley, 94720, CA, USA\n\nSandhya Susarla,\u00a0Shanglin Hsu\u00a0&\u00a0Ramamoorthy Ramesh\n\nDepartmento de Ciencias de la Tierra y F\u00edsica de la Materia Condensada, Universidad de Cantabria, Cantabria Campus Internacional Santander, Santander, 39005, Spain\n\nFernando G\u00f3mez-Ortiz,\u00a0Pablo Garc\u00eda-Fern\u00e1ndez\u00a0&\u00a0Javier Junquera\n\nMaterials Research Centre, Indian Institute of Science, Bangalore, 560012, Karnataka, India\n\nSujit Das\n\nDepartment of Materials Science & Engineering, University of California, Berkeley, 94720, CA, USA\n\nPiush Behera\n\nDepartment of Physics, University of California, Berkeley, Berkeley, 94720, CA, USA\n\nRamamoorthy Ramesh\n\nDepartment of Physics, Rice University, Houston, 77005, TX, USA\n\nRamamoorthy Ramesh\n\nDepartment of Materials Science and Nanoengineering, Houston, 77005, TX, USA\n\nRamamoorthy Ramesh\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nS.S., R.R., and C.O. conceived the idea, and designed the experiments. S.S. analyzed the 4D STEM datasets, made the figures and wrote the initial draft of the manuscript. S.L.H. performed the 4D STEM experiments. B.H.S. provided inputs for the scripts of the 4D STEM analysis. F.G.O., P.G.F., and J.J. helped in providing intellectual inputs regarding the origin of chiral domain walls and triple points. S.D. and P.B.\u00a0conducted thin film fabrication and preliminary\u00a0structural characterization. P.B. and P.E. participated in revising the manuscript. R.R. and C.O. supervised the project.\n\nCorrespondence to\n                Sandhya Susarla, Ramamoorthy Ramesh or Colin Ophus.",
+            "section_image": []
+        },
+        {
+            "section_name": "Ethics declarations",
+            "section_text": "The authors declare no competing interests.",
+            "section_image": []
+        },
+        {
+            "section_name": "Peer review",
+            "section_text": "Nature Communications thanks the anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.",
+            "section_image": []
+        },
+        {
+            "section_name": "Additional information",
+            "section_text": "Publisher\u2019s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.",
+            "section_image": []
+        },
+        {
+            "section_name": "Rights and permissions",
+            "section_text": "Open Access  This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article\u2019s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article\u2019s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.\n\nReprints and permissions",
+            "section_image": []
+        },
+        {
+            "section_name": "About this article",
+            "section_text": "Susarla, S., Hsu, S., G\u00f3mez-Ortiz, F. et al. The emergence of three-dimensional chiral domain walls in polar vortices.\n                    Nat Commun 14, 4465 (2023). https://doi.org/10.1038/s41467-023-40009-2\n\nDownload citation\n\nReceived: 04 February 2023\n\nAccepted: 07 July 2023\n\nPublished: 25 July 2023\n\nVersion of record: 25 July 2023\n\nDOI: https://doi.org/10.1038/s41467-023-40009-2\n\nAnyone you share the following link with will be able to read this content:\n\nSorry, a shareable link is not currently available for this article.\n\n\n\n\n                            Provided by the Springer Nature SharedIt content-sharing initiative\n                        ",
+            "section_image": [
+                "https://data:image/svg+xml;base64,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"
+            ]
+        }
+    ]
+}

022065845b0149dd1bde836e14da3448f70c4603767a53533c6448baf134c8c3/metadata.json

@@ -0,0 +1,153 @@
+{
+    "title": "Crowding results from optimal integration of visual targets with contextual information",
+    "pre_title": "Crowding results from optimal integration of visual targets with contextual information",
+    "journal": "Nature Communications",
+    "published": "30 September 2022",
+    "supplementary_0": [
+        {
+            "label": "Peer Review File",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-022-33508-1/MediaObjects/41467_2022_33508_MOESM1_ESM.pdf"
+        },
+        {
+            "label": "Reporting Summary",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-022-33508-1/MediaObjects/41467_2022_33508_MOESM2_ESM.pdf"
+        }
+    ],
+    "supplementary_1": [
+        {
+            "label": "Source Data",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-022-33508-1/MediaObjects/41467_2022_33508_MOESM3_ESM.zip"
+        }
+    ],
+    "supplementary_2": NaN,
+    "source_data": [
+        "/articles/s41467-022-33508-1#ref-CR62",
+        "/articles/s41467-022-33508-1#Sec14"
+    ],
+    "code": [
+        "https://doi.org/10.5281/zenodo.6460723",
+        "/articles/s41467-022-33508-1#ref-CR62"
+    ],
+    "subject": [
+        "Decision",
+        "Object vision"
+    ],
+    "license": "http://creativecommons.org/licenses/by/4.0/",
+    "preprint_pdf": "https://www.researchsquare.com/article/rs-1296243/v1.pdf?c=1664622479000",
+    "research_square_link": "https://www.researchsquare.com//article/rs-1296243/v1",
+    "nature_pdf": "https://www.nature.com/articles/s41467-022-33508-1.pdf",
+    "preprint_posted": "01 Mar, 2022",
+    "research_square_content": [
+        {
+            "section_name": "Abstract",
+            "section_text": "Crowding is the inability to recognize peripheral objects in clutter, usually considered a fundamental low-level bottleneck to object recognition. Here we advance and test an alternative hypothesis, that crowding, like \u201cserial dependence\u201d, results from optimizing strategies that exploit redundancies in natural scenes. This notion leads to several testable predictions: (1) crowding should be greatest for unreliable targets and reliable flankers; (2) crowding-induced biases should be maximal when target and flankers have similar orientations, falling off for differences around 20\u00b0; (3) flanker interference should be associated with higher precision in orientation judgements, leading to lower overall error rate; (4) effects should be maximal when the orientation of the target is near that of the average of the flankers, rather than to that of individual flankers. All these effects were verified, and well simulated with ideal-observer models that maximize performance. The results suggest that while crowding can impact strongly on object recognition, it is best understood not as a processing bottleneck, but as a consequence of efficient exploitation of the spatial redundancies of the natural world.",
+            "section_image": []
+        },
+        {
+            "section_name": "Additional Declarations",
+            "section_text": "There is NO Competing Interest.",
+            "section_image": []
+        }
+    ],
+    "nature_content": [
+        {
+            "section_name": "Abstract",
+            "section_text": "Crowding is the inability to recognize an object in clutter, usually considered a fundamental low-level bottleneck to object recognition. Here we advance and test an alternative idea, that crowding, like predictive phenomena such as serial dependence, results from optimizing strategies that exploit redundancies in natural scenes. This notion leads to several testable predictions: crowding should be greatest for unreliable targets and reliable flankers; crowding-induced biases should be maximal when target and flankers have similar orientations, falling off for differences around 20\u00b0; flanker interference should be associated with higher precision in orientation judgements, leading to lower overall error rate; effects should be maximal when the orientation of the target is near that of the average of the flankers, rather than to that of individual flankers. Each of these predictions were supported, and could be simulated with ideal-observer models that maximize performance. The results suggest that while crowding can affect object recognition, it may be better understood not as a processing bottleneck, but as a consequence of efficient exploitation of the spatial redundancies of the natural world.",
+            "section_image": []
+        },
+        {
+            "section_name": "Introduction",
+            "section_text": "Crowding is the inability to recognize and identify objects in clutter, despite their being clearly visible, and recognizable when presented in isolation1 (see examples in Fig.\u00a01a). It is particularly elevated in the periphery, impacts on many important daily tasks, such as face recognition and reading (for reviews see2,3,4), to the extent it has been considered a major bottleneck to object recognition.\n\na Crowding is a visual phenomenon where items that can be easily identified in isolation are not identifiable if surrounded by similar items. The P and hand symbol on the right are difficult to recognize while fixating the central red dots. b Stimuli employed in this experiment. Observers judged the orientation of a peripheral target (the central oval), which was flanked above and below by oval stimuli. Two conditions were tested: a rounded target with elongated flankers (Low reliability target, high reliability flankers, blue at left) or an elongated target with rounded ovals (red at right). In the main condition the centre-to-centre distance of flankers and targets was 5.5\u2009deg, and eccentricity 26 degrees, leading to a Bouma ratio of 0.21.\n\nThere are several diagnostic criteria for crowding, the most important being that it scales linearly with eccentricity, such that the minimal spacing between centres of targets and flanking elements supporting uncrowded vision is equal to roughly half the target eccentricity (Bouma\u2019s law5). Another is that flankers similar (in colour, shape or orientation) to the target crowd more effectively than dissimilar ones6,7,8,9. Crowding is stronger in the upper than the lower visual field10, and for radial than for tangential flankers11.\n\nMost popular current models of crowding involve some form of compulsory pooling (or substitution) of targets with flankers. For example, Parkes and colleagues12 showed that while the orientation of a Gabor patch cannot be determined when embedded in flankers, it does influence the perceived orientation of the ensemble: hence it is merged with the flankers, rather than suppressed. This is reinforced by several studies showing that the targets can take on characteristics of the flanker stimuli13,14,15. The compulsory integration could occur in higher cortical areas, such as V42,16,17 or V218,19, which have large receptive fields, appropriately sized to account for Bouma\u2019s law.\n\nHowever, compulsory integration is vague and does not explain all the known facts about crowding. For example, flankers that are similar in size, colour or orientation cause more crowding than dissimilar ones9,20,21. More difficult to explain are the recent demonstrations of Herzog and colleagues22 of \u201cuncrowding\u201d, where the addition of extra flanking stimuli around the flankers can reduce drastically their crowding effect, particularly if the extra flankers group with the original flankers to form coherent objects. These data do not fit easily with compulsory integration, even with appropriate linear filtering, which could in principle account for other effects, such as orientation or size selectivity.\n\nCrowding has been studied for decades, and usually considered to be a defect in the system, \u201can essential bottleneck to object perception\u201d23. Certainly, it impacts heavily on object recognition in tasks like letter or face recognition: but is it possible that it may reflect processes that are in principle advantageous to perception? Perception is strongly affected by contextual information, particularly temporal context, where recent and longer term perceptual history has been shown to exert a major influence on current perception24,25,26,27. While the role of context and experience has been appreciated for some time28,29, it has become particularly topical in recent years within the framework of Bayesian analysis. This approach has revealed an interesting phenomenon termed \u201cserial dependence\u201d, where the appearance of many important attributes of a stimulus (including orientation, numerosity, facial identity, beauty etc) are biased towards previously viewed stimuli25,26,30,31. Counterintuitively, these consistent biases in perception have been shown to reflect an efficient perceptual strategy, exploiting temporal redundancies in natural viewing to reduce overall reproduction errors, despite the biases26,32,33.\n\nCould crowding also be a consequence of efficient integration processes that exploit spatial (rather than temporal) redundancies to improve performance? We investigate this possibility by studying crowding with a paradigm similar to one used in serial dependence studies. If, like serial dependence, crowding is a by-product of efficient redundancy-reducing mechanisms, it should display several specific signature characteristics. One is that crowding-induced biases should be stronger for targets that are unreliably perceived, and for flankers that are reliably perceived. In addition, crowding should follow the signature pattern seen in serial dependence, highest when the orientations of target and flankers are similar, then steadily falling off. We provide evidence for these characteristics qualitatively and quantitatively, and show that crowding, while leading to biases, also improves overall performance. The results fit well with models simulating intelligent combination of signals from a small receptive field centred on the target with signals from a much larger integration region, following the same rules that govern serial dependence. On this view crowding is not merely a defect, or bottleneck, in the system, but the unavoidable consequence of efficient exploitation of spatial redundancies of the natural world.",
+            "section_image": [
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-022-33508-1/MediaObjects/41467_2022_33508_Fig1_HTML.png"
+            ]
+        },
+        {
+            "section_name": "Results",
+            "section_text": "To test if visual crowding follows the rules of optimal integration, which well describe serial dependence25,33, we measured crowding with an orientation reproduction task. Participants reproduced the orientation of oval stimuli, which were either elongated (aspect ratio 1: 2.8) or rounded (1: 1.4). Targets were presented 26\u00b0 to the right of fixation, and vertically flanked by similar oval stimuli, elongated if the target was rounded, and vice versa (see Fig.\u00a01b). The orientation of the target was either 35\u00b0 or 55\u00b0 (at random). The orientations of the two flankers were yoked together, and varied randomly over a range of \u2009\u00b1\u200945\u00b0 from target orientation. The clear prediction from models of efficient integration32,34 (see Eq.\u00a010 & 15) is that the effects of crowding will be stronger for the unreliable targets and reliable flankers than vice versa, which we test with rounded targets and elongated flankers. The reasons are explained formally in the modelling section, but the intuition is that the rounded stimuli have less reliable orientation signals and therefore benefit more from integration with contextual information, especially if it is reliable.\n\nFigure\u00a02a shows the bias in target reproduction as a function of difference in flanker orientation. Both sets of stimuli show positive, assimilative effects of the flankers, with positive flanker orientation causing positive biases and negative flankers negative bias. The rounded targets show the strongest contextual effects of crowding, with peak biases varying by up to \u2009\u00b1\u20095.1\u00b0, compared with\u2009\u00b1\u20091.9\u00b0 for the elongated targets. Furthermore, the pattern of bias follows closely that predicted and observed in serial dependence studies33, varying non-linearly with the difference between target and flanker orientation, increasing to a maximum around \u2009\u00b1\u200920\u00b0, then decreasing. These data are well fit by derivative of gaussian functions (Eq.\u00a015, light-coloured lines), commonly used in serial dependence studies25, and expected from a causal inference model (see modelling section35). The dark lines show the predictions of another Bayesian model (Eq.\u00a010), which has also proven successful with serial dependence data26,33. While the models are detailed later, it is worth noting that they are almost entirely anchored by data, down to a simple scaling factor, suggesting that the data are consistent with ideal behaviour.\n\na Average response bias (response minus target orientation) as a function of the orientation of two identical flankers. Low reliability (rounded) targets in blue, high reliability (elongated) in red. Positive biases refer to clockwise response errors, and positive orientation differences indicate that flankers are clockwise with respect to the target (so the biases are assimilative). Error bars show \u00b1\u20091 SEM. N\u2009=\u200910 observers. Dark lines show predictions from an ideal-observer Bayesian model which scales the action of flankers according to their reliability and orientation difference (Eq.\u00a010 of model section). Light blue and red curves show predictions for the causal inference model that doses flanker and target information according to their reliability and the probability of originating from a common cause (Eq.\u00a015 of model section). b Response standard deviation as a function of the orientation of two identical flankers, together with model predictions. Error bars show \u00b1\u20091 SEM. N\u2009=\u200910 observers. Colour coding as in A. Isolated squares and diamonds show results of a control experiment measuring orthogonal flankers (ort\u2014diamonds) and an unflanked baseline (B\u2014squares). N\u2009=\u20098 observers. c Total response error parsed as response standard deviation plotted against bias errors for the two conditions. Dashed circles indicate regions with identical RMS Error, given by the Pythagorean sum of the two types of error. RMSE varies with orientation, and is least around 0\u00b0, when target and flankers coincide. Source data are provided as a Source Data file.\n\nAnother important prediction is that the contextual effects should improve performance. On a reproduction task of this sort, errors can be broadly divided into two orthogonal categories, average accuracy (inverse bias) and precision (inverse scatter about the mean value). Figure\u00a02a reports average bias (inaccuracy), while Fig.\u00a02b plots reproduction scatter (imprecision, given by root-variance of reproduction trials), as a function of orientation difference. As expected, at all orientation differences, scatter is lower for the elongated than the rounded targets. However, for both targets, particularly the rounded targets, the scatter decreased as the difference between target and flanker orientation decreased to be minimal when the test and flankers had matched orientation (the condition that produces maximal crowding).\n\nTo reinforce the idea that performance is at its best when flanker and target are identical, we ran a separate experiment to test two new crucial conditions in the elongated target condition: orthogonal flankers and an unflanked baseline (open diamonds and squares in Fig.\u00a02b). Scatter for these two conditions was very similar, and more than when the flankers were present. This shows that the flankers actually improved precision, even compared with the isolated target condition.\n\nFigure\u00a02c plots standard deviation and bias on a two-dimensional plot, with points connected to follow the change in orientation. On this plot, total root-mean squared error is given by the Pythagorean sum of scatter and bias, the radial distance from the origin. For the points with flanker orientation most distant from the target (near\u2009\u00b1\u200945\u00b0), the total error is around 15\u00b0. Between these extremes, total error falls off, despite the constant bias. When the flankers and targets have similar orientations, the error falls to around 11\u00b0, evidence that \u201ccrowding\u201d improves overall performance, by this measure.\n\nIf the effects shown in Fig.\u00a02 represent visual crowding, they should depend on critical spacing between target and flankers, and follow Bouma\u2019s law1. We therefore measured the effects as a function of target-flanker spacing, for 5 participants. Figure\u00a03 shows the data for the rounded targets with elongated flankers (which show the strongest effects). For the two smallest spacings (5.5 and 7.5\u2009deg), bias showed the characteristic S-shaped dependency on the orientation of the flankers. For the larger spacings (11.0 and 16.6\u2009deg), however, the effect was much reduced and even inverted at 11\u2009deg. As before, the curves are fit by a derivative of gaussian function (Eq.\u00a018), which is the product of a linear regression (illustrated by dashed line in Fig.\u00a03a) and a gaussian. The best fitting slope of this regression is an estimate of the weight given to the flankers when judging orientation. Figure\u00a03b plots the fitted weight as a function of target-flanker spacing (lower abscissa), with the upper abscissa showing the normalized target-flanker distance, the distance between target and flanker centres divided by the eccentricity (26\u2009deg). The weight drops from 0.5 to 0 for normalized target-flanker distances between 0.3 and 0.4, broadly in line with the literature, suggesting that the effects observed here relate to crowding.\n\na Response bias as function of flanker orientation for various target-flanker distances leading to four different normalized target-flanker distances (distance between flanker and target centres divided by eccentricity). Data are obtained from N\u2009=\u20095 observers and are fit with a derivative of gaussian function with free parameters (Eq.\u00a018). b Weight of the flankers (maximal slope of the curves in panel a) as a function of the normalized target-flanker distance (colour-code as before). Negative weights imply a repulsive, rather than attractive effect of flankers on target. Error bars show \u00b1\u20091 SEM. Source data are provided as a Source Data file.\n\nThe results so far show that integration is not obligatory, but depends on the reliability of both target and flankers. They are also in line with previous studies showing that effects are maximal when targets are most similar to flankers. A remaining question is how the flankers integrate with the target: each separately, or after combination with each other. Figure\u00a04 illustrates two possibilities (see also modelling section). One is a feedforward model where the target integrates independently with low-level, high-resolution neural representations of each of the flankers. The other depicts integration with a broader representation including both flankers, potentially implemented through recurrent feedback.\n\na Rationale for investigating the site of action of the flankers. They could either act independently on the target (illustrated by purple arrows in top left panel), or first pooled within a larger RF, which in turn biases the target (illustrated by the large yellow circle and arrow in bottom left panel). b Predictions for the two hypotheses. If the flankers act independently, when one is locked at +\u200915\u00b0 and the other free to vary, the pattern should be similar to that of the main experiment (centre close to 0\u00b0), but raised because of the action of the locked flanker (purple curve). If flankers are first integrated at a more global stage, maximal effect is expected when all the elements in the larger operator average 0\u00b0. Since one of the flankers is locked at +\u200915\u00b0, this occurs when the other flanker is \u221215\u00b0, leading to a leftward shift of the curve of the main experiment (yellow curve).\n\nTo distinguish between these two plausible possibilities, we measured target bias with the orientation of the two flankers varying independently. Specifically, one flanker (randomly top or bottom) was always oriented +\u200915\u00b0 from the target, while the other varied randomly over the range. The logic is that the gaussian function windowing the contextual effect should be centred where the orientations of target and context coincide. If the integration occurs directly between the target and individual flankers, then the maximum effects should occur when the variable flanker coincides with the target; on the other hand, if the integration is with a broader representation including both flankers, maximum integration should occur when the flanker mean is zero, which occurs when the variable flanker is \u221215\u00b0. These predictions are illustrated in Fig.\u00a04b: note that the individual flanker effect also predicts the curve to be higher at all flanker orientations, as the fixed flanker will exert a constant effect at all orientations of the variable flanker.\n\nThe results for the rounded targets with elongated flankers are shown Fig.\u00a05a. The biases clearly follow the signature pattern, well fit by a derivative of gaussian function. The centre of the function is \u221210.8\u00b0, closer to the \u221215\u00b0 predicted by integration with the average orientation of the flankers, than to 0\u00b0 predicted by the individual flanker model. The mean height of the function is 0.5\u00b0, close to that observed in the previous experiment (\u22120.9\u00b0), while the individual-flanker integration model predicts a constant average bias 4.7\u00b0. Figure\u00a05b shows the scatter for this experiment, which was reduced over the region of bias, well described by an inverted Gaussian with center at \u221214.26\u00b0, again close to the \u221215\u00b0 predicted by the average orientation of the flankers.\n\na Biasing errors as function of a single flanker orientation, while the other flanker was locked at +\u200915\u00b0. Error bars show \u00b11 SEM. Colours and conventions as for Fig.\u00a02. N\u2009=\u200913 observers. Thick dark lines refer to the ideal observer model (Eq.\u00a010), thick light blue lines to the causal inference model (Eq.\u00a015). Thin dashed lines show best-fitting derivative of gaussian, with all parameters free to vary (Eq.\u00a018). b Response standard deviation as a function of the variable flanker orientation. Conventions as in panel a. c Histogram of the centres of the gaussian derivative for 1000 bootstrap fits. Source data are provided as a Source Data file.\n\nTo test significance, we bootstrapped the data 1000 times (sampling with replacement) and measured the centre of the gaussian derivative on each iteration. The results plotted in the histogram of Fig.\u00a05c show that on only 16 out of 1000 iterations (1.6%) was the centre closer to 0\u00b0 (individual flanker prediction) than to \u221215\u00b0 (joint-flanker prediction). This leads to a likelihood ratio (Bayes factor) of 984/16\u2009=\u200961.5, strong evidence in favour of the joint-flanker-integration model.\n\nWe propose two plausible models to explain the pattern of data. Both are motivated by principles of \u201coptimal cue integration\u201d commonly used in multi-sensory perception34,36, which predict optimal combination of information from multiple sources after appropriate weighting to minimize overall root-mean-square error. The first is based on an ideal-observer model successfully used to model serial dependence26, the second on a \u201ccausal-inference\u201d model of multi-sensory integration35. Both models predict the data well.\n\nThe ideal observer model selects the appropriate weight to assign to the flankers in order to minimize the total error in the reproduction task26.\n\nTotal RMS error (E) can be decomposed into bias (B) and precision (scatter standard deviation: S), whose squares sum to give total squared error:\n\nThe ideal responses (R) in a pooling model can be expressed as a linear weighted combination of internal representation of target (T) and flankers (F1 and F2), each weighted by \\({w}_{1}\\)and \\({w}_{2}\\).\n\nAs the two flankers of this study had the same aspect ratio they should be weighted equally, (\\({w}_{1}={w}_{2}=w\\)), so Eq.\u00a02 simplifies to:\n\nThe mean of the responses (\\({\\mu }_{R}\\)) is a simple linear combination of the means of flankers and target (\\({\\mu }_{1}\\), \\({\\mu }_{2}\\) and \\({\\mu }_{T}\\)).\n\nBias is the difference between the mean estimated response (\\({\\mu }_{R}\\)) and real orientation, \\({x}_{T}\\); \\(B={\\mu }_{R}-{x}_{T}\\). Using Eq.\u00a04 and considering that the average target representation (\\({\\mu }_{T}\\)) should be unbiased and coincide with target (\\({\\mu }_{T}\\,{={x}}_{T}\\)) it follows that:\n\nThe term \\({\\mu }_{1}+{\\mu }_{2}-{2\\mu }_{T}\\)can be rearranged as \\(2(({\\,\\mu }_{1}+{\\mu }_{2})/2-{\\mu }_{T})\\) which is twice the distance between the average of the flanker representations, \\(({\\mu }_{1}+{\\mu }_{2})/2\\), and the target representation \\({\\mu }_{T}\\). For convenience we define:\n\nso that Eq.\u00a05 becomes:\n\nVariance of the linear combination of the flankers and target is itself a linear combination of the flanker and target variances (\\({\\sigma }_{F}^{2}\\) and \\({\\sigma }_{T}^{2}\\)) with the squared coefficients\n\nFrom Eqs.\u00a01, 7 and 8 it follows that \\({RMSE}\\) can be written as:\n\nSince RMSE is a function of second order of \\(w\\), it is minimized when \\(w=\\frac{-b}{2a}\\), so the optimal weight (\\({w}_{{opt}}\\)) is obtained at:\n\nThis equation has much in common with all Bayesian-like integrations used in multi-sensory research and serial dependence: the weight depends directly on target variance \\({\\sigma }_{T}^{2}\\), so targets of low reliability (inverse variance) benefit more from integration, resulting in higher weighting to the flankers. Increase in flanker variance (\\({\\sigma }_{F}^{2}\\)) has the opposite effect.\n\nThe term \\({2d}^{2}\\) is fundamental for the signature function found in serial dependence literature, as the weight of the flankers will decrease with angular difference between target and average flanker orientation. This ensures that contextual cues are used only if they are plausibly similar to the target26,32,33. Importantly, the point that will ensure maximal weight of the flankers is when the target coincides with the average of the flankers (i.e. \\({d}^{2}=0\\)).\n\nEquations\u00a03 and 10 define the optimal dependence of flankers. However, the flankers are not the only contextual information that may influence responses. Two obvious (and related) examples are serial dependence and \u201cregression to the mean\u201d, both leading to a tendency to underestimate 55\u00b0 and overestimate 35\u00b0. Given that the average reproduction of the two angles was 50.5\u00b0 and 37.2\u00b0 respectively, we assume that the regression to the mean (\\({{{{{\\rm{\\rho }}}}}}\\)) was a factor of 0.67 (calculated as (50.5\u221237.2)/20). Thus the final estimate of response R needs to be multiplied by this factor, and a constant weighted bias summed, pulling the responses towards the average (45\u00b0). We also include an additional scaling factor (\\(\\alpha\\)) comprising any further unspecified influences or sub-optimal behaviour:\n\nAn alternative model prescribes that an optimal blend of information is obtained by maximum likelihood combination of two sources (assuming that the two curves originate from the same cause), multiplied by the probability that the two sources originate from the same cause35. Within this framework maximal interaction between cues occurs when the two sources coincide, where the weight assigned to the two cues is the well known formula used in sensory integration literature34,36 (see also Eq.\u00a010):\n\nThe probability of the two sources originating from a common cause can be calculated using Bayes\u2019 Theorem, as demonstrated in35. Assuming gaussian probability distribution functions (with centres at \\({\\mu }_{A}\\) and \\({\\mu }_{B}\\) and variances \\({\\sigma }_{A}^{2}\\) and \\({\\sigma }_{B}^{2}\\)), the solution is solvable analytically35:\n\nwhere \\({\\mu }_{P}\\) and \\({\\sigma }_{P}^{2}\\)) are the mean and variance of the a-prior likelihood of there being one cause (the prior, also gaussian). If no prior knowledge is available (\\({\\sigma }_{P}^{2}\\to \\infty\\)) Eq.\u00a013 simplifies to\n\nThis is a gaussian peaking when the distributions of the two cues coincide (\\({\\mu }_{A}={\\mu }_{B}\\)) and falling off with a space constant related to the sum of their variances (\\({\\sigma }_{A}^{2}+{\\sigma }_{B}^{2}\\)).\n\nIn the specific case of our experiment we can map the two sources of information to the flanker compound (a gaussian with centre at \\({\\mu }_{C}=\\left({\\mu }_{1}+{\\mu }_{2}\\right)/2\\), variance \\({\\sigma }_{C}^{2}={\\sigma }_{F}^{2}/2\\)) and the target (assumed gaussian with centre \\({\\mu }_{T}\\), and variance \\({\\sigma }_{T}^{2}.\\) Putting together Eqs.\u00a012 and 14, the bias (difference between the response and the target) is given by:\n\nWhich is a derivative of gaussian as a function of average flanker orientation \\({\\mu }_{C}\\).\n\nIt also follows that response scatter is minimized only when the system considers a common cause likely (Eq.\u00a014), predicting U-shaped (gaussian) plots of Figs.\u00a02b and 5b.\n\nAs above, we need to incorporate regression to the mean (\\({{{{{\\rm{\\rho }}}}}}\\) =0.67) and to allow for suboptimal behaviour to this end we introduce two free parameters that regulate the amplitude of the dependency on the flankers (\\(\\beta\\)) and the breadth of the region of interaction (\\(\\gamma\\)) so that the average bias is:\n\nInterestingly, comparable behaviour is obtained if, instead of constructing a system which multiplies probabilities as in35, one considers a system that measures the similarity between two distributions via their point-by-point product of the distributions and takes either the peak or area under the distribution.\n\nThe product of gaussians is itself a gaussian, is centred at (\\(\\frac{{\\mu }_{B}{\\sigma }_{A}^{2}+{{\\mu }_{A}\\sigma }_{B}^{2}}{{\\sigma }_{A}^{2}+{\\sigma }_{B}^{2}}\\)), has variance (\\(\\frac{{\\sigma }_{A}^{2}{\\sigma }_{B}^{2}}{{\\sigma }_{A}^{2}+{\\sigma }_{B}^{2}}\\)) and peak at:\n\nSo the peak embeds the same behaviour of Eq.\u00a014. It is easy to demonstrate that also the area under the curve follows the same gaussian dependency on the distance between cues as the area of a gaussian is equivalent to the peak (Eq.\u00a016) times the standard deviation of the curve (\\(\\sqrt{\\frac{{\\sigma }_{A}^{2}{\\sigma }_{B}^{2}}{{\\sigma }_{A}^{2}+{\\sigma }_{B}^{2}}}\\)) and a constant factor \\(1/\\sqrt{2\\pi }\\) all of which are constant once the distributions have known width and thus reduce to a scaling factor.\n\nThe predictions of the two modelling approaches are overlayed on the data of Figs.\u00a02 and 5 with dark and light colours. To minimize degrees of freedom we derived the values of sensory reliability from the data of Fig.\u00a02b, assuming that the extreme points (\u00b1\u200930\u00b0 and \u00b1\u200945\u00b0) give baseline data, not influenced by flanker integration: this is 13.2 for rounded targets (blue symbols), and 10.0 for elongated targets (red symbols).\n\nWe implemented the ideal observer model (Eq.\u00a011) with only a scaling constant (\\(\\alpha\\)), which allows for sub-optimal behaviour. These fits are particularly good for the rounded targets (with largest effects), with R2 of 0.93 and 0.71 (for bias and scatter), and 0.37 and 0.74 for elongated targets) and come about assuming \\(\\alpha\\)\u2009=\u20090.7 and 0.5 for the two conditions. One of the key features of the ideal observer model is that it reduces RMSE by leveraging on all available information. Thus it predicts the Global Integration of Fig.\u00a04, with centres of the Gaussian derivatives close to \u221215\u00b0. Besides capturing this key feature, the model also provides good quantitative fits to the data of Fig.\u00a05a with R2 of 0.73 and average fits to those of Fig.\u00a05b 0.76 for bias and scatter respectively (\\(\\alpha\\)=0.57).\n\nWe used the same reliability values from Fig.\u00a02b to implement the \u201coptimal causality gating model\u201d35, the derivative of gaussian function plotted with light colours in Figs.\u00a02 and 5. The sensory reliabilities fix both the maximal slope of the curve (see Eq.\u00a012) and the width of the region of interaction (see Eq.\u00a014). Assuming the same sensory precisions as above (13.2 and 10.0 for the two types of stimuli) maximal slopes should be 0.78 and 0.53 for the two conditions. Considering regression to the mean, which caps the possibility of detecting the weight of the flankers to about 0.7, the model predictions are 0.52 and 0.35, still larger than the real data (0.37 and 0.10). Also the widths (28.4 and 28.6) are larger than those predicted by Eq.\u00a014 (18.8 and 19.2). For this reason we allowed two scaling factors, one enabling lower weighting of the context (\\(\\beta\\)) and the other modulating the width (\\(\\gamma\\)). Setting \\(\\beta\\)\u2009=\u20090.71 and \\(\\gamma\\)\u2009=\u20091.51 led to good fits with R2\u2009=\u20090.97 and 0.75 for the low reliability target (bias and scatter curves), and 0.73 and 0.84 for the high reliability target (\\(\\beta\\)\u2009=\u20090.28 and \\(\\gamma\\)\u2009=\u20091.48). As with the other model, the prediction in Experiment 2 is for large pooling of all available cues, thus the prediction is that of a centre at \u221215\u00b0. This model also provides good fits for response bias (R2\u2009=\u20090.89) and acceptable fits for response scatter (R2\u2009=\u20090.78\\(,\\beta\\)\u2009=\u20090.73 and \\(\\gamma\\)\u2009=\u20090.97).",
+            "section_image": [
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-022-33508-1/MediaObjects/41467_2022_33508_Fig2_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-022-33508-1/MediaObjects/41467_2022_33508_Fig3_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-022-33508-1/MediaObjects/41467_2022_33508_Fig4_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-022-33508-1/MediaObjects/41467_2022_33508_Fig5_HTML.png"
+            ]
+        },
+        {
+            "section_name": "Discussion",
+            "section_text": "The results of this study suggest an alternative interpretation of visual crowding: that it is a by-product of efficient Bayesian processes, which lead in general to improved perceptual performance, minimizing RMS error. We tested and provided evidence supporting several key predictions of this idea. Firstly, crowding, measured as flanker-induced orientation bias, was greatest when targets had the weakest orientation signals (least reliability) and flankers had the strongest, most reliable, signals as predicted from most models of optimal cue combination34,36. The magnitude of the bias varied with the difference of target and flanker orientation, following the predicted non-linear pattern, increasing to a maximum of around 15\u00b0, then falling off for larger orientation differences. Importantly, the interaction of the flankers and target was associated with a reduction in response scatter (increased precision), which led to a reduction in total RMS error, an index of improved performance. Finally, the results suggest that the bias does not result from direct interactions with individual flankers, but from interaction with a representation of the average orientations of the two flankers. All these results were predicted by optimal feature combination principles, and quantitatively well modelled an ideal-observer model that minimizes reproduction errors.\n\nOur results are consistent with previous observations showing that crowding depends on relative strength or saliency of target and flankers, with high contrast targets more immune to crowding, and configurations embedded in noise leading to stronger assimilative effects37,38. Our stimuli were matched for luminance and contrast, but differed in strengths of orientation signals (determined by aspect ratio), leading to different reliabilities, suggesting that the crucial variables are not low level properties but the reliabilities that they afford. Our results also agree with the myriad of experiments showing that similarities in shape (in this case orientation) cause maximum crowding6,21,39,40, and offer an explanation of why.\n\nThese results are clearly difficult to reconcile with simple models of obligatory integration12,41, indicating that these mechanisms are more sophisticated and selective than previously envisaged. Passive integration systems may be tweaked to explain the stronger effects for more elongated flankers (such as having more Fourier energy at that orientation), but cannot easily explain the fall off in crowding effects when the difference exceeds 15\u00b0. Any basic integrator would necessarily combine orientation energy of all angles, not only similar angles. On the other hand, the flexible integrator models proposed here (Eqs.\u00a010 and 15) predict both the pattern and the magnitude of the results. Furthermore, the final experiment suggests that this intelligent orientation-dependent integration is unlikely to occur directly within a higher order cell itself, as the orientation-dependent integration function aligns with the average of two disparate flankers, rather than with each individual flanker. This suggests that the integration is between the target and a broad representation that includes both flankers. Mechanisms operating directly between target and individual flankers, such as the proposed \u201clocal association field\u201d42 between neurones of similar tuning43, are not consistent with the results of Fig.\u00a05, which shows that flankers are first combined with each other before exerting their effects on the target.\n\nCombination of target and a broad representation of both flankers could be implemented in several ways. One physiologically plausible mechanism would be feedback from mid-level areas, such as V218,19 or V42,16,17, which have large receptive fields, integrating over a wide area. These cells could contain information of both flankers (as well as the target), which could be fed back to low levels (e.g. V1) to integrate flexibly with finer representations of the target. Within this framework the fine-grain target information is not lost, but combined with broad contextual information in an optimal manner to improve performance. This is analogous to the process of serial dependence, where higher-level representations of perceptual history (often termed Bayesian priors) are generated at mid- to high-levels of analysis, but feed back onto fairly low processing levels44. Similar processes could evoke crowding, integrating over space rather than time.\n\nInteresting, the spatial selectivity of serial dependence seems to be spatiotopic, in external rather than retinal coordinates25,45. Crowding has also been shown to be spatiotopically selective46,47. Spatiotopic selectivity is a signature of high-level and functionally complex processing, indicating that both crowding and serial dependence involve moderately high levels of analysis.\n\nSimilarity between target and flankers is a major diagnostic criterion of crowding, and the current study suggests a reason for this. The interaction between target and flankers is determined by two main factors: relative reliability (highly related to salience) of target and flankers and, importantly, by target-flanker similarity (\\(d\\) in Eq.\u00a010, \\(\\left({\\mu }_{C}-{\\mu }_{T}\\right)\\) in Eq.\u00a015). This explains why the biasing effects are maximal for similarly oriented flankers, and steadily fall off. Formally, this behaviour derives from theoretical minimization of total RMS errors, explained in detail in the modelling section, but readily understood intuitively. RMS Error comprises two orthogonal factors, accuracy (average bias) and precision (scatter around the mean), whose Pythagorean sum yields total error. Thus, while the contextual effects do lead to inaccuracies (biases), these are more than offset by the increased precision decrease in response scatter (Fig.\u00a02c). Clearly, if the effects were to increase continuously with orientation, then the bias would become large, and offset the reduction in scatter, leading to increased error: integration is therefore efficient only over a limited range. Note that the efficiency-driven ideal model gives good fits simultaneous to both bias and scatter data with only one free parameter, a scaling factor. This comes out at around 0.7 after taking into account other known phenomena of orientation judgements, such as regression to the mean48,49.\n\nOur data and modelling also shed light on why it is usually the overall stimulus configuration promotes crowding, rather than individual flankers working in isolation. Regardless of the assumptions leading to the two models (Ideal Observer or Causal Inference), they both benefit from accumulation of relevant information, and hence the context as to be ascertained by pooling both flankers. A direct consequence of this is that it is their combined value that affects the stimulus.\n\nMaximum crowding occurs when flankers and targets are most similar; yet our results also showed that precision is maximal when orientations coincide, seemingly contrary to a vast body of much literature reporting poor performance at that point6,7,21,50,51. However, this apparent discrepancy depends critically on how performance is measured. As mentioned above, our technique measured separately the accuracy and precision. Response scatter (a measure of imprecision) was lowest when the orientations coincided, as predicted by our models. RMSE (comprising both precision and accuracy) was also lowest at this point. Other standard performance measures will not necessarily show this pattern. For example, measures of \u201cpercent correct\u201d will be poor when there is a bias, and will not be improved by high precision (minimal scatter around the incorrect bias). One crowding study that measured separately bias and precision in an orientation task (as we did) found similar results to ours, with precision highest when target and flankers coincide (see Fig.\u00a04a in52). Calculations from their data show that RMSE was also lowest when orientations matched, about half that of when they differed by 45\u00b0. Their data were collected at 3.7\u2009deg eccentricity (compared to our 26\u2009deg), showing that the results reported here generalize to lower eccentricities. Other studies measuring standard deviation of responses do not confirm so closely our results53, but there are differences in the paradigms used (such as using forced choice rather than reproduction techniques and other details of the display sequence).\n\nThe current experiment shows that under conditions of crowding, information about the target is not necessarily lost. This is consistent with a good deal of previous evidence (see reference54 for review), including studies showing that it can affect the ensemble judgment12, can cause adaptation10 and that crowding-induced biases may not affect grasping55. Even more dramatic are the demonstrations that increasing flanker length56 or adding additional flankers22 can decrease or eliminate crowding. Our study employed simple, well controlled stimuli to allow quantitative prediction and measurement of crowding-effects, similar to the studies with serial dependence studies. Thus they do not readily relate to the clever uncrowding studies of Herzog and colleagues. However, it is not difficult to envisage extensions to the model incorporating grouping principles within the rules of integration, in the spirit of the general principles of our model: flexible, \u201cintelligent\u201d combination of signals, rather than a rigid integration via \u201crectify and sum\u201d or similar rules18.\n\nWe are not the first to propose that crowding may be beneficial to vision. Parkes et al.12 suggested that compulsory integration could be a by-product of ensemble perception, the ability to judge some average, global property, such as orientation: what is lost in the individual perception may be gained by a perception of the gist of the ensemble. However, when examined closely, this idea did not hold up, as ensemble perception and crowding follow different psychophysical rules, suggesting they are different processes57. It would be interesting to see whether applying a similar approach to ensemble perception (asking observers to report average rather than target orientation) would help further to explore the commonalities and differences of the two processes.\n\nReaders may find it paradoxical that we are claiming that crowding, which impacts heavily on so many fundamental aspects of daily life, such as face recognition and reading, could be considered in any sense \u201coptimal\u201d or \u201cefficient\u201d. Clearly, optimality depends on what is being optimized. Our models lead to minimization of total root-mean-square error, a criterion used in many branches of engineering and science, and becoming increasingly popular in neuroscience, particular motor and perceptual research34,58,59,60,61. However, minimizing scatter and total error can also lead to misperceptions, or illusions such as the \u201cventriloquist effect\u201d36 or the \u201chollow mask illusion\u201d29. Similarly, minimizing total error may be \u201cideal\u201d for some basic tasks, but can lead to biasing errors that impact strongly on face and letter recognition.\n\nCrowding occurs for many object properties, including contrast, motion and colour8,21,53. Here we have demonstrated clear signatures of optimality in crowding of orientation, a fundamental feature for shape perception, object recognition and reading. However, it is not certain that the effect will generalize to all forms of crowding. It would be interesting to extend the paradigm to other examples, such as colour or motion53, to test whether crowding of these features cause the same form of optimal integration. Similarly, it would be interesting to measure integration under conditions where crowding is minimal, such as in central viewing, to test whether crowding and integration are causally related.\n\nIn summary, the current study suggests that crowding may be analogous to serial dependence, an index of predictive coding-like processes, pointing to similar function and mechanisms. As serial dependence has been shown to exploit temporal redundancies to maximize performance, crowding may also reflect similar exploitation of redundancies over space. It is worth noting that while the rules governing crowding are flexible, leading to improved performance, crowding remains completely obligatory: no effort of will or deployment of attention can allow us to resolve the crowded objects, or to ignore the contextual effects of the orientated flankers. Indeed, while our proposed pooling process is flexible and \u201cintelligent\u201d, it remains automatic, not subject to voluntary control. This is similar to many of the experience-driven perceptual illusions, such as the \u201chollow mask illusion\u201d29: no effort of will can cause us to see the inside of a hollow mask as concave, we always see the convex face. However, while visual crowding remains an obligatory limitation to object recognition, we conclude that like the effects of temporal context and experience, it is best understood not as a defect or bottleneck of the system, but the consequence of efficient exploitation of spatial redundancies of the natural world.",
+            "section_image": []
+        },
+        {
+            "section_name": "Methods",
+            "section_text": "Experimental procedures are in line with the declaration of Helsinki and approved by the local ethics committee (Commissione per l\u2019Etica della Ricerca, University of Florence, n. 111 7 July 2020). Written informed consent was obtained from each participant, which included consent to process, preserve and publish the data in anonymous form.\n\nNineteen participants with normal or corrected-to-normal vision were recruited (aged 18\u201355 years, mean age\u2009=\u200934, 10 females).\n\nThe stimuli, illustrated in Fig.\u00a01a, were generated with Psychtoolbox for MATLAB (R2016b; MathWorks). They comprised an oval-shaped visual target flanked by oval-shaped upper and lower visual flankers, displayed 26\u2009deg eccentric from the fixation point, with the target close to the horizontal meridian (vertical position was slightly varied from trial to trial to avoid pre-allocation of attention to the target) and flankers 5.5\u2009deg away from the target. Both target and flankers were sketches of oval shapes, defined by 12 dark grey dots (diameter 0.3\u2009deg, 1.4\u2009deg inter-dots distant, 16.8\u2009deg perimeter), presented against a uniform grey background. The target was orientated either +\u200935\u00b0 or +\u200955\u00b0 (clockwise) from the vertical, and flanker orientation randomly chosen in steps of 5\u00b0 from \u221245\u00b0 to +\u200945\u00b0 with respect to the target orientation. The two flankers were 5.5\u2009deg from target, leading to a Bouma ratio of 0.2. We manipulated the reliability of orientation information of target and flanker stimuli by using two different aspect ratios, 2.8 (axes 3.48 and 1.23\u2009deg) and 1.4 (axes 3.19 and 2.28\u2009deg), illustrated in Fig.\u00a01a. The more elongated target was always associated with more rounded flankers, and vice versa. In each experimental session of the three experiments, the two target-flanker combinations were shown both kinds of stimuli in random order.\n\nStimuli were displayed on a linearized 22\u201d LCD monitor (resolution 1920\u2009\u00d7\u20091080 pixels, refresh rate 60\u2009Hz). Observers were positioned 57\u2009cm from the monitor, in a quiet room with dim lighting, and maintained fixation on a small (0.35\u2009deg) black central dot. After a random delay from the observer initiating the trial, the stimulus was displayed for 167\u2009ms. Then a thin rotatable white bar (0.05\u2009\u00d7\u20095\u2009deg with a gaussian profile) was presented at the fixation point with random orientation, and observers matched its orientation to that of the target by mouse control. In the first two experiments, the orientation of the two flankers was yoked, while in the third, one flanker was always +15\u00b0 (clockwise) while the other varied from \u221245\u00b0 to +45\u00b0. In the second experiment, the target-flanker distance varied, being 5.5, 7.5, 11.0 and 16.6\u2009deg, leading to a normalized target-flanker distance of 0.21, 0.27, 0.4, 0.6.\n\nTen observers (6 females, mean age\u2009=\u200936) participated in the first experiment, five in the second (3 females, mean age\u2009=\u200940), thirteen in the third (7 females, mean age\u2009=\u200934). They contributed for a total of 10699 trials for the first experiment, 14377 for the second (spread across the four flanker-target distances) and 16574 for the last.\n\nAs a control experiment, aimed at measuring response standard deviation, we repeated the experiment probing two baseline conditions either with 90\u00b0 flankers or unflanked. As in this experiment often there would have been a recognizable unflanked target we increased the number of possible target orientations so that they spanned from (22.5\u00b0 to 67.5\u00b0) either clockwise from vertical or counter clockwise from vertical. Eight observers (5 females, mean age\u2009=\u200935) participated to this extra batch of data, contributing for a total of 889 trials.\n\nResponses occurred out from the range between 0.5 and 3\u2009seconds after the stimulus offset were removed (for a total of 15.9% trials across the 3 experiments), as were responses with reproduction error greater than 35\u00b0 (6.9% of trials).\n\nFor each target and relative orientation of the flanker, we calculated the average constant error (bias, positive meaning clockwise) and scatter (computing residuals separately for each observer and averaging them). We then averaged the values for the two targets. Bias functions were fitted by a derivative of gaussian function, which can be considered to be a gaussian of width s multiplied by a straight line of slope a [or w], which can be considered the weighting given to the flankers: 1 means the flankers are weighted equally to the target. Bias is given by:\n\nWhere \\(\\theta\\) is orientation difference, \\(m\\) the centre, and \\(b\\) the vertical offset of the function. \\(a\\), \\(b\\) and \\(m\\) were free to vary.\n\nScatter (\\(S\\)) was defined as the average root variance in each condition. The variation with orientation a gaussian function in the form:\n\nWhere \\({b}\\) is the baseline at high orientation differences and \\(a\\) is the amplitude of the Gaussian. As Bias and Scatter likely originate from the same process, we yoked the parameter \\(s\\) to best fit both curves.\n\nFurther information on research design is available in the\u00a0Nature Research Reporting Summary linked to this article.",
+            "section_image": []
+        },
+        {
+            "section_name": "Data availability",
+            "section_text": "The processed data needed to evaluate the conclusions in the paper are available as a Source Data file. The raw data used in this study are available in the Zenodo database under accession code (10.5281/zenodo.6460723)62.\u00a0Source data are provided with this paper.",
+            "section_image": []
+        },
+        {
+            "section_name": "Code availability",
+            "section_text": "The MATLAB source codes that were used to generate the datasets and analyse the results are available at a dedicated Zenodo repository (https://doi.org/10.5281/zenodo.6460723)62.",
+            "section_image": []
+        },
+        {
+            "section_name": "References",
+            "section_text": "Bouma, H. Interaction effects in parafoveal letter recognition. Nature 226, 177\u2013178 (1970).\n\nADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nPelli, D. G. & Tillman, K. A. The uncrowded window of object recognition. Nat. Neurosci. 11, 1129\u20131135 (2008).\n\nCAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWhitney, D. & Levi, D. M. Visual crowding: a fundamental limit on conscious perception and object recognition. Trends Cogn. Sci. 15, 160\u2013168 (2011).\n\nPubMed\u00a0\n    PubMed Central\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nStrasburger, H., Rentschler, I. & Juttner, M. Peripheral vision and pattern recognition: a review. J. Vis. 11, 13 (2011).\n\nPubMed\u00a0\n    PubMed Central\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBouma, H. Visual interference in the parafoveal recognition of initial and final letters of words. Vis. Res. 13, 767\u2013782 (1973).\n\nADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nAndriessen, J. J. & Bouma, H. Eccentric vision: Adverse interactions between line segments. Vis. Res. 16, 71\u201378 (1976).\n\nCAS\u00a0\n    PubMed\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGheri, C., Morgan, M. J. & Solomon, J. A. The relationship between search efficiency and crowding. Perception 36, 1779\u20131787 (2007).\n\nPubMed\u00a0\n    PubMed Central\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nPoder, E. & Wagemans, J. Crowding with conjunctions of simple features. J. Vis. 7, 23 21\u201312 (2007).\n\nMATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKennedy, G. J. & Whitaker, D. The chromatic selectivity of visual crowding. J. Vis. 10, 15 (2010).\n\nPubMed\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHe, S., Cavanagh, P. & Intriligator, J. Attentional resolution and the locus of visual awareness. Nature 383, 334\u2013337 (1996).\n\nADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nToet, A. & Levi, D. M. The two-dimensional shape of spatial interaction zones in the parafovea. Vis. Res. 32, 1349\u20131357 (1992).\n\nCAS\u00a0\n    PubMed\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nParkes, L., Lund, J., Angelucci, A., Solomon, J. A. & Morgan, M. Compulsory averaging of crowded orientation signals in human vision. Nat. Neurosci. 4, 739\u2013744 (2001).\n\nCAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBalas, B., Nakano, L. & Rosenholtz, R. A summary-statistic representation in peripheral vision explains visual crowding. J. Vis. 9, 11\u201318 (2009). 13.\n\nMATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLevi, D. M. & Carney, T. Crowding in peripheral vision: why bigger is better. Curr. Biol. 19, 1988\u20131993 (2009).\n\nCAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGreenwood, J. A., Bex, P. J. & Dakin, S. C. Crowding changes appearance. Curr. Biol. 20, 496\u2013501 (2010).\n\nCAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMotter, B. C. & Simoni, D. A. The roles of cortical image separation and size in active visual search performance. J. Vis. 7, 6 1\u201315 (2007).\n\nMATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMotter, B. C. Central V4 receptive fields are scaled by the V1 cortical magnification and correspond to a constant-sized sampling of the V1 surface. J. Neurosci. 29, 5749\u20135757 (2009).\n\nCAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nFreeman, J. & Simoncelli, E. P. Metamers of the ventral stream. Nat. Neurosci. 14, 1195\u20131201 (2011).\n\nCAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHe, D., Wang, Y. & Fang, F. The critical role of V2 population receptive fields in visual orientation crowding. Curr. Biol. 29, 2229\u20132236 e2223 (2019).\n\nCAS\u00a0\n    PubMed\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nNazir, T. A. Effects of lateral masking and spatial precueing on gap-resolution in central and peripheral vision. Vis. Res. 32, 771\u2013777 (1992).\n\nCAS\u00a0\n    PubMed\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKooi, F. L., Toet, A., Tripathy, S. P. & Levi, D. M. The effect of similarity and duration on spatial interaction in peripheral vision. Spat. Vis. 8, 255\u2013279 (1994).\n\nCAS\u00a0\n    PubMed\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nManassi, M., Sayim, B. & Herzog, M. H. When crowding of crowding leads to uncrowding. J. Vis. 13, 10 (2013).\n\nPubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLevi, D. M. Crowding\u2013an essential bottleneck for object recognition: A mini-review. Vis. Res. 48, 635\u2013654 (2008).\n\nADS\u00a0\n    PubMed\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nChopin, A. & Mamassian, P. Predictive properties of visual adaptation. Curr. Biol. 22, 622\u2013626 (2012).\n\nCAS\u00a0\n    PubMed\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nFischer, J. & Whitney, D. Serial dependence in visual perception. Nat. Neurosci. 17, 738\u2013743 (2014).\n\nCAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nCicchini, G. M., Anobile, G. & Burr, D. C. Compressive mapping of number to space reflects dynamic encoding mechanisms, not static logarithmic transform. Proc. Natl Acad. Sci. USA 111, 7867\u20137872 (2014).\n\nADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nPascucci, D. et al. Laws of concatenated perception: Vision goes for novelty, decisions for perseverance. PLoS Biol. 17, e3000144 (2019).\n\nCAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHelmholtz, H. V. Handbuch der physiologischen Optik. (Voss, 1867).\n\nGregory, R. L. Eye and brain; the psychology of seeing. (McGraw-Hill, 1966).\n\nLiberman, A., Fischer, J. & Whitney, D. Serial dependence in the perception of faces. Curr. Biol. 24, 2569\u20132574 (2014).\n\nCAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nTaubert, J., Van der Burg, E. & Alais, D. Love at second sight: Sequential dependence of facial attractiveness in an on-line dating paradigm. Sci. Rep. 6, 22740 (2016).\n\nADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBurr, D. & Cicchini, G. M. Vision: efficient adaptive coding. Curr. Biol. 24, R1096\u2013R1098 (2014).\n\nCAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nCicchini, G. M., Mikellidou, K. & Burr, D. C. The functional role of serial dependence. Proc. Biol. Sci. 285 https://doi.org/10.1098/rspb.2018.1722 (2018).\n\nErnst, M. O. & Banks, M. S. Humans integrate visual and haptic information in a statistically optimal fashion. Nature 415, 429\u2013433 (2002).\n\nADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKording, K. P. et al. Causal inference in multisensory perception. PLoS One 2, e943 (2007).\n\nADS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nAlais, D. & Burr, D. The ventriloquist effect results from near-optimal bimodal integration. Curr. Biol. 14, 257\u2013262 (2004).\n\nCAS\u00a0\n    PubMed\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nFelisberti, F. M., Solomon, J. A. & Morgan, M. J. The role of target salience in crowding. Perception 34, 823\u2013833 (2005).\n\nPubMed\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMareschal, I., Morgan, M. J. & Solomon, J. A. Cortical distance determines whether flankers cause crowding or the tilt illusion. J. Vis. 10, 13 (2010).\n\nPubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nEster, E. F., Zilber, E. & Serences, J. T. Substitution and pooling in visual crowding induced by similar and dissimilar distractors. J. Vis. 15, 15 11 14 https://doi.org/10.1167/15.1.4 (2015).\n\nBernard, J. B. & Chung, S. T. The dependence of crowding on flanker complexity and target-flanker similarity. J. Vis. 11 https://doi.org/10.1167/11.8.1 (2011).\n\nRosenholtz, R., Yu, D. & Keshvari, S. Challenges to pooling models of crowding: Implications for visual mechanisms. J. Vis. 19, 15 (2019).\n\nPubMed\u00a0\n    PubMed Central\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nField, D. J., Hayes, A. & Hess, R. F. Contour integration by the human visual system: Evidence for a local \u201cassociation field\u201d. Vis. Res. 33, 173\u2013193 (1993).\n\nCAS\u00a0\n    PubMed\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nTs\u2019o, D. Y., Gilbert, C. D. & Wiesel, T. N. Relationships between horizontal interactions and functional architecture in cat striate cortex as revealed by cross-correlation analysis. J. Neurosci. 6, 1160\u20131170 (1986).\n\nPubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nCicchini, G. M., Benedetto, A. & Burr, D. C. Perceptual history propagates down to early levels of sensory analysis. Curr. Biol. 31, 1245\u20131250 e1242 (2021).\n\nCAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMikellidou, K., Cicchini, G. M. & Burr, D. C. Perceptual history acts in world-centred coordinates. Iperception 12, 20416695211029301 (2021).\n\nPubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHarrison, W. J. & Bex, P. J. Integrating retinotopic features in spatiotopic coordinates. J. Neurosci. 34, 7351\u20137360 (2014).\n\nCAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nFornaciai, M., Binda, P. & Cicchini, G. M. Trans-saccadic integration of orientation information. J. Vis. 18, 9 (2018).\n\nPubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nJazayeri, M. & Shadlen, M. N. Temporal context calibrates interval timing. Nat. Neurosci. 13, 1020\u20131026 (2010).\n\nCAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nCicchini, G. M., Arrighi, R., Cecchetti, L., Giusti, M. & Burr, D. C. Optimal encoding of interval timing in expert percussionists. J. Neurosci. 32, 1056\u20131060 (2012).\n\nCAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWilkinson, F., Wilson, H. R. & Ellemberg, D. Lateral interactions in peripherally viewed texture arrays. J. Opt. Soc. Am. A Opt. Image Sci. Vis. 14, 2057\u20132068 (1997).\n\nADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nChung, S. T., Levi, D. M. & Legge, G. E. Spatial-frequency and contrast properties of crowding. Vis. Res. 41, 1833\u20131850 (2001).\n\nCAS\u00a0\n    PubMed\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSolomon, J. A., Felisberti, F. M. & Morgan, M. J. Crowding and the tilt illusion: toward a unified account. J. Vis. 4, 500\u2013508 (2004).\n\nPubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGreenwood, J. A. & Parsons, M. J. Dissociable effects of visual crowding on the perception of color and motion. Proc. Natl Acad. Sci. USA. 117, 8196\u20138202 (2020).\n\nADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nManassi, M. & Whitney, D. Multi-level crowding and the paradox of object recognition in clutter. Curr. Biol. 28, R127\u2013R133 (2018).\n\nCAS\u00a0\n    PubMed\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBulakowski, P. F., Post, R. B. & Whitney, D. Visuomotor crowding: the resolution of grasping in cluttered scenes. Front Behav. Neurosci. 3, 49 (2009).\n\nPubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBanks, W. P., Larson, D. W. & Prinzmetal, W. Asymmetry of visual interference. Percept. Psychophys. 25, 447\u2013456 (1979).\n\nCAS\u00a0\n    PubMed\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBulakowski, P. F., Post, R. B. & Whitney, D. Reexamining the possible benefits of visual crowding: dissociating crowding from ensemble percepts. Atten. Percept. Psychophys. 73, 1003\u20131009 (2011).\n\nPubMed\u00a0\n    PubMed Central\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHillis, J. M., Ernst, M. O., Banks, M. S. & Landy, M. S. Combining sensory information: mandatory fusion within, but not between, senses. Science 298, 1627\u20131630 (2002).\n\nADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKersten, D., Mamassian, P. & Yuille, A. Object perception as Bayesian inference. Annu Rev. Psychol. 55, 271\u2013304 (2004).\n\nPubMed\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKording, K. P. & Wolpert, D. M. Bayesian integration in sensorimotor learning. Nature 427, 244\u2013247 (2004).\n\nADS\u00a0\n    PubMed\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nTrommershauser, J., Gepshtein, S., Maloney, L. T., Landy, M. S. & Banks, M. S. Optimal compensation for changes in task-relevant movement variability. J. Neurosci. 25, 7169\u20137178 (2005).\n\nPubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nCicchini, G. M. (Dataset for the manuscript \u201cCrowding results from optimal integration of visual targets with contextual information\u201d, Zenodo, https://doi.org/10.5281/zenodo.6460723, (2022).\n\nDownload references",
+            "section_image": []
+        },
+        {
+            "section_name": "Acknowledgements",
+            "section_text": "This work was supported by Horizon 2020 European Research Council Advanced Grant GenPercept No. 832813 (to D.C.B.), Italian Ministry of Education PRIN2017 Grants 2017SBCPZY, and FLAG-ERA Joint Transnational Call 2019 Grant DOMINO.",
+            "section_image": []
+        },
+        {
+            "section_name": "Author information",
+            "section_text": "Institute of Neuroscience, CNR, via Moruzzi, 1, 56124, Pisa, Italy\n\nGuido Marco Cicchini,\u00a0Giovanni D\u2019Errico\u00a0&\u00a0David Charles Burr\n\nDepartment of Neurosciences, Psychology, Drug Research and Child Health, University of Florence, viale Pieraccini, 6, 50139, Firenze, Italy\n\nDavid Charles Burr\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nG.M.C. designed the experiment and carried out data collection and analysis. G.D.E. carried out data collection and analysis. D.C.B. designed the experiment and wrote the manuscript. All authors viewed and approved the submitted version of the draft.\n\nCorrespondence to\n                David Charles Burr.",
+            "section_image": []
+        },
+        {
+            "section_name": "Ethics declarations",
+            "section_text": "The authors declare no competing interests.",
+            "section_image": []
+        },
+        {
+            "section_name": "Peer review",
+            "section_text": "Nature Communications thanks the anonymous reviewers for their contribution to the peer review of this work.\u00a0Peer reviewer reports are available.",
+            "section_image": []
+        },
+        {
+            "section_name": "Additional information",
+            "section_text": "Publisher\u2019s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.",
+            "section_image": []
+        },
+        {
+            "section_name": "Source data",
+            "section_text": "",
+            "section_image": []
+        },
+        {
+            "section_name": "Rights and permissions",
+            "section_text": "Open Access  This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article\u2019s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article\u2019s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.\n\nReprints and permissions",
+            "section_image": []
+        },
+        {
+            "section_name": "About this article",
+            "section_text": "Cicchini, G.M., D\u2019Errico, G. & Burr, D.C. Crowding results from optimal integration of visual targets with contextual information.\n                    Nat Commun 13, 5741 (2022). https://doi.org/10.1038/s41467-022-33508-1\n\nDownload citation\n\nReceived: 23 February 2022\n\nAccepted: 16 September 2022\n\nPublished: 30 September 2022\n\nVersion of record: 30 September 2022\n\nDOI: https://doi.org/10.1038/s41467-022-33508-1\n\nAnyone you share the following link with will be able to read this content:\n\nSorry, a shareable link is not currently available for this article.\n\n\n\n\n                            Provided by the Springer Nature SharedIt content-sharing initiative\n                        ",
+            "section_image": [
+                "https://data:image/svg+xml;base64,PHN2ZyBoZWlnaHQ9IjgxIiB3aWR0aD0iNTciIHhtbG5zPSJodHRwOi8vd3d3LnczLm9yZy8yMDAwL3N2ZyI+PGcgZmlsbD0ibm9uZSIgZmlsbC1ydWxlPSJldmVub2RkIj48cGF0aCBkPSJtMTcuMzUgMzUuNDUgMjEuMy0xNC4ydi0xNy4wM2gtMjEuMyIgZmlsbD0iIzk4OTg5OCIvPjxwYXRoIGQ9Im0zOC42NSAzNS40NS0yMS4zLTE0LjJ2LTE3LjAzaDIxLjMiIGZpbGw9IiM3NDc0NzQiLz48cGF0aCBkPSJtMjggLjVjLTEyLjk4IDAtMjMuNSAxMC41Mi0yMy41IDIzLjVzMTAuNTIgMjMuNSAyMy41IDIzLjUgMjMuNS0xMC41MiAyMy41LTIzLjVjMC02LjIzLTIuNDgtMTIuMjEtNi44OC0xNi42Mi00LjQxLTQuNC0xMC4zOS02Ljg4LTE2LjYyLTYuODh6bTAgNDEuMjVjLTkuOCAwLTE3Ljc1LTcuOTUtMTcuNzUtMTcuNzVzNy45NS0xNy43NSAxNy43NS0xNy43NSAxNy43NSA3Ljk1IDE3Ljc1IDE3Ljc1YzAgNC43MS0xLjg3IDkuMjItNS4yIDEyLjU1cy03Ljg0IDUuMi0xMi41NSA1LjJ6IiBmaWxsPSIjNTM1MzUzIi8+PHBhdGggZD0ibTQxIDM2Yy01LjgxIDYuMjMtMTUuMjMgNy40NS0yMi40MyAyLjktNy4yMS00LjU1LTEwLjE2LTEzLjU3LTcuMDMtMjEuNWwtNC45Mi0zLjExYy00Ljk1IDEwLjctMS4xOSAyMy40MiA4Ljc4IDI5LjcxIDkuOTcgNi4zIDIzLjA3IDQuMjIgMzAuNi00Ljg2eiIgZmlsbD0iIzljOWM5YyIvPjxwYXRoIGQ9Im0uMiA1OC40NWMwLS43NS4xMS0xLjQyLjMzLTIuMDFzLjUyLTEuMDkuOTEtMS41Yy4zOC0uNDEuODMtLjczIDEuMzQtLjk0LjUxLS4yMiAxLjA2LS4zMiAxLjY1LS4zMi41NiAwIDEuMDYuMTEgMS41MS4zNS40NC4yMy44MS41IDEuMS44MWwtLjkxIDEuMDFjLS4yNC0uMjQtLjQ5LS40Mi0uNzUtLjU2LS4yNy0uMTMtLjU4LS4yLS45My0uMi0uMzkgMC0uNzMuMDgtMS4wNS4yMy0uMzEuMTYtLjU4LjM3LS44MS42Ni0uMjMuMjgtLjQxLjYzLS41MyAxLjA0LS4xMy40MS0uMTkuODgtLjE5IDEuMzkgMCAxLjA0LjIzIDEuODYuNjggMi40Ni40NS41OSAxLjA2Ljg4IDEuODQuODguNDEgMCAuNzctLjA3IDEuMDctLjIzcy41OS0uMzkuODUtLjY4bC45MSAxYy0uMzguNDMtLjguNzYtMS4yOC45OS0uNDcuMjItMSAuMzQtMS41OC4zNC0uNTkgMC0xLjEzLS4xLTEuNjQtLjMxLS41LS4yLS45NC0uNTEtMS4zMS0uOTEtLjM4LS40LS42Ny0uOS0uODgtMS40OC0uMjItLjU5LS4zMy0xLjI2LS4zMy0yLjAyem04LjQtNS4zM2gxLjYxdjIuNTRsLS4wNSAxLjMzYy4yOS0uMjcuNjEtLjUxLjk2LS43MnMuNzYtLjMxIDEuMjQtLjMxYy43MyAwIDEuMjcuMjMgMS42MS43MS4zMy40Ny41IDEuMTQuNSAyLjAydjQuMzFoLTEuNjF2LTQuMWMwLS41Ny0uMDgtLjk3LS4yNS0xLjIxLS4xNy0uMjMtLjQ1LS4zNS0uODMtLjM1LS4zIDAtLjU2LjA4LS43OS4yMi0uMjMuMTUtLjQ5LjM2LS43OC42NHY0LjhoLTEuNjF6bTcuMzcgNi40NWMwLS41Ni4wOS0xLjA2LjI2LTEuNTEuMTgtLjQ1LjQyLS44My43MS0xLjE0LjI5LS4zLjYzLS41NCAxLjAxLS43MS4zOS0uMTcuNzgtLjI1IDEuMTgtLjI1LjQ3IDAgLjg4LjA4IDEuMjMuMjQuMzYuMTYuNjUuMzguODkuNjdzLjQyLjYzLjU0IDEuMDNjLjEyLjQxLjE4Ljg0LjE4IDEuMzIgMCAuMzItLjAyLjU3LS4wNy43NmgtNC4zNmMuMDcuNjIuMjkgMS4xLjY1IDEuNDQuMzYuMzMuODIuNSAxLjM4LjUuMjkgMCAuNTctLjA0LjgzLS4xM3MuNTEtLjIxLjc2LS4zN2wuNTUgMS4wMWMtLjMzLjIxLS42OS4zOS0xLjA5LjUzLS40MS4xNC0uODMuMjEtMS4yNi4yMS0uNDggMC0uOTItLjA4LTEuMzQtLjI1LS40MS0uMTYtLjc2LS40LTEuMDctLjctLjMxLS4zMS0uNTUtLjY5LS43Mi0xLjEzLS4xOC0uNDQtLjI2LS45NS0uMjYtMS41MnptNC42LS42MmMwLS41NS0uMTEtLjk4LS4zNC0xLjI4LS4yMy0uMzEtLjU4LS40Ny0xLjA2LS40Ny0uNDEgMC0uNzcuMTUtMS4wNy40NS0uMzEuMjktLjUuNzMtLjU4IDEuM3ptMi41LjYyYzAtLjU3LjA5LTEuMDguMjgtMS41My4xOC0uNDQuNDMtLjgyLjc1LTEuMTNzLjY5LS41NCAxLjEtLjcxYy40Mi0uMTYuODUtLjI0IDEuMzEtLjI0LjQ1IDAgLjg0LjA4IDEuMTcuMjNzLjYxLjM0Ljg1LjU3bC0uNzcgMS4wMmMtLjE5LS4xNi0uMzgtLjI4LS41Ni0uMzctLjE5LS4wOS0uMzktLjE0LS42MS0uMTQtLjU2IDAtMS4wMS4yMS0xLjM1LjYzLS4zNS40MS0uNTIuOTctLjUyIDEuNjcgMCAuNjkuMTcgMS4yNC41MSAxLjY2LjM0LjQxLjc4LjYyIDEuMzIuNjIuMjggMCAuNTQtLjA2Ljc4LS4xNy4yNC0uMTIuNDUtLjI2LjY0LS40MmwuNjcgMS4wM2MtLjMzLjI5LS42OS41MS0xLjA4LjY1LS4zOS4xNS0uNzguMjMtMS4xOC4yMy0uNDYgMC0uOS0uMDgtMS4zMS0uMjQtLjQtLjE2LS43NS0uMzktMS4wNS0uN3MtLjUzLS42OS0uNy0xLjEzYy0uMTctLjQ1LS4yNS0uOTYtLjI1LTEuNTN6bTYuOTEtNi40NWgxLjU4djYuMTdoLjA1bDIuNTQtMy4xNmgxLjc3bC0yLjM1IDIuOCAyLjU5IDQuMDdoLTEuNzVsLTEuNzctMi45OC0xLjA4IDEuMjN2MS43NWgtMS41OHptMTMuNjkgMS4yN2MtLjI1LS4xMS0uNS0uMTctLjc1LS4xNy0uNTggMC0uODcuMzktLjg3IDEuMTZ2Ljc1aDEuMzR2MS4yN2gtMS4zNHY1LjZoLTEuNjF2LTUuNmgtLjkydi0xLjJsLjkyLS4wN3YtLjcyYzAtLjM1LjA0LS42OC4xMy0uOTguMDgtLjMxLjIxLS41Ny40LS43OXMuNDItLjM5LjcxLS41MWMuMjgtLjEyLjYzLS4xOCAxLjA0LS4xOC4yNCAwIC40OC4wMi42OS4wNy4yMi4wNS40MS4xLjU3LjE3em0uNDggNS4xOGMwLS41Ny4wOS0xLjA4LjI3LTEuNTMuMTctLjQ0LjQxLS44Mi43Mi0xLjEzLjMtLjMxLjY1LS41NCAxLjA0LS43MS4zOS0uMTYuOC0uMjQgMS4yMy0uMjRzLjg0LjA4IDEuMjQuMjRjLjQuMTcuNzQuNCAxLjA0Ljcxcy41NC42OS43MiAxLjEzYy4xOS40NS4yOC45Ni4yOCAxLjUzcy0uMDkgMS4wOC0uMjggMS41M2MtLjE4LjQ0LS40Mi44Mi0uNzIgMS4xM3MtLjY0LjU0LTEuMDQuNy0uODEuMjQtMS4yNC4yNC0uODQtLjA4LTEuMjMtLjI0LS43NC0uMzktMS4wNC0uN2MtLjMxLS4zMS0uNTUtLjY5LS43Mi0xLjEzLS4xOC0uNDUtLjI3LS45Ni0uMjctMS41M3ptMS42NSAwYzAgLjY5LjE0IDEuMjQuNDMgMS42Ni4yOC40MS42OC42MiAxLjE4LjYyLjUxIDAgLjktLjIxIDEuMTktLjYyLjI5LS40Mi40NC0uOTcuNDQtMS42NiAwLS43LS4xNS0xLjI2LS40NC0xLjY3LS4yOS0uNDItLjY4LS42My0xLjE5LS42My0uNSAwLS45LjIxLTEuMTguNjMtLjI5LjQxLS40My45Ny0uNDMgMS42N3ptNi40OC0zLjQ0aDEuMzNsLjEyIDEuMjFoLjA1Yy4yNC0uNDQuNTQtLjc5Ljg4LTEuMDIuMzUtLjI0LjctLjM2IDEuMDctLjM2LjMyIDAgLjU5LjA1Ljc4LjE0bC0uMjggMS40LS4zMy0uMDljLS4xMS0uMDEtLjIzLS4wMi0uMzgtLjAyLS4yNyAwLS41Ni4xLS44Ni4zMXMtLjU1LjU4LS43NyAxLjF2NC4yaC0xLjYxem0tNDcuODcgMTVoMS42MXY0LjFjMCAuNTcuMDguOTcuMjUgMS4yLjE3LjI0LjQ0LjM1LjgxLjM1LjMgMCAuNTctLjA3LjgtLjIyLjIyLS4xNS40Ny0uMzkuNzMtLjczdi00LjdoMS42MXY2Ljg3aC0xLjMybC0uMTItMS4wMWgtLjA0Yy0uMy4zNi0uNjMuNjQtLjk4Ljg2LS4zNS4yMS0uNzYuMzItMS4yNC4zMi0uNzMgMC0xLjI3LS4yNC0xLjYxLS43MS0uMzMtLjQ3LS41LTEuMTQtLjUtMi4wMnptOS40NiA3LjQzdjIuMTZoLTEuNjF2LTkuNTloMS4zM2wuMTIuNzJoLjA1Yy4yOS0uMjQuNjEtLjQ1Ljk3LS42My4zNS0uMTcuNzItLjI2IDEuMS0uMjYuNDMgMCAuODEuMDggMS4xNS4yNC4zMy4xNy42MS40Ljg0LjcxLjI0LjMxLjQxLjY4LjUzIDEuMTEuMTMuNDIuMTkuOTEuMTkgMS40NCAwIC41OS0uMDkgMS4xMS0uMjUgMS41Ny0uMTYuNDctLjM4Ljg1LS42NSAxLjE2LS4yNy4zMi0uNTguNTYtLjk0LjczLS4zNS4xNi0uNzIuMjUtMS4xLjI1LS4zIDAtLjYtLjA3LS45LS4ycy0uNTktLjMxLS44Ny0uNTZ6bTAtMi4zYy4yNi4yMi41LjM3LjczLjQ1LjI0LjA5LjQ2LjEzLjY2LjEzLjQ2IDAgLjg0LS4yIDEuMTUtLjYuMzEtLjM5LjQ2LS45OC40Ni0xLjc3IDAtLjY5LS4xMi0xLjIyLS4zNS0xLjYxLS4yMy0uMzgtLjYxLS41Ny0xLjEzLS41Ny0uNDkgMC0uOTkuMjYtMS41Mi43N3ptNS44Ny0xLjY5YzAtLjU2LjA4LTEuMDYuMjUtMS41MS4xNi0uNDUuMzctLjgzLjY1LTEuMTQuMjctLjMuNTgtLjU0LjkzLS43MXMuNzEtLjI1IDEuMDgtLjI1Yy4zOSAwIC43My4wNyAxIC4yLjI3LjE0LjU0LjMyLjgxLjU1bC0uMDYtMS4xdi0yLjQ5aDEuNjF2OS44OGgtMS4zM2wtLjExLS43NGgtLjA2Yy0uMjUuMjUtLjU0LjQ2LS44OC42NC0uMzMuMTgtLjY5LjI3LTEuMDYuMjctLjg3IDAtMS41Ni0uMzItMi4wNy0uOTVzLS43Ni0xLjUxLS43Ni0yLjY1em0xLjY3LS4wMWMwIC43NC4xMyAxLjMxLjQgMS43LjI2LjM4LjY1LjU4IDEuMTUuNTguNTEgMCAuOTktLjI2IDEuNDQtLjc3di0zLjIxYy0uMjQtLjIxLS40OC0uMzYtLjctLjQ1LS4yMy0uMDgtLjQ2LS4xMi0uNy0uMTItLjQ1IDAtLjgyLjE5LTEuMTMuNTktLjMxLjM5LS40Ni45NS0uNDYgMS42OHptNi4zNSAxLjU5YzAtLjczLjMyLTEuMy45Ny0xLjcxLjY0LS40IDEuNjctLjY4IDMuMDgtLjg0IDAtLjE3LS4wMi0uMzQtLjA3LS41MS0uMDUtLjE2LS4xMi0uMy0uMjItLjQzcy0uMjItLjIyLS4zOC0uM2MtLjE1LS4wNi0uMzQtLjEtLjU4LS4xLS4zNCAwLS42OC4wNy0xIC4ycy0uNjMuMjktLjkzLjQ3bC0uNTktMS4wOGMuMzktLjI0LjgxLS40NSAxLjI4LS42My40Ny0uMTcuOTktLjI2IDEuNTQtLjI2Ljg2IDAgMS41MS4yNSAxLjkzLjc2cy42MyAxLjI1LjYzIDIuMjF2NC4wN2gtMS4zMmwtLjEyLS43NmgtLjA1Yy0uMy4yNy0uNjMuNDgtLjk4LjY2cy0uNzMuMjctMS4xNC4yN2MtLjYxIDAtMS4xLS4xOS0xLjQ4LS41Ni0uMzgtLjM2LS41Ny0uODUtLjU3LTEuNDZ6bTEuNTctLjEyYzAgLjMuMDkuNTMuMjcuNjcuMTkuMTQuNDIuMjEuNzEuMjEuMjggMCAuNTQtLjA3Ljc3LS4ycy40OC0uMzEuNzMtLjU2di0xLjU0Yy0uNDcuMDYtLjg2LjEzLTEuMTguMjMtLjMxLjA5LS41Ny4xOS0uNzYuMzFzLS4zMy4yNS0uNDEuNGMtLjA5LjE1LS4xMy4zMS0uMTMuNDh6bTYuMjktMy42M2gtLjk4di0xLjJsMS4wNi0uMDcuMi0xLjg4aDEuMzR2MS44OGgxLjc1djEuMjdoLTEuNzV2My4yOGMwIC44LjMyIDEuMi45NyAxLjIuMTIgMCAuMjQtLjAxLjM3LS4wNC4xMi0uMDMuMjQtLjA3LjM0LS4xMWwuMjggMS4xOWMtLjE5LjA2LS40LjEyLS42NC4xNy0uMjMuMDUtLjQ5LjA4LS43Ni4wOC0uNCAwLS43NC0uMDYtMS4wMi0uMTgtLjI3LS4xMy0uNDktLjMtLjY3LS41Mi0uMTctLjIxLS4zLS40OC0uMzctLjc4LS4wOC0uMy0uMTItLjY0LS4xMi0xLjAxem00LjM2IDIuMTdjMC0uNTYuMDktMS4wNi4yNy0xLjUxcy40MS0uODMuNzEtMS4xNGMuMjktLjMuNjMtLjU0IDEuMDEtLjcxLjM5LS4xNy43OC0uMjUgMS4xOC0uMjUuNDcgMCAuODguMDggMS4yMy4yNC4zNi4xNi42NS4zOC44OS42N3MuNDIuNjMuNTQgMS4wM2MuMTIuNDEuMTguODQuMTggMS4zMiAwIC4zMi0uMDIuNTctLjA3Ljc2aC00LjM3Yy4wOC42Mi4yOSAxLjEuNjUgMS40NC4zNi4zMy44Mi41IDEuMzguNS4zIDAgLjU4LS4wNC44NC0uMTMuMjUtLjA5LjUxLS4yMS43Ni0uMzdsLjU0IDEuMDFjLS4zMi4yMS0uNjkuMzktMS4wOS41M3MtLjgyLjIxLTEuMjYuMjFjLS40NyAwLS45Mi0uMDgtMS4zMy0uMjUtLjQxLS4xNi0uNzctLjQtMS4wOC0uNy0uMy0uMzEtLjU0LS42OS0uNzItMS4xMy0uMTctLjQ0LS4yNi0uOTUtLjI2LTEuNTJ6bTQuNjEtLjYyYzAtLjU1LS4xMS0uOTgtLjM0LTEuMjgtLjIzLS4zMS0uNTgtLjQ3LTEuMDYtLjQ3LS40MSAwLS43Ny4xNS0xLjA4LjQ1LS4zMS4yOS0uNS43My0uNTcgMS4zem0zLjAxIDIuMjNjLjMxLjI0LjYxLjQzLjkyLjU3LjMuMTMuNjMuMi45OC4yLjM4IDAgLjY1LS4wOC44My0uMjNzLjI3LS4zNS4yNy0uNmMwLS4xNC0uMDUtLjI2LS4xMy0uMzctLjA4LS4xLS4yLS4yLS4zNC0uMjgtLjE0LS4wOS0uMjktLjE2LS40Ny0uMjNsLS41My0uMjJjLS4yMy0uMDktLjQ2LS4xOC0uNjktLjMtLjIzLS4xMS0uNDQtLjI0LS42Mi0uNHMtLjMzLS4zNS0uNDUtLjU1Yy0uMTItLjIxLS4xOC0uNDYtLjE4LS43NSAwLS42MS4yMy0xLjEuNjgtMS40OS40NC0uMzggMS4wNi0uNTcgMS44My0uNTcuNDggMCAuOTEuMDggMS4yOS4yNXMuNzEuMzYuOTkuNTdsLS43NC45OGMtLjI0LS4xNy0uNDktLjMyLS43My0uNDItLjI1LS4xMS0uNTEtLjE2LS43OC0uMTYtLjM1IDAtLjYuMDctLjc2LjIxLS4xNy4xNS0uMjUuMzMtLjI1LjU0IDAgLjE0LjA0LjI2LjEyLjM2cy4xOC4xOC4zMS4yNmMuMTQuMDcuMjkuMTQuNDYuMjFsLjU0LjE5Yy4yMy4wOS40Ny4xOC43LjI5cy40NC4yNC42NC40Yy4xOS4xNi4zNC4zNS40Ni41OC4xMS4yMy4xNy41LjE3LjgyIDAgLjMtLjA2LjU4LS4xNy44My0uMTIuMjYtLjI5LjQ4LS41MS42OC0uMjMuMTktLjUxLjM0LS44NC40NS0uMzQuMTEtLjcyLjE3LTEuMTUuMTctLjQ4IDAtLjk1LS4wOS0xLjQxLS4yNy0uNDYtLjE5LS44Ni0uNDEtMS4yLS42OHoiIGZpbGw9IiM1MzUzNTMiLz48L2c+PC9zdmc+"
+            ]
+        },
+        {
+            "section_name": "This article is cited by",
+            "section_text": "Nature Communications (2025)\n\nNature Communications (2025)",
+            "section_image": []
+        },
+        {
+            "section_name": "Associated content",
+            "section_text": "",
+            "section_image": []
+        }
+    ]
+}

0226534d35058c645a89e4e44377c8b2e4f25f5dddefbf9c78c868b60659db7b/metadata.json

@@ -0,0 +1,154 @@
+{
+    "title": "Changes in limiting factors for forager population dynamics in Europe across the last glacial-interglacial transition",
+    "pre_title": "Changes in limiting factors for forager population dynamics in Europe across the Last Glacial-Interglacial Transition",
+    "journal": "Nature Communications",
+    "published": "06 September 2022",
+    "supplementary_0": [
+        {
+            "label": "Supplementary Information",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-022-32750-x/MediaObjects/41467_2022_32750_MOESM1_ESM.pdf"
+        },
+        {
+            "label": "Peer Review File",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-022-32750-x/MediaObjects/41467_2022_32750_MOESM2_ESM.pdf"
+        },
+        {
+            "label": "Reporting Summary",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-022-32750-x/MediaObjects/41467_2022_32750_MOESM3_ESM.pdf"
+        }
+    ],
+    "supplementary_1": [
+        {
+            "label": "Source Data",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-022-32750-x/MediaObjects/41467_2022_32750_MOESM4_ESM.xlsx"
+        }
+    ],
+    "supplementary_2": NaN,
+    "source_data": [
+        "/articles/s41467-022-32750-x#ref-CR26",
+        "https://d-place.org/contributions/Binford",
+        "https://pandoradata.earth/dataset/radiocarbon-palaeolithic-europe-database-v28",
+        "https://www.worldclim.org/data/worldclim21.html",
+        "https://doi.org/10.6084/m9.figshare.c.4673120.v2",
+        "/articles/s41467-022-32750-x#Sec13"
+    ],
+    "code": [
+        "https://doi.org/10.5281/zenodo.6962693",
+        "/articles/s41467-022-32750-x#ref-CR101"
+    ],
+    "subject": [
+        "Archaeology",
+        "Biogeography",
+        "Biological anthropology",
+        "Palaeoecology",
+        "Social anthropology"
+    ],
+    "license": "http://creativecommons.org/licenses/by/4.0/",
+    "preprint_pdf": "https://www.researchsquare.com/article/rs-1173690/v1.pdf?c=1662548796000",
+    "research_square_link": "https://www.researchsquare.com//article/rs-1173690/v1",
+    "nature_pdf": "https://www.nature.com/articles/s41467-022-32750-x.pdf",
+    "preprint_posted": "22 Dec, 2021",
+    "research_square_content": [
+        {
+            "section_name": "Abstract",
+            "section_text": "Population dynamics set the framework for human genetic and cultural evolution. For foragers, demographic and environmental changes correlate strongly, although the causal relations between different environmental variables and human responses through time and space likely varied. Building on the notion of limiting factors, namely that the scarcest resource regulates population size, we present a statistical approach to identify the dominant climatic constraints for hunter-gatherer population densities and then hindcast their changing dynamics in Europe for the period between 20kyBP to 8kyBP. Limiting factors shifted from temperature-related variables during the Pleistocene to a regional mosaic of limiting factors in the Holocene. This spatiotemporal variation suggests that hunter-gatherers needed to overcome very different adaptive challenges in different parts of Europe, and that these challenges vary over time. The signatures of these changing adaptations may be visible archaeologically. In addition, the spatial disaggregation of limiting factors from the Pleistocene to the Holocene coincides with and may partly explain the diversification of the cultural geography at this time.",
+            "section_image": []
+        },
+        {
+            "section_name": "Additional Declarations",
+            "section_text": "There is NO Competing Interest.",
+            "section_image": []
+        }
+    ],
+    "nature_content": [
+        {
+            "section_name": "Abstract",
+            "section_text": "Population dynamics set the framework for human genetic and cultural evolution. For foragers, demographic and environmental changes correlate strongly, although the causal relations between different environmental variables and human responses through time and space likely varied. Building on the notion of limiting factors, namely that at any one time, the scarcest resource caps population size, we present a statistical approach to identify the dominant climatic constraints for hunter-gatherer population densities and then hindcast their changing dynamics in Europe for the period between 21,000 to 8000 years ago. Limiting factors shifted from temperature-related variables (effective temperature) during the Pleistocene to a regional mosaic of limiting factors in the Holocene dominated by temperature seasonality and annual precipitation. This spatiotemporal variation suggests that hunter-gatherers needed to overcome very different adaptive challenges in different parts of Europe and that these challenges varied over time. The signatures of these changing adaptations may be visible archaeologically. In addition, the spatial disaggregation of limiting factors from the Pleistocene to the Holocene coincided with and may partly explain the diversification of the cultural geography at this time.",
+            "section_image": []
+        },
+        {
+            "section_name": "Introduction",
+            "section_text": "As the link between exogenous environmental factors and organismal physiology, demography is vital for understanding evolution, including cultural evolution1. The relevance of past demography for understanding culture change in prehistory has long been recognised2,3. Demographic conditions impinge on cultural transmission4,5,6 but are also clearly implicated in the boom-and-bust patterns of population fluctuations\u2014including periodic extirpations-\u2014suggested to have characterised the demographic histories of prehistoric foragers and incipient farmers in many regions7,8,9,10. Numerous recent studies have focused on the drivers of population expansion to explain the pattern and timing of human colonisation using a variety of ecological comparative approaches11,12 (but see refs. 13, 14, for a discussion of points of concern of such approaches). As foragers have a high intrinsic growth rate, however,\u00a0population increase is the default demographic trajectory in the absence of cultural or environmental constraints. Yet, past populations did not grow substantially, making it particularly germane to understand the factors that curtailed population growth15,16. The approach adopted here builds on the central theorem that population sizes would, at any one time, be regulated by the scarcest resource: the limiting factor17.\n\nForagers of the recent past persisted in various\u00a0environments, from the frigid Arctic to tropical rainforests. Each environment offered particular opportunities but also posed specific challenges. Cultural practices and technology provided a means of buffering against or even overcoming some of these challenges. Nevertheless, environmental factors have also been shown to directly, albeit broadly, constrain human palaeodemography. Several earlier studies have pointed to temperature or seasonality as crucial drivers of forager demography at global or continental scales18,19. However, the specific factors that would have capped or even depressed population size are likely to have varied in space and time. Only by understanding these limiting factors can we begin to conduct targeted investigations of how specific forager populations may have overcome them via population-specific genetic adaptations, behavioural modifications or the \u2018extra-somatic adaptions\u201920 that are so characteristic of human culture.\n\nThis study focused on forager palaeodemography in Europe from the Last Glacial Maximum (Greenland Stadial 2, GS2) to 8000 years before present (kyBP), a climatically volatile period also known as the Last Glacial-Interglacial Transition21. Previous studies have identified population growth and expansion patterns using various methods commonly used in ecological analyses12,22,23,24. Correlations between temperature and overall population density have been identified, suggesting overall increases in energy availability as the key driver of the increase in human population size following the end of GS218. However, regional population collapses have also been suggested to have occurred asynchronously and in different places9,25. These regional patterns raise the question of which specific limiting factors acted on forager populations and how these limiting factors varied over space and time.\n\nLike many related studies, we begin with the global ethnographic hunter-gatherer dataset originally assembled by Binford and now digitally available26,27. We couple this to a suite of quantile generalised additive models (qGAMs) that describe changes in the maximum (90th percentile), mean (50th percentile), and minimum (10th percentile) population density as a univariate function of environmental variables related to the effect of energy or water availability, productivity, and annual limits and variability (Table\u00a01). A strong correlation between predictor variables is not a major obstacle in our analyses28 for two reasons. Firstly, our aim is not to determine the best variables to predict population density. Second, we do not compare the relative effects of evaluated predictors nor establish the variable contributing the most to the deviations from the regional population density average.\n\nHere, instead of simultaneously\u00a0evaluating groups of variables, as traditionally done in regression analyses using Binford\u2019s ethnographic data11,12,22,29, our analyses focus on defining the possible main driver (i.e. the individual variable)\u00a0that acted as a limiting constraint (as suggested by refs. 30, 31). Specifically, we generated 1000 different models for each predictor-quantile computation using 70% of Binford\u2019s data. Observations to build each model were selected using an h-block cross-validation approach (ref. 32; see Methods). It is essential to highlight that these models seek not to predict past population densities across regions but to reveal the limiting effects of climate on this vital variable as determined by comparing the absolute effects of the evaluated variables. In this way, our analyses serve as a prerequisite for identifying the specific role of human cultural practices and innovations to buffer and even overcome such limiting factors.\n\nBased on this analysis, we hindcast hunter-gatherer population densities between GS2 and 8kyBP in 500-year intervals using the downscaled centennial average conditions of each predictor derived from a transient climatic simulation (SynTrace-2133). We only predict population densities for those climatic variables associated with the best-performing univariate qGAM models where performance was defined using a cross-validation approach (see Methods). Population density for each cell within each 500-years interval is then determined as the minimum predicted density for that cell. Population size for each period was determined by summing the products between population size and the area of each grid over the ice-free areas. We define the limiting environmental factor as the variable predicting the lowest mean population density at a given place and time and explore the dynamic changes in these factors as a function of the quantile used to make our predictions. This approach allows us to query the spatial dynamics of forager limiting factors across the Last Glacial-Interglacial Transition and derive specific hypotheses as to which selection pressures acted most strongly on different forager communities in Late Pleistocene and Early Holocene Europe.\n\nOur analysis demonstrates that the limiting factors for forager population densities showed marked differences in space and time. Temperature-related variables were the main limiting factors during the Pleistocene, whereas a regional mosaic of limiting factors characterised the Early Holocene. Furthermore, our model reveals geographic differences in the limiting factors between Fennoscandia, Southern, Central, and Eastern Europe. The spatiotemporal variation in limiting factors suggests that hunter-gatherers needed to overcome very different adaptive challenges in different parts of Europe across this period of climatic and environmental change.",
+            "section_image": []
+        },
+        {
+            "section_name": "Results and discussion",
+            "section_text": "The relation between the environmental factors explored here (Table\u00a01) and population density assessed using qGAMs (Fig.\u00a01) was negative for temperature seasonality (TS); positive for effective temperature (ET), net primary productivity (NPP), and temperature of the coldest month (MCM); unimodal for the temperature of the warmest month (MWM); and asymptotic for potential evapotranspiration (PET). Total annual precipitation (TAP) and monthly limits showed an overall flat trend. In all cases, our qGAMs performed significantly better than a model without predictors (hereafter termed the mean model), as determined by the significant deviance explained (Table\u00a01).\n\nQuantile Generalised Additive Models (qGAM) describing the relation between environmental factors and population density for A Effective temperature, B Potential evapotranspiration, C Net primary productivity, D Mean temperature of the coldest month, E Mean temperature warmest month, F Temperature seasonality, G) Total annual precipitation, H Precipitation driest month, I Precipitation wettest month and J Precipitation seasonality. Lines on each panel show the 10th percentiles (purple lines), 50th percentiles (red lines), and 90th percentiles (blue lines). Solid lines show the mean predicted values, and dashed bands indicate the 95% confidence intervals of each model prediction.\n\nIn our 50th percentile qGAMs, most of the environmental variables produced models that explain, on average, between 10% and 49% of the population density variation among ethnographic foraging societies (Table\u00a01). The five environmental variables with the highest model performance compared to a mean model (deviance explained in Table\u00a01; see Methods) were MCM, TS, ET, NPP and MWM. MCM represents the effect of yearly limiting conditions (= winter mortality) on ecological performance and hence demographic trends34,35. The other two temperature variables (ET, TS) relate to spatiotemporal energy variation34. Lastly, net primary productivity (NPP) reflects environmental productivity in the form of the availability of plant resources, where high values lead to larger population densities, as already suggested by a plethora of earlier studies36,37,38. These variables display high correlations (Pearson correlations range between 0.83 and 0.96), supporting the notion that they broadly reflect energy availability and variability effects on forager demography. Other variables related to environmental productivity (PET) have lower yet somewhat similar predictive accuracy (Table\u00a01). We use the correlation between predictors to underpin our grouping of individual variables within suites of possible explanatory mechanisms (as listed in Table\u00a01). This correlation between variables is not in itself a problem in our analytical paradigm as we are building univariate models. Furthermore, as our analyses aim to define the possible mechanism acting as a limiting constraint, considering the absolute effect of individual variables at a time allows us to capture their effects better.\n\nBesides the well-known limitations of using foragers of the recent past for reconstructing prehistoric social and demographic conditions13, the issues of model truncation and non-analogy of climatic conditions present themselves as major potential caveats. Climatic non-analogy here refers to the problem of projecting models beyond the domain for which they have been calibrated39,40,41. Model truncation refers to the incomplete characterisation of hunter-gatherer populations\u2019 total climate space42,43,44 and has been a long-noted limitation of ethnographic analogies for prehistoric foragers45. Yet, it has already been shown that the dataset assembled by Binford is not critically biased in terms of forager niche space26. However, we see that MWM, under current conditions, does not fully represent the environmental states of the evaluated period\u2019s early (Pleistocene) time points. Therefore, MWM was not considered in our population density estimation and limiting factors. For the other variables, we do not see either truncation or severe non-analogy: the climate space observed at different moments during the 21-8kyBP period shows broad overlaps with the climate space used to develop our qGAMs (Fig.\u00a02). Therefore, our models are not unduly extrapolating into environmental regions where there is no clear indication of how population density changes as a function of evaluated climatic variables. By the same token, it is necessary to highlight that the distributions of some palaeoclimatic conditions\u2014including all those with the highest predictive values in our models\u2014are skewed towards the lower end of contemporary values. This\u00a0skewness is especially pronounced for the Pleistocene and variables such as maximum temperatures (Fig.\u00a02), affecting our inferential power on changes in population densities at these extremes. Therefore,\u00a0our hindcast population densities are gross overestimations, especially for the Pleistocene, where temperature-related variables dominate as limiting factors.\n\nConvergence between current climatic conditions (hashed density plots) and paleoclimatic conditions at four different periods (coloured density plots) for A Effective temperature, B Potential evapotranspiration, C Net primary productivity, D Mean temperature of the coldest month, E Mean temperature warmest month, F Temperature seasonality, G Total annual precipitation, H Precipitation driest month, I Precipitation wettest month and J Precipitation seasonality. Paleoclimatic periods are Greenland Stadial 2, Greenland Interstadial 1, Greenland Stadial 1 and Early Holocene. Panels show the density plots for each evaluated variable (Variables are explained in Table\u00a01).\n\nBased on our 50th percentile qGAM models, Europe\u2019s estimated climate-limited human population size was smallest at 21kyBP (~117,500 individuals) and largest at 8kyBP (~213,900 individuals). Based on the 90th and 10th percentile models, these climate-limited human population sizes could have been as high as ~625,000 individuals and as low as ~28,300 individuals at 21kyBP. At 8kyBP, a\u00a0climatically limited population size could have been as high as ~1,111,000 individuals and as low as ~52,000 individuals. These estimates represent the overall continental-scale human population size to be expected if climatic conditions were the only factor affecting the number of individuals and under conditions where all available space was, in fact, occupied.\n\nAlso, based on our 50th percentile qGAM model and using a viable population density threshold of 0.2 individuals/100\u2009km2, we show that at the warmest point of Greenland Interstadial 1 (~14.7kyBP; GI1), Europe\u2019s human population size estimated by our model was ~155,000 individuals; a number that decreased to ~143,000 individuals at the coldest point of Greenland Stadial 1 (~11.7kyBP; GS1). The overall occupied area (number of inhabited cells), based on our 50th percentile qGAM model, was 62% of the ice-free region at the end of the GS2 (~21kyBP), increasing to ~74% during GI1 and reaching its nadir (~93%) by 8kyBP. Forager land-use was evidently extensive, however, including many largely empty spaces46. Still, at continental and centennial-to-millennial scales, overall population growth is suggested. As indicated by a strong correlation (Pearson\u2019s rho\u2009=\u20090.7 p\u2009=\u20090.007)-\u2014and despite the differences in the temporal aggregation and types of data utilised\u2014there is a robust alignment in trends between our estimates of population size and independently derived archaeological occupation proxies (red lines; Fig.\u00a03).\n\nEstimated average population density for all of Europe is based on the average of grid-based population densities (minimum predicted density for a cell) in ice-free regions (black line\u2014top panel) based on the 50th percentile generalised additive models (qGAM). Here solid black lines show the mean predicted values and black dashed bands indicate each model prediction\u2019s 95% confidence intervals. Estimates for 90th and 10th quantiles are presented in Supplementary Fig.\u00a01. These estimates are compared to archaeological population proxy based on the number of calibrated radiocarbon dates for Europe between 21 and 11kyBP extracted from ref. 12.\u00a0Summaries of the Radiocarbon Palaeolithic Europe Database v28105 (red\u2014top panel), and core area calculations (cf.24 population density mean [solid blue line] and upper/lower estimates [blue area limits] based on the Cologne Protocol (blue\u2014top panel). The\u00a0bottom panels show the changes in continental averages of the environmental variables used to determine our population density estimate (minimum across all variables). For each environmental predictor, solid lines show the continental mean, and dashed bands indicate the 95% confidence intervals.\n\nDuring the evaluated period, the mean population density in the inhabited area varied between 1.6 and 1.8 persons per 100km2 (GS2\u2009=\u20091.63 p/100\u2009km2; GI1\u2009=\u20091.70 p/100\u2009km2; GS1\u2009=\u20091.61p/100\u2009km2; EHol\u2009=\u20091.78p/100\u2009km2; Fig.\u00a03). Mean temporal patterns were similar to those predicted based on 90th and 10th percentile qGAM models (Supplementary Fig.\u00a01). Although the temporal patterns in average population density derived from our limiting-factor analysis are similar to those of core area estimates by ref. 24 (blue areas, Fig.\u00a03), these do not match numerically due to our focus on 50th percentile population densities. Our population density estimates are consistent with those suggested by ref. 12, and more recently by ref. 11.\n\nThe estimated pattern of human population density (Fig.\u00a03) indicates a population expansion starting almost 3ky after the ice sheet began to recede from its maximum extent at 21kyBP. Evaluating the spatially explicit predictions of our model (Fig.\u00a04 and Supplementary Figs.\u00a02,3), we find that at the end of the GS2, hunter-gatherer societies in Europe extended\u2014at appreciable potential densities\u2014as far north as central France, southern Germany and southern parts of modern-day Ukraine with steep negative gradients northwards. This distribution\u00a0pattern is consistent with the\u00a0archaeological evidence for the recolonisation of Europe47,48,49,50. Our models also suggest that by the end of the GS2, a relatively large proportion of the European continent may have been at least sporadically visited (~62%; Fig.\u00a04A, B), with the Mediterranean region up to the northern Alpine foreland showing population densities up to five individuals/100\u2009km2. This restricted occurrence pattern is also supported by the archaeological record46. Our\u00a0model also\u00a0indicates a persistent southwest-northeast gradient of decreasing population densities in this southern region, with the most populated areas occurring in the Iberian Peninsula specifically and the Mediterranean region more broadly (Fig.\u00a04A, B). Recent archaeological analyses suggest early exploratory dispersals northwards between 19\u201318kyBP51, which may correspond to the slight increase in modelled population densities. Subsequently, a more sustained recolonisation of the continent gathered pace from ~17kyBP (Fig.\u00a04A), reaching up to Scandinavia by the onset of GS1 (~12.8kyBP, Fig.\u00a04C). Earlier archaeological52,53 and modelling studies23 have already suggested that this colonisation was rapid but also that it proceeded in several steps where both climate and landforms served as barriers to expansion23. Our results contribute to this discussion by highlighting that different climate variables limited human dispersal for a given location and that these limits changed over time.\n\nPredictions are based on 50th percentile generalised additive models (qGAM), and any human presence is coloured. Areas in grey represent the glacier extent as derived by ICE-6G-C104. Panels from top to bottom and left to right are Greenland Stadial-2, Greenland Interstadial-1, Greenland Stadial-1, Holocene initiation and Early Holocene.\n\nUsing our limiting factor approach, we improve our understanding of demographic mechanisms in Late Pleistocene and Early Holocene European hunter-gatherer societies by highlighting the spatiotemporal changes in the main factor restricting population density (Figs.\u00a05, 6 and Supplementary Fig.\u00a04). Our modelled population density estimates can be linked to regional or local narratives or empirical tests of changes in occurrences and population sizes (e.g. refs. 26, 54). The changes in limiting factors suggested in our models can be divided into three periods. The first period spans from the termination of GS2 to the onset of interstadial warming at around 15kyBP. During this period, energy availability measured as ET was the main factor limiting population density across most of Europe (~30% of cells; Figs.\u00a05 and 6A). Likewise, TAP was also a strong limiting factor (~30% of cells; Figs.\u00a05, 6A). Furthermore, limitations imposed by winter temperatures could also be considered as likely limiting factors based on estimates of average conditions on a continental scale (Fig.\u00a06B). The range of experienced temperature conditions, represented by ET, can thus be seen as the major limiting factor shaping human population density in Europe between GS2 and the initiation of warming associated with GI1 (Figs.\u00a05, 6A). With temperature-related variables as the overwhelming limiting factor during this period, the emergence of sophisticated sewing techniques55 and pyrotechnology56 likely facilitated the persistence and even moderate expansion of populations at this time.\n\nPredictions based on 50th percentile generalised additive models (qGAM). Areas in grey represent the glacier extent as derived by ICE-6G-C104. Panels from top to bottom and left to right are Greenland Stadial-2, Greenland Interstadial-1, Greenland Stadial-1, Holocene initiation and Early Holocene.\n\nProportion of the ice-free area of Europe where each variable was estimated to be the factor limiting population density (A) and estimated population size based on the mean environmental conditions for each century (B).\n\nThe second period covers the rapid warming (GI1) as well as cooling (GS1) events between 14.7kyBP to 11.7kyBP. During this period, the overall importance of ET steadily decreased, with both TS (~32% of cells) and TAP becoming the main factors limiting population density (Figs.\u00a05 and 6A). Limitations imposed by winter temperatures remained in force in some parts of central Europe (Fig.\u00a05). The decrease of ET as a limiting factor indicates that during this period of rapid change, it was not temperature but its variability (measured as TS) and energy availability (due to the link between PET/TAP and productivity; refs. 57, 58) that determined human population density in Europe. Our models suggest that overall population densities increased (Fig.\u00a03), although a temporary reduction associated with GS1 cooling also stands clear.\n\nThe last period encompasses the early part of the Holocene from its onset at 11.7ky to 8kyBP. Here, TS and TAP were the main limiting factors (~60% of cells; Fig.\u00a06A), while the effect of ET became marginal (Fig.\u00a06A). These patterns indicate a complete shift from experienced temperature conditions to annual variability and available resources as the main limiting factors of European forager population densities during the Holocene. Such a shift is interesting as the Early Holocene also witnessed a significant reorganisation of forager socio-ecological systems towards more varied use of resources and more pronounced territoriality focused on spatially circumscribed and regionally available resources. It also saw a widespread shift from immediate-return to delayed-return economies increasingly characterised by a focus on food storage through smoking, roasting and fermentation that necessitated considerable investments in time and resources59,60,61. Furthermore, this shift also aligns with the idea that decreasing territory sizes and more marked boundary formation directly relate to the spatiotemporal dynamics of resource availability62.\n\nThe regional disaggregation of patterns in limiting factors shows strong differences between Fennoscandia, Southern, Central, and Eastern Europe (Figs.\u00a04 and 5). These patterns persist over time, with regional shifts linked to temperature change as a key feature. In Fennoscandia and the British Isles, ET was the main limiting factor for most of the Late Pleistocene. This pattern changed after the onset of the Holocene when TS and TAP became the dominant limiting factors. In Eastern and Western Europe, ET was the main limiting factor at the end of the GS2 but was replaced by TAP in the west and TS at the onset of the GI1. During GS1 and the Early Holocene, the main limiting factors were TAP and TS. In southern Europe, especially in the Mediterranean, TAP was the main limiting factor from GS2 onwards, supporting the idea that humans were closely tied to water resources in mid-latitudes63.\n\nOur analyses show that the main limiting factors constraining forager population densities across the Last Glacial-Interglacial Transition in Europe changed markedly over space (Fig.\u00a05) and time (Fig.\u00a06A). With these detailed dynamics in hand, we can now return to the archaeological record in search of material culture proxies that may have allowed these past communities to overcome these limiting factors64,65,66. These may have related to water availability (= containers) in the Mediterranean and are predicted to relate to temperature (= clothing or pyrotechnology) in higher latitudes. Certainly, for the latter two technologies, recent analyses suggest that the diversity and complexity of sewing technologies55, shelters67 and fireplace construction56 and use68,69 dates to the periods in which temperature-related variables acted as limiting factors. Conversely, where such technologies are absent in the archaeological record, we can also begin to think about population vulnerability to climatic factors at regional levels. Especially in higher latitudes, population fluctuations may have been pronounced at the sub-centennial scale, to the point of local population extirpations9,70.\n\nFinally, the marked shift in limiting factors at the onset of the Holocene may be reflected in the observed shift away from energy and heat conservation\u00a0technologies towards resource access and harvesting behaviours and technologies\u00a0via processing, and that the increased spatial heterogeneity of limiting factors also engendered stronger cultural differentiation at a regional scale. The spatiotemporal dynamics of resource availability directly impact land-use, mobility, territoriality and the formation of information networks in foragers62,71. In line with these expectations, regional cultural signatures became more pronounced\u00a0in the Holocene, and borders between different cultural zones were eventually more strongly articulated. These patterns could be seen as a response to the fundamental shift in limiting factors we have identified in our models.\n\nSeeking correlations between environmental variables and past human population densities is not a new endeavour. Following recent calls for more theoretically-informed rather than mere statistical explorations of this relationship13, we highlight that while the environment can be said to strongly constrain forager lifeways, precisely which aspects of the environment do so at any one place and time must be expected to vary. Our approach offers a way to infer the hierarchy of limiting factors and hence provides a spatiotemporal hypothesis for major selection pressures acting on forager populations in the past.\n\nIndependent palaeodemographic estimates broadly support our models, but many questions remain. Climate models, for instance, only indirectly capture the interaction of human population dynamics with changes in biodiversity and ecosystem compositions. In addition, the match between modelled population densities and the field-validated presence of Late Pleistocene/Early Holocene populations is not equally robust everywhere. These deviations may stimulate targeted field-testing that assesses whether and why population densities periodically fell short of or exceeded modelled values. In conjunction with legacy data derived from archives and the literature, such fieldwork can also shed light on the specific strategies these past foragers employed to mitigate the risks posed by specific limiting factors.\n\nSmall-scale societies have various adaptive options at their disposal (see ref. 72), most of which can be captured through archaeological proxies73,74,75. Our limiting factor model provides an explicit spatiotemporal hypothesis as to which risk mitigation measures could have been in use at any given time and place. By comparing our climate-driven estimates of population density to those estimated using the archaeological record we can identify potential mismatches suggestive of declines or increases not predicted by our models. By then identifying if shelters, fire or projectile technology, among other technologies, provided a given group with the capacity to buffer and perhaps even overcome climatically determined demographic limitations, we can better assess the adaptive role of specific cultural adaptations. The successful testing of such hypotheses would then shed light on these populations\u2019 resilience and adaptation\u2014or lack of it\u2014during this climatically and environmentally tumultuous time. Finally, the marked shifts in dominant limiting factors identified in our models map onto results of Late Pleistocene/Early Holocene Earth System tipping points recently discussed by Brovkin and colleagues76. Analogous to anthropogenic warming in the present, these periods of rapid and substantive climatic change are likely to have created challenges for contemporaneous forager populations. In an effort to align archaeological perspectives on climate change with the quandaries of our time (cf. ref. 77), future research would be well-advised to focus on such periods of major systemic transitions.",
+            "section_image": [
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-022-32750-x/MediaObjects/41467_2022_32750_Fig1_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-022-32750-x/MediaObjects/41467_2022_32750_Fig2_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-022-32750-x/MediaObjects/41467_2022_32750_Fig3_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-022-32750-x/MediaObjects/41467_2022_32750_Fig4_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-022-32750-x/MediaObjects/41467_2022_32750_Fig5_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-022-32750-x/MediaObjects/41467_2022_32750_Fig6_HTML.png"
+            ]
+        },
+        {
+            "section_name": "Methods",
+            "section_text": "We use ethnographic data on terrestrially adapted, mobile hunter-gatherers and their climatic space26 to construct a series of statistical models that predict hunter-gatherer population density based on one of ten climatic predictors (see Table\u00a01 for rezoning and source). While there are important caveats13, this approach builds on multiple ethnographic studies showing a link between climate on the one hand and hunter-gatherer diet, mobility, and demography on the other14,71,78,79,80,81,82. This statistical connection is the basis of recent studies focused on building complex multivariate models of population dynamics11,12,29. A benefit of our statistical approach is that it overcomes some significant limitations, such as lack of quantitative population size data based on the archaeological record itself or genetic data, each associated with its own limitations (as reviewed in refs. 2, 12,). Also, this univariate approach allows us to evaluate the absolute effects of multiple correlated environmental variables in the same study. Due to the nature of this study, it was not subject to ethics approval.\n\nWe omitted four observation classes in the original ethnographic dataset in defining the association between hunter-gatherer population density and climatic predictors. First, we removed observations associated with food producers. Second, sedentary populations or those that reside at a single location for >1 year. Third, populations using aquatic resources (>30% of their dietary protein comes from aquatic environments, as defined in refs. 83, 84,). Forth, we excluded all observations related to horse-riding populations. The filters employed here correspond to those used by ref. 12 to maximise the match between ethnographic data and the current knowledge of the highly mobile and overwhelmingly terrestrially oriented lifestyles of Late Pleistocene/Early Holocene hunter-gatherers in Europe. The implemented filters are less restrictive than those used by other studies that have sought to reconstruct forager population dynamics during this time85 and thus allow for a relatively large degree of behavioural variation. This is important given that increasing evidence of marine and lacustrine resource use is emerging for at least certain times and regions in Late Pleistocene Europe86,87,88, and that a marked diversification characterises the resource base of Early Holocene foragers. Finally, these filters remove any population using external supplements to their hunter-gatherer lifestyle, resulting in a database including information on 159 populations.\n\nWe used climate data on historical averages (1970\u20132000) for 19 climate variables (Table\u00a01) to build our ethnography-based population density models. These were obtained from the Worldclim version 2.189 at a 10-ArcMin resolution. Importantly, we used Worldclim data instead of climatic variables directly available from Binford\u2019s dataset to ensure comparability between climatic variables not in the database (i.e. net primary productivity, yearly limits). Equally importantly, this approach prevents any estimation biases due to differences between the data used to define climate-density relations and paleoclimatic surfaces used to estimate population density changes and limiting factors over time.\n\nInitially, we model how population densities of hunter-gatherer communities change along current environmental gradients using quantile generalised additive models (qGAMs). Modelling such dynamics using qGAMs offers a transparent way to determine the nonlinear changes in different percentiles of a response variable (= population densities) to one or multiple environmental variables and to understand the response-predictor dynamics outside of the mean of the data, making it useful in understanding outcomes that are non-normally distributed and that have nonlinear relationships with predictor variables30,31,90. This approach is commonly used in the ecological literature to determine the likelihood of occurrence or abundance of a given species under a particular environmental regime91,92,93,94,95 but has never before been applied to human palaeodemography.\n\nIn contrast to previous studies evaluating past human population density changes, we do not consider the synergies between multiple climatic variables when describing the relation between population densities and climate. We reasoned that using multiple predictors simultaneously in one model results in the evaluation of relative effects, and that these are contingent on the variables used\u00a0in the model. Furthermore, these relative effects do not capture the limiting effects exerted by a given variable, but the partial contributions of each evaluated variable to the deviation from the regional mean population density. We therefore instead focus on the individual effects of evaluated variables on different percentiles of population densities (90th, 50th and 10th percentiles) to (i) identify the most pronounced limiting factor determining palaeodemographic patterns and (ii) assess how the climatically possible maximum, average and minimum population densities changed in space and time. The shapes of the relation between population densities as a function of environmental variables were consistent for different percentiles (Fig.\u00a01).\n\nThe population density estimates derived from the ethnographic data followed a log-normal distribution, so these were log transformed for subsequent analyses, and a gaussian response distribution was used in our qGAMs models. Annual and yearly limit precipitation variables were similarly transformed. The ability of each of the evaluated variables to predict hunter-gatherer population densities was determined using the mean deviance explained (1\u2009\u2212\u2009(residual deviance/null deviance)), which in the context of this study is identical to a GLM unadjusted R2 96,97. These were calculated both for the whole dataset and using a 1000-fold cross-validation approach. We used an h-block approach32 to select our training and testing datasets to avoid any possible spatial autocorrelation in the datasets, as Binford\u2019s ethnographic database is known to consist of populations some of which are ecologically, demographically, and culturally related. Our h-block design uses the range of the population density variogram to define the minimum distance two observations can be apart (4500\u2009km) to be included in our training dataset. All models and prediction accuracy estimates were implemented in R (version 3.698) using the mgcv (version 1.8.2499) and qgam (version 1.3.2100) packages. Code to replicate these analyses is available from101.\n\nThe monthly average temperature and annual precipitation values for Europe for the 21ky to 8kyBP period come from the CCSM3 SynTrace paleoclimate simulations102. These were bias-corrected and downscaled to 0.5\u00b0\u2009\u00d7\u20090.5\u00b0 following the methods described by ref. 103. The paleoclimatic simulation data used here were originally generated to evaluate changes in European and North American fossil pollen data and vegetation novelty since the Last Glacial Maximum33. Source climate surfaces were aggregated to 500-years intervals from the original decadal averages of monthly values. We only considered locations not covered by ice, using a mask from the ICE-6G gridded data product104. As CCSM3 SynTrace paleoclimate simulations do not contain estimates of uncertainty, we cannot assess how variability in paleoclimatic model outcomes would propagate to our estimates of population density and limiting factors.\n\nTo generate population density estimates for each evaluated percentile within each variable/500-year interval, we used the average of the predictions resulting from running the 1000\u2009h-blocked qGAM models generated for each variable under the conditions of each 500-year interval. To mitigate any truncation or non-analogy artefacts, only those qGAM models where all historical means were contained under current conditions were projected into past climatic conditions. Population density for each cell within each 500-year interval point in time is then determined as the minimum predicted density for that cell. To calculate the human population size in Europe during every century, we multiplied the predicted population density in each cell where the predicted population density was above 0.2 individuals per 100\u2009km2 (the lowest densities recorded in the ethnographic dataset) by the land area of the corresponding cell to arrive at per cell population size. We then summed these values to arrive at the total population size for each century. As our objective was to establish the climatic variable that imposed the strongest constraints on hunter-gatherer population density at any one time, we determined for each of the three evaluated percentiles the variable estimating the lowest population density for a given cell at each evaluated time-period to be the limiting factor (the scarcest resource that would then limit population size cf. ref. 17). For each evaluated time-period, we summarised the proportion of the available land area (i.e. land area not covered by ice) where each of the assessed variables was determined to be the limiting factor.\n\nUncertainties in population density, size, occupied area, and limiting factor estimates were determined using a cross-validation approach, where model fitting was iterated 1000 times using a random sample (70%) of the ethnographic and climate data at each time step. Each model was used to hindcast population densities, estimate the percentage of inhabited land area and human population size, and define the relevant limiting factor. Uncertainty in continental-scale estimates of population densities, occupied area and population size was determined using 95% confidence intervals. The variable selected as the limiting factor in most cross-validation folds was selected as the limiting factor.\n\nTo assess the validity of our population density estimations, we use the International Union for Quaternary Science (INQUA) Radiocarbon Palaeolithic Europe Database v28105. Changes in the density of records are a useful continental-scale proxy measurement of prehistoric population size changes are ommonly used to describe prehistoric human population dynamics trends106,107,108,109,110,111. We extracted 14C dates from the INQUA Radiocarbon Palaeolithic Europe Database, aggregating these to the closest 1000 years in order to determine the match between our qGAM-derived populations' density estimates and those derived from the frequencies of radiocarbon dates between 21kyBP and 10kyBP, closely following ref. 12. This approach allowed validating our hindcasted estimates of absolute prehistoric population density since our model is not archaeologically informed, avoiding any possible circularity between model development and validation.\n\nWe also used site-based estimates of population density as derived using the Cologne Protocol by ref. 24. We focus on estimates of extended interconnected socio-economic areas (Core Areas) for five unequal time bands between 25kyBP and 11.7KyBP. Although ultimately also based on Binford26, these estimates present independently derived estimates of population density for Late Pleistocene populations in Europe.\n\nAll maps were generated using R (version 3.6; ref. 98) and the packages raster112 (version 3.4-13) and maptools113 (version 1.1-2). Code to replicate these is available from ref. 101.\n\nFurther information on research design is available in the\u00a0Nature Research Reporting Summary linked to this article.",
+            "section_image": []
+        },
+        {
+            "section_name": "Data availability",
+            "section_text": "All the datasets used in this study are publicly available. The \u2018Binford\u2019 ethnographic database26 is available from the Database of Places, Language, Culture and Environment (D-PLACE; https://d-place.org/contributions/Binford). Current and Late Quaternary environmental datasets are publicly available from the associated references. International Union for Quaternary Science (INQUA) Radiocarbon Palaeolithic Europe Database v28 is available from https://pandoradata.earth/dataset/radiocarbon-palaeolithic-europe-database-v28. Contemporary climate databases are available from WorldClim V.2.1. project (https://www.worldclim.org/data/worldclim21.html) and Late Pleistocene climate sources are available at https://doi.org/10.6084/m9.figshare.c.4673120.v2.\u00a0Source data are provided with this paper.",
+            "section_image": []
+        },
+        {
+            "section_name": "Code availability",
+            "section_text": "Code and subset of the data used in this study is available at https://doi.org/10.5281/zenodo.6962693101.",
+            "section_image": []
+        },
+        {
+            "section_name": "References",
+            "section_text": "Metcalf, C. J. & Pavard, S. Why evolutionary biologists should be demographers. Trends Ecol. Evol. 22, 205\u2013212 (2007).\n\nArticle\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nFrench, J. C., Riris, P., Fernandez-Lopez de Pablo, J., Lozano, S. & Silva, F. A manifesto for palaeodemography in the twenty-first century. Philos. Trans. R. Soc. Lond. B Biol. Sci. 376, 20190707 (2021).\n\nArticle\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nFrench, J. C. Demography and the Palaeolithic archaeological record. J. Archaeol. Method Th. 23, 150\u2013199 (2016).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHenrich, J. Demography and cultural evolution: how adaptive cultural processes can produce maladaptive losses - The Tasmanian case. Am. Antiquity 69, 197\u2013214 (2004).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nPowell, A., Shennan, S. & Thomas, M. G. Late Pleistocene demography and the appearance of modern human behavior. Science 324, 1298\u20131301 (2009).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nShennan, S. Demography and cultural innovation: a model and its implications for the emergence of modern human culture. Camb. Archaeol. J. 11, 5\u201316 (2001).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nJorgensen, E. K. The palaeodemographic and environmental dynamics of prehistoric Arctic Norway: an overview of human-climate covariation. Quat. Int. 549, 36\u201351 (2020).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nJorgensen, E. K. & Riede, F. Convergent catastrophes and the termination of the Arctic Norwegian Stone Age: a multi-proxy assessment of the demographic and adaptive responses of mid-Holocene collectors to biophysical forcing. Holocene 29, 1782\u20131800 (2019).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nRiede, F. Lateglacial and Postglacial Pioneers in Northern Europe (Archaeopress, 2014).\n\nTallavaara, M. & Sepp\u00e4, H. Did the mid-Holocene environmental changes cause the boom and bust of hunter-gatherer population size in eastern Fennoscandia? Holocene 22, 215\u2013225 (2011).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKavanagh, P. H. et al. Hindcasting global population densities reveals forces enabling the origin of agriculture. Nat. Hum. Behav. 2, 478\u2013484 (2018).\n\nArticle\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nTallavaara, M., Luoto, M., Korhonen, N., Jarvinen, H. & Seppa, H. Human population dynamics in Europe over the Last Glacial Maximum. Proc. Natl Acad. Sci. USA 112, 8232\u20138237 (2015).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBliege Bird, R. & Codding, B. F. Promise and peril of ecological and evolutionary modelling using cross-cultural datasets. Nat. Ecol. Evol. 6, 1\u20133 (2021).\n\nHamilton, M. J. & Tallavaara, M. Statistical inference, scale and noise in comparative anthropology. Nat. Ecol. Evol. 6, 122 (2022).\n\nArticle\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGurven, M. D. & Davison, R. J. Periodic catastrophes over human evolutionary history are necessary to explain the forager population paradox. Proc. Natl Acad. Sci. USA 116, 12758\u201312766 (2019).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nTallavaara, M. & Jorgensen, E. K. Why are population growth rate estimates of past and present hunter-gatherers so different? Philos. T R Soc. B 376, 20190708 (2021).\n\nBlackman, F. F. Optima and limiting factors. With two diagrams in the text. Ann. Bot. Lond. 19, 281\u2013296 (1905).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMaier, A. et al. Cultural evolution and environmental change in Central Europe between 40 and 15 ka. Quat. Int. 581-582, 225\u2013240 (2021).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nZhu, D., Galbraith, E. D., Reyes-Garcia, V. & Ciais, P. Global hunter-gatherer population densities constrained by influence of seasonality on diet composition. Nat. Ecol. Evol. 5, 1536 (2021).\n\nArticle\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBinford, L. R. Archaeology as anthropology. Am. Antiquity 28, 217\u2013225 (1962).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLowe, J. J. et al. Synchronisation of palaeoenvironmental events in the North Atlantic region during the Last Termination: a revised protocol recommended by the INTIMATE group. Quat. Sci. Rev. 27, 6\u201317 (2008).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBocquet-Appel, J. P., Demars, P. Y., Noiret, L. & Dobrowsky, D. Estimates of upper Palaeolithic meta-population size in Europe from archaeological data. J. Archaeol. Sci. 32, 1656\u20131668 (2005).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nFort, J., Pujol, T. & Cavalli-Sforza, L. L. Palaeolithic populations and waves of advance (Human range expansions). Camb. Archaeol. J. 14, 53\u201361 (2004).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSchmidt, I. et al. Approaching prehistoric demography: proxies, scales and scope of the Cologne Protocol in European contexts. Philos. Trans. R. Soc. Lond. B Biol. Sci. 376, 20190714 (2021).\n\nArticle\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nde Pablo, J. F. L. et al. Palaeodemographic modelling supports a population bottleneck during the Pleistocene-Holocene transition in Iberia. Nat. Commun. 10, 1872 (2019).\n\nBinford, L. R. Constructing Frames of Reference: An Analytical Method for Archaeological Theory Building Using Ethnographic and Environmental Data Sets. (Univ. California Press, 2019).\n\nJohnson, A. L. Exploring adaptive variation among hunter-gatherers with Binford\u2019s frames of reference. J. Archaeol. Res. 22, 1\u201342 (2014).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGraham, M. H. Confronting multicollinearity in ecological multiple regression. Ecology 84, 2809\u20132815 (2003).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nTallavaara, M., Eronen, J. T. & Luoto, M. Productivity, biodiversity, and pathogens influence the global hunter-gatherer population density. Proc. Natl Acad. Sci. USA 115, 1232\u20131237 (2018).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nCade, B. S. & Noon, B. R. A gentle introduction to quantile regression for ecologists. Front Ecol. Environ. 1, 412\u2013420 (2003).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nCade, B. S., Terrell, J. W. & Schroeder, R. L. Estimating effects of limiting factors with regression quantiles. Ecology 80, 311\u2013323 (1999).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBurman, P., Chow, E. & Nolan, D. A cross-validatory method for dependent data. Biometrika 81, 351\u2013358 (1994).\n\nArticle\u00a0\n    MathSciNet\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBurke, K. D. et al. Differing climatic mechanisms control transient and accumulated vegetation novelty in Europe and eastern North America. Philos. Trans. R. Soc. Lond. B Biol. Sci. 374, 20190218 (2019).\n\nArticle\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nCurrie, D. J. Energy and large-scale patterns of animal-species and plant-species richness. Am. Nat. 137, 27\u201349 (1991).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nFranklin, J. Mapping Species Distributions: Spatial Inference and Prediction (Cambridge Univ. Press, 2010).\n\nHarcourt, A. Human Biogeography (Univ. California Press, 2012).\n\nMarlowe, F. W. Hunter-gatherers and human evolution. Evol. Anthropol. 14, 54\u201367 (2005).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBelovsky, G. E. An optimal foraging-based model of hunter-gatherer population-dynamics. J. Anthropol. Archaeol. 7, 329\u2013372 (1988).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWilliams, J. W. & Jackson, S. T. Novel climates, no-analog communities, and ecological surprises. Front. Ecol. Environ. 5, 475\u2013482 (2007).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nOhlemuller, R. Climate. Running out of climate space. Science 334, 613\u2013614 (2011).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWilliams, J. W., Jackson, S. T. & Kutzbach, J. E. Projected distributions of novel and disappearing climates by 2100 AD. Proc. Natl Acad. Sci. USA 104, 5738\u20135742 (2007).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWarren, D. L., Glor, R. E. & Turelli, M. Environmental niche equivalency versus conservatism: quantitative approaches to niche evolution. Evolution 62, 2868\u20132883 (2008).\n\nArticle\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBroennimann, O. et al. Measuring ecological niche overlap from occurrence and spatial environmental data. Glob. Ecol. Biogeogr. 21, 481\u2013497 (2012).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWarren, D. L., Cardillo, M., Rosauer, D. F. & Bolnick, D. I. Mistaking geography for biology: inferring processes from species distributions. Trends Ecol. Evol. 29, 572\u2013580 (2014).\n\nArticle\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWobst, H. M. The archaeo-ethnology of hunter-gatherers or the tyranny of the ethnographic record in archaeology. Am Antiquity 43, 303\u2013309 (1978).\n\nMaier, A. et al. Demographic estimates of hunter-gatherers during the Last Glacial Maximum in Europe against the background of palaeoenvironmental data. Quat. Int. 425, 49\u201361 (2016).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nRiede, F. Oxford Handbook of the Archaeology and Anthropology of Hunter-Gatherers (Oxford Univ. Press, 2014).\n\nJochim, M., Herhahn, C. & Starr, H. The Magdalenian colonization of southern Germany. Am. Anthropol. 101, 129\u2013142 (1999).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nArts, N. & Deeben, J. On the Northwestern Border of Late Magdalenian Territory: Ecology and Archaeology of Early Late Glacial Band Societies in Northwestern Europe. In Late Glacial in Central Europe. Culture and\u00a0Environment. (eds Burdukiewicz, J. M. & Kobusiewicz, M.) (Polska Akademia Nauk, Warszawa 1987).\n\nMaier, A. Population and settlement dynamics from the Gravettian to the Magdalenian. Mitteilungen der Ges. f.\u00fcr. Urgesch. 26, 83\u2013101 (2017).\n\n\n                    Google Scholar\u00a0\n                \n\nMaier, A., Liebermann, C. & Pfeifer, S. J. Beyond the Alps and Tatra Mountains-the 20-14 ka repopulation of the northern mid-latitudes as inferred from palimpsests deciphered with keys from Western and Central Europe. J. Paleolit. Archaeol. 3, 398\u2013452 (2020).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGamble, C., Davies, W., Pettitt, P. & Richards, M., Climate change. and evolving human diversity in Europe during the last glacial. Philos. Trans. R. Soc. Lond. B Biol. Sci. 359, 243\u2013253 (2004).\n\nArticle\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHousley, R. A., Gamble, C. S., Street, M. & Pettitt, P. Proceedings of the Prehistoric Society. (Cambridge Univ. Press).\n\nBellwood, P. S. First Farmers: the Origins of Agricultural Societies. (Blackwell, Oxford 2005).\n\nd\u2019Errico, F. et al. The origin and evolution of sewing technologies in Eurasia and North America. J. Hum. Evol. 125, 71\u201386 (2018).\n\nArticle\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMoseler, F. Brandstrukturen im sp\u00e4ten Magdal\u00e9nien: Betrieb, Nutzung und Funktion (Verlag des R\u00f6misch-Germanischen Zentralmuseums, 2020).\n\nSimova, I. & Storch, D. The enigma of terrestrial primary productivity: measurements, models, scales and the diversity-productivity relationship. Ecography 40, 239\u2013252 (2017).\n\nRosenzweig, M. L. Net primary productivity of terrestrial communities - prediction from climatological data. Am. Nat. 102, 67 (1968).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nJensen, H. J. & M\u00f8berg, T. Et r\u00f8geri fra \u00e6ldre stenalder ved B\u00f8lling S\u00f8? Midtjyske Fortaellinger 2007, 51\u201362 (2008).\n\nHolst, D. Hazelnut economy of early Holocene hunter-gatherers: a case study from Mesolithic Duvensee, northern Germany. J. Archaeol. Sci. 37, 2871\u20132880 (2010).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBoethius, A. Something rotten in Scandinavia: the world\u2019s earliest evidence of fermentation. J. Archaeol. Sci. 66, 169\u2013180 (2016).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nDyson\u2010Hudson, R. & Smith, E. A. Human territoriality: an ecological reassessment. Am. Anthropol. 80, 21\u201341 (1978).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nFinlayson, C. The water optimisation hypothesis and the human occupation of the mid-latitude belt in the Pleistocene. Quat. Int 300, 22\u201331 (2013).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLaland, K. N. & Brown, G. R. Niche construction, human behavior, and the adaptive-lag hypothesis. Evol. Anthropol. 15, 95\u2013104 (2006).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLaland, K. N. & O\u2019Brien, M. J. Niche construction theory and archaeology. J. Archaeol. Method Th. 17, 303\u2013322 (2010).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nRiede, F. Handbook of Evolutionary Research in Archaeology (Springer, 2019).\n\nJ\u00f6ris, O. & Terberger, T. Zur Rekonstruktion eines Zeltes mit Trapezf\u00f6rmigem Grundriss am Magdal\u00e9nien-Fundplatz G\u00f6nnersdorf/Mittelrhein: Eine\u00bb Quadratur des Kreises \u00ab? Arch.\u00e4ologisches Korrespondenzblatt 31, 163\u2013172 (2001).\n\n\n                    Google Scholar\u00a0\n                \n\nSalomon, H., Vignaud, C., Lahlil, S. & Menguy, N. Solutrean and Magdalenian ferruginous rocks heat-treatment: accidental and/or deliberate action? J. Archaeol. Sci. 55, 100\u2013112 (2015).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nNakazawa, Y., Straus, L. G., Gonzalez-Morales, M. R., Solana, D. C. & Saiz, J. C. On stone-boiling technology in the Upper Paleolithic: behavioral implications from an Early Magdalenian hearth in El Miron Cave, Cantabria, Spain. J. Archaeol. Sci. 36, 684\u2013693 (2009).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nPedersen, J., Maier, A. & Riede, F. A punctuated model for the colonisation of the Late Glacial margins of northern Europe by Hamburgian hunter-gatherers. Quart.\u00e4r. 65, 85\u2013104 (2018).\n\n\n                    Google Scholar\u00a0\n                \n\nWhallon, R. Social networks and information: non-\u201cutilitarian\u201d mobility among hunter-gatherers. J. Anthropol. Archaeol. 25, 259\u2013270 (2006).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLeal Filho, W. et al. Impacts of climate change to African indigenous communities and examples of adaptation responses. Nat. Commun. 12, 6224 (2021).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHeitz, C. F., Hinz, M., Laabs, J. & Hafner, A. Mobility as resilience capacity in northern Alpine Neolithic settlement communities. Archaeol. Rev. Camb. 36, 75\u2013106 (2021).\n\n\n                    Google Scholar\u00a0\n                \n\nRiede, F., Oetelaar, G. A. & VanderHoek, R. From crisis to collapse in hunter-gatherer societies. A comparative investigation of the cultural impacts of three large volcanic eruptions on past hunter-gatherers. Crisis to Collapse\u2013The Archaeology of Social Breakdown. Louvain-la-Neuve: UCL Presses Universitaires De Louvian 23\u201339 (2017).\n\nHalstead, P., O\u2019Shea, J. & O\u2019Shea, J. M. Bad Year Economics: Cultural Responses to Risk and Uncertainty. (Cambridge Univ. Press, 2004).\n\nBrovkin, V. et al. Past abrupt changes, tipping points and cascading impacts in the Earth system. Nat. Geosci. 14, 550\u2013558 (2021).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBurke, A. et al. The archaeology of climate change: the case for cultural diversity. Proc Natl Acad Sci USA 118, e2108537118 (2021).\n\nBinford, L. R. Willow smoke and dogs tails - Hunter-gatherer settlement systems and archaeological site formation. Am. Antiquity 45, 4\u201320 (1980).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBirdsell, J. B. Some environmental and cultural factors influencing the structuring of Australian aboriginal populations. Am. Nat. 87, 171\u2013207 (1953).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKelly, R. L. The Lifeways of Hunter-Gatherers: The Foraging Spectrum (Cambridge Univ. Press, 2013).\n\nPenington, R. Hunter-gatherer demography. In Hunter-Gatherers: An Interdisciplinary Perspective. (eds. Panter-Brick, C., Layton, R. H. & Rowley-Conwy, P.) (Cambridge University Press, Cambridge, 2001).\n\nWobst, H. M. Locational relationships in Paleolithic society. J. Hum. Evol. 5, 49\u201358 (1976).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nRichards, M. P., Pettitt, P. B., Stiner, M. C. & Trinkaus, E. Stable isotope evidence for increasing dietary breadth in the European mid-Upper Paleolithic. Proc. Natl Acad. Sci. USA 98, 6528\u20136532 (2001).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nDrucker, D. & Bocherens, H. Carbon and nitrogen stable isotopes as tracers of change in diet breadth during Middle and Upper Palaeolithic in Europe. Int J. Osteoarchaeol. 14, 162\u2013177 (2004).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKretschmer, I. Demographische Untersuchungen zu Bev\u00f6lkerungsdichten, Mobilit\u00e4t und Landnutzungsmustern im sp\u00e4ten Jungpal\u00e4olithikum (Verlag Marie Leidorf GmbH, 2015).\n\nLangley, M. C. & Street, M. Long range inland-coastal networks during the Late Magdalenian: evidence for individual acquisition of marine resources at Andernach-Martinsberg, German Central Rhineland. J. Hum. Evol. 64, 457\u2013465 (2013).\n\nArticle\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLanczont, M. et al. Late Glacial environment and human settlement of the Central Western Carpathians: a case study of the Nowa Biala 1 open-air site (Podhale Region, southern Poland). Quat. Int 512, 113\u2013132 (2019).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nCziesla, E. Robbenjagd in Brandenburg? Gedanken zur Verwendung gro\u00dfer Widerhakenspitzen. Ethnographisch-archaologische Z. 48, 1\u201348 (2007).\n\n\n                    Google Scholar\u00a0\n                \n\nFick, S. E. & Hijmans, R. J. WorldClim 2: new 1\u2010km spatial resolution climate surfaces for global land areas. Int. J. Climatol. 37, 4302\u20134315 (2017).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLe Cook, B. & Manning, W. G. Thinking beyond the mean: a practical guide for using quantile regression methods for health services research. Shanghai Arch. Psychiatry 25, 55 (2013).\n\nPubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nYee, T. W. & Mitchell, N. D. Generalized additive-models in plant ecology. J. Veg. Sci. 2, 587\u2013602 (1991).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGuisan, A., Edwards, T. C. & Hastie, T. Generalized linear and generalized additive models in studies of species distributions: setting the scene. Ecol. Model 157, 89\u2013100 (2002).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nFewster, R. M., Buckland, S. T., Siriwardena, G. M., Baillie, S. R. & Wilson, J. D. Analysis of population trends for farmland birds using generalized additive models. Ecology 81, 1970\u20131984 (2000).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nDrexler, M. & Ainsworth, C. H. Generalized additive models used to predict species abundance in the Gulf of Mexico: an ecosystem modeling tool. PLos ONE 8, e64458 (2013).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMoisen, G. G. & Frescino, T. S. Comparing five modelling techniques for predicting forest characteristics. Ecol. Model 157, 209\u2013225 (2002).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWood, S. N. Generalized Additive Models: An Introduction with R (CRC Press, 2006).\n\nZuur, A. F. A Beginner\u2019s Guide to Generalized Additive Models with R (Highland Statistics Limited, 2012).\n\nTeam, R. C. R: a language and environment for statistical computing. (2013).\n\nWood, S. N. Generalized Additive Models: An Introduction with R (CRC Press, 2017).\n\nFasiolo, M., Wood, S. N., Zaffran, M., Nedellec, R. & Goude, Y. Fast calibrated additive quantile regression. J. Am. Stat. Assoc. 116, 1402\u20131412 (2021).\n\nArticle\u00a0\n    MathSciNet\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nAlejoOrdonez/PaleoPopDen: (Version NatCommV0) [Computer software]. Zenodo. https://doi.org/10.5281/ZENODO.6962693 (2022).\n\nLiu, Z. et al. Transient simulation of last deglaciation with a new mechanism for Bolling-Allerod warming. Science 325, 310\u2013314 (2009).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLorenz, D. J., Nieto-Lugilde, D., Blois, J. L., Fitzpatrick, M. C. & Williams, J. W. Downscaled and debiased climate simulations for North America from 21,000 years ago to 2100AD. Sci. Data 3, 160048 (2016).\n\nArticle\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nPeltier, W. R., Argus, D. & Drummond, R. Space geodesy constrains ice age terminal deglaciation: the global ICE\u20106G_C (VM5a) model. J. Geophys. Res. Solid Earth 120, 450\u2013487 (2015).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nVermeersch, P. M. European population changes during the Marine Isotope Stages 2 and 3. Quat. Int 137, 77\u201385 (2005).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGamble, C., Davies, W., Pettitt, P., Hazelwood, L. & Richards, M. The archaeological and genetic foundations of the European population during the late glacial: Implications for \u2018agricultural thinking\u2019. Camb. Archaeol. J. 15, 193\u2013223 (2005).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSteele, J. Radiocarbon dates as data: quantitative strategies for estimating colonization front speeds and event densities. J. Archaeol. Sci. 37, 2017\u20132030 (2010).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nShennan, S. et al. Regional population collapse followed initial agriculture booms in mid-Holocene Europe. Nat. Commun. 4, 2486 (2013).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSurovell, T. A., Finley, J. B., Smith, G. M., Brantingham, P. J. & Kelly, R. Correcting temporal frequency distributions for taphonomic bias. J. Archaeol. Sci. 36, 1715\u20131724 (2009).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWilliams, A. N. The use of summed radiocarbon probability distributions in archaeology: a review of methods. J. Archaeol. Sci. 39, 578\u2013589 (2012).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKelly, R. L., Surovell, T. A., Shuman, B. N. & Smith, G. M. A continuous climatic impact on Holocene human population in the Rocky Mountains. Proc. Natl Acad. Sci. USA 110, 443\u2013447 (2013).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHijmans, R. J. et al. Package \u2018raster\u2019. R package 734, (2015).\n\nLewin-Koh, N. J. et al. Package \u2018maptools\u2019. Internet: http://cran.r-project.org/web/packages/maptools/maptools.pdf (2012).\n\nThornthwaite, C. W. An approach toward a rational classification of climate. Geogr. Rev. 38, 55\u201394 (1948).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLieth, H.Primary Productivity of the Biosphere (Springer, 1975).\n\nDownload references",
+            "section_image": []
+        },
+        {
+            "section_name": "Acknowledgements",
+            "section_text": "A.O. was supported by the AUFF Starting Grant (AUFF-F-2018-7-8). F.R.\u2019s contribution is part of CLIOARCH, an ERC Consolidator Grant project that has received funding from the European Research Council (ERC) under the European Union\u2019s Horizon 2020 research and innovation programme (grant agreement No. 817564).",
+            "section_image": []
+        },
+        {
+            "section_name": "Author information",
+            "section_text": "Center for Biodiversity Dynamics in a Changing World, Aarhus University, Aarhus, Denmark\n\nAlejandro Ordonez\u00a0&\u00a0Felix Riede\n\nDepartment of Biology, Aarhus University, Aarhus, Denmark\n\nAlejandro Ordonez\n\nCenter for Sustainable Landscapes under Global Change, Aarhus University, Aarhus, Denmark\n\nAlejandro Ordonez\n\nDepartment of Archaeology and Heritage Studies, Aarhus University, Aarhus, Denmark\n\nFelix Riede\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nA.O.: Conceptualisation; Methodology; Formal analysis; Resources; writing\u2014original draft, writing\u2014review and editing; Visualisation. F.R.: Conceptualisation; Methodology; writing\u2014original draft, writing\u2014review and editing.\n\nCorrespondence to\n                Alejandro Ordonez.",
+            "section_image": []
+        },
+        {
+            "section_name": "Ethics declarations",
+            "section_text": "The authors declare no competing interests.",
+            "section_image": []
+        },
+        {
+            "section_name": "Peer review",
+            "section_text": "Nature Communications thanks Ron Pinhasi, Miikka Tallavaara and Dan Zhu for their contribution to the peer review of this work.\u00a0Peer reviewer reports are available.",
+            "section_image": []
+        },
+        {
+            "section_name": "Additional information",
+            "section_text": "Publisher\u2019s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.",
+            "section_image": []
+        },
+        {
+            "section_name": "Source data",
+            "section_text": "",
+            "section_image": []
+        },
+        {
+            "section_name": "Rights and permissions",
+            "section_text": "Open Access  This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article\u2019s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article\u2019s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.\n\nReprints and permissions",
+            "section_image": []
+        },
+        {
+            "section_name": "About this article",
+            "section_text": "Ordonez, A., Riede, F. Changes in limiting factors for forager population dynamics in Europe across the last glacial-interglacial transition.\n                    Nat Commun 13, 5140 (2022). https://doi.org/10.1038/s41467-022-32750-x\n\nDownload citation\n\nReceived: 15 December 2021\n\nAccepted: 16 August 2022\n\nPublished: 06 September 2022\n\nVersion of record: 06 September 2022\n\nDOI: https://doi.org/10.1038/s41467-022-32750-x\n\nAnyone you share the following link with will be able to read this content:\n\nSorry, a shareable link is not currently available for this article.\n\n\n\n\n                            Provided by the Springer Nature SharedIt content-sharing initiative\n                        ",
+            "section_image": [
+                "https://data:image/svg+xml;base64,PHN2ZyBoZWlnaHQ9IjgxIiB3aWR0aD0iNTciIHhtbG5zPSJodHRwOi8vd3d3LnczLm9yZy8yMDAwL3N2ZyI+PGcgZmlsbD0ibm9uZSIgZmlsbC1ydWxlPSJldmVub2RkIj48cGF0aCBkPSJtMTcuMzUgMzUuNDUgMjEuMy0xNC4ydi0xNy4wM2gtMjEuMyIgZmlsbD0iIzk4OTg5OCIvPjxwYXRoIGQ9Im0zOC42NSAzNS40NS0yMS4zLTE0LjJ2LTE3LjAzaDIxLjMiIGZpbGw9IiM3NDc0NzQiLz48cGF0aCBkPSJtMjggLjVjLTEyLjk4IDAtMjMuNSAxMC41Mi0yMy41IDIzLjVzMTAuNTIgMjMuNSAyMy41IDIzLjUgMjMuNS0xMC41MiAyMy41LTIzLjVjMC02LjIzLTIuNDgtMTIuMjEtNi44OC0xNi42Mi00LjQxLTQuNC0xMC4zOS02Ljg4LTE2LjYyLTYuODh6bTAgNDEuMjVjLTkuOCAwLTE3Ljc1LTcuOTUtMTcuNzUtMTcuNzVzNy45NS0xNy43NSAxNy43NS0xNy43NSAxNy43NSA3Ljk1IDE3Ljc1IDE3Ljc1YzAgNC43MS0xLjg3IDkuMjItNS4yIDEyLjU1cy03Ljg0IDUuMi0xMi41NSA1LjJ6IiBmaWxsPSIjNTM1MzUzIi8+PHBhdGggZD0ibTQxIDM2Yy01LjgxIDYuMjMtMTUuMjMgNy40NS0yMi40MyAyLjktNy4yMS00LjU1LTEwLjE2LTEzLjU3LTcuMDMtMjEuNWwtNC45Mi0zLjExYy00Ljk1IDEwLjctMS4xOSAyMy40MiA4Ljc4IDI5LjcxIDkuOTcgNi4zIDIzLjA3IDQuMjIgMzAuNi00Ljg2eiIgZmlsbD0iIzljOWM5YyIvPjxwYXRoIGQ9Im0uMiA1OC40NWMwLS43NS4xMS0xLjQyLjMzLTIuMDFzLjUyLTEuMDkuOTEtMS41Yy4zOC0uNDEuODMtLjczIDEuMzQtLjk0LjUxLS4yMiAxLjA2LS4zMiAxLjY1LS4zMi41NiAwIDEuMDYuMTEgMS41MS4zNS40NC4yMy44MS41IDEuMS44MWwtLjkxIDEuMDFjLS4yNC0uMjQtLjQ5LS40Mi0uNzUtLjU2LS4yNy0uMTMtLjU4LS4yLS45My0uMi0uMzkgMC0uNzMuMDgtMS4wNS4yMy0uMzEuMTYtLjU4LjM3LS44MS42Ni0uMjMuMjgtLjQxLjYzLS41MyAxLjA0LS4xMy40MS0uMTkuODgtLjE5IDEuMzkgMCAxLjA0LjIzIDEuODYuNjggMi40Ni40NS41OSAxLjA2Ljg4IDEuODQuODguNDEgMCAuNzctLjA3IDEuMDctLjIzcy41OS0uMzkuODUtLjY4bC45MSAxYy0uMzguNDMtLjguNzYtMS4yOC45OS0uNDcuMjItMSAuMzQtMS41OC4zNC0uNTkgMC0xLjEzLS4xLTEuNjQtLjMxLS41LS4yLS45NC0uNTEtMS4zMS0uOTEtLjM4LS40LS42Ny0uOS0uODgtMS40OC0uMjItLjU5LS4zMy0xLjI2LS4zMy0yLjAyem04LjQtNS4zM2gxLjYxdjIuNTRsLS4wNSAxLjMzYy4yOS0uMjcuNjEtLjUxLjk2LS43MnMuNzYtLjMxIDEuMjQtLjMxYy43MyAwIDEuMjcuMjMgMS42MS43MS4zMy40Ny41IDEuMTQuNSAyLjAydjQuMzFoLTEuNjF2LTQuMWMwLS41Ny0uMDgtLjk3LS4yNS0xLjIxLS4xNy0uMjMtLjQ1LS4zNS0uODMtLjM1LS4zIDAtLjU2LjA4LS43OS4yMi0uMjMuMTUtLjQ5LjM2LS43OC42NHY0LjhoLTEuNjF6bTcuMzcgNi40NWMwLS41Ni4wOS0xLjA2LjI2LTEuNTEuMTgtLjQ1LjQyLS44My43MS0xLjE0LjI5LS4zLjYzLS41NCAxLjAxLS43MS4zOS0uMTcuNzgtLjI1IDEuMTgtLjI1LjQ3IDAgLjg4LjA4IDEuMjMuMjQuMzYuMTYuNjUuMzguODkuNjdzLjQyLjYzLjU0IDEuMDNjLjEyLjQxLjE4Ljg0LjE4IDEuMzIgMCAuMzItLjAyLjU3LS4wNy43NmgtNC4zNmMuMDcuNjIuMjkgMS4xLjY1IDEuNDQuMzYuMzMuODIuNSAxLjM4LjUuMjkgMCAuNTctLjA0LjgzLS4xM3MuNTEtLjIxLjc2LS4zN2wuNTUgMS4wMWMtLjMzLjIxLS42OS4zOS0xLjA5LjUzLS40MS4xNC0uODMuMjEtMS4yNi4yMS0uNDggMC0uOTItLjA4LTEuMzQtLjI1LS40MS0uMTYtLjc2LS40LTEuMDctLjctLjMxLS4zMS0uNTUtLjY5LS43Mi0xLjEzLS4xOC0uNDQtLjI2LS45NS0uMjYtMS41MnptNC42LS42MmMwLS41NS0uMTEtLjk4LS4zNC0xLjI4LS4yMy0uMzEtLjU4LS40Ny0xLjA2LS40Ny0uNDEgMC0uNzcuMTUtMS4wNy40NS0uMzEuMjktLjUuNzMtLjU4IDEuM3ptMi41LjYyYzAtLjU3LjA5LTEuMDguMjgtMS41My4xOC0uNDQuNDMtLjgyLjc1LTEuMTNzLjY5LS41NCAxLjEtLjcxYy40Mi0uMTYuODUtLjI0IDEuMzEtLjI0LjQ1IDAgLjg0LjA4IDEuMTcuMjNzLjYxLjM0Ljg1LjU3bC0uNzcgMS4wMmMtLjE5LS4xNi0uMzgtLjI4LS41Ni0uMzctLjE5LS4wOS0uMzktLjE0LS42MS0uMTQtLjU2IDAtMS4wMS4yMS0xLjM1LjYzLS4zNS40MS0uNTIuOTctLjUyIDEuNjcgMCAuNjkuMTcgMS4yNC41MSAxLjY2LjM0LjQxLjc4LjYyIDEuMzIuNjIuMjggMCAuNTQtLjA2Ljc4LS4xNy4yNC0uMTIuNDUtLjI2LjY0LS40MmwuNjcgMS4wM2MtLjMzLjI5LS42OS41MS0xLjA4LjY1LS4zOS4xNS0uNzguMjMtMS4xOC4yMy0uNDYgMC0uOS0uMDgtMS4zMS0uMjQtLjQtLjE2LS43NS0uMzktMS4wNS0uN3MtLjUzLS42OS0uNy0xLjEzYy0uMTctLjQ1LS4yNS0uOTYtLjI1LTEuNTN6bTYuOTEtNi40NWgxLjU4djYuMTdoLjA1bDIuNTQtMy4xNmgxLjc3bC0yLjM1IDIuOCAyLjU5IDQuMDdoLTEuNzVsLTEuNzctMi45OC0xLjA4IDEuMjN2MS43NWgtMS41OHptMTMuNjkgMS4yN2MtLjI1LS4xMS0uNS0uMTctLjc1LS4xNy0uNTggMC0uODcuMzktLjg3IDEuMTZ2Ljc1aDEuMzR2MS4yN2gtMS4zNHY1LjZoLTEuNjF2LTUuNmgtLjkydi0xLjJsLjkyLS4wN3YtLjcyYzAtLjM1LjA0LS42OC4xMy0uOTguMDgtLjMxLjIxLS41Ny40LS43OXMuNDItLjM5LjcxLS41MWMuMjgtLjEyLjYzLS4xOCAxLjA0LS4xOC4yNCAwIC40OC4wMi42OS4wNy4yMi4wNS40MS4xLjU3LjE3em0uNDggNS4xOGMwLS41Ny4wOS0xLjA4LjI3LTEuNTMuMTctLjQ0LjQxLS44Mi43Mi0xLjEzLjMtLjMxLjY1LS41NCAxLjA0LS43MS4zOS0uMTYuOC0uMjQgMS4yMy0uMjRzLjg0LjA4IDEuMjQuMjRjLjQuMTcuNzQuNCAxLjA0Ljcxcy41NC42OS43MiAxLjEzYy4xOS40NS4yOC45Ni4yOCAxLjUzcy0uMDkgMS4wOC0uMjggMS41M2MtLjE4LjQ0LS40Mi44Mi0uNzIgMS4xM3MtLjY0LjU0LTEuMDQuNy0uODEuMjQtMS4yNC4yNC0uODQtLjA4LTEuMjMtLjI0LS43NC0uMzktMS4wNC0uN2MtLjMxLS4zMS0uNTUtLjY5LS43Mi0xLjEzLS4xOC0uNDUtLjI3LS45Ni0uMjctMS41M3ptMS42NSAwYzAgLjY5LjE0IDEuMjQuNDMgMS42Ni4yOC40MS42OC42MiAxLjE4LjYyLjUxIDAgLjktLjIxIDEuMTktLjYyLjI5LS40Mi40NC0uOTcuNDQtMS42NiAwLS43LS4xNS0xLjI2LS40NC0xLjY3LS4yOS0uNDItLjY4LS42My0xLjE5LS42My0uNSAwLS45LjIxLTEuMTguNjMtLjI5LjQxLS40My45Ny0uNDMgMS42N3ptNi40OC0zLjQ0aDEuMzNsLjEyIDEuMjFoLjA1Yy4yNC0uNDQuNTQtLjc5Ljg4LTEuMDIuMzUtLjI0LjctLjM2IDEuMDctLjM2LjMyIDAgLjU5LjA1Ljc4LjE0bC0uMjggMS40LS4zMy0uMDljLS4xMS0uMDEtLjIzLS4wMi0uMzgtLjAyLS4yNyAwLS41Ni4xLS44Ni4zMXMtLjU1LjU4LS43NyAxLjF2NC4yaC0xLjYxem0tNDcuODcgMTVoMS42MXY0LjFjMCAuNTcuMDguOTcuMjUgMS4yLjE3LjI0LjQ0LjM1LjgxLjM1LjMgMCAuNTctLjA3LjgtLjIyLjIyLS4xNS40Ny0uMzkuNzMtLjczdi00LjdoMS42MXY2Ljg3aC0xLjMybC0uMTItMS4wMWgtLjA0Yy0uMy4zNi0uNjMuNjQtLjk4Ljg2LS4zNS4yMS0uNzYuMzItMS4yNC4zMi0uNzMgMC0xLjI3LS4yNC0xLjYxLS43MS0uMzMtLjQ3LS41LTEuMTQtLjUtMi4wMnptOS40NiA3LjQzdjIuMTZoLTEuNjF2LTkuNTloMS4zM2wuMTIuNzJoLjA1Yy4yOS0uMjQuNjEtLjQ1Ljk3LS42My4zNS0uMTcuNzItLjI2IDEuMS0uMjYuNDMgMCAuODEuMDggMS4xNS4yNC4zMy4xNy42MS40Ljg0LjcxLjI0LjMxLjQxLjY4LjUzIDEuMTEuMTMuNDIuMTkuOTEuMTkgMS40NCAwIC41OS0uMDkgMS4xMS0uMjUgMS41Ny0uMTYuNDctLjM4Ljg1LS42NSAxLjE2LS4yNy4zMi0uNTguNTYtLjk0LjczLS4zNS4xNi0uNzIuMjUtMS4xLjI1LS4zIDAtLjYtLjA3LS45LS4ycy0uNTktLjMxLS44Ny0uNTZ6bTAtMi4zYy4yNi4yMi41LjM3LjczLjQ1LjI0LjA5LjQ2LjEzLjY2LjEzLjQ2IDAgLjg0LS4yIDEuMTUtLjYuMzEtLjM5LjQ2LS45OC40Ni0xLjc3IDAtLjY5LS4xMi0xLjIyLS4zNS0xLjYxLS4yMy0uMzgtLjYxLS41Ny0xLjEzLS41Ny0uNDkgMC0uOTkuMjYtMS41Mi43N3ptNS44Ny0xLjY5YzAtLjU2LjA4LTEuMDYuMjUtMS41MS4xNi0uNDUuMzctLjgzLjY1LTEuMTQuMjctLjMuNTgtLjU0LjkzLS43MXMuNzEtLjI1IDEuMDgtLjI1Yy4zOSAwIC43My4wNyAxIC4yLjI3LjE0LjU0LjMyLjgxLjU1bC0uMDYtMS4xdi0yLjQ5aDEuNjF2OS44OGgtMS4zM2wtLjExLS43NGgtLjA2Yy0uMjUuMjUtLjU0LjQ2LS44OC42NC0uMzMuMTgtLjY5LjI3LTEuMDYuMjctLjg3IDAtMS41Ni0uMzItMi4wNy0uOTVzLS43Ni0xLjUxLS43Ni0yLjY1em0xLjY3LS4wMWMwIC43NC4xMyAxLjMxLjQgMS43LjI2LjM4LjY1LjU4IDEuMTUuNTguNTEgMCAuOTktLjI2IDEuNDQtLjc3di0zLjIxYy0uMjQtLjIxLS40OC0uMzYtLjctLjQ1LS4yMy0uMDgtLjQ2LS4xMi0uNy0uMTItLjQ1IDAtLjgyLjE5LTEuMTMuNTktLjMxLjM5LS40Ni45NS0uNDYgMS42OHptNi4zNSAxLjU5YzAtLjczLjMyLTEuMy45Ny0xLjcxLjY0LS40IDEuNjctLjY4IDMuMDgtLjg0IDAtLjE3LS4wMi0uMzQtLjA3LS41MS0uMDUtLjE2LS4xMi0uMy0uMjItLjQzcy0uMjItLjIyLS4zOC0uM2MtLjE1LS4wNi0uMzQtLjEtLjU4LS4xLS4zNCAwLS42OC4wNy0xIC4ycy0uNjMuMjktLjkzLjQ3bC0uNTktMS4wOGMuMzktLjI0LjgxLS40NSAxLjI4LS42My40Ny0uMTcuOTktLjI2IDEuNTQtLjI2Ljg2IDAgMS41MS4yNSAxLjkzLjc2cy42MyAxLjI1LjYzIDIuMjF2NC4wN2gtMS4zMmwtLjEyLS43NmgtLjA1Yy0uMy4yNy0uNjMuNDgtLjk4LjY2cy0uNzMuMjctMS4xNC4yN2MtLjYxIDAtMS4xLS4xOS0xLjQ4LS41Ni0uMzgtLjM2LS41Ny0uODUtLjU3LTEuNDZ6bTEuNTctLjEyYzAgLjMuMDkuNTMuMjcuNjcuMTkuMTQuNDIuMjEuNzEuMjEuMjggMCAuNTQtLjA3Ljc3LS4ycy40OC0uMzEuNzMtLjU2di0xLjU0Yy0uNDcuMDYtLjg2LjEzLTEuMTguMjMtLjMxLjA5LS41Ny4xOS0uNzYuMzFzLS4zMy4yNS0uNDEuNGMtLjA5LjE1LS4xMy4zMS0uMTMuNDh6bTYuMjktMy42M2gtLjk4di0xLjJsMS4wNi0uMDcuMi0xLjg4aDEuMzR2MS44OGgxLjc1djEuMjdoLTEuNzV2My4yOGMwIC44LjMyIDEuMi45NyAxLjIuMTIgMCAuMjQtLjAxLjM3LS4wNC4xMi0uMDMuMjQtLjA3LjM0LS4xMWwuMjggMS4xOWMtLjE5LjA2LS40LjEyLS42NC4xNy0uMjMuMDUtLjQ5LjA4LS43Ni4wOC0uNCAwLS43NC0uMDYtMS4wMi0uMTgtLjI3LS4xMy0uNDktLjMtLjY3LS41Mi0uMTctLjIxLS4zLS40OC0uMzctLjc4LS4wOC0uMy0uMTItLjY0LS4xMi0xLjAxem00LjM2IDIuMTdjMC0uNTYuMDktMS4wNi4yNy0xLjUxcy40MS0uODMuNzEtMS4xNGMuMjktLjMuNjMtLjU0IDEuMDEtLjcxLjM5LS4xNy43OC0uMjUgMS4xOC0uMjUuNDcgMCAuODguMDggMS4yMy4yNC4zNi4xNi42NS4zOC44OS42N3MuNDIuNjMuNTQgMS4wM2MuMTIuNDEuMTguODQuMTggMS4zMiAwIC4zMi0uMDIuNTctLjA3Ljc2aC00LjM3Yy4wOC42Mi4yOSAxLjEuNjUgMS40NC4zNi4zMy44Mi41IDEuMzguNS4zIDAgLjU4LS4wNC44NC0uMTMuMjUtLjA5LjUxLS4yMS43Ni0uMzdsLjU0IDEuMDFjLS4zMi4yMS0uNjkuMzktMS4wOS41M3MtLjgyLjIxLTEuMjYuMjFjLS40NyAwLS45Mi0uMDgtMS4zMy0uMjUtLjQxLS4xNi0uNzctLjQtMS4wOC0uNy0uMy0uMzEtLjU0LS42OS0uNzItMS4xMy0uMTctLjQ0LS4yNi0uOTUtLjI2LTEuNTJ6bTQuNjEtLjYyYzAtLjU1LS4xMS0uOTgtLjM0LTEuMjgtLjIzLS4zMS0uNTgtLjQ3LTEuMDYtLjQ3LS40MSAwLS43Ny4xNS0xLjA4LjQ1LS4zMS4yOS0uNS43My0uNTcgMS4zem0zLjAxIDIuMjNjLjMxLjI0LjYxLjQzLjkyLjU3LjMuMTMuNjMuMi45OC4yLjM4IDAgLjY1LS4wOC44My0uMjNzLjI3LS4zNS4yNy0uNmMwLS4xNC0uMDUtLjI2LS4xMy0uMzctLjA4LS4xLS4yLS4yLS4zNC0uMjgtLjE0LS4wOS0uMjktLjE2LS40Ny0uMjNsLS41My0uMjJjLS4yMy0uMDktLjQ2LS4xOC0uNjktLjMtLjIzLS4xMS0uNDQtLjI0LS42Mi0uNHMtLjMzLS4zNS0uNDUtLjU1Yy0uMTItLjIxLS4xOC0uNDYtLjE4LS43NSAwLS42MS4yMy0xLjEuNjgtMS40OS40NC0uMzggMS4wNi0uNTcgMS44My0uNTcuNDggMCAuOTEuMDggMS4yOS4yNXMuNzEuMzYuOTkuNTdsLS43NC45OGMtLjI0LS4xNy0uNDktLjMyLS43My0uNDItLjI1LS4xMS0uNTEtLjE2LS43OC0uMTYtLjM1IDAtLjYuMDctLjc2LjIxLS4xNy4xNS0uMjUuMzMtLjI1LjU0IDAgLjE0LjA0LjI2LjEyLjM2cy4xOC4xOC4zMS4yNmMuMTQuMDcuMjkuMTQuNDYuMjFsLjU0LjE5Yy4yMy4wOS40Ny4xOC43LjI5cy40NC4yNC42NC40Yy4xOS4xNi4zNC4zNS40Ni41OC4xMS4yMy4xNy41LjE3LjgyIDAgLjMtLjA2LjU4LS4xNy44My0uMTIuMjYtLjI5LjQ4LS41MS42OC0uMjMuMTktLjUxLjM0LS44NC40NS0uMzQuMTEtLjcyLjE3LTEuMTUuMTctLjQ4IDAtLjk1LS4wOS0xLjQxLS4yNy0uNDYtLjE5LS44Ni0uNDEtMS4yLS42OHoiIGZpbGw9IiM1MzUzNTMiLz48L2c+PC9zdmc+"
+            ]
+        },
+        {
+            "section_name": "This article is cited by",
+            "section_text": "Nature Communications (2025)\n\nNature Communications (2025)\n\nJournal of Archaeological Method and Theory (2025)\n\nScientific Reports (2024)\n\nScientific Reports (2024)",
+            "section_image": []
+        }
+    ]
+}

02381398ad3c168dad4fb3121959ba38b9f694fc8f5d7aeae79ef0842f4b081a/metadata.json

@@ -0,0 +1,153 @@
+{
+    "title": "Assessing GPT-4 for cell type annotation in single-cell RNA-seq analysis",
+    "pre_title": "Reference-free and cost-effective automated cell type annotation with GPT-4 in single-cell RNA-seq analysis",
+    "journal": "Nature Methods",
+    "published": "25 March 2024",
+    "supplementary_0": [
+        {
+            "label": "Reporting Summary",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41592-024-02235-4/MediaObjects/41592_2024_2235_MOESM1_ESM.pdf"
+        },
+        {
+            "label": "Peer Review File",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41592-024-02235-4/MediaObjects/41592_2024_2235_MOESM2_ESM.pdf"
+        },
+        {
+            "label": "Supplementary Table",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41592-024-02235-4/MediaObjects/41592_2024_2235_MOESM3_ESM.xlsx"
+        }
+    ],
+    "supplementary_1": NaN,
+    "supplementary_2": NaN,
+    "source_data": [
+        "https://azimuth.hubmapconsortium.org/",
+        "/articles/s41592-024-02235-4#ref-CR5",
+        "https://gtexportal.org/home/datasets",
+        "/articles/s41592-024-02235-4#ref-CR5",
+        "/articles/s41592-024-02235-4#ref-CR6",
+        "https://figshare.com/articles/dataset/HCL_DGE_Data/7235471",
+        "/articles/s41592-024-02235-4#ref-CR7",
+        "https://figshare.com/s/865e694ad06d5857db4b",
+        "https://zenodo.org/record/7813151",
+        "https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE132465",
+        "https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE131907",
+        "https://cells.ucsc.edu/?ds=tabula-sapiens",
+        "/articles/s41592-024-02235-4#ref-CR12",
+        "/articles/s41592-024-02235-4#Sec2"
+    ],
+    "code": [
+        "https://github.com/Winnie09/GPTCelltype",
+        "https://doi.org/10.5281/zenodo.8317406",
+        "/articles/s41592-024-02235-4#ref-CR22",
+        "https://github.com/Winnie09/GPTCelltype_Paper",
+        "https://doi.org/10.5281/zenodo.8317410",
+        "https://zenodo.org/record/8317410",
+        "/articles/s41592-024-02235-4#ref-CR23"
+    ],
+    "subject": [
+        "Computational models",
+        "Gene expression profiling",
+        "Machine learning",
+        "Software",
+        "Transcriptomics"
+    ],
+    "license": "http://creativecommons.org/licenses/by/4.0/",
+    "preprint_pdf": "https://www.researchsquare.com/article/rs-2824971/v1.pdf?c=1711451133000",
+    "research_square_link": "https://www.researchsquare.com//article/rs-2824971/v1",
+    "nature_pdf": "https://www.nature.com/articles/s41592-024-02235-4.pdf",
+    "preprint_posted": "01 May, 2023",
+    "research_square_content": [
+        {
+            "section_name": "Abstract",
+            "section_text": "Cell type annotation is an essential step in single-cell RNA-seq analysis. However, it is a time-consuming process that often requires expertise in collecting canonical marker genes and manually annotating cell types. Automated cell type annotation methods typically require the acquisition of high-quality reference datasets and the development of additional pipelines. We demonstrate that GPT-4, a highly potent large language model, can automatically and accurately annotate cell types by utilizing marker gene information generated from standard single-cell RNA-seq analysis pipelines. Evaluated across hundreds of tissue types and cell types, GPT-4 generates cell type annotations exhibiting strong concordance with manual annotations, and has the potential to considerably reduce the effort and expertise needed in cell type annotation.Biological sciences/Computational biology and bioinformatics/SoftwareBiological sciences/Computational biology and bioinformatics/Computational modelsBiological sciences/Computational biology and bioinformatics/Literature mining",
+            "section_image": []
+        },
+        {
+            "section_name": "Additional Declarations",
+            "section_text": "There is NO Competing Interest.",
+            "section_image": []
+        },
+        {
+            "section_name": "Supplementary Files",
+            "section_text": "supptable1.csvSupplementary Table 1",
+            "section_image": []
+        }
+    ],
+    "nature_content": [
+        {
+            "section_name": "Abstract",
+            "section_text": "Here we demonstrate that the large language model GPT-4 can accurately annotate cell types using marker gene information in single-cell RNA sequencing analysis. When evaluated across hundreds of tissue and cell types, GPT-4 generates cell type annotations exhibiting strong concordance with manual annotations. This capability can considerably reduce the effort and expertise required for cell type annotation. Additionally, we have developed an R software package GPTCelltype for GPT-4\u2019s automated cell type annotation.",
+            "section_image": []
+        },
+        {
+            "section_name": "Main",
+            "section_text": "Cell type annotation is a fundamental step in single-cell RNA sequencing (scRNA-seq) analysis. This process is often laborious and time-consuming, requiring a human expert to compare genes highly expressed in each cell cluster with canonical cell type marker genes. Although automated cell type annotation methods have been developed (Supplementary Table 1), manual annotation using marker genes remains widely used.\n\nGenerative pre-trained transformers (GPT), including GPT-3.5 and GPT-4, are large language models designed for language understanding and generation. Recent studies have demonstrated their effectiveness in biomedical contexts1,2. In this Brief Communication, we hypothesize that GPT-4 can accurately annotate cell types, transitioning the annotation process from manual to a semi- or even fully automated procedure (Fig. 1a). GPT-4 offers cost-efficiency and seamless integration into existing single-cell analysis pipelines such as Seurat3, avoiding the need for building additional pipelines and collecting high-quality reference datasets. The vast training data of GPT-4 enables broader applications across various tissues and cell types, and its chatbot nature allows for user-driven annotation refinement (Fig. 1a,b).\n\na, Comparison of cell type annotations by human experts, GPT-4, and other automated methods. b, Example of GPT-4 annotating human prostate cells with increasing granularity. c, Example of GPT-4 annotating single, mixed and new cell types.\n\nWe systematically assessed GPT-4\u2019s cell type annotation performance across ten datasets4,5,6,7,8,9,10,11,12, covering five species and hundreds of tissue and cell types, and including both normal and cancer samples (Supplementary Table 2). GPT-4 was queried using GPTCelltype, a software tool we developed (Methods). For competing methods, we evaluated GPT-3.5, a prior version of GPT-4, and CellMarker2.013, SingleR14 and ScType15, which are automatic cell type annotation methods that provide references applicable to a large number of tissues (Methods and Supplementary Table 1). Cell type annotations by GPT-4 or competing methods were evaluated based on their agreement with manual annotations provided by the original studies. The degree of agreement was measured using a numeric score (Methods). Supplementary Table 3 presents an example of evaluating GPT-4 cell type annotations in human prostate tissue, and details of all cell type annotations and their evaluation results are included in Supplementary Table 4.\n\nWe first explored different factors that may affect the annotation accuracy of GPT-4 (Fig. 2a and Supplementary Table 5). We found that GPT-4 performs best when using the top ten differential genes, and when differential genes are derived using the two-sided Wilcoxon test. GPT-4 exhibits similar accuracy across various prompt strategies, including a basic prompt strategy, a chain-of-thought16-inspired prompt strategy that includes reasoning steps, and a repeated prompt strategy (Methods). In subsequent analyses, both GPT-4 and GPT-3.5 used the basic prompt strategy with the top ten differential genes obtained from Wilcoxon test as inputs for applicable datasets.\n\na, Average agreement scores for varying numbers of top differential genes, statistical tests for differential analysis, and prompt strategies. b, Proportion of cell types with varying agreement levels in each study and tissue, most abundant broad cell types, malignant cells, different cell population sizes, and major cell types versus cell subtypes. c, log2-transformed ratio of type I (COL1A1 and COL1A2) and II (COL2A1) collagen gene expression. d,e, Comparison of average agreement scores (d) and running times (e). In e, n\u2009=\u200959 for GPT-4 and GPT-3.5 and n\u2009=\u200936 for ScType and SingleR. Each boxplot shows the distribution (center: median; bounds of box: first and third quartiles; bounds of whiskers: data points within 1.5\u00d7 interquartile range from the box; minima; maxima) of running time. f, Financial cost of querying GPT-4 API versus cell type numbers. g, GPT-4\u2019s performance in identifying mixed/single cell types and known/unknown cell types, and under different subsampling and noise levels in multiple simulation rounds (dots). h, Reproducibility of GPT-4 annotations. i, Consistency of agreement scores between two versions of GPT-4.\n\nGPT-4\u2019s annotations fully or partially match manual annotations in over 75% of cell types in most studies and tissues (Fig. 2b), demonstrating its competency in generating expert-comparable cell type annotations. This agreement is particularly high for marker genes from literature searches, with at least 70% fully match rate in most tissues. Though lower for genes identified by differential analysis, the agreement remains high. However, results from datasets published before September 2021 should be interpreted cautiously as they predate GPT-4\u2019s training cutoff. GPT-4 performs better for immune cells like granulocytes compared to other cell types (Fig. 2b). It identifies malignant cells in colon and lung cancer datasets but struggles with B lymphoma, potentially due to a lack of distinct gene sets. The identification of malignant cells could benefit from other approaches such as copy number variation9. Performance dips slightly in small cell populations comprising no more than ten cells (Fig. 2b), possibly due to the limited available information. GPT-4 annotations fully match manual annotations more frequently in major cell types (for example, T cells) than in subtypes (for example, CD4 memory T cells), while over 75% of subtypes still achieve full or partial matches (Fig. 2b).\n\nThe low agreement between GPT-4 and manual annotations in some cell types does not necessarily imply that GPT-4\u2019s annotation is incorrect. For instance, cell types classified as stromal cells include fibroblasts and osteoblasts expressing type I collagen genes, and chondrocytes expressing type II collagen genes. For cells manually annotated as stromal cells, GPT-4 assigns cell type annotations with higher granularity (for example, fibroblasts and osteoblasts), resulting in partial matches and a lower agreement. For cell types that are manually annotated as stromal cells but identified by GPT-4 as fibroblasts or osteoblasts, type I collagen genes show substantially higher expression than type II collagen genes (Fig. 2c). This agrees with the pattern observed in cells manually annotated as chondrocytes, fibroblasts, and osteoblasts (Fig. 2c), suggesting that GPT-4 provides more accurate cell type annotations for stromal cells.\n\nGPT-4 substantially outperforms other methods based on average agreement scores (Methods and Fig. 2d). Using GPTCelltype as the interface, GPT-4 is also notably faster (Fig. 2e), partly due to its utilization of differential genes from the standard single-cell analysis pipelines such as Seurat3. Given the integral role of these pipelines, we regard the differential genes as immediately available for GPT-4. In contrast, other methods like SingleR and ScType require additional steps to reprocess the gene expression matrices. Compared to other methods that are free of charge, GPT-4 incurs a $20 monthly fee for using online web portal. Cost of GPT-4 API is linearly correlated with the number of queried cell types and does not exceed $0.1 for all queries in this study (Fig. 2f).\n\nWe further assessed GPT-4\u2019s robustness in complex real data scenarios (Fig. 1c) with simulated datasets (Methods). GPT-4 can distinguish between pure and mixed cell types with 93% accuracy, and differentiate between known and unknown cell types with 99% accuracy (Fig. 2g). When the input gene set includes fewer genes or is contaminated with noise, GPT-4\u2019s performance decreases but remains high (Fig. 2g). These results demonstrate GPT-4\u2019s robustness in various scenarios.\n\nFinally, we assessed the reproducibility of GPT-4\u2019s annotations using prior simulation studies (Methods). GPT-4 generated identical annotations for the same marker genes in 85% of cases (Fig. 2h), indicating high reproducibility. Annotations of two GPT-4 versions showed identical agreement scores in most cases, with a Cohen\u2019s \u03ba of 0.65, demonstrating substantial consistency (Fig. 2i).\n\nWhile GPT-4 excels in cell type annotation, which surpasses existing methods, there are limitations to consider. Firstly, the undisclosed nature of GPT-4\u2019s training corpus makes verifying the basis of its annotations challenging, thus requiring human evaluation to ensure annotation quality and reliability. Secondly, human involvement in the optional fine-tuning of the model may affect reproducibility due to subjectivity and could limit the scalability of the model in large datasets. Thirdly, high noise levels in scRNA-seq data and unreliable differential genes can adversely affect GPT-4\u2019s annotations. Lastly, over-reliance on GPT-4 risks artificial intelligence hallucination. We recommend validation of GPT-4\u2019s cell type annotations by human experts before proceeding with downstream analyses.\n\nWhile this study focuses on the standard version of GPT-4, fine-tuning GPT-4 with high-quality reference marker gene lists could further improve cell type annotation performance, utilizing services such \u2018GPTs\u2019 provided by OpenAI.",
+            "section_image": [
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41592-024-02235-4/MediaObjects/41592_2024_2235_Fig1_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41592-024-02235-4/MediaObjects/41592_2024_2235_Fig2_HTML.png"
+            ]
+        },
+        {
+            "section_name": "Methods",
+            "section_text": "For the HuBMAP Azimuth project, manually annotated cell types and their marker genes were downloaded from the Azimuth website (https://azimuth.hubmapconsortium.org/). Azimuth provides cell type annotations for each tissue at different granularity levels. We selected the level of granularity with the fewest number of cell types, provided that there are more than ten cell types within that level. Details of how marker genes were generated are not reported by Azimuth.\n\nFor the GTEx5 dataset, manually annotated cell types, differential gene lists and gene expression matrices were downloaded directly from the publication5. In the original study, gene expression raw counts were library-size-normalized and log-transformed after adding a pseudocount of 1 with SCANPY17. ComBat18 was used to account for the protocol- and sex-specific effects with SCANPY17. Welch\u2019s t-test was then performed to identify differential genes that compare one cell type against the rest. For each cell type, genes were ranked increasingly by P values, and genes with the same P values were further ranked decreasingly by t-statistics. Top 10, 20 and 30 differential genes were used in this study. Lists of marker genes through literature search and the corresponding cell types were downloaded from the same study5, and only cell types with at least five marker genes were used.\n\nFor the HCL6 dataset, manually annotated cell types, differential gene lists and the gene expression matrix were downloaded directly from the publication6. In the original study, gene expression raw counts underwent a batch removal process to facilitate cross-tissue comparison and were subsequently normalized by library size and log-transformed after adding a pseudocount of 1. Two-sided Wilcoxon rank-sum test was then performed to identify differential genes comparing one cell type against the rest using Seurat3. Differential genes were further selected by log fold change larger than 0.25, Bonferroni-adjusted P value smaller than 0.1, and expressed in at least 15% of cells in either population. For each cell type, genes were ranked increasingly by P values, and genes with the same P values were further ranked decreasingly by two-sided Wilcoxon test statistics. Top 10, 20 and 30 differential genes were used in this study.\n\nFor the Mouse Cell Atlas (MCA)7 dataset, manually annotated cell types, differential gene lists and gene expression matrix were downloaded directly from the publication6. In the original study, gene expression raw counts underwent a batch removal process to facilitate cross-tissue comparison, and Seurat3 was used to perform preprocessing and differential analysis. For each cell type, genes were ranked increasingly by P values, and genes with the same P values were further ranked decreasingly by log fold change. Top 10, 20 and 30 differential genes were used in this study.\n\nFor non-model mammal dataset12, manually annotated cell types and lists of marker genes through literature search were downloaded directly from the original study.\n\nFor Tabula Sapiens (TS)8, B-cell lymphoma (BCL)9, lung cancer11 and colon cancer10 datasets, manually annotated cell types and raw gene expression count matrices were downloaded directly from original studies. Raw counts were normalized by library size and log-transformed after adding a pseudocount of 1. Seurat FindAllMarkers() function with default settings was used to obtain differential genes by comparing one cell type with the rest within each tissue. Briefly, genes with at least 0.25 log fold change between two cell populations and detected in at least 10% of cells in either cell population were retained. Two-sided Wilcoxon rank-sum test was then performed for differential analysis. In addition, two-sided two-sample t-test was also performed for differential analysis using the FindAllMarkers() function with default settings. For each cell type, genes were ranked increasingly by P values, and genes with the same P values were further ranked decreasingly by log fold changes. Top 10, 20 and 30 differential genes were used in this study.\n\nAll GPT-4 (13 June 2023 version) and GPT-3.5 (13 June 2023 version) cell type annotations in this study were performed using GPTCelltype, an R software package we developed as an interface for GPT models. GPTCelltype takes marker genes or top differential genes as input, and automatically generates prompt message using the following template with the basic prompt strategy:\n\n\u2018Identify cell types of TissueName cells using the following markers separately for each row. Only provide the cell type name. Do not show numbers before the name. Some can be a mixture of multiple cell types.\\n GeneList\u2019.\n\nHere \u2018TissueName\u2019 is a variable that will be replaced with the actual name of the tissue (for example, human prostate), and \u2018GeneList\u2019 is a list of marker genes or top differential genes. Genes for the same cell population are joined by comma (,), and gene lists for different cell populations are separated by the newline character (\\n). GPT-4 or GPT-3.5 was then queried using the generated prompt message through OpenAI API, and the returned information was parsed and converted to cell type annotations.\n\nFor chain-of-thought prompt strategy, the following sentence was added to the beginning of the message generated by the basic prompt strategy: \u2018Because CD3 gene is a marker gene of T cells, if CD3 gene is included in the marker gene list of an unknown cell type, the cell type is likely to be T cells, a subtype of T cells, or a mixed cell type containing T cells\u2019.\n\nFor repeated prompt strategy, GPT-4 was queried with the basic prompt strategy repeatedly for five times. The annotation result that appears most frequently among the five queries was selected as the final cell type annotation.\n\nGPT-4 (23 March 2023 version) cell type annotations were performed by manually copying and pasting prompt messages to GPT-4 online web interface (https://chat.openai.com/). The prompt message was constructed using the following template:\n\n\u2018Identify cell types of TissueName cells using the following markers. Identify one cell type for each row. Only provide the cell type name. \\n GeneList\u2019.\n\nComputationally identified differential genes in eight scRNA-seq datasets and canonical marker genes identified through literature search in two datasets were used as inputs to GPT-4 and GPT-3.5 (Supplementary Table 2). Cell type annotation for HCL and MCA was performed and evaluated once by aggregating all tissues, similar to the original studies. In other studies, cell type annotation was performed and evaluated within each tissue.\n\nSingleR14 (version 1.4.1) R package was used to perform cell type annotations with default settings. For HCL and MCA datasets, the gene expression matrices after batch effect removal, library size normalization and log transformation across all tissues were used as input. For all other datasets, SingleR was performed separately within each tissue, and the input is the log-transformed and library-size normalized gene expression matrix. The built-in Human Primary Cell Atlas reference19 was used as the reference dataset for all SingleR annotations. SingleR generates single-cell level cell type annotations by returning an assignment score matrix for each single cell and each cell type label in the reference. To convert single-cell level annotations to cell-cluster level annotations, for each manually annotated cell type, we assigned the reference label with assignment scores summed across all single cells in that manually annotated cell type as the predicted cell type annotation.\n\nScType15 (version 1.0) R package was used to perform cell type annotations with default settings. To meet the need for computational efficiency when working with large datasets, we developed an in-house version of ScType. We utilized vectorization to optimize the most time-consuming steps, while still generating the same output of the original ScType software. The input gene expression matrices to ScType were the same as used in SingleR described above. The built-in cell type marker database was used as the reference for all ScType annotations. Manually annotated cell types were treated as cell clusters and given as inputs to ScType. ScType directly generates cluster-level cell type annotations.\n\nCellMarker2.0 (ref. 13) only provides an online user interface and does not have a software implementation. We used the exact same marker gene sets or top ten differential gene sets identified by two-sided Wilcoxon tests for GPT-4 and GPT-3.5 cell type annotations as inputs of CellMarker2.0.\n\nCell type annotations by GPT-4 or competing methods were compared to manual annotations provided by the original studies. Each manually or automatically identified cell type annotation was assigned an unambiguous cell ontology (CL) name20 and a broad cell type name when applicable. A pair of manually and automatically identified cell type annotations was classified as \u2018fully match\u2019 if they have the same annotation term or available CL cell ontology name, \u2018partially match\u2019 if they have the same or subordinate (for example, fibroblast and stromal cell) broad cell type name but different annotations and CL cell ontology names, and \u2018mismatch\u2019 if they have different broad cell type names, annotations and CL cell ontology names.\n\nTo facilitate comparison, we assigned agreement scores of 1, 0.5 and 0 to cases of \u2018fully match\u2019, \u2018partially match\u2019 and \u2018mismatch\u2019 respectively, and calculated average scores within each dataset across cell types and tissues.\n\nTo generate simulation datasets, we used canonical cell type markers through GTEx literature search of human breast cells, the top ten differential genes from the human colon cancer dataset, and the top ten differential genes from the vasculature tissue of the TS dataset as templates. Simulation studies were performed separately for the three tissue types.\n\nTo generate simulation datasets of mixed cell types, marker genes for each mixed cell type were created by combining the marker gene lists of two randomly selected cell types. Ten mixed cell types were generated in each simulation iteration. Additionally, we incorporated the original cell type markers of ten randomly chosen cell types as negative controls of single cell types. This entire simulation process was repeated five times. Subsequently, GPT-4 was queried using these simulated marker gene lists, and its performance in differentiating between mixed and single cell types was assessed.\n\nTo generate simulation datasets of unknown cell types, we compiled a list of all human genes using the Bioconductor org.Hs.eg.db package21. In each simulation iteration, ten simulated unknown cell types were generated. The marker genes for each unknown cell type were produced by combining ten randomly selected human genes. Additionally, we included ten real cell types and their marker genes as negative controls of known cell types, similar to the previous simulation study. This entire simulation process was repeated five times. Subsequently, GPT-4 was queried using these simulated marker gene lists, and its performance in distinguishing between known and unknown cell types was assessed.\n\nTo generate simulation datasets with partial marker gene information, we randomly subsampled 25%, 50% or 75% of the original marker genes. The simulation process was repeated five times. Subsequently, GPT-4 was queried using these subsampled marker gene lists, and the performance was assessed by agreement scores.\n\nTo generate simulation datasets with contaminated information, we added randomly selected human genes to the original marker gene list. The numbers of randomly selected genes are 25%, 50% or 75% of the number of original marker genes. The simulation process was repeated five times. Subsequently, GPT-4 was queried using these subsampled marker gene lists, and the performance was assessed by agreement scores.\n\nWe assessed the reproducibility of GPT-4 responses by leveraging the repeated querying of GPT-4 with identical marker gene lists of the same negative control cell types in simulation studies. For each cell type, reproducibility is defined as the proportion of instances in which GPT-4 generates the most prevalent cell type annotation. For instance, in the case of vascular endothelial cells, GPT-4 produces \u2018endothelial cells\u2019 eight times and \u2018blood vascular endothelial cells\u2019 once. Consequently, the most prevalent cell type annotation is \u2018endothelial cells\u2019, and the reproducibility is calculated as \\(\\frac{8}{9}=0.89\\).\n\nAccording to information provided by OpenAI, the application programming interface (API) cost for running GPT-4 13 June 2023 version is $0.03 for every thousand input tokens and $0.06 for every thousand output tokens. For each query, we obtained i and o, which represent the numbers of input tokens and output tokens respectively, through the OpenAI API. The total API financial cost is thus calculated as $(0.00003i\u2009+\u20090.00006o).\n\nFurther information on research design is available in the Nature Portfolio Reporting Summary linked to this article.",
+            "section_image": []
+        },
+        {
+            "section_name": "Data availability",
+            "section_text": "The data used in this manuscript are all downloaded from publicly available data sources. Specifically, HubMAP Azimuth data were downloaded from the Azimuth website (https://azimuth.hubmapconsortium.org/). GTEx manually annotated cell types and differential gene lists were downloaded from the supplementary materials of the original study5. GTEx gene expression matrix was downloaded from the GTEx website (https://gtexportal.org/home/datasets). Marker genes from literature search were downloaded from the supplementary materials of the original study5. HCL manually annotated cell types and differential gene lists were downloaded from the supplementary materials of the original study6. HCL gene expression matrix was downloaded from figshare (https://figshare.com/articles/dataset/HCL_DGE_Data/7235471). MCA manually annotated cell types and differential gene lists were downloaded from the supplementary materials of the original study7. MCA gene expression matrix was downloaded from figshare (https://figshare.com/s/865e694ad06d5857db4b). BCL gene expression matrix and manually annotated cell types were downloaded from Zenodo (https://zenodo.org/record/7813151). Colon cancer gene expression matrix and manually annotated cell types were downloaded from GEO under accession number GSE132465. Lung cancer gene expression matrix and manually annotated cell types were downloaded from GEO under accession number GSE131907. TS gene expression matrix and manually annotated cell types were downloaded from UCSC Cell Browser (https://cells.ucsc.edu/?ds=tabula-sapiens). Marker genes and cell type annotations for the non-model mammal dataset were downloaded from the supplementary materials of the original study12. All relevant information about data is described in Methods. All data generated in this study are included in the supplementary tables.",
+            "section_image": []
+        },
+        {
+            "section_name": "Code availability",
+            "section_text": "The GPTCelltype package (v.1.0.0) is provided as an open-source software package with a detailed user manual available in the GitHub repository at https://github.com/Winnie09/GPTCelltype. The software is released in Zenodo under https://doi.org/10.5281/zenodo.8317406 for all versions (ref. 22). All codes to reproduce the presented analyses are publicly available in the GitHub repository at https://github.com/Winnie09/GPTCelltype_Paperand also in Zenodo under https://doi.org/10.5281/zenodo.8317410 (https://zenodo.org/record/8317410) (ref. 23). R version 4.0.2 was used to perform the analyses in the manuscript.",
+            "section_image": []
+        },
+        {
+            "section_name": "References",
+            "section_text": "Hou, W. et al. GeneTuring tests GPT models in genomics. Preprint at bioRxiv https://doi.org/10.1101/2023.03.11.532238 (2023).\n\nHou, W. et al. GPT-4V exhibits human-like performance in biomedical image classification. Preprint at bioRxiv https://doi.org/10.1101/2023.12.31.573796 (2024).\n\nHao, Y. et al. Integrated analysis of multimodal single-cell data. Cell 184, 3573\u20133587 (2021).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHuBMAP Consortium. The human body at cellular resolution: the NIH Human Biomolecular Atlas Program. Nature 574, 187\u2013192 (2019).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nEraslan, G. et al. Single-nucleus cross-tissue molecular reference maps toward understanding disease gene function. Science 376, eabl4290 (2022).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHan, X. et al. Construction of a human cell landscape at single-cell level. Nature 581, 303\u2013309 (2020).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHan, X. et al. Mapping the mouse cell atlas by microwell-seq. Cell 172, 1091\u20131107 (2018).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nThe Tabula Sapiens Consortium. The Tabula Sapiens: a multiple-organ, single-cell transcriptomic atlas of humans. Science 376, eabl4896 (2022).\n\nArticle\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLiu, N. et al. Single-cell landscape of primary central nervous system diffuse large B-cell lymphoma. Cell Discov. 9, 55 (2023).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLee, H.-O. et al. Lineage-dependent gene expression programs influence the immune landscape of colorectal cancer. Nat. Genet. 52, 594\u2013603 (2020).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKim, N. et al. Single-cell RNA sequencing demonstrates the molecular and cellular reprogramming of metastatic lung adenocarcinoma. Nat. Commun. 11, 2285 (2020).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nChen, D. et al. Single cell atlas for 11 non-model mammals, reptiles and birds. Nat. Commun. 12, 7083 (2021).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHu, C. et al. CellMarker 2.0: an updated database of manually curated cell markers in human/mouse and web tools based on scRNA-seq data. Nucleic Acids Res. 51, D870\u2013D876 (2023).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nAran, D. et al. Reference-based analysis of lung single-cell sequencing reveals a transitional profibrotic macrophage. Nat. Immunol. 20, 163\u2013172 (2019).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nIanevski, A. et al. Fully-automated and ultra-fast cell-type identification using specific marker combinations from single-cell transcriptomic data. Nat. Commun. 13, 1246 (2022).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWei, J. et al. Chain-of-thought prompting elicits reasoning in large language models. Adv. Neural Inf. Process. Syst. 35, 24824\u201324837 (2022).\n\n\n                    Google Scholar\u00a0\n                \n\nWolf, F. A. et al. SCANPY: large-scale single-cell gene expression data analysis. Genome Biol. 19, 15 (2018).\n\nArticle\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLeek, J. T. et al. The sva package for removing batch effects and other unwanted variation in high-throughput experiments. Bioinformatics 28, 882\u2013883 (2012).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMabbott, N. A. et al. An expression atlas of human primary cells: inference of gene function from coexpression networks. BMC Genomics 14, 632 (2013).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nC\u00f4t\u00e9, R. G. et al. A new Ontology Lookup Service at EMBL-EBI. BMC Bioinforma. 7, 97 (2006).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGentleman, R. C. et al. Bioconductor: open software development for computational biology and bioinformatics. Genome Biol. 5, R80 (2004).\n\nArticle\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHou, W. et al. GPTCelltype R software package. Zenodo https://doi.org/10.5281/zenodo.8317406 (2023).\n\nHou, W. et al. Repository of code to reproduce the analysis in this study. Zenodo https://doi.org/10.5281/zenodo.8317410 (2023).\n\nDownload references",
+            "section_image": []
+        },
+        {
+            "section_name": "Acknowledgements",
+            "section_text": "Z.J. was supported by the National Institutes of Health under award number U54AG075936 and by the Whitehead Scholars Program at Duke University School of Medicine. W.H. was partially supported by the National Institute Of General Medical Sciences of the National Institutes of Health under award number R35GM150887 and by the General Fund at Columbia University Department of Biostatistics. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.",
+            "section_image": []
+        },
+        {
+            "section_name": "Author information",
+            "section_text": "Department of Biostatistics, Columbia University Mailman School of Public Health, New York City, NY, USA\n\nWenpin Hou\n\nDepartment of Biostatistics and Bioinformatics, Duke University School of Medicine, Durham, NC, USA\n\nZhicheng Ji\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nW.H. and Z.J. conceived the study, conducted the analysis and wrote the manuscript.\n\nCorrespondence to\n                Wenpin Hou or Zhicheng Ji.",
+            "section_image": []
+        },
+        {
+            "section_name": "Ethics declarations",
+            "section_text": "The authors declare no competing interests.",
+            "section_image": []
+        },
+        {
+            "section_name": "Peer review",
+            "section_text": "Nature Methods thanks Qin Ma and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Primary Handling Editor: Lin Tang, in collaboration with the Nature Methods team. Peer reviewer reports are available.",
+            "section_image": []
+        },
+        {
+            "section_name": "Additional information",
+            "section_text": "Publisher\u2019s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.",
+            "section_image": []
+        },
+        {
+            "section_name": "Rights and permissions",
+            "section_text": "Open Access  This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article\u2019s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article\u2019s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.\n\nReprints and permissions",
+            "section_image": []
+        },
+        {
+            "section_name": "About this article",
+            "section_text": "Hou, W., Ji, Z. Assessing GPT-4 for cell type annotation in single-cell RNA-seq analysis.\n                    Nat Methods 21, 1462\u20131465 (2024). https://doi.org/10.1038/s41592-024-02235-4\n\nDownload citation\n\nReceived: 16 April 2023\n\nAccepted: 05 March 2024\n\nPublished: 25 March 2024\n\nVersion of record: 25 March 2024\n\nIssue date: August 2024\n\nDOI: https://doi.org/10.1038/s41592-024-02235-4\n\nAnyone you share the following link with will be able to read this content:\n\nSorry, a shareable link is not currently available for this article.\n\n\n\n\n                            Provided by the Springer Nature SharedIt content-sharing initiative\n                        ",
+            "section_image": [
+                "https://data:image/svg+xml;base64,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"
+            ]
+        },
+        {
+            "section_name": "This article is cited by",
+            "section_text": "Journal of Translational Medicine (2025)\n\nGenome Biology (2025)\n\nJournal of Translational Medicine (2025)\n\nNature Methods (2025)\n\nNature Machine Intelligence (2025)",
+            "section_image": []
+        }
+    ]
+}

02ed829a2b0d25d6841e8c42a5b49a127bb3f49651373ce20617894c907bc9bb/metadata.json

@@ -0,0 +1,147 @@
+{
+    "title": "Human neutralizing antibodies target a conserved lateral patch on H7N9 hemagglutinin head",
+    "pre_title": "Allosteric Neutralization by Human H7N9 Antibodies",
+    "journal": "Nature Communications",
+    "published": "27 May 2024",
+    "supplementary_0": [
+        {
+            "label": "Supplementary Information",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-024-48758-4/MediaObjects/41467_2024_48758_MOESM1_ESM.pdf"
+        },
+        {
+            "label": "Peer Review File",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-024-48758-4/MediaObjects/41467_2024_48758_MOESM2_ESM.pdf"
+        },
+        {
+            "label": "Reporting Summary",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-024-48758-4/MediaObjects/41467_2024_48758_MOESM3_ESM.pdf"
+        }
+    ],
+    "supplementary_1": [
+        {
+            "label": "Source Data",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-024-48758-4/MediaObjects/41467_2024_48758_MOESM4_ESM.xlsx"
+        }
+    ],
+    "supplementary_2": NaN,
+    "source_data": [
+        "https://www.ncbi.nlm.nih.gov/nuccore/OR901962,OR901963,OR901964,OR901965,OR901966,OR901967,OR901968,OR901969",
+        "https://www.ebi.ac.uk/emdb/EMD-41422",
+        "https://www.ebi.ac.uk/emdb/EMD-41441",
+        "https://www.rcsb.org/structure/8TNL",
+        "https://www.rcsb.org/structure/8TOA",
+        "/articles/s41467-024-48758-4#Sec28"
+    ],
+    "code": [],
+    "subject": [
+        "Antibodies",
+        "Cryoelectron microscopy",
+        "Influenza virus"
+    ],
+    "license": "http://creativecommons.org/licenses/by/4.0/",
+    "preprint_pdf": "https://www.researchsquare.com/article/rs-3429355/v1.pdf?c=1716894450000",
+    "research_square_link": "https://www.researchsquare.com//article/rs-3429355/v1",
+    "nature_pdf": "https://www.nature.com/articles/s41467-024-48758-4.pdf",
+    "preprint_posted": "06 Nov, 2023",
+    "research_square_content": [
+        {
+            "section_name": "Abstract",
+            "section_text": "The avian influenza A virus H7N9 causes severe human infections with more than 30% fatality despite the use of neuraminidase inhibitors. Currently there is no H7N9-specific prevention or treatment for humans. From a 2013 H7N9 convalescent case occurred in Hong Kong, we isolated four H7 hemagglutinin (HA)-reactive monoclonal antibodies (mAbs) by single B cell cloning, with three mAbs directed to the HA globular head domain (HA1) and one to the HA stem region (HA2). Two clonally related HA1-directed mAbs, H7.HK1 and H7.HK2, potently neutralized H7N9 and protected mice from a lethal H7N9/AH1 challenge. Cryo-EM structures revealed that H7.HK1 and H7.HK2 bind to a \u03b214-centered surface partially overlapping with the antigenic site D of HA1 and disrupt the 220-loop that makes hydrophobic contacts with sialic acid on the adjacent protomer, thus affectively blocking viral entry. The more potent mAb H7.HK2 retained full HA1 binding and neutralization capacity to later H7N9 isolates from 2016-2017, which is consistent with structural data showing that the antigenic mutations of 2016-2017 from the 2013 H7N9 only occurred at the periphery of the mAb epitope. The HA2-directed mAb H7.HK4 lacked neutralizing activity but protected mice from the lethal H7N9/AH1 challenge when engineered to mouse IgG2a enabling Fc effector function in mice. Used in combination with H7.HK2 at a suboptimal dose, H7.HK4 augmented mouse protection. Our data demonstrated an allosteric mechanism of mAb neutralization and augmented protection against H7N9 when a HA1-directed neutralizing mAb and a HA2-directed non-neutralizing mAb were combined.Biological sciences/Immunology/Adaptive immunity/Humoral immunity/AntibodiesBiological sciences/Microbiology/Virology/Influenza virus",
+            "section_image": []
+        },
+        {
+            "section_name": "Additional Declarations",
+            "section_text": "Yes there is potential Competing Interest.\nAn U.S. provisional patent titled \u201cHuman Protective Neutralizing and Non-neutralizing Antibodies and Their Use against Influenza Viruses\u201d was filed with filing No. 63/578,505 and XW, MJ, NCM, HL, DDH, KY, KKT, PDK, and LS as co-inventors.",
+            "section_image": []
+        }
+    ],
+    "nature_content": [
+        {
+            "section_name": "Abstract",
+            "section_text": "Avian influenza A virus H7N9 causes severe human infections with >30% fatality. Currently, there is no H7N9-specific prevention or treatment for humans. Here, from a 2013 H7N9 convalescent case in Hong Kong, we isolate four hemagglutinin (HA)-reactive monoclonal antibodies (mAbs), with three directed to the globular head domain (HA1) and one to the stalk domain (HA2). Two clonally related HA1-directed mAbs, H7.HK1 and H7.HK2, potently neutralize H7N9 and protect female mice from lethal H7N9/AH1 challenge. Cryo-EM structures reveal that H7.HK1 and H7.HK2 bind to a \u03b214-centered surface and disrupt the 220-loop that makes hydrophobic contacts with sialic acid on an adjacent protomer, thereby blocking viral entry. Sequence analysis indicates the lateral patch targeted by H7.HK1 and H7.HK2 to be conserved among influenza subtypes. Both H7.HK1 and H7.HK2 retain HA1 binding and neutralization capacity to later H7N9 isolates from 2016\u20132017, consistent with structural data showing that the antigenic mutations during this timeframe occur at their epitope peripheries. The HA2-directed mAb H7.HK4 lacks neutralizing activity but when used in combination with H7.HK2 moderately augments female mouse protection. Overall, our data reveal antibodies to a conserved lateral HA1 supersite that confer neutralization, and when combined with a HA2-directed non-neutralizing mAb, augment protection.",
+            "section_image": []
+        },
+        {
+            "section_name": "Introduction",
+            "section_text": "H7N9 is an avian influenza A group 2 virus first transmitted to humans in the spring of 2013 most likely through live poultry market exposure in China1,2,3. The virus reemerged in the fall of 2013 and in the winter of later years, with the largest epidemic reported as the 5th wave in 2016\u201320174,5,6. Though there is limited evidence for human-to-human transmission, a few mutations in the hemagglutinin (HA) gene of the virus might be sufficient to overcome its inefficiency for human transmission7,8,9,10. Like other influenza virus infections, the most common treatments against H7N9 are neuraminidase inhibitors Tamiflu (i.e., oseltamivir) and Relena (i.e., zanamivir), but oseltamivir-resistant strains have emerged11,12,13. Intravenous (i.v.) zanamivir, though not clinically approved, has been used on a compassionate basis in some severe cases because of favorable pharmacokinetics and in vitro susceptibility against oseltamivir-resistant strains14,15. Other antiviral treatment includes the endonuclease inhibitor Xofluza (i.e., baloxavir marboxil) that targets the viral polymerase and has been shown effective in mice against H7N9 lethal challenges16. Despite the use of neuraminidase inhibitors and endonuclease inhibitor, H7N9 case-fatality rate remains higher than 30%, and currently there is no licensed vaccine against H7N9 for humans. Concerns for a potential major outbreak warrant the development of human monoclonal antibodies (mAbs) against H7N9.\n\nBecause HA is the major target for influenza neutralizing antibodies, H7-reactive human mAbs have been isolated and characterized from H7N9 acute infections17, convalescent cases18, and H7N9 experimental vaccinees19,20,21. The binding sites of these mAbs have been mapped to the HA globular head (HA1) and stem (HA2) domains. A subset of HA1-directed mAbs potently neutralized H7N9 and protected mice from H7N9 challenges at doses of 0.3, 1, 5\u2009mg/kg or higher17,18,19,20. These HA1-directed mAbs typically neutralized H7N9 by direct interference with or around the receptor (sialic acid) binding site17,19,22. These epitopes correspond to the antigenic sites A and B as previously mapped on the surface of H3 HA23,24,25. Of note, significant antigenic drift has been documented in the HA gene of 2016\u20132017 H7N9 from the initial 2013 isolates17,26,27. For example, Huang et al. isolated 17 neutralizing mAbs from four cases infected in 2013 and 2014, yet only three of these mAbs were active against viral isolates from 2016 to 201717. A broad mAb FluA-20 targeting the HA1 trimer interface did not mediate neutralization in vitro, but protected mice from viral challenges by disrupting HA trimers and inhibiting cell-to-cell spread of virus21. HA2-directed mAbs have also demonstrated neutralizing activity against divergent subtypes28,29,30,31,32,33,34,35, although typically not as robust in neutralizing activity when compared to HA1 mAbs. A few HA2 mAbs, neutralizing or not, protected mice from H7N9 challenges at 5\u2009mg/kg20, especially when engineered as mouse IgG2a, which has the highest Fc-mediated effector functions in mouse36. However, previous studies have not tested the combination of two or more mAbs that target different regions of H7N9 HA.\n\nIn the post COVID-19 era, preparedness for future pandemics has become a high priority, as exemplified by the science community closely monitoring a bird flu (avian influenza A H5N1) outbreak in US cows37. Here, we aim to facilitate the development of human mAbs against H7N9, which has also been considered one of the most serious pandemic threats. We obtained peripheral blood mononuclear cells (PBMCs) from a 2013 H7N9 convalescent case in Hong Kong with the virus isolated as A/Hong Kong/470129/2013 H7N914. The course of this infection lasted for about one month and the treatment required extracorporeal membrane oxygenation (ECMO) and i.v. zanamivir14. Development of plasma neutralizing antibodies was evident at recovery. The PBMC sample used to isolate mAbs was collected one year post recovery. From the isolated mAbs, we not only demonstrate potent H7N9 neutralization, but use cryo-EM analysis to delineate a conserved lateral site-of-vulnerability in the HA head, a finding of vaccine relevance.",
+            "section_image": []
+        },
+        {
+            "section_name": "Results",
+            "section_text": "For H7-specific mAb isolation, we purchased a soluble recombinant H7 HA protein based on A/Shanghai/2/2013 H7N9 for biotinylation, followed by streptavidin-PE conjugation. With this H7-PE bait, we stained 5 million PBMCs from the H7N9_HK2013 donor and sorted a total of 68 IgG+ B cells (defined as CD3-CD19+CD20+IgG+) that are H7-PE+ (Fig.\u00a01a). Most of the sorted cells were at the borderline of H7-PE staining, but a few stained brightly for H7-PE. From the sorted B cells, we performed single B cell RT-PCR and recovered four H7-reactive mAbs\u2014namely, H7.HK1, H7.HK2, H7.HK3, and H7.HK4.\n\na FACS depicting the staining and selection of H7-specific B cells from donor H7N9_HK2013 PBMCs. SSC-A, side scatter area; FSC-A, forward scatter area. b ELISA binding curves of the indicated mAbs to soluble recombinant H7N9 HA, with and without Endo H treatment, to H7N9 HA1 from 2013, 2016, and 2017, to H7N7 HA, and to 6 non-H7 HA or HA1. c Western blot of a cleaved H7 HA (molecular mass of 43\u2009kDa for HA1 and 30\u2009kDa for HA2) with mAb H7.HK2 or H7.HK4. d Neutralization curves of H7.HK mAbs against H7N9 2013 and 2017 pseudo viruses in MDCK cells. Data shown are mean\u2009\u00b1\u2009SEM. Source data are provided in the Source Data file. Similar results were independently reproduced at least once.\n\nMeasured by ELISA, the four reconstituted mAbs bound tightly to the H7N9 HA antigen used for H7-PE staining (Fig.\u00a01b). Pre-treating the H7N9 HA with Endoglycosidase H (Endo H) had no effect on the mAb binding profiles, indicating that these mAbs do not rely on H7 glycans for binding. After switching the ELISA coating antigen to HA1 of the matching H7N9 HA from A/Shanghai/2/2013, the binding curves of H7.HK1, H7.HK2, and H7.HK3 were fully retained, indicating that these mAbs bind to the globular head domain HA1; in contrast, H7.HK4 lost binding to H7N9 HA1, suggesting that its binding epitope is likely located in the HA2 stem domain (Fig.\u00a01b). Because of the documented antigenic drift for 2016\u20132017 H7N9 isolates, we also tested the mAb binding to HA1s from A/Guangdong/17SF003/2016 and A/Hong Kong/125/2017 H7N9. The binding curves of H7.HK1, H7.HK2, and H7.HK3 to both 2016 and 2017 HA1s were fully retained, and H7.HK4 did not bind to any HA1s. In addition, all four mAbs bound tightly to a recombinant H7N7 HA antigen based on A/Netherlands/219/2003 H7N7 (Fig.\u00a01b). To non-H7 HA proteins, H7.HK1 and H7.HK2 did not react with any of the tested non-H7 HA; H7.HK3 cross-reacted with H15N8 HA, and H7.HK4 cross-reacted with H10N8 and H15N8 HAs (Fig.\u00a01b), which sequence-wise are the closest to H7 in group 2 influenza HA genes38. Western blot of a cleaved HA based on A/Shanghai/1/2013 H7N9 confirmed that H7.HK2 binds to HA1 and H7.HK4 binds to HA2 (Fig.\u00a01c).\n\nUsing expression plasmids separately encoding H7 and N9 genes from A/Shanghai/4664T/2013 to pseudotype with HIV-1 NL4-3\u0394env.Luc\u00a0backbone39, we generated the H7N9 2013 pseudo virus and tested mAb neutralization by a luciferase readout from single round infection of Madin-Darby Canine Kidney (MDCK) cells (Fig.\u00a01d). H7.HK1 and H7.HK2 each potently neutralized the H7N9 2013 pseudo virus with an IC50 of 20\u2009ng/mL, while the other two mAbs H7.HK3 and H7.HK4 did not neutralize at up to 10\u2009\u03bcg/mL (Fig.\u00a01d, Table\u00a01). Similarly, we generated pseudo virus using the H7 from A/Hong Kong/125/2017 H7N9. Both H7.HK1 and H7.HK2 retained their neutralization titers against the H7N9 2017 pseudo virus with an IC50 of 30\u2009ng/mL, while the other two mAbs H7.HK3 and H7.HK4 did not neutralize (Fig.\u00a01d, Table\u00a01). We further assessed the mAb neutralization against three live replicating H7N9 viruses, A/Anhui/1/2013, A/Zhejiang/DTID-ZJU01/2013, and the donor\u2019s autologous isolate A/Hong Kong/470129/2013, for multiple rounds of infection in MDCK cells. Scored by the presence of cytopathic effect, mAbs H7.HK1 and H7.HK2 neutralized all three H7N9 live isolates with IC50s ranging 0.26\u20131.0\u2009\u03bcg/mL; however, they did not neutralize any non-H7N9 influenza isolates tested, indicating that these mAbs are specific to H7N9 (Table\u00a01). The other two mAbs H7.HK3 and H7.HK4 did not neutralize any of the tested H7N9 and therefore were not tested against non-H7N9 viruses. The neutralization IC50s of H7.HK1 and H7.HK2 using pseudo viruses were about 10-fold more potent than those using live replicating viruses, suggesting that the pseudo virus neutralization is more sensitive thus useful for initial screening of neutralizing mAbs, which could then be confirmed with live replicating viruses. Similar differences in IC50 values have been reported for other HA-reactive mAbs tested by both pseudo virus and live replicating virus29.\n\nThree previous mAbs representing the best from each corresponding study were compared to H7.HK2 for ELISA binding to H7 antigens and neutralization of H7N9 pseudo viruses. Cloned from plasmablasts of an acute infection in 2013\u2013201417, mAb L4A-14 is directed at the receptor-binding site (RBS) and bound 2013 H7N9 HA1 and HA similarly to H7.HK2, retained full binding to 2016 and 2017 HA1s (Supplementary Fig.\u00a01a) but lost vast majority of binding to H7N7 HA (Supplementary Fig.\u00a01b). Derived from EBV transformed B cells after vaccination19, another RBS-directed mAb H7.167 bound less well than H7.HK2 to 2013 H7N9 HA1 and HA and lost substantial binding to 2016 and 2017 HA1s and to H7N7 HA. Cloned from plasmablasts after vaccination20, the non-RBS mAb 07-5F01 bound 2013, 2016, and 2017 H7N9 HA1s similarly to H7.HK2 and fully retained reactivity to H7N7 HA. Evaluated by pseudo virus neutralization (Supplementary Fig.\u00a01c), the RBS-directed mAbs H7.167 and L4A-14 neutralized the 2013 H7N9 weaker than H7.HK2. L4A-14 neutralized the 2017 strain slightly better than 2013 H7N9, but H7.167\u2019s activity was further reduced by the 2017 virus. The non-RBS mAb 07-5F01 neutralized both 2013 and 2017 H7N9 viruses with comparable potency to H7.HK2. Hence, evaluated by H7 antigen binding and pseudo virus neutralization, H7.HK2 is superior to the two best previous RBS-directed mAbs L4A-14 and H7.167 and matches the one best previous non-RBS mAb 07-5F01 against H7N9.\n\nSequence analysis revealed that all four H7.HK mAbs are IgG1 (Table\u00a02). H7.HK1 and H7.HK2 are clonal variants using IGHV4-59 for heavy chain with 8-10% somatic hypermutation (SHM) and a complementarity-determining region (CDR) H3 of 11 amino acids according to the Kabat and\u00a0Chothia definition40,41,42, and IGKV2-28 for light chain with 6% SHM and a CDR L3 of 9 amino acids. Though clonally related, H7.HK1 and H7.HK2 share only 3 out of 13-15 amino acid SHMs in the heavy chain V-gene and 1 out of 8 amino acid SHMs in the light chain V-gene (Supplementary Fig.\u00a02). A putative N-linked glycosylation site is present in the light chain CDR L1 of H7.HK1 and H7.HK2. H7.HK3 uses IGHV7-4-1 for heavy chain with 7% SHM and a CDR H3 of 14 amino acids, and IGKV1-5 for light chain with 5% SHM and a CDR L3 of 8 amino acids. A putative N-linked glycosylation site is also present in H7.HK3 at the heavy chain CDR H2. H7.HK4 uses IGHV4-61 for heavy chain with 7% SHM and a CDR H3 of 13 amino acids, and IGKV1-16 for light chain with 5% SHM and a CDR L3 of 9 amino acids (Table\u00a02, Supplementary Fig.\u00a02).\n\nFor structural analysis, we expressed a soluble, disulfide-stabilized, and fully cleaved H7 HA trimer by transient transfection of Expi293F cells. H7.HK1 and H7.HK2 bound the H7 HA trimer tightly, H7.HK3 bound less well, and H7.HK4 did not bind at all (Supplementary Fig.\u00a03a). As expected, the three previous neutralizing mAbs all bound the H7 HA trimer tightly, with H7.167 showing weaker binding (Supplementary Fig.\u00a03b). We next generated the antibody fragments for antigen binding (Fabs) of H7.HK1 and H7.HK2 to bind the H7 HA trimer. We froze grids containing the Fab:HA complexes and determined cryo-EM structures of each Fab bound to an H7 HA trimer (Supplementary Fig.\u00a04). A resolution of 3.62\u2009\u00c5 for H7.HK1 and 3.69\u2009\u00c5 for H7.HK2 was achieved (Fig.\u00a02a, Supplementary Fig.\u00a05, Table\u00a0S1). These complex structures demonstrate that H7.HK1 and H7.HK2 are highly superimposable (Fig.\u00a02b) and their interactions with H7 are centered at \u03b214 and extended to the surfaces of \u03b210 and \u03b219 (Fig.\u00a02c). This \u03b214-targeting surface partially overlaps with the antigenic site D towards sites A and B as previously mapped on H323,25. Analysis of the H7.HK1 epitope demonstrates that most interactions are driven by the heavy chain and consist of seven hydrogen bonds (Y52:E121, R94:G124, G99:S167, D100:T126, Y100a:T165, Y100a:S167, S100c:T126) and one salt bridge (H53:E121) (Fig.\u00a02d). The light chain is less involved in binding, making only one hydrogen bond (Y49:Q163) and weak hydrophobic interactions (Fig.\u00a02e). The light chains of both H7.HK1 and H7.HK2 are glycosylated in CDR L1; this glycan plays no role in binding, but there is good density to support its presence. The epitope of H7.HK2 is similar to that of H7.HK1, only differing in slight contacts on the periphery (Supplementary Fig.\u00a06a). In addition, nearly all hydrogen bonds are conserved between the two antibodies (Supplementary Fig.\u00a06b). However, the substitution of F56S in CDR L2 of H7.HK2 results in an additional hydrogen bond with HA G129. This substitution also shifts the orientation of H7.HK2 CDR L2 slightly so that Y49 interacts with T165 for H7.HK2 instead of Q163 for H7.HK1 (Supplementary Fig.\u00a06c). Finally, as H53 is substituted with tyrosine in the heavy chain of H7.HK2, it does not make the H53:E121 salt bridge.\n\na Cryo-EM structures of H7.HK1 and H7.HK2 bound to H7 HA in the head region. b Top view of alignment of H7.HK1 and H7.HK2 complex structures. c Surface presentation of the H7.HK1 epitope (orange) on H7 HA1, with interacting CDRs shown. d H7.HK1 heavy chain forms seven hydrogen bonds and one salt bridge with H7 HA1. e H7.HK1 light chain forms one additional hydrogen bond with H7 HA1, and the interactions are stabilized by hydrophobic residues on the periphery of the light chain interface. f Modeling published structures of H7 HA1-binding antibodies (PDB: 6II4, 6II8, 6II9, 5V2A) onto the H7.HK1 bound structure, with an escape mutation R57K (green) reported for mAb 07-5F01. Competition ELISA with biotinylated H7.HK2 binding to the H7 HA monomer, in which unlabeled competing mAbs were titrated at increasing concentrations to evaluate the effect on H7.HK2 binding. Source data is provided in the Source Data file. g Sequence analysis of N\u2009=\u20091,483 H7 HA1s revealed a conserved lateral patch that largely overlaps with the H7.HK1 epitope. h Modeling the binding site of human receptor analogue LSTc (red) based on a previous crystal structure (PDB: 4BSE) onto H7 from the H7.HK1 complex, showing that H7.HK1 does not compete with sialic acid on the adjacent protomer (black). Alignment of the H7.HK1 complex with a previous crystal structure of H7 (PDB: 4BSE) shows that the 220-loop (pink) required for sialic acid binding (G218-G228) is disordered and would clash with the H7.HK1 light chain if it were present. Green asterisk symbol denotes the <2\u2009\u00c5 clash between the CDR L1 N28 and the predicted location of P221 on HA1.\n\nTo analyze the mechanism of neutralization, we first compared the binding site of H7.HK1 to that of four other H7-reactive antibodies with published structures, L4A-14, L4B-18, L3A-44 (PDB: 6II4, 6II8, 6II9)17 and H7.167 (PDB: 5V2A)19. This analysis demonstrates that the binding site of H7.HK1 is almost completely distinct from that of these previously published antibodies, which compete for the RBS (Fig.\u00a02f). The binding site of H7.HK1 is also distant from that of 07-5F01, which was mapped to an escape mutation R65K (corresponding to R57K here by H3 numbering) of HA120. Direct competition ELISA applying biotin labeled H7.HK2 to bind H7 HA did not detect any effective competition by H7.HK3, H7.HK4, L4A-14, H7.167, and 07-5F01 (Fig.\u00a02f), confirming the unique location of H7.HK1 and H7.HK2 epitopes. Further analysis of N\u2009=\u20091,483 H7 HA1 amino acid sequences from the Global Initiative on Sharing All Influenza Data (GISAID) revealed a conserved lateral patch (Fig.\u00a02g) similar to what was initially identified in H1 viruses43.\n\nStrikingly, the lateral patch epitope of H7.HK1 was distal to the RBS of the protomer with which it interacted and closer to the RBS on the adjacent protomer. To further examine the relationship between the mAb binding site and RBS, the human receptor analogue Sialylneolacto-N-tetraose c (LSTc) was modeled into the RBS of H7 (PDB: 4BSE)44 in the H7.HK1 complex. Interestingly, no steric clashes were observed between H7.HK1 and sialic acid bound to the adjacent protomer, and no mAb interaction with RBS (Fig.\u00a02g). However, the HA 220-loop (G218-G228), which makes hydrophobic contacts with sialic acid has no density present in the structure of H7.HK1 or H7.HK2 bound to HA, suggesting that these antibody binding causes the\u00a0220-loop to become disordered. All previously mentioned H7 structures (Fig.\u00a02f), as well as an additional cryo-EM structure in which Fab 1D12 is bound to the stem region of H7 HA (PDB: 6WXL)35 have consistent electron density accounting for this loop. Alignments of the H7.HK1 complex structure with the crystal structure of H7 HA bound to LSTc (PDB: 4BSE)44 demonstrate where the 220-loop would be when receptor is bound and that the light chain of H7.HK1 would clash with this loop (Fig.\u00a02h), further supporting that H7.HK1 and H7.HK2 act by causing the\u00a0220-loop to become disordered, thus preventing its interactions with the sialic acid receptor. The HA1 trimer interface mAb FluA-20 interacts with the non-RBS side of 220-loop on the protomer it interacts with21. To our knowledge, this mechanism of neutralization employed by H7.HK1 and H7.HK2\u2014disrupting the 220-loop on H7 trimer\u2014is distinct from previously reported HA1-directed H7N9 neutralizing mAbs, which all directly compete with sialic acid for binding to HA on the protomer they interact with17,19,21,22,28.\n\nSince the H7N9 HA gene has significantly evolved and changed in 2016\u20132017 compared to that of 2013 (with up to 12 amino acid substitutions in HA1), we examined the locations of mutated residues in the epitopes of H7.HK1 and H7.HK2 that consist of 32 contacting residues in HA1 for both mAbs (Supplementary Fig.\u00a07a). There are three mutations in the binding site of H7.HK1 and H7.HK2\u2014namely, A122T/P, S128N, and R172K, appeared in 2016\u20132017 compared to the 2013 H7N9, and all three mutations are located at one side edge of the epitopes (Supplementary Fig.\u00a07b), thus not altering the mAb interactions with HA1. This analysis is consistent with the intact binding of H7.HK1 and H7.HK2 to both 2016 and 2017 HA1s aligned to the 2013 HA1 (Fig.\u00a01b) and the mAbs\u2019 retention of neutralization against the H7N9 2017 pseudo virus (Fig.\u00a01d).\n\nWe next assessed the prophylactic and therapeutic effect of H7.HK mAbs as human IgG1 in a mouse lethal challenge model. To assess mAb prophylactic effect, BALB/c mice (n\u2009=\u20095\u201310 per group from 1 to 2 experiments) were injected intraperitoneally (i.p.) with human H7N9 mAbs one day before intranasal (i.n.) challenge of 10-fold 50% lethal dose (10 LD50) of A/Anhui/1/2013 H7N9 virus. Given 100\u2009\u03bcg per mouse (equivalent to 5\u2009mg/kg), the neutralizing mAbs H7.HK1 and H7.HK2 each fully protected mice without apparent weight loss (Fig.\u00a03a); given 20\u2009\u03bcg per mouse (equivalent to 1\u2009mg/kg), H7.HK2 still fully protected mice from death (defined as \u226520% weight loss), with up to 8% average weight loss; H7.HK1 protected 7 out of 10 mice from death, with up to 12% average weight loss for mice that survived (Fig.\u00a03a). By day 2 post challenge, the weight preservation was significantly better in mice receiving 20\u2009\u03bcg of H7.HK1 or H7.HK2 than mice receiving the placebo mAb or phosphate buffered saline (PBS). Mice receiving the non-neutralizing mAbs H7.HK3 or H7.HK4 (100\u2009\u03bcg or 20\u2009\u03bcg) were not protected and showed no difference from placebo mAb and PBS controls (Fig.\u00a03a).\n\na Female mice were i.p. injected 100\u2009\u03bcg (equivalent of 5\u2009mg/kg) or 20\u2009\u03bcg (equivalent of 1\u2009mg/kg) of the indicated mAbs (as human IgG1 unless otherwise specified) one day before viral challenge; % survival (<20% weight loss) and % body weight of survived mice were plotted over time. b Female mice were i.p. injected 100\u2009\u03bcg of the indicated mAbs one day after viral challenge; % survival and % body weight of survived mice were plotted over time. Arrows indicate the time when mAbs were administered. Control groups of a non-H7 placebo mAb and PBS were included. Data for each group were combined from 1 to 2 experiments and shown as mean \u2012 SEM. Source data with P values from two-sided unpaired student\u2019s t-test are provided in the Source Data file. Asterisk symbols denote P\u2009<\u20090.05, and # denote P\u2009<\u20090.1.\n\nSince anti-HA2 mAbs have demonstrated Fc-mediated protection against influenza45, we converted the anti-HA2 non-neutralizing mAb H7.HK4 to mouse IgG2a\u2014an isotype that mediates strong Fc effector function in mice, and tested it for prophylaxis in the mouse challenge model, along with mouse IgG1, which lacks Fc effector function in mice36. Given 100\u2009\u03bcg per mouse, H7.HK4 mouse\u00a0IgG2a but not IgG1 protected 4 out of 5 mice from death, with up to 17% average weight loss for mice that survived (Fig.\u00a03a). By day 3 post challenge, the weight preservation was significantly better in mice receiving H7.HK4 mouse\u00a0IgG2a than mice receiving H7.HK4 mouse\u00a0IgG1 or placebo mouse\u00a0IgG2a. Though survived, mice receiving 100\u2009\u03bcg H7.HK4 mouse\u00a0IgG2a lost more weight than those receiving 20\u2009\u03bcg neutralizing mAbs H7.HK1 or H7.HK2 (Fig.\u00a03a), indicating less prophylaxis efficiency for H7.HK4 than H7.HK1 and H7.HK2.\n\nSince the H7.HK2 and H7.HK4 mAbs bind to different sites on the HA and protect through different mechanisms, we tested the combination of suboptimal dose of 20\u2009\u03bcg H7.HK2 (as human IgG1) with 100\u2009\u03bcg H7.HK4 mouse\u00a0IgG2a in the mouse challenge model, using 20\u2009\u03bcg H7.HK2 (as human IgG1) with 100\u2009\u03bcg H7.HK4 mouse\u00a0IgG1 as a control. Compared to this control group, which protected 9 out of 10 mice from death and lost up to 11% body weight for mice that survived, the combination of 20\u2009\u03bcg of H7.HK2 (as human IgG1) with 100\u2009\u03bcg H7.HK4 mouse\u00a0IgG2a fully protected mice from death, with only up to 7% weight loss, and the weight difference was statistically significant between these two groups since day 3 post challenge (Fig.\u00a03a), indicating a beneficial role of H7.HK4 in mAb combination regimen. Overlaying the survival and body weight data of the 20\u2009\u03bcg H7.HK2 alone group from the previous experiment (Fig.\u00a03a), H7.HK2 in combination with H7.HK4 mouse\u00a0IgG2a did not improve the body weight trough from day 4\u20136 post challenge as both groups fully protected mice from death with up to 7-8% weight loss; the mAb combination demonstrated a statistical trend and then significance for improved recovery of weight loss starting on day 7 post challenge (Fig.\u00a03a).\n\nTo assess mAb therapeutic effects, we first i.n. challenged mice (n\u2009=\u20095\u201310 per group from 1 to 2 experiments) with 10 LD50 of A/Anhui/1/2013 H7N9 virus, waited for one day, and then on day 1 post challenge i.p. injected mice with 100\u2009\u03bcg H7.HK1 or H7.HK2 as human IgG1, or H7.HK4 as mouse\u00a0IgG2a (Fig.\u00a03b). Twelve and 13 out of 15 mice receiving 100\u2009\u03bcg H7.HK1 or H7.HK2 one day after viral challenge initially lost weight similarly to placebo and PBS controls but then started to recover on day 5 after challenge. Therefore, the neutralizing mAbs H7.HK1 and H7.HK2 showed both prophylactic and therapeutic efficacies in the mouse lethal challenge model. None of the 5 mice receiving 100\u2009\u03bcg H7.HK4 mouse\u00a0IgG2a one day after challenge survived (Fig.\u00a03b), indicating that H7.HK4 as mouse\u00a0IgG2a demonstrated measurable prophylactic effect but not therapeutic efficacy.\n\nThus far, there have been two other published structures of lateral patch-binding antibodies on the HA head, 045-09-2B05 (PDB: 7MEM) and Fab6649 (PDB: 5W6G)\u2014both of which bind H143,46. Comparing 045-09-2B05, Fab6649, and H7.HK1 demonstrated diverse angles of approach and different heavy and light chain orientations towards the lateral patch (Fig.\u00a04a). The epitopes of these three antibodies were all centered on the lateral patch, with the composite footprint of all three defining a lateral patch supersite of vulnerability (Fig.\u00a04b). The epitope of H7.HK1 was most similar to that of Fab6649. However, the light chain of H7.HK1 was in a slightly higher position on the head of HA than the heavy chain of Fab6649 (Fig.\u00a04a, b), which allowed CDR\u00a0L1 to clash with the 220-loop. H7.HK1 heavy chain overlapped with the light chain of 045-09-2B05, and the heavy chain of 045-09-2B05 occupied an epitope distinct from H7.HK1, H7.HK2, and Fab6649. The epitopes of 045-09-2B05 and Fab6649 had modest overlap, centered around the conserved lateral patch (Fig.\u00a04c colored in magenta). Of this overlapping epitope surface, there were four residues conserved between H1 and H7 (positions E121, S/T126, Y168, and R/K172). The overall structure of this site of vulnerability was also conserved between H1 and H7 (Fig.\u00a04c). Comparison of the interactions between Fab6649, 045-09-2B05, H7.HK1, H7.HK2, and these four conserved HA\u00a0residues revealed different modes of recognition for each antibody (Fig.\u00a04d). Thus, the lateral patch supersite, as defined by Fab6649, 045-09-2B05, H7.HK1, and H7.HK2, was composed in part by residues that were conserved in H1 and H7 and could be targeted via diverse chemistries and modes of recognition.\n\na Representation of H7.HK1, Fab6649, and 045-09-2B05 bound to their respective HAs indicates diverse angles of approach and heavy/light chain orientations towards the lateral patch supersite. b Comparison of epitopes of H7.HK1, Fab6649, and 045-09-2B05 centering on the lateral patch defines the lateral patch supersite (blue). c A subset of epitope surface, centered on the lateral patch, overlaps between H7.HK1, Fab6649, and 045-09-2B05 (magenta). Of shared epitope surface, a subset of epitope residues is conserved (positions 121, 126, 168, and 172 by H3 numbering). d The four lateral patch antibodies contact conserved residues using diverse chemistry. Table displays the type of interaction between each residue and antibody.",
+            "section_image": [
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-024-48758-4/MediaObjects/41467_2024_48758_Fig1_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-024-48758-4/MediaObjects/41467_2024_48758_Fig2_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-024-48758-4/MediaObjects/41467_2024_48758_Fig3_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-024-48758-4/MediaObjects/41467_2024_48758_Fig4_HTML.png"
+            ]
+        },
+        {
+            "section_name": "Discussion",
+            "section_text": "Already endemic, adapted, and evolved in humans for 11 years, H7N9 continues to pose risk and infect humans exposed to infected poultry in China. While the current risk to public health is low, the pandemic potential of H7N9 is especially concerning if it were to gain the ability of sustained human-to-human transmission. Based on its biological features such as dual affinity for avian and human receptors, high case-fatality rate, resistance to neuraminidase inhibitors, and lack of pre-existing immunity in the human populations, there is an immediate need and interest to develop human mAb prophylaxis and therapeutics against H7N9, to which a specific treatment or licensed vaccine (for humans) is not available.\n\nIn this study, we identified two HA1-directed clonally related human mAbs H7.HK1 and H7.HK2 that neutralized H7N9 with potencies and mouse protection efficacies (prophylactic and therapeutic) in line with the best of previous H7N9 mAbs. Specifically, a combined phage library from three H7N9 convalescent cases yielded a single neutralizing mAb clone18, represented by the best clonal member, HNIgGA6, which neutralized H7N9 and protected mice against a lethal challenge at 5\u2009mg/kg with up to about 10% weight loss18. Likewise, from a study of four H7N9 acutely infected cases, the best mAb L4A-14 cloned from plasmablasts protected mice against a lethal challenge at 10\u2009mg/kg with up to about 10% weight loss17. The most potent mAb H7.167 from a study of EBV transformed B cells from five representative H7N9 experimental vaccinees neutralized H7N9 and protected mice against a sub-lethal challenge of H7-PR8 at 1.65\u2009mg/kg without apparent weight loss19. The best HA1-directed neutralizing mAb 07-5F01 from a study of H7N9 experimental vaccinees\u2019 plasmablasts protected mice against a lethal challenge at 0.3\u2009mg/kg without apparent weight loss20. The broad HA1 trimer interface mAb FluA-20 from a healthy donor with extensive influenza vaccinations lacked in vitro neutralization but protected mice against a sub-lethal challenge of H7-PR8 at 10\u2009mg/kg without apparent weight loss21. In comparison, H7.HK1 and H7.HK2 protected mice against a lethal challenge at 1\u2009mg/kg with up to 12% weight loss.\n\nWe have also structurally defined the epitopes of H7.HK1 and H7.HK2 to the \u03b214-centered surface of H7 HA1, partially overlapping with antigenic site D, targeting a lateral patch43 rather than the commonly targeted RBS and trimer interface by previous H7N9 mAbs28. Structural comparisons demonstrated that H7.HK1 and H7.HK2 interacted with H7 differently from L4A-14, H7.167, 07-5F01, and FluA-20. By escape mutations, a previous H3 neutralizing mAb D1-8 was mapped to the lower part of antigenic site D towards site E47; this epitope partially overlaps with the H7.HK1 and H7.HK2 epitope described here. However, without structural data, the action of neutralization by D1-8 cannot be determined. Importantly, D1-8 does not react to H7, and likewise, H7.HK1 and H7.HK2 do not react to H3. Hence, D1-8 cannot replace the anti-H7N9 function of H7.HK1 and H7.HK2. The unique \u03b214-targeting epitope on the HA1 lateral patch would render H7.HK1 and H7.HK2 favorable candidates for combination prophylaxis and therapy against H7N9 to augment protection efficacy and increase the genetic barrier for viral escape. This is supported by data showing no competition between H7.HK1 and H7.HK2 with two RBS and one non-RBS directed antibodies (Fig.\u00a02f). Indeed, the H1 lateral patch has been a promising target for next-generation H1 vaccine strategy as previous studies showed that H1 lateral patch-binding antibodies are abundant in humans, and that they react broadly with H1 viruses and cross react with some H3 viruses46,48. Our work here demonstrates that the lateral patch is a viable epitope for H7 vaccine and therapeutic antibody development as well. Previously reported lateral patch antibodies are all restricted to expressing IGHV3 or IGHV4-39 genes and often had a Y-x-R motif in CDR\u00a0H343,46. H7.HK1 and H7.HK2 are derived from IGHV4-59 and do not contain a Y-x-R motif in CDR\u00a0H3. In addition, unlike other reported lateral patch antibodies, H7.HK1 and H7.HK2 disrupt the structure of the 220-loop of the H7 RBS. Therefore, the lateral patch binding site of HA1 is expanded to H7 and composed in part with residues conserved between H1 and H7, which makes it a supersite of vulnerability that could be targeted by more diverse antibodies than previously recognized.\n\nH7N9 has evolved over time and its HA gene has significantly changed in 2016\u20132017 compared to that of 2013. Consequently, most neutralizing mAbs isolated from individuals infected or vaccinated with the 2013 H7 HA lost reactivity to 2016\u20132017 isolates, requiring updated H7 immunogens for mAb and vaccine development17. We show that three mutations appeared in 2016\u20132017 are located at the periphery of the H7.HK1 and H7.HK2 epitopes and confirmed that the binding profiles of H7.HK1 and H7.HK2 are intact to both 2016 and 2017 HA1s as compared to 2013 HA1. We also showed that H7.HK1 and H7.HK2 retained their neutralization titers against the H7N9 2017 pseudo virus. The previous RBS-directed mAb HNIgGA6 was shown to lose reactivity to V186G and L226Q mutations18 that are present in A/Netherlands/219/2003 H7N7 and A/Guangdong/17SF003/2016 H7N9, respectively. Evaluated by both antigen binding and pseudo virus neutralization, H7.HK2 is superior to the two best previous RBS-directed mAbs L4A-14 and H7.167 and matches the one best previous non-RBS mAb 07-5F01 against H7N9.\n\nLastly, we tested a suboptimal dose of H7.HK2 combining with the HA2-directed non-neutralizing mAb H7.HK4 against mouse lethal challenge. Compared to HA1 (head domain of HA), the HA2 (stalk domain) is genetically more conserved. Hence, HA2-directed mAbs typically display broader recognition of HA subtypes than HA1-directed mAbs. This is indeed the case for H7.HK4, i.e., in addition to H7N9 and H7N7, it also recognized the HAs from H10N8 and H15N8, to which both H7.HK1 and H7.HK2 had no reactivity. When converted to mouse IgG2a enabling Fc effector function in mice, H7.HK4 demonstrated moderate prophylactic protection at 5\u2009mg/kg and augmented mouse protection of H7.HK2, supporting the inclusion of HA2-directed antibodies in a mAb combination regimen against H7N9.\n\nIn summary, from a 2013 H7N9 convalescent case occurring in Hong Kong, we isolated two clonally related HA1-directed neutralizing mAbs H7.HK1 and H7.HK2 that demonstrated prophylactic and therapeutic efficacies in a mouse lethal challenge model. Cryo-EM structures revealed a \u03b214-centered site of vulnerability targeted by H7.HK1 and H7.HK2, those being the first reported antibodies to bind to the H7 lateral patch. Recognition of this conserved epitope facilitates near full binding and neutralization capacity of H7.HK1 and H7.HK2 to the later 2016\u20132017 H7N9 isolates. This unique epitope at the lateral patch of HA1 renders H7.HK1 and H7.HK2 favorable candidates for combination prophylaxis and therapy against H7N9, which may include multiple HA1-directed neutralizing mAbs targeting different epitopes and benefit from the inclusion of HA2-directed mAbs as well.",
+            "section_image": []
+        },
+        {
+            "section_name": "Methods",
+            "section_text": "A blood specimen was collected from the H7N9_HK2013 patient about one year after recovery from a hospitalized severe H7N9 infection. Written informed consent was obtained from the patient. Each specimen is unique and cannot be replaced once processed. The study was approved by the Institutional Review Board (IRB) of the University of Hong Kong and the Hospital Authority (Reference number: UW-13-265).\n\nExpression plasmids encoding the H7 hemagglutinin and N9 neuraminidase based on A/Shanghai/4664T/2013 H7N9 strain were obtained from Dr. Jianqing Xu39. Codon-optimized gene encoding the H7 hemagglutinin of A/Hong Kong/125/2017 H7N9 was synthesized (Twist Bioscience) and cloned into pcDNA3.1 (Invitrogen). HIV-1 NL4-3\u0394env.Luc.R-E-\u00a0backbone was obtained through the NIH HIV Reagent Program as contributed by Dr. Nathaniel Landau. These plasmids were used to co-transfect 293T cells to generate H7N9 pseudo viruses. All live replicating influenza A viruses used in this study were isolated from patients and include A/Hong Kong/470129/2013 H7N914, A/Zhejiang/DTID-ZJU01/2013 H7N93, A/Anhui/1/2013 H7N9 (obtained from the China Center for Disease Control and Prevention), A/Vietnam/1194/2004 H5N1, A/Hong Kong/459094/2010 H5N1, A/Hong Kong/1073/1999 H9N2, A/Hong Kong/415742/2009 H1N1, and A/Hong Kong/400500/2015 H3N2. DNA sequences encoding the variable regions of previous H7N9 mAbs L4A-14 (PDB: 6II4), H7.167 (PDB: 5V2A), and 07-5F01 (GenBank KU987563 and KU987564) were synthesized (Twist Bioscience) and cloned into the corresponding human gamma, kappa, and lambda chain expression vectors described49,50,51. Full IgG1 was expressed by co-transfecting Expi293F cells with equal amounts of paired heavy and light chain plasmids and purified using recombinant protein A agarose (Thermo Fisher). The non-H7N9 placebo mAb used in this study, AD358_n1, has been described51 and is specific to HIV-1 gp120. Human embryonic kidney 293 cell line, of which the sex is female, is the parental cell for 293T and Expi293F cell lines. 293T was obtained from ATCC (Cat. No. CRL-11268, Clone 17) and maintained as adherent cells in complete DMEM medium at 37\u2009\u00b0C. 293T is highly transfectable and contains SV40 T-antigen. Expi293F was obtained from Thermo Fisher (Cat. No. A14527) and adapted to suspension culture in Expi293 Expression Medium at 37\u2009\u00b0C. Madin-Darby Canine Kidney (MDCK) cell line, of which the sex is female, was obtained from ATCC (Cat. No. CCL-34) and maintained as adherent cells in complete DMEM medium at 37\u2009\u00b0C. The vendors provided certificates of analysis for the cell lines. No further authentication was performed on cell lines used in this study.\n\nA soluble recombinant HA \u0394TM antigen based on A/Shanghai/2/2013 H7N9 (Immune Technologies, Cat. No. IT-003-0074\u0394TMp) was biotinylated via EZ-Link (Thermo Fisher, Cat. No. A39256), followed by streptavidin mediated conjugation of phycoerythrin (PE) (Invitrogen, Cat. No. SA10041). The patient PBMCs were stained with an antibody cocktail to CD3-PE-CF594 (BD Biosciences, Cat. No. 562406, Clone SP34-2), CD19-PE-Cy7 (BioLegend, Cat. No. 302216, Clone HIB19), CD20-APC-Cy7 (BioLegend, Cat. No. 302314, Clone 2H7), IgG-FITC (BD Biosciences, Cat. No. 555786, Clone G18-145), and IgM-V450 (BD Biosciences, Cat. No. 561286, Clone G20-127). In addition, live/dead yellow stain (Invitrogen, Cat. No. L34968) was used to exclude dead cells. All staining antibodies were purchased from vendors providing Quality Certificates and further validated on healthy human blood donor PBMCs purchased from New York Blood Center. After washing, H7-PE+\u2009B cells were sorted using a multi-laser MoFlo sorter (Beckman Coulter, Jersey City, NJ). Fluorescence compensation was performed with anti-mouse Ig kappa CompBeads (BD Biosciences, Cat. No. 552843). Individual B cells were sorted into a 96-well PCR plate, each well containing 20\u2009\u03bcL lysis buffer composed of 0.5\u2009\u03bcL RNaseOut (Invitrogen, Cat. No. 10777019), 5\u2009\u03bcL 5x first-strand buffer, 1.25\u2009\u03bcL 0.1\u2009M DTT, and 0.0625\u2009\u03bcL Igepal (Sigma, St. Louis, MO). The PCR plate with sorted cells was frozen on dry-ice and then stored at \u221280\u2009\u00b0C. The total cell sample passing through the sorter was analyzed with FlowJo 10.0 (TreeStar, Cupertino, CA).\n\nFrom each sorted cell, the variable regions of IgG heavy and light chains were amplified by RT-PCR and cloned into expression vectors as previously described51. Briefly, frozen plates with single B\u00a0cell RNA were thawed at room temperature, and RT was carried out by adding into each well 3\u2009\u03bcL random hexamers at 150\u2009ng/\u03bcL (Gene Link, Cat. No. 26-4000-03), 2\u2009\u03bcL dNTP (each at 10\u2009mM), and 1\u2009\u03bcL SuperScript II (Invitrogen, Cat. No. 18064022), followed by incubation at 42\u2009\u00b0C for 2\u2009h. We note that these RT parameters may be suboptimal to those described previously49,50. After RT, 25\u2009\u03bcL water was added to each well to dilute cDNA, and the cDNA plates were stored at \u221220\u2009\u00b0C for later use. The variable regions of heavy, kappa, and lambda chains were amplified independently by nested PCR in 50\u2009\u03bcL, using 5\u2009\u03bcL cDNA as template, with HotStarTaq Plus DNA polymerase (Qiagen, Cat. No. 203605) and primer mixes as described49,52. Cycler parameters were 94\u2009\u00b0C for 5\u2009m, 50 cycles of 94\u2009\u00b0C for 30\u2009s, 52\u201355\u2009\u00b0C for 30\u2009s, and 72\u2009\u00b0C for 1\u2009m, followed by 72\u2009\u00b0C for 10\u2009m. The PCR amplicons were subjected to direct Sanger sequencing, and the antibody sequences were analyzed using IMGT/V-QUEST. Selected PCR sequences that gave productive gamma, kappa, and lambda chain rearrangements were re-amplified with custom primers containing unique restriction digest sites and cloned into the corresponding human gamma, kappa, and lambda chain expression vectors as described49,50,51. Full IgG1 was expressed by co-transfecting Expi293F cells with equal amounts of paired heavy and light chain plasmids and purified using recombinant protein A agarose (Thermo Fisher).\n\nH7N9 \u0394TM HA and HA1 based on A/Shanghai/2/2013, HA1s based on A/Guangdong/17SF003/ 2016, A/Hong Kong/125/2017, and H7N7 \u0394TM HA based on A/Netherlands/219/2003 were purchased from Immune Technologies (Cat. No. IT-003-0074\u0394TMp, IT-003-0074p, IT-003-0075p, IT-003-0076p, IT-003-0081\u0394TMp). Other non-H7 \u0394TM HA proteins were purchased from Sino Biological (Cat. No. 40787-V08H, 11688-V08H, 40868-V08B, 40359-V08B, 11720-V08H). ELISA plates were coated with HA or HA1 antigens at 2\u2009\u03bcg/mL in PBS at 4\u2009\u00b0C overnight. For Endo H treatment, the required amount of antigen was diluted in 10x buffer and mixed with 1\u2009\u03bcL Endo H (New England BioLabs, Cat. No. P0702S) at 37\u2009\u00b0C for 1\u2009h; an equal amount of antigen (untreated) was processed under identical condition without Endo H. Both treated and untreated antigens were diluted in PBS to coat ELISA plates. Coated plates were blocked with 1% BSA (bovine serum albumin) in PBS, followed by incubation with serially diluted mAbs at 37\u2009\u00b0C for 1\u2009h. Horseradish peroxidase (HRP)-conjugated goat anti-human IgG Fc (Jackson ImmunoResearch, Cat. No. 109-035-098) was added at 1:10,000 at 37\u2009\u00b0C for 1\u2009h. All ELISA incubation volumes were 100\u2009\u03bcL/well except that 200\u2009\u03bcL/well was used for blocking. Plates were washed between steps with 0.1% Tween 20 in PBS and developed with 3,3\u2032,5,5\u2032-tetramethylbenzidine (TMB) (Novex, Life Technologies), with 1\u2009M H2SO4 as terminator and read at 450\u2009nm. For competitive ELISA, plates were coated with 2\u2009\u03bcg/mL of H7N9 \u0394TM HA. After blocking, serial dilutions of competing mAbs were added in 50\u2009\u03bcL of blocking buffer, followed by addition of 50\u2009\u03bcL of biotin labeled H7.HK2 at 100\u2009ng/mL. After incubation at 37\u2009\u00b0C for 1\u2009h, the plates were washed and then incubated with 250\u2009ng/mL of streptavidin-HRP (Jackson ImmunoResearch, Cat. No. 016-030-084) at ambient temperature for 30\u2009m before development with TMB as described above. Similar results were independently reproduced at least once.\n\nCleaved HA protein from A/Shanghai/1/2013 H7N9 (HA1\u2009+\u2009HA2, cleavage) (Sino Biological, Cat. No. 40104-V08H4) was added at 1\u2009\u00b5g with 4x SDS loading buffer with reduced reagent and heated at 70\u2009\u00b0C for 10\u2009m. The protein was separated on NuPAGE 4-12% Bis-Tris gel with MOPS running buffer (Invitrogen) and transferred to PVDF membrane semi-dry with the Bio-Rad trans-blot turbo transfer system. The membrane was blocked in 2% skim milk in PBS-T, followed by incubation with mAb H7.HK2 or H7.HK4 as primary antibody at 1\u2009\u00b5g/mL in blocking buffer at 4\u2009\u00b0C overnight. HRP-conjugated goat anti-human IgG Fc (Jackson ImmunoResearch, Cat. No. 109-035-098) was used as secondary antibody at 1:10,000 in blocking buffer at room temperature for 1\u2009h. The immunoreactive band was detected with ECL reagent (Thermo Fisher, Cat. No. A38555). Similar results were independently reproduced at least once.\n\nH7N9 neutralization was first measured with a single-round infection of MDCK cells using pseudo viruses expressing the H7 gene from A/Shanghai/4664T/2013 H7N9 or A/Hong Kong/125/2017 H7N9, and the N9 gene from A/Shanghai/4664T/2013 H7N9, pseudotyped with the HIV-1 NL4-3\u0394env.Luc.R-E- backbone39. In 96-well plate, 70\u2009\u03bcL of antibody-virus mixture were incubated at 37\u2009\u00b0C for 1\u2009h in triplicate wells before transferring to pre-seeded MDCK cells, followed by the addition of 35\u2009\u03bcL of medium containing\u00a0DEAE-dextran at a final concentration of 10\u2009\u03bcg/mL. To keep assay conditions constant, sham medium was used in place of antibody in control wells. Infection levels were determined 2 days later with Bright-Glo luciferase assay system (Promega, Cat. No. E1501). Neutralization curves were fitted by a 5-parameter nonlinear regression built in Prism 9.5.1 (GraphPad Software, La Jolla, CA). The 50% inhibitory titers (IC50s) were reported as the antibody concentrations required to inhibit infection by 50%. H7N9 neutralization was next measured using live replicating influenza viruses to infect MDCK cells as described53. Briefly, serially diluted mAbs were incubated with 100 TCID50 (50% tissue culture infective dose) of an influenza virus at 37\u2009\u00b0C for 2\u2009h, and 100 \u03bcL virus-mAb mixture was added to MDCK cells. After 1\u2009h incubation, the virus-mAb mixture was removed, and minimum-essential medium with 2 \u03bcg/mL L-1-tosylamide-2-phenylethylchloromethyl ketone-treated trypsin (TPCK-trypsin) was added to each well. The plates were then incubated for 72\u2009h, and cytopathic effects were recorded. The mAb concentration that protected 50% of 5 replicate wells from cytopathology was reported as IC50. Similar results were independently reproduced at least once.\n\nSoluble, disulfide-stabilized, fully cleaved H7 HA trimers were produced by transient co-transfection of plasmids encoding H7 HA (H7 SH13 DS2 6R) and Furin of Expi293F cells (Life Technologies) using Turbo293 transfection Reagent (Speed biosystem). After 5 days at 37\u2009\u00b0C, culture supernatants were harvested by centrifugation and concentrated 5-fold by Tangential Flow Filtration. The recombinant HA trimer was captured by Ni-NTA (Sigma-Aldrich) through a C-terminal 6xHis-tag. The imidazole eluant was combined 1:1 (v/v) with saturated ammonium sulfate, centrifuged at 4\u2009\u00b0C, and pellet removed. The supernatant was dialyzed against 500\u2009mM NaCl, 50\u2009mM Tris pH 8, and purified by size exclusion chromatography on a Superdex 200 Increase 10/300 GL column (Cytiva).\n\nHuman mAb Fab fragments were produced by digestion of the full IgG antibodies with immobilized Papain (ThermoFisher) equilibrated with 25\u2009mM phosphate, 150\u2009mM NaCl, pH 10, and 2\u2009mM EDTA for 3\u2009h. The resulting Fabs were purified from the cleaved Fc domain by affinity chromatography using protein A. Fab purity was analyzed by SDS-PAGE. All Fabs were buffer-exchanged into 25\u2009mM phosphate, 150\u2009mM NaCl, pH 7.0 prior to cryo-EM experiments.\n\nTo determine the structures of H7.HK1 and H7.HK2 with H7 HA trimer, trimer was mixed with the antibody Fab at 1 to 1.2 molar ratio at a final total protein concentration of ~1\u2009mg/mL and adjusted to a final concentration of 0.005% (w/v) n-Dodecyl \u03b2-D-maltoside (DDM) to prevent preferred orientation and aggregation during vitrification. Cryo-EM grids were prepared by applying 3\u2009\u03bcL of sample to a freshly glow discharged carbon-coated copper grid (CF 1.2/1.3 300 mesh). The sample was vitrified in liquid ethane using a Vitrobot Mark IV with a wait time of 30\u2009s, a blot time of 3\u2009s, and a blot force of 0. Cryo-EM data were collected on a Titan Krios operating at 300\u2009keV, equipped with a K3 detector (Gatan) operating in counting mode. Data were acquired using Leginon54. The dose was fractionated over 50 raw frames. For all structures, the movie frames were aligned and dose-weighted55 using cryoSPARC 3.456; the CTF estimation, particle picking, 2D classifications, ab initio model generation, heterogeneous refinements, homogeneous 3D refinements and non-uniform refinement calculations were carried out using cryoSPARC 3.456.\n\nFor structural determination, a model of the antibody Fab was generated using SAbPred57. The Fab model and the crystal structure of an H7 HA mutant (PDB: 6IDD)10 was docked into the cryo-EM density map using UCSF Chimera58 to build an initial model of the complex. The model was then manually rebuilt to the best fit into the density using Coot59 and refined using Phenix60. Interface calculations were performed using PISA61. Structures were analyzed and figures were generated using PyMOL (http://www.pymol.org) and UCSF Chimera58. Final model statistics are summarized in Table\u00a0S1.\n\nSearching the Global Initiative on Sharing All Influenza Data (GISAID) with H7Nx returned a total of N\u2009=\u20091511 H7 HA sequences, with N\u2009=\u20092 H7N2, N\u2009=\u20092 H7N3, N\u2009=\u20092 H7N4, N\u2009=\u200954 H7N7, and N\u2009=\u20091451 H7N9. After removing N\u2009=\u200928 duplicate or defective sequences, N\u2009=\u20091483 H7 amino acid sequences were aligned using Clustal62, and the sequence conservation analysis was performed using AL2CO63.\n\nThe mouse prophylactic and therapeutic studies were approved by the Committee on the Use of Live Animals in Teaching and Research (CULATR) of the University of Hong Kong (Reference number: 4011-16) and conducted in biosafety level 3 animal facilities as described previously64,65,66. Female BALB/c mice were imported from Harlan UK Ltd, UK, and those of 6\u20138 weeks of age were obtained from the Laboratory Animal Unit of the University of Hong Kong. Mice were housed at temperatures between 22 to 25\u2009\u00b0C with dark/light cycles and given access to standard pellet feed and water ad libitum. For prophylactic study, one day before virus inoculation, each mouse was administered with 100\u2009\u00b5L of mAb at 1\u2009mg/mL intraperitoneally. For therapeutic study, infected mice were administered with 100\u2009\u00b5L of mAb at 1\u2009mg/mL intraperitoneally at day 1 post viral challenge. Mice in the control groups were administered with either PBS or with a non-H7N9 mAb. On the day of virus infection, each mouse was inoculated with 10 LD50 (40 \u03bcL) of H7N9/AH1 virus through intranasal route. Virus inoculation was performed under ketamine (100\u2009mg/kg) and xylazine (10\u2009mg/kg) anesthesia. The mice were monitored for 14 days with disease severity score and body weight recorded daily. Disease severity were scored as follow: Score 0, apparently healthy; Score 1, mild disease symptom with ruffled fur but still active; Score 2, medium disease symptom with ruffled fur, reduced activity and no weight gain; Score 3, severe disease symptoms with ruffled fur, hunched posture, labored breathing and weight loss; Score 4, moribund being very inactive, showing difficulty moving around and accessing to food and water, and weight loss. The predefined humane endpoints were either a weight loss of \u226520% or a disease severity score of 4. Mice were euthanized if the humane endpoints were reached. Each study group included 5\u201315 randomly allocated female mice for calculating survival rates and statistically significant differences based on the viral challenge model previously established with only female mice64,65,66. The investigators were blinded to animal group allocation during data collection.\n\nGraphPad Prism 9.5.1 was used to plot the ELISA data using sigmoidal dose-response with variable slope for curve fitting and neutralization data using 5-parameter nonlinear regression for curve fitting. All quantitative data are presented as mean\u2009\u00b1\u2009standard error (SEM). GraphPad Prism 9.5.1 was used to plot the mouse survival curves. Two-sided unpaired student\u2019s t-test in GraphPad Prism 9.5.1 and Microsoft 365 Excel version 2404 was used for comparisons between mouse groups. P values\u2009<\u20090.05 were considered statistically significant; P values of 0.05-0.10 were considered statistical trends.\n\nFurther information on research design is available in the\u00a0Nature Portfolio Reporting Summary linked to this article.",
+            "section_image": []
+        },
+        {
+            "section_name": "Data availability",
+            "section_text": "Sequences of the heavy and light chain variable regions of the four H7N9 human mAbs have been deposited in GenBank under accession # OR901962 \u2013 OR901969. The cryo-EM reconstruction of H7.HK1 and H7.HK2 Fabs in complex with H7 SH13 DS2 6R HA has been deposited in the Electron Microscopy Data Bank as EMD-41422 and EMD-41441 and the Protein Data Bank as PDB: 8TNL and 8TOA. Materials will be made available to researchers with appropriate materials transfer agreements (MTAs). All inquiries should be sent to the corresponding authors.\u00a0Source data are provided with this paper.",
+            "section_image": []
+        },
+        {
+            "section_name": "References",
+            "section_text": "Gao, R. et al. Human infection with a novel avian-origin influenza A (H7N9) virus. N. Engl. J. Med. 368, 1888\u20131897 (2013).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBao, C. J. et al. Live-animal markets and influenza A (H7N9) virus infection. N. Engl. J. Med. 368, 2337\u20132339 (2013).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nChen, Y. et al. Human infections with the emerging avian influenza A H7N9 virus from wet market poultry: clinical analysis and characterisation of viral genome. Lancet 381, 1916\u20131925 (2013).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSu, S. et al. Epidemiology, evolution, and pathogenesis of H7N9 influenza viruses in five epidemic waves since 2013 in China. Trends Microbiol. 25, 713\u2013728 (2017).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWang, X. et al. Epidemiology of avian influenza A H7N9 virus in human beings across five epidemics in mainland China, 2013-17: an epidemiological study of laboratory-confirmed case series. Lancet Infect. Dis. 17, 822\u2013832 (2017).\n\nArticle\u00a0\n    MathSciNet\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nQi, W. et al. Emergence and adaptation of a novel highly pathogenic H7N9 influenza virus in birds and humans from a 2013 human-infecting low-pathogenic ancestor. J. Virol. 92, e00921\u201317 (2018).\n\nArticle\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nZhou, J. et al. Biological features of novel avian influenza A (H7N9) virus. Nature 499, 500\u2013503 (2013).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nShi, Y. et al. Structures and receptor binding of hemagglutinins from human-infecting H7N9 influenza viruses. Science 342, 243\u2013247 (2013).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nXu, R. et al. Preferential recognition of avian-like receptors in human influenza A H7N9 viruses. Science 342, 1230\u20131235 (2013).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nXu, Y. et al. Avian-to-human receptor-binding adaptation of Avian H7N9 influenza virus hemagglutinin. Cell Rep. 29, 2217\u20132228 e5 (2019).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHu, Y. et al. Association between adverse clinical outcome in human disease caused by novel influenza A H7N9 virus and sustained viral shedding and emergence of antiviral resistance. Lancet 381, 2273\u20132279 (2013).\n\nArticle\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHai, R. et al. Influenza A(H7N9) virus gains neuraminidase inhibitor resistance without loss of in vivo virulence or transmissibility. Nat. Commun. 4, 2854 (2013).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMarjuki, H. et al. Neuraminidase mutations conferring resistance to oseltamivir in influenza A(H7N9) viruses. J. Virol. 89, 5419\u20135426 (2015).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nTo, K. K. et al. Unique reassortant of influenza A(H7N9) virus associated with severe disease emerging in Hong Kong. J. Infect. 69, 60\u201368 (2014).\n\nArticle\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHo, P. L., Sin, W. C., Chan, J. F., Cheng, V. C. & Chan, K. H. Severe influenza A H7N9 pneumonia with rapid virological response to intravenous zanamivir. Eur. Respir. J. 44, 535\u2013537 (2014).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKiso, M., Yamayoshi, S., Furusawa, Y., Imai, M. & Kawaoka, Y. Treatment of highly pathogenic H7N9 virus-infected mice with baloxavir marboxil. Viruses 11, 1066 (2019).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHuang, K. A. et al. Structure-function analysis of neutralizing antibodies to H7N9 influenza from naturally infected humans. Nat. Microbiol. 4, 306\u2013315 (2019).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nChen, Z. et al. Human monoclonal antibodies targeting the haemagglutinin glycoprotein can neutralize H7N9 influenza virus. Nat. Commun. 6, 6714 (2015).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nThornburg, N. J. et al. H7N9 influenza virus neutralizing antibodies that possess few somatic mutations. J. Clin. Invest. 126, 1482\u20131494 (2016).\n\nArticle\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHenry Dunand, C. J. et al. Both neutralizing and non-neutralizing human H7N9 influenza vaccine-induced monoclonal antibodies confer protection. Cell Host Microbe 19, 800\u2013813 (2016).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBangaru, S. et al. A Site of Vulnerability on the Influenza Virus Hemagglutinin Head Domain Trimer Interface. Cell 177, 1136\u20131152.e18 (2019).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWang, J. et al. Characterization of two human monoclonal antibodies neutralizing influenza A H7N9 viruses. J. Virol. 89, 9115\u20139118 (2015).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWebster, R. G. & Laver, W. G. Determination of the number of nonoverlapping antigenic areas on Hong Kong (H3N2) influenza virus hemagglutinin with monoclonal antibodies and the selection of variants with potential epidemiological significance. Virology 104, 139\u2013148 (1980).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSkehel, J. J. et al. A carbohydrate side chain on hemagglutinins of Hong Kong influenza viruses inhibits recognition by a monoclonal antibody. Proc. Natl Acad. Sci. USA 81, 1779\u20131783 (1984).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nPopova, L. et al. Immunodominance of antigenic site B over site A of hemagglutinin of recent H3N2 influenza viruses. PLoS ONE 7, e41895 (2012).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nZhu, W. et al. Biological characterisation of the emerged highly pathogenic avian influenza (HPAI) A(H7N9) viruses in humans, in mainland China, 2016 to 2017. Eur. Surveill. 22, 30533 (2017).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKile, J. C. et al. Update: increase in human infections with novel asian lineage avian influenza A(H7N9) viruses during the fifth epidemic - China, October 1, 2016-August 7, 2017. MMWR Morb. Mortal. Wkly Rep. 66, 928\u2013932 (2017).\n\nArticle\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nJiao, C., Wang, B., Chen, P., Jiang, Y. & Liu, J. Analysis of the conserved protective epitopes of hemagglutinin on influenza A viruses. Front Immunol. 14, 1086297 (2023).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nCorti, D. et al. A neutralizing antibody selected from plasma cells that binds to group 1 and group 2 influenza A hemagglutinins. Science 333, 850\u2013856 (2011).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nJoyce, M. G. et al. Vaccine-induced antibodies that neutralize group 1 and group 2 influenza A viruses. Cell 166, 609\u2013623 (2016).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSui, J. et al. Structural and functional bases for broad-spectrum neutralization of avian and human influenza A viruses. Nat. Struct. Mol. Biol. 16, 265\u2013273 (2009).\n\nArticle\u00a0\n    MathSciNet\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKallewaard, N. L. et al. Structure and function analysis of an antibody recognizing all influenza A subtypes. Cell 166, 596\u2013608 (2016).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nEkiert, D. C. et al. A highly conserved neutralizing epitope on group 2 influenza A viruses. Science 333, 843\u2013850 (2011).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nFriesen, R. H. et al. A common solution to group 2 influenza virus neutralization. Proc. Natl Acad. Sci. USA 111, 445\u2013450 (2014).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nCheung, C. S. et al. Structure of an influenza group 2-neutralizing antibody targeting the hemagglutinin stem supersite. Structure 30, 993\u20131003.e6 (2022).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nNimmerjahn, F. & Ravetch, J. V. Divergent immunoglobulin g subclass activity through selective Fc receptor binding. Science 310, 1510\u20131512 (2005).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKozlov, M. & Mallapaty, S. Bird flu outbreak in US cows: why scientists are concerned. Nature https://doi.org/10.1038/d41586-024-01036-1 (2024).\n\nArticle\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGamblin, S. J. & Skehel, J. J. Influenza hemagglutinin and neuraminidase membrane glycoproteins. J. Biol. Chem. 285, 28403\u201328409 (2010).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nQiu, C. et al. Safe pseudovirus-based assay for neutralization antibodies against influenza A(H7N9) virus. Emerg. Infect. Dis. 19, 1685\u20131687 (2013).\n\nArticle\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nChothia, C. & Lesk, A. M. Canonical structures for the hypervariable regions of immunoglobulins. J. Mol. Biol. 196, 901\u2013917 (1987).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nChothia, C. et al. Conformations of immunoglobulin hypervariable regions. Nature 342, 877\u2013883 (1989).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nAl-Lazikani, B., Lesk, A. M. & Chothia, C. Standard conformations for the canonical structures of immunoglobulins. J. Mol. Biol. 273, 927\u2013948 (1997).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nRaymond, D. D. et al. Conserved epitope on influenza-virus hemagglutinin head defined by a vaccine-induced antibody. Proc. Natl Acad. Sci. USA 115, 168\u2013173 (2018).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nXiong, X. et al. Receptor binding by an H7N9 influenza virus from humans. Nature 499, 496\u2013499 (2013).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nDiLillo, D. J., Tan, G. S., Palese, P. & Ravetch, J. V. Broadly neutralizing hemagglutinin stalk-specific antibodies require FcgammaR interactions for protection against influenza virus in vivo. Nat. Med. 20, 143\u2013151 (2014).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGuthmiller, J. J. et al. First exposure to the pandemic H1N1 virus induced broadly neutralizing antibodies targeting hemagglutinin head epitopes. Sci. Transl. Med. 13, eabg4535 (2021).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBenjamin, E. et al. A broadly neutralizing human monoclonal antibody directed against a novel conserved epitope on the influenza virus H3 hemagglutinin globular head. J. Virol. 88, 6743\u20136750 (2014).\n\nArticle\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGuthmiller, J. J. et al. Polyreactive broadly neutralizing B cells are selected to provide defense against pandemic threat influenza viruses. Immunity 53, 1230\u20131244.e5 (2020).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nTiller, T. et al. Efficient generation of monoclonal antibodies from single human B cells by single cell RT-PCR and expression vector cloning. J. Immunol. Methods 329, 112\u2013124 (2008).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWu, X. et al. Rational design of envelope identifies broadly neutralizing human monoclonal antibodies to HIV-1. Science 329, 856\u2013861 (2010).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nJia, M., Lu, H., Markowitz, M., Cheng-Mayer, C. & Wu, X. Development of broadly neutralizing antibodies and their mapping by monomeric gp120 in human immunodeficiency virus type 1-infected humans and simian-human immunodeficiency virus SHIVSF162P3N-infected macaques. J. Virol. 90, 4017\u20134031 (2016).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nScheid, J. F. et al. Sequence and structural convergence of broad and potent HIV antibodies that mimic CD4 binding. Science 333, 1633\u20131637 (2011).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nTo, K. K. et al. High titer and avidity of nonneutralizing antibodies against influenza vaccine antigen are associated with severe influenza. Clin. Vaccin. Immunol. 19, 1012\u20131018 (2012).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nCheng, A. et al. Leginon: New features and applications. Protein Sci. 30, 136\u2013150 (2021).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nZheng, S. Q. et al. MotionCor2: anisotropic correction of beam-induced motion for improved cryo-electron microscopy. Nat. Methods 14, 331\u2013332 (2017).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nPunjani, A., Rubinstein, J. L., Fleet, D. J. & Brubaker, M. A. cryoSPARC: algorithms for rapid unsupervised cryo-EM structure determination. Nat. Methods 14, 290\u2013296 (2017).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nDunbar, J. et al. SAbPred: a structure-based antibody prediction server. Nucleic Acids Res. 44, W474\u2013W478 (2016).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nPettersen, E. F. et al. UCSF Chimera-a visualization system for exploratory research and analysis. J. Comput Chem. 25, 1605\u20131612 (2004).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nCasanal, A., Lohkamp, B. & Emsley, P. Current developments in Coot for macromolecular model building of Electron Cryo-microscopy and Crystallographic Data. Protein Sci. 29, 1069\u20131078 (2020).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nAdams, P. D. et al. Recent developments in the PHENIX software for automated crystallographic structure determination. J. Synchrotron Radiat. 11, 53\u201355 (2004).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKrissinel, E. & Henrick, K. Inference of macromolecular assemblies from crystalline state. J. Mol. Biol. 372, 774\u2013797 (2007).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nThompson, J. D., Higgins, D. G. & Gibson, T. J. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 22, 4673\u20134680 (1994).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nPei, J. & Grishin, N. V. AL2CO: calculation of positional conservation in a protein sequence alignment. Bioinformatics 17, 700\u2013712 (2001).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nTo, K. K. et al. Recombinant influenza A virus hemagglutinin HA2 subunit protects mice against influenza A(H7N9) virus infection. Arch. Virol. 160, 777\u2013786 (2015).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nZhao, H. et al. A novel peptide with potent and broad-spectrum antiviral activities against multiple respiratory viruses. Sci. Rep. 6, 22008 (2016).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nZhao, H. et al. Novel residues in the PA protein of avian influenza H7N7 virus affect virulence in mammalian hosts. Virology 498, 1\u20138 (2016).\n\nArticle\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nDownload references",
+            "section_image": []
+        },
+        {
+            "section_name": "Acknowledgements",
+            "section_text": "We thank the patient for donating blood for the study. We thank Reda Rawi and Jeffrey C. Boyington for design of H7 SH13 DS2 6R used for structural analysis. Cryo-EM data were collected at the Columbia University Cryo-Electron Microscopy Center. We thank Shuofeng Yuan and Vincent Poon for assistance with the animal experiments. This study is supported by U.S. Department of Defense contract No. W911NF-14-C-0001 to D.D.H. and X.W., by Health@InnoHK, Innovation and Technology Commission of Hong Kong to K.Y. and K.K.T., by donations from Richard Yu and Carol Yu, Shaw Foundation Hong Kong, Michael Seak-Kan Tong, The Hui Ming, Hui Hoy and Chow Sin Lan Charity Fund Limited, Chan Yin Chuen Memorial Charitable Foundation, Marina Man-Wai Lee, Jessie and George Ho Charitable Foundation, Kai Chong Tong, Tse Kam Ming Laurence, Foo Oi Foundation Limited, Betty Hing-Chu Lee, and Ping Cham So to K.Y. and K.K.T., by Bill and Melinda Gates Foundation grants FNIH SHAP19IUFV and INV-016167 to L.S., and by U.S. National Institutes of Health, National Institute of Allergy and Infectious Disease, Intramural Research Program of the Vaccine Research Center\u00a0ZIA IA0005022 to P.D.K.",
+            "section_image": []
+        },
+        {
+            "section_name": "Author information",
+            "section_text": "These authors contributed equally: Manxue Jia, Hanjun Zhao, Nicholas C. Morano.\n\nAaron Diamond AIDS Research Center, Columbia University Vagelos College of Physicians and Surgeons, New York, NY, 10032, USA\n\nManxue Jia,\u00a0Nicholas C. Morano,\u00a0Hong Lu,\u00a0Jordan E. Becker,\u00a0David D. Ho,\u00a0Peter D. Kwong,\u00a0Lawrence Shapiro\u00a0&\u00a0Xueling Wu\n\nState Key Laboratory for Emerging Infectious Diseases, Carol Yu Centre for Infection, Department of Microbiology, School of Clinical Medicine, Li Ka Shing Faculty of Medicine, University of Hong Kong, Pokfulam, Hong Kong Special Administrative Region, China\n\nHanjun Zhao,\u00a0Yin-Ming Lui,\u00a0Kwok-Yung Yuen\u00a0&\u00a0Kelvin Kai-Wang To\n\nCentre for Virology, Vaccinology and Therapeutics, Hong Kong Science and Technology Park, Sha Tin, Hong Kong Special Administrative Region, China\n\nHanjun Zhao,\u00a0Kwok-Yung Yuen\u00a0&\u00a0Kelvin Kai-Wang To\n\nDepartment of Biochemistry, Zuckerman Mind Brain Behavior Institute, Columbia University, New York, NY, 10027, USA\n\nNicholas C. Morano,\u00a0Jordan E. Becker\u00a0&\u00a0Lawrence Shapiro\n\nVaccine Research Center, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD, 20892, USA\n\nHaijuan Du,\u00a0Peter D. Kwong\u00a0&\u00a0Lawrence Shapiro\n\nDepartment of Clinical Microbiology and Infection, University of Hong Kong-Shenzhen Hospital, Shenzhen, Guangdong, 518053, China\n\nKwok-Yung Yuen\u00a0&\u00a0Kelvin Kai-Wang To\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nConceptualization: X.W., D.D.H., and K.Y. Methodology: X.W., K.K.T., and L.S. Investigation: M.J., H.Z., N.C.M., H.L., Y.L., H.D., and J.E.B. Visualization: X.W. and N.C.M. Funding acquisition: D.D.H., X.W., K.Y., K.K.T., L.S., and P.D.K. Project administration: X.W. and K.K.T. Supervision: X.W., K.K.T., K.Y., P.D.K., and L.S. Writing\u2014original draft: X.W., K.K.T., and N.C.M. Writing\u2014review & editing: X.W., K.K.T., M.J., H.Z., N.C.M., K.Y., D.D.H., P.D.K., and L.S.\n\nCorrespondence to\n                Kelvin Kai-Wang To or Xueling Wu.",
+            "section_image": []
+        },
+        {
+            "section_name": "Ethics declarations",
+            "section_text": "An U.S. provisional patent titled \u201cHuman Protective Neutralizing and Non-neutralizing Antibodies and Their Use against Influenza Viruses\u201d was filed with filing No. 63/650,342 and X.W., M.J., N.C.M., H.L., D.D.H., K.Y., K.K.T., and L.S. as co-inventors; the remaining authors declare no competing interests. The authors declare no other competing interests.",
+            "section_image": []
+        },
+        {
+            "section_name": "Peer review",
+            "section_text": "Nature Communications thanks the anonymous reviewers for their contribution to the peer review of this work. A peer review file is available.",
+            "section_image": []
+        },
+        {
+            "section_name": "Additional information",
+            "section_text": "Publisher\u2019s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.",
+            "section_image": []
+        },
+        {
+            "section_name": "Source data",
+            "section_text": "",
+            "section_image": []
+        },
+        {
+            "section_name": "Rights and permissions",
+            "section_text": "Open Access  This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article\u2019s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article\u2019s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.\n\nReprints and permissions",
+            "section_image": []
+        },
+        {
+            "section_name": "About this article",
+            "section_text": "Jia, M., Zhao, H., Morano, N.C. et al. Human neutralizing antibodies target a conserved lateral patch on H7N9 hemagglutinin head.\n                    Nat Commun 15, 4505 (2024). https://doi.org/10.1038/s41467-024-48758-4\n\nDownload citation\n\nReceived: 27 October 2023\n\nAccepted: 13 May 2024\n\nPublished: 27 May 2024\n\nVersion of record: 27 May 2024\n\nDOI: https://doi.org/10.1038/s41467-024-48758-4\n\nAnyone you share the following link with will be able to read this content:\n\nSorry, a shareable link is not currently available for this article.\n\n\n\n\n                            Provided by the Springer Nature SharedIt content-sharing initiative\n                        ",
+            "section_image": [
+                "https://data:image/svg+xml;base64,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"
+            ]
+        },
+        {
+            "section_name": "This article is cited by",
+            "section_text": "Nature Communications (2025)",
+            "section_image": []
+        }
+    ]
+}

039c199243e3ff42a62092628fd75c4b2179e5ca70bda3bac1533b18d7195199/peer_review/peer_review.md

@@ -0,0 +1,157 @@
+Dominant control of temperature on (sub-)tropical soil carbon turnover
+
+Corresponding Author: Dr Vera Dorothee Meyer
+
+This file contains all reviewer reports in order by version, followed by all author rebuttals in order by version.
+
+Version 0:
+
+Reviewer comments:
+
+Reviewer #1
+
+(Remarks to the Author)
+The paper presents analysis of radiocarbon dating of carbon compounds in marine sediment cores to infer the average residence time and age of soil carbon in the supplying catchment. The data is retrieved from the Nile River delta and spans the last deglaciation and Holocene. The analysis of the authors suggests drastic changes in residence time/turnover over this time frame, exceeding those of current land surface models. This in turn is relevant as soil carbon turnover is a major determinant of the trajectory of soil carbon under global change, possibly creating large positive feedbacks in the climate system under anthropogenic warming.
+I believe the method and the analysis is intriguing. I admit to understanding little of technical part of the sediment analyses, the lipid compounds, and their relationship with the bulk soil carbon and its soil turnover, and hope that another reviewer can provide an assessment. My expertise lies in the biogeochemistry of the carbon cycle, and their representation in land surface models.
+I have a few comments that I think need to be addressed or at least would help me to gain some clarity.
+
+1) My confusion about age: The author refer to age in different way. At times it is the time until present and sometimes it is the pre-depositional age. I hope the authors find a way to clarify this. For example use 'pre-depositional age' (or a similar expression) consequently. Similarly, could you explain the difference between turnover and age for the benefit of the reader (and me ☺). The formula for turnover is the mean residence time and therefore represents also the mean age of the soil, but clearly here and in ref 30 in the paper they are different.
+2) Presentation of turnover vs. temperature: Many models use a Q10 formulation, an exponential function. Hence the plot of ln(tau) vs. temperature would reveal a Q10 (the 3) slope of such a plot is ln(Q10)/10). Hence the Q10 could be inferred (about 11 here, if I am correct) which is way higher than in models (about 2-3), including LPJ.
+4) Offset and c3/c4 vegetation transition. Glacial/interglacial change with distinct temperature, moisture and CO2 changes creates shifts in vegetation types, and particularly in in c3 vs. c4 photosynthetic pathway, which is especially pronounced in warm areas with seasonal precipitation (i.e. subtropical). With fractionation against 14C being very different between C4 and C3, this could considerably affect offset? Perhaps this was considered in this study, but I don’t see it discussed here. If omitted, a 20 permil fractionation difference would (according to my -perhaps flawed- calculation) translate into a an about 160 years offset, which would be in the vicinity of the differences in turnover times reported here. If 13C was measured along 14C (maybe it was measured but not shown?), maybe this could help disentangle this effect?
+5) Error propagation: Does the calculation of tau include potential errors for equation 3 and 4? The relationship between lipid age and tau has some error to, and so it may be propagated in this work.
+
+Reviewer #2
+
+(Remarks to the Author)
+Review and revise the manuscript
+Review conclusion: Acceptance after major revision
+The article uses the radiocarbon dating method derived from plant derived lipids, combined with temperature and rainfall reconstructed from sediment cores of the eastern Mediterranean that received terrestrial materials from the Nile River Basin, to investigate the dominant control of temperature on carbon turnover in subtropical and tropical soils over the past 18000
+years. The research work is solid, the content is rich, the methods are advanced, and it has good innovation. I am very interested in this part of the content, but it is not easy to read the article. The most important issue is that the main research question and conclusion are not very clear, and the writing logic may need to be reorganized for readers to better understand the content. The second question is whether the definition of carbon in the article is reasonable and accurate, which is very important. At least in my opinion, I don’t understand what redefining carbon means? And how to determine, does it refer to the rate or quantity of carbon outflow and inflow, how to calculate it is not clearly stated in the article? Then I still don’t understand why the reservoir age shift of leaf wax biomarkers can be used to calculate average soil carbon. This part may require more explanation and clarification. The research is very interesting and meaningful, but there are still some issues that need to be modified and resolved. The specific opinions are as follows:
+1. What is the meaning of redefining carbon in Line 30? Can’t we directly use organic carbon content? How to determine f, does it refer to the rate or quantity of carbon outflow and inflow, how to calculate it, and what is the unit?
+2. Why Lines 33-35 said that the impact of water climate on soil carbon turnover time plays a strong controlling role in low latitude climate, more important than temperature? I don’t understand. Is there any theoretical support?
+3. Lines 49-55: The two biomarkers (e.g., long-chain n-alkanes and n-alkanoic acids) were a bit suddenly proposed here, I suggest to add some descriptions for explaining why you choose these two biomarkers? As I am studying modern processes and mechanisms of these two biomarkers for more than ten years, I totally support that you use these two biomarkers to demonstrate the responses of soil carbon cycle to climate changes, but more explanations are needed here. Please refer some references (e.g., n-alkane for Liu and An, 2020 and Liu et al., 2022; ESR; n-alkanoic acids for Liu et al., 2024; SCES).
+4. Line 107-108: Why n-alkanoic acids reflect a local signals while n-alkane provide a more catchment-integrating signal? We know these two biomarkers originate the same precursor (i.e., acyl-ACP).
+5. Does the topic consider adding marine sediments? The research is not focused on traditional soils.
+6. At the beginning, the research site was in the Nile Delta region, and later it moved to tropical and subtropical regions as well as global analysis. I don’t know how this changed in between. Can the Nile Delta region represent subtropical regions?
+7. I didn’t understand the part about the time in the article. It started in the environmental changes in the Nile Delta region in the previous 18 years, and then it was before glaciers until the New Century. How did it change during this period? What did the environmental changes in the Nile Delta region in the previous 18 years mean? What is the connection to the other content? I don’t quite understand?
+8. In my opinion, Lines 124-129 has no theoretical support.
+9. Lines 135-136: the hydrogen isotopic composition of paleo precipitation (δDp), serve as a common proxy for the amount of rainfall? This needs a detailed explanation, because our modern investigation showed that δDp is mainly controlled by temperature at the global scale, instead of rainfall amount, only in some intensive monsoon regions, the δDp is dominately controlled by rainfall amount, so please supplement some modern data for supporting this inference.
+10. Why do you say that in Lines 162-163? Is there any data support for non-tropical and subtropical areas? Doesn’t this change exist in non-tropical and non-subtropical regions?
+11. Should different vegetation zones in Figure 1 be marked with different colors.
+12. It would be better to mark the year on the horizontal axis in Figure 2.
+13. It seems unreasonable to use surface temperature instead of sea surface temperature in the eastern Mediterranean in Figure 3.
+credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.
+In cases where reviewers are anonymous, credit should be given to 'Anonymous Referee' and the source.
+The images or other third party material in this Peer Review File are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
+To view a copy of this license, visit https://creativecommons.org/licenses/by/4.0/
+REVIEWER COMMENTS
+
+Reviewer #1 (Remarks to the Author):
+
+The paper presents analysis of radiocarbon dating of carbon compounds in marine sediment cores to infer the average residence time and age of soil carbon in the supplying catchment. The data is retrieved from the Nile River delta and spans the last deglaciation and Holocene. The analysis of the authors suggests drastic changes in residence time/turnover over this time frame, exceeding those of current land surface models. This in turn is relevant as soil carbon turnover is a major determinant of the trajectory of soil carbon under global change, possibly creating large positive feedbacks in the climate system under anthropogenic warming.
+I believe the method and the analysis is intriguing. I admit to understanding little of technical part of the sediment analyses, the lipid compounds, and their relationship with the bulk soil carbon and its soil turnover, and hope that another reviewer can provide an assessment. My expertise lies in the biogeochemistry of the carbon cycle, and their representation in land surface models.
+I have a few comments that I think need to be addressed or at least would help me to gain some clarity.
+
+Dear reviewer,
+
+Thank you very much for your positive feedback on our manuscript and for considering it a relevant piece of work. We feel that your comments are very constructive and help to improve our work. Below, we specify how we address your points in the revised version of our manuscript and we hope that clarity improved by the changes made. The lines numbers given refer to the track-changes version of the manuscript.
+
+1) My confusion about age: The author refer to age in different way. At times it is the time until present and sometimes it is the pre-depositional age. I hope the authors find a way to clarify this. For example use ‘pre-depositional age’ (or a similar expression) consequently. Similarly, could you explain the difference between turnover and age for the benefit of the reader (and me 😊). The formula for turnover is the mean residence time and therefore represents also the mean age of the soil, but clearly here and in ref 31 in the paper they are different.
+
+Reply: You are right saying that turnover time means residence time of organic matter in soils and theoretically should equal the mean age of the soil. However, turnover time assumes that all carbon is cycling homogenously, which oversimplifies soil carbon dynamics and the complexity of the soil organic carbon. It has been shown that the estimate of residence time based on radiocarbon dating of bulk soil organic carbon is higher than calculations based on the steady-state assumption and using the ratio of carbon stock size over NPP (\( t_{soil} \); eq.1 in our manuscript) as documented e.g. in Shi et al. (2020)[ref. 51]. The reason is as follows: Organic matter is a complex mixture of fast cycling, labile fractions and slow-cycling refractory fractions. That means some decompose very quickly within years or decades while other refractory components survive over millennia. Therefore, slow-cycling components accumulate in soils dominating the soil organic carbon pool. Accordingly, the age of soil organic carbon determined by radiocarbon dating represents the older, slow cycling fractions. By contrast, the calculation of tsoil using NPP and the carbon stock size is biased towards fast cycling components as “most net primary production cycles through relatively small soil carbon pools on timescales of years to decades” as note by Shi et al. (2021).
+
+We added a short paragraph to lines 238-263 to clarify this.
+2) Presentation of turnover vs. temperature: Many models use a Q10 formulation, an exponential function. Hence the plot of ln(tau) vs. temperature would reveal a Q10 (the slope of such a plot is ln(Q10)/10). Hence the Q10 could be inferred (about 11 here, if I am correct) which is way higher than in models (about 2-3), including LPJ.
+
+Reply: That is a good point. We are aware of the \( Q_{10} \) formulation commonly used to express temperature sensitivity. You are correct that \( Q_{10} \) inferred from our data is 11 and is higher than the values most models operate with. We decided that this is a relevant and intriguing observation and the revised version contains a short discussion of \( Q_{10} \) (lines 200-228 and 312-328). An additional figure is also provided to demonstrate the effect of different \( Q_{10} \) on the carbon loss rate from soils (Figure 5).
+
+4) Offset and c3/c4 vegetation transition. Glacial/interglacial change with distinct temperature, moisture and CO\(_2\) changes creates shifts in vegetation types, and particularly in in c3 vs. c4 photosynthetic pathway, which is especially pronounced in warm areas with seasonal precipitation (i.e. subtropical). With fractionation against \( ^{14}\mathrm{C} \) being very different between C4 and C3, this could considerably affect offset? Perhaps this was considered in this study, but I don’t see it discussed here. If omitted, a 20 permil fractionation difference would (according to my -perhaps flawed-calculation) translate into a an about 160 years offset, which would be in the vicinity of the differences in turnover times reported here. If \( ^{13}\mathrm{C} \) was measured along \( ^{14}\mathrm{C} \) (maybe it was measured but not shown?), maybe this could help disentangle this effect?
+
+Reply: In \( ^{14}\mathrm{C} \) analysis, it is a standard procedure to correct the data (\( ^{14}\mathrm{C}/^{12}\mathrm{C} \) ratios) for changes in the \( ^{13}\mathrm{C}/^{12}\mathrm{C} \) ratio by normalizing the data to a \( \delta^{13}\mathrm{C} \) value of -25 permill (PDB), which is representative of pre-industrial wood. This correction allows to compare samples from different environments (Stuiver and Pollach, 1977). As for paleo records, this correction removes potential biases introduced from changing \( ^{13}\mathrm{C} \) ratios through time (Stuiver and Pollach, 1977). During \( ^{14}\mathrm{C} \) measurements using the MICADAS, \( ^{13}\mathrm{C}/^{12}\mathrm{C} \) is analyzed along with \( ^{14}\mathrm{C}/^{12}\mathrm{C} \) ratios and the correction is automatically performed by the processing software of the system (BATS; Mollenhauer et al., 2021). All Fmc values reported in our manuscript thus are \( ^{13}\mathrm{C} \)-corrected and any influence from changing C4/C3 plant ratios can be excluded.
+
+To avoid that questions arise regarding the impact of isotopic fractionation in our \( ^{14}\mathrm{C} \) data we added a few lines to the method section mentioning the \( ^{13}\mathrm{C} \) correction (458-460).
+
+5) Error propagation: Does the calculation of tau include potential errors for equation 3 and 4? The relationship between lipid age and tau has some error to, and so it may be propagated in this work.
+
+Reply: Yes, the calculated \( \tau_{soil} \) includes errors from both equations. (Note that in the revised version, these equations are numbered as 4 and 5). Uncertainties from \( ^{14}\mathrm{C}_{sample} \) (AMS measurement uncertainty and errors introduced from the blank correction) and t (deposition age/core chronology) are considered for \( ^{14}\mathrm{C}_{initial} \) (eq 4). From eq. 5 we used the error of the slope (40.1 ± 3.9) which is given in ref 31.
+Reviewer #2 (Remarks to the Author):
+
+Review and revise the manuscript
+Review conclusion: Acceptance after major revision
+The article uses the radiocarbon dating method derived from plant derived lipids, combined with temperature and rainfall reconstructed from sediment cores of the eastern Mediterranean that received terrestrial materials from the Nile River Basin, to investigate the dominant control of temperature on carbon turnover in subtropical and tropical soils over the past 18000 years. The research work is solid, the content is rich, the methods are advanced, and it has good innovation. I am very interested in this part of the content, but it is not easy to read the article. The most important issue is that the main research question and conclusion are not very clear, and the writing logic may need to be reorganized for readers to better understand the content. The second question is whether the definition of carbon in the article is reasonable and accurate, which is very important. At least in my opinion, I don’t understand what redefining carbon means? And how to determine, does it refer to the rate or quantity of carbon outflow and inflow, how to calculate it is not clearly stated in the article? Then I still don’t understand why the reservoir age shift of leaf wax biomarkers can be used to calculate average soil carbon. This part may require more explanation and clarification. The research is very interesting and meaningful, but there are still some issues that need to be modified and resolved. The specific opinions are as follows:
+
+Dear Reviewer,
+
+thank you very much for your positive review of our manuscript. We are very delighted to read that you find our research very interesting and meaningful and that you think it has good innovation. As we understand from your review, the most important point to you is that our main research goals and conclusions need to be more clearly stated. We address this point in our new version of the manuscript and rephrased parts of the introduction (lines 42-52) and the concluding paragraph (330-343). We also extended a few sentences throughout the manuscript to improve the elaborations about our methods, i.e. the justification why we selected the biomarkers and how we can deduce mean soil carbon turnover times (lines 72-90; 186-195).
+
+Please, find below how we address each of your points. The line numbers refer to the track-changes version of the manuscript.
+
+1. What is the meaning of redefining carbon in Line 30? Can't we directly use organic carbon content? How to determine f, does it refer to the rate or quantity of carbon outflow and inflow, how to calculate it, and what is the unit?
+
+Reply: We are not sure what exactly you mean by “redefining carbon”. In line 30 we show the formula commonly used to determine the mean turnover time - or residence time - of carbon in soils. This formula was not developed by us and therefore neither something new is defined here nor is something re-defined. We added a citation to the formula to clarify that this formula is background knowledge that has already existed for long time (ref. 6).
+
+As for your questions regarding the units: Carbon stock is the organic carbon content per unit area and is often given in kgC m\(^{-2}\) in both, field experiments and data (see Figure 4e). To determine soil organic content, several methods exist to analyze samples from the field in the laboratory. One example is the dry combustion method using an Elementar Analyzer. Besides analysis in the lab data from remote sensing are used. As mentioned in line 32, influx and carbon efflux are equal in steady states (in equilibrium conditions). So, turnover time of soil carbon can be calculated using either influx or efflux. Therefore, it is common practice to determine the turnover time using the net primary
+production (NPP), i.e. the carbon influx. For determining NPP the increase in plant biomass per unit area and time is measured in the field, and using remote sensing or modeling approaches. NPP is often expressed in kgC m^{-2} yr^{-1}.
+
+We added the units to lines 31 and 32 in order to clarify.
+
+2. Why Lines 36-38 said that the impact of water climate on soil carbon turnover time plays a strong controlling role in low latitude climate, more important than temperature? I don't understand. Is there any theoretical support?
+
+Reply: This statement is based on observations from refs. 4,12,13 which are given in these lines. For example, in ref. 4 (Carvalhais et al., 2014), ecosystem turnover times are correlated against temperature and precipitation to investigate regional differences in the dependence of turnover to these two climate variables. The results suggest that relatively strong relationships with precipitation tend to especially occur in the low to mid latitudes. Carvalhais et al. (2014) concluded that in these regions, precipitation overrides temperature effects as relatively weak correlations with temperature were observed at these places. One reason for this pattern may be that in these regions, temperature variability is small throughout the year while rainfall variability is large throughout. However, the interaction of temperature and precipitation effects on turnover times are still not fully understood. We refrain from elaborating and speculating about potential mechanisms that may explain this observation in the introduction as this would distract from the major focus of our study. By finding that temperature is the major driver of \( \tau_{soil} \) over the past 18 thousand years the focus is set on the effect of temperature.
+
+3. Lines 68-72: The two biomarkers (e.g., long-chain n-alkanes and n-alkanoic acids) were a bit suddenly proposed here, I suggest to add some descriptions for explaining why you choose these two biomarkers? As I am studying modern processes and mechanisms of these two biomarkers for more than ten years, I totally support that you use these two biomarkers to demonstrate the responses of soil carbon cycle to climate changes, but more explanations are needed here. Please refer some references (e.g., n-alkane for Liu and An, 2020 and Liu et al., 2022; ESR; n-alkanoic acids for Liu et al., 2024; SCES).
+
+Reply: To improve the justification for the selection of the biomarkers, we provide more details in our introduction of the biomarkers in lines 68-95. We added more information about their origin and their application as biomarkers for continental environmental changes as well as about the method for deducing \( \tau_{soil} \) from CSRA of the n-alkanoic acids. We also provide a new reference (Eglinton and Eglinton, 2008 => ref. 24), which summarizes the application of molecular proxies –including plant-wax lipids (alkanes, alcohols and acids) in paleoclimatology. By deciding to provide this single reference instead of the three you suggested, we are able to stay within the limit of 70 references. Having extended this paragraph, we hope that the proposition of the biomarkers now appears more gradual and coherent.
+
+4. Line 151-153: Why n-alkanoic acids reflect a local signal while n-alkane provide a more catchment-integrating signal? We know these two biomarkers originate the same precursor (i.e., acyl-ACP).
+Reply: The precursor molecule of both compounds is acyl-ACP. However, in this case we refer to the geographical origin rather than the biogeochemical one. The reasoning for this conclusion that n-alkanes and n-alkanoic acids have different source areas is given in detail in Meyer et al. (2024) which is also cited in line 151 (ref. 23). A key finding of this study is that δD of n-alkanes and n-alkanoic acids (\( \delta D_{wax} \)) differ substantially from each other (Figure 1). The different trends throughout the past 18 kyrs attest to different rainfall regimes and consequently different source areas within the Nile catchment. When comparing the \( \delta D_{wax} \) data from core GeoB7702-3 to \( \delta D_{wax} \) records from Lakes Tana and Victoria (sources of the Nile River) and to data from the Red Sea (Mediterranean realm) it becomes clear that n-alkanoic acids reflect rainfall in the Mediterranean winter rainfall zone (Nile delta region) while the n-alkanes also reflect the African summer monsoon rains. To illustrate this, we provide Figure 1, which is from Meyer et al. (2024) (Records a-e are of interest). Today, the Mediterranean winter rains supply a small band along the coast while monsoon rains influence the catchment area south of ~12°N. The hyperarid desert where vegetation is extremely scarce lies in between. Accordingly, Meyer et al., 2024 conclude that the n-alkanoic acids provide a regional signal from the delta (Mediterranean rainfall zone) and that n-alkanes provide a more catchment integrating signal. This provenance pattern probably owes to differences in the vulnerability of the two compounds towards degradation. n-alkanoic acids are relatively labile compared to n-alkanes and probably get more efficiently degraded during transport. Most likely fatty acids from the headwaters did not reach the Mediterranean Sea throughout the last 18 kyrs. As elaborated in Meyer et al. (2024) similar observations regarding provenance of leaf wax biomarkers were previously made in other large river catchments, e.g. the Ganga-Brahmaputra and Congo Rivers (Galy et al., 2011; Hemingway et al., 2016).
+
+As we cite Meyer et al. (2024) in lines 151, we did not change the paragraph. If interested, the reader can easily access the detailed discussion of the provenance patterns following the reference (ref. 23).
+
+![δDwax records from Lakes Tana, Victoria, Tanganyika, and GeoB7702-3, showing different source areas and rainfall zones](page_1012_682_393_627.png)
+
+Figure 1: (a)-(c) \( \delta D_{wax} \) records from Lakes Tana (Costa et al., 2014), Victoria (Berke et al., 2012) and Tanganyika (Tierney et al., 2008) (d) \( \delta D_{wax} \) n-alkanoic acids (orange: n-C26:0; green: n-C28:0; Meyer et al., 2024) along with \( \delta D_{wax} \) n-alkane (blue: n-C31; Castañeda et al., 2016) from core GeoB7702-3 and \( \delta D_{wax} \) based on the n-C30:0 alkanoic acid from core DSDP 5017-1, Dead Sea (black; Tierney et al., 2022). (e) Ratio of tetra and penta-methylated brGDGTs from the Nile deep sea fan (Ménot et al., 2020) (f) oxygen isotopic composition of the planktic foraminifera species Globorotalioides ruber alba from the Levantine Basin (Revel et al., 2010; 2015). The light grey and dark grey shadings mark the episodes of the AHP and the “Green Sahara”, and their optimum. LGM: Last Glacial Maximum; HS1: Heinrich Stadial 1; B/A: Bølling-Allerød interstadial; YD: Younger Dryas
+5. Does the topic consider adding marine sediments? The research is not focused on traditional soils.
+
+Reply: Our work is based on a marine sediment core but this fact does not matter for our endeavor. Despite being marine, the core forms a suitable archive for reconstructing continental changes as we choose terrigenous biomarkers for radiocarbon analysis and the reconstruction of \( \tau_{soil} \). As the n-C26:0, n-C28:0 alkanoic acids and the n-C31 alkane are specific biomarkers for terrestrial higher plants and a bias from marine algal or bacterial biomass can be ruled out.
+
+6. At the beginning, the research site was in the Nile Delta region, and later it moved to tropical and subtropical regions as well as global analysis. I don't know how this changed in between. Can the Nile Delta region represent subtropical regions?
+
+Reply: The delta is the main origin of the n-alkanoic acids as concluded in Meyer et al. (2024) (ref. 23) and accordingly the signals of the n-alkanoic acids reflect environmental change in the Nile delta in the first place. In line 152 we state that the n-alkanes provide catchment integrating signals. We find that the \( ^{14}C \) signals in n-alkanoic acids and n-alkanes are very similar (supplementary figure S2). This observation is the key to justify the extrapolation of the \( \tau_{soil} \) results to the entire Nile catchment. From the similarity we conclude that changes in soil carbon turnover must have been similar in the entire catchment (line 162-165).
+
+The Nile catchment is vast and extends over the northeastern African tropics and subtropics. So it is directly representative of northeastern Africa. The extrapolation to the global scale is made because CSRA data from (sub-) tropical areas in Asia (Ganga-Brahmaputra catchment; ref. 16) suggest similar magnitudes in change of \( \tau_{soil} \) (lines 283-286; Table 1 in the manuscript). In view of these similarities in two datasets from the two continents, we consider it likely that these changes in \( \tau_{soil} \) may have occurred globally across the (sub-) tropics. However, more data from other regions (e.g. South America or Southeast Asia) is necessary to verify this assumption.
+
+We extended the sentence in lines 286-289 to clarify our justification for the extrapolation to the global (sub-) tropics. We also rewrote the introduction of the Nile catchment to stress that it is representative of a relatively large (sub-)tropical area in northeastern Africa to improve the justification of the selection of this study area (lines 56-61).
+
+7. I didn't understand the part about the time in the article. It started in the environmental changes in the Nile Delta region in the previous 18 years, and then it was before glaciers until the New Century. How did it change during this period? What did the environmental changes in the Nile Delta region in the previous 18 years mean? What is the connection to the other content? I don't quite understand?
+
+Reply: Our manuscript is about the last 18 thousand years and discusses changes in \( \tau_{soil} \) during the last deglaciation. When referring to the time before present we express the time in kiloyears before present (18 kyrs BP). Perhaps the confusion arose because we wrote “18,000 years” in the abstract but “18 kyrs” in the following main text. In the new version of the article we indicate that “18,000 years” is termed “18 kyrs” in the following (line 47).
+
+8. In my opinion, Lines 168-175 has no theoretical support.
+
+Reply: When revising our manuscript, this paragraph was removed.
+9. Lines 194-195: the hydrogen isotopic composition of paleo precipitation (δDp), serve as a common proxy for the amount of rainfall? This needs a detailed explanation, because our modern investigation showed that δDp is mainly controlled by temperature at the global scale, instead of rainfall amount, only in some intensive monsoon regions, the δDp is dominantly controlled by rainfall amount, so please supplement some modern data for supporting this inference.
+
+Reply: It is correct that temperature exerts strong control on the δD of precipitation on a global scale. In low latitude regions where temperature variations are often small but variability in rainfall is large – such as monsoon regions – the amount effect is the dominant control on δD of precipitation. This is the reason why δD of n-alkanes and n-alcanoic acids (δD_{wax}) has often been applied as a proxy for rainfall variability in Africa to reconstruct changes in the African monsoon (Castaneda et al., 2016; Berke et al., 2012; Tierney et al., 2008 and many more). The δD_{wax} from our n-alcanoic acids is interpreted as a signal from the Nile delta region (Meyer et al., 2024 => ref. 23) and is outside the area of influence of the monsoonal rains. As for the coastal areas of North Africa that are influenced by Mediterranean winter rainfall, Tierney et al. (2017) as well as Goldsmith et al. (2017; 2019) show that the amount effect exerts dominant control. The δD_{wax} record from the n-alcanoic acids that we use in our current study is intensively discussed in our previous paper that focusses on rainfall variability in the Nile delta region during the last deglaciation (Meyer et al.; 2024 => ref. 23). Below we show a figure from this study showing the δD_{wax} of n-alkanoic acids and n-alkanes together with the TEX_{86}-based SST record. The development of δD_{wax} of the n-alkanoic acids clearly differs from the development of SST which is a strong indication that the amount effect dominated over temperature during the past 18,000 years. Accordingly, we consider the δDp record (which is deduced from the δD_{wax} by correcting it for changes in the abundance of C3 and C4 plants via the δ^{13}C_{wax}) a robust proxy for rainfall amount. So the use of this record to investigate the link between precipitation and soil-carbon turnover times in our present study is justified.
+
+For the revised version of the manuscript, we extended the paragraph encompassing lines 191-195 to strengthen the selection of δDp as proxy of rainfall amount in the Nile river catchment.
+Figure 2: Figure adopted from Meyer et al. (2024). δDwax from three plant-wax homologues is plotted together with SST reconstructions from the Eastern Mediterranean (data from Castañeda et al. (2010) and summer insolation (Berger and Loutre, 1991). δDwox is corrected for changes in ice-volume.
+
+10. Why do you say that in Lines 277? Is there any data support for non-tropical and subtropical areas? Doesn't this change exist in non-tropical and non-subtropical regions?
+
+Reply: The conclusion in line 277 was made because there are similarities between two datasets from two continents, i.e., the Ganga-Brahmaputra catchment (Asia) and the Nile catchment (Africa). Specifically, it is the order of magnitude of change in \( \tau_{soil} \) that occurred since the last glacial that is similar at both sites. Data from two river catchments that reflect sub-tropical to tropical areas on two continents make it very likely that the changes in \( \tau_{soil} \) that we infer for the Nile catchment are not a regional phenomenon. Therefore, we hypothesize that similar changes were common across the sub-tropics globally. However, more data is needed to confirm this hypothesis (e.g. from South America, western Africa or Southeast Asia).
+
+As for the non-tropical areas, the retreat of permafrost and ice sheets probably had substantial effects on carbon storage, turnover times and release of CO\(_2\) into the atmosphere. The retreat of permafrost from these regions is considered one of the mechanisms fueling the deglacial rise in atmospheric CO\(_2\) and the concurrent decline in atmospheric radiocarbon content. Therefore, many previous studies focused on the northern high latitudes (e.g., Ciais et al., 2013; Winterfeld et al., 2018; Meyer et al., 2019 and more) using radiocarbon data and modeling. By contrast, only very few data exist for the subtropics and tropics (to our best knowledge, Hein et al., 2020 (ref.16), and ours) and very little is known about soil-carbon cycling and the impact on atmospheric CO\(_2\) and Δ\(^{14}\)C. Addressing this gap of knowledge is one of our major research goals. We added a small paragraph to the introduction to clarify this (lines 48-52). In our paper we do not address the extra-tropics as they are beyond the scope of our study.
+11. Should different vegetation zones in Figure 1 be marked with different colors.
+
+Reply: We decided to leave the figure as is since we intend to highlight the location and extent of the Nile catchment. We think that this focus would get lost, if the vegetation zones were colored. No action taken.
+
+12. It would be better to mark the year on the horizontal axis in Figure 2.
+
+Reply: Thank you for the suggestion. However, no changes were made as it is common practice to plot kyrs on the x-axis in paleo-studies. We feel indicating years instead of kyrs would make the axis labels more difficult to read, as it would require displaying numbers with up to 5 digits.
+
+13. It seems unreasonable to use surface temperature instead of sea surface temperature in the eastern Mediterranean in Figure 3
+
+Reply: Perhaps this is a misunderstanding. We do not use surface air temperatures (SAT) instead of sea surface temperature (SST). In fact, it is the other way around because no SAT records are available. If there were some, we would have used them instead of SST because SAT would be the best parameter to approach soil temperature in the Nile catchment. As explained in the caption of figure 3 in the initial version of the manuscript we think this that SST still is an appropriate record to use. It is very likely that SST and SAT in the Nile delta region developed similarly due to heat exchange between the sea surface and the overlying air. Other work from the Mediterranean realm concludes the same. For example, Bar-Mathews et al. (2003) summarize in their paper that similarities between \( \delta^{18}O \) records from terrestrial (speleothems) and marine (sediments, foraminifera) archives suggest that SST and land temperatures developed similarly during the last deglaciation and the Holocene.
+
+During the revision we shifted the justification for the selection of the SST record from the figure caption to the main text (lines 186-191) as we felt it this information was a bit hidden previously.

03cad287deb044c05daa550c40716d2819e5af19adf18b7b6e72878c97733fe6/metadata.json

@@ -0,0 +1,176 @@
+{
+    "title": "Vanishing weekly hydropeaking cycles in American and Canadian rivers",
+    "pre_title": "Vanishing weekly hydropeaking cycles in American and Canadian rivers",
+    "journal": "Nature Communications",
+    "published": "09 December 2021",
+    "supplementary_0": [
+        {
+            "label": "Supplementary Information",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-021-27465-4/MediaObjects/41467_2021_27465_MOESM1_ESM.pdf"
+        },
+        {
+            "label": "Peer Review File",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-021-27465-4/MediaObjects/41467_2021_27465_MOESM2_ESM.pdf"
+        },
+        {
+            "label": "Description of Additional Supplementary Files",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-021-27465-4/MediaObjects/41467_2021_27465_MOESM3_ESM.pdf"
+        },
+        {
+            "label": "Supplementary Data 1",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-021-27465-4/MediaObjects/41467_2021_27465_MOESM4_ESM.xlsx"
+        },
+        {
+            "label": "Supplementary Data 2",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-021-27465-4/MediaObjects/41467_2021_27465_MOESM5_ESM.xlsx"
+        },
+        {
+            "label": "Supplementary Data 3",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-021-27465-4/MediaObjects/41467_2021_27465_MOESM6_ESM.xlsx"
+        },
+        {
+            "label": "Supplementary Data 4",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-021-27465-4/MediaObjects/41467_2021_27465_MOESM7_ESM.xlsx"
+        }
+    ],
+    "supplementary_1": NaN,
+    "supplementary_2": NaN,
+    "source_data": [
+        "/articles/s41467-021-27465-4#MOESM3",
+        "http://www.cehq.gouv.qc.ca/hydrometrie/historique_donnees/info_validite.htm",
+        "https://data.usbr.gov/",
+        "https://waterdata.usgs.gov/nwis",
+        "https://wateroffice.ec.gc.ca",
+        "https://www.ibwc.gov/Water_Data/",
+        "/articles/s41467-021-27465-4#MOESM6"
+    ],
+    "code": [
+        "https://doi.org/10.5281/zenodo.5646458"
+    ],
+    "subject": [
+        "Hydrology"
+    ],
+    "license": "http://creativecommons.org/licenses/by/4.0/",
+    "preprint_pdf": "https://www.researchsquare.com/article/rs-441563/v1.pdf?c=1639070231000",
+    "research_square_link": "https://www.researchsquare.com//article/rs-441563/v1",
+    "nature_pdf": "https://www.nature.com/articles/s41467-021-27465-4.pdf",
+    "preprint_posted": "20 Apr, 2021",
+    "research_square_content": [
+        {
+            "section_name": "Abstract",
+            "section_text": "Sub-daily and weekly flow cycles termed \u2018hydropeaking\u2019 are common features in regulated rivers worldwide. Weekly flow periodicity arises from fluctuating hydropower demand and production tied to socioeconomic activity, typically with higher consumption during weekdays followed by reductions on weekends. Here, we propose a novel weekly hydropeaking index to quantify the 1920-2019 intensity and prevalence of weekly hydropeaking cycles at 400 sites across the United States of America and Canada. A robust weekly hydropeaking signal exists at 1.1% of sites starting in 1920, peaking at 17.0% in 1963, and diminishing to 3.2% in 2019, marking a 21st century decline in hydropeaking intensity. We propose this decline may be tied to recent, above-average precipitation, socioeconomic shifts, alternative energy production, and legislative and policy changes impacting water management in regulated systems. Vanishing weekly hydropeaking cycles may offset some of the prior deleterious ecohydrological impacts from hydropeaking in highly regulated rivers.HydrologyClimatologyRenewable ResourcesMarine and Freshwater EcologyCanadaUnited States of AmericaFlow RegulationHuman InterventionHydropeakingHydropowerStreamflow",
+            "section_image": []
+        },
+        {
+            "section_name": "Figures",
+            "section_text": "Figure 1Figure 2Figure 3Figure 4",
+            "section_image": [
+                "https://assets-eu.researchsquare.com/files/rs-441563/v1/3d38a5351a903399cfa95ef2.jpg%3FmaxDims%3D150x150&w=256&q=75.png",
+                "https://assets-eu.researchsquare.com/files/rs-441563/v1/1b201da0144fe2909b4002b4.jpg%3FmaxDims%3D150x150&w=256&q=75.png",
+                "https://assets-eu.researchsquare.com/files/rs-441563/v1/c9109fb958d149bc9434b585.jpg%3FmaxDims%3D150x150&w=256&q=75.png",
+                "https://assets-eu.researchsquare.com/files/rs-441563/v1/4ea2a196f797149159fa12bd.jpg%3FmaxDims%3D150x150&w=256&q=75.png"
+            ]
+        },
+        {
+            "section_name": "Supplementary Files",
+            "section_text": "SupplementaryInformation.pdfSupplementary InformationSupplementaryTable2.xlsxSupplementary Table 2SupplementaryTable3.xlsxSupplementary Table 3WHITimeSeries.xlsxSupplementary Data Table",
+            "section_image": []
+        }
+    ],
+    "nature_content": [
+        {
+            "section_name": "Abstract",
+            "section_text": "Sub-daily and weekly flow cycles termed \u2018hydropeaking\u2019 are common features in regulated rivers worldwide. Weekly flow periodicity arises from fluctuating electricity demand and production tied to socioeconomic activity, typically with higher consumption during weekdays followed by reductions on weekends. Here, we propose a weekly hydropeaking index to quantify the 1920\u20132019 intensity and prevalence of weekly hydropeaking cycles at 500 sites across the United States of America and Canada. A robust weekly hydropeaking signal exists at 1.8% of sites starting in 1920, peaking at 18.9% in 1963, and diminishing to 3.1% in 2019, marking a 21st century decline in weekly hydropeaking intensity. We propose this decline may be tied to recent, above-average precipitation, socioeconomic shifts, alternative energy production, and legislative and policy changes impacting water management in regulated systems. Vanishing weekly hydropeaking cycles may offset some of the prior deleterious ecohydrological impacts from hydropeaking in highly regulated rivers.",
+            "section_image": []
+        },
+        {
+            "section_name": "Introduction",
+            "section_text": "In 2019, the United States of America (USA) and Canada generated a combined 674 TWh of hydroelectricity from a total 184\u2009GW of installed capacity, ranking them with China and Brazil in the four largest global producers of hydroelectricity1. With the proliferation of dam and reservoir construction during the 20th and early 21st centuries2,3, many of the two countries\u2019 main rivers are now moderately or strongly affected by fragmentation, regulation and/or diversions4,5,6. With increasing demands for renewable sources of energy, additional generating capacity is being developed or planned across Canada. This includes the 1,100\u2009MW Site C Dam on the Peace River in northeastern British Columbia (BC), the 824\u2009MW Muskrat Falls development on the lower Churchill River in Labrador, and the 695\u2009MW Keeyask Generating Station on the Nelson River in northern Manitoba1, with its first of seven units becoming operational in February 2021.\n\nWhile overall demand for electricity continues to increase, consumption patterns vary depending on socioeconomic activity, short-term weather conditions, seasonal climate fluctuations and long-term climate trends7,8. In Canada, the winter season usually incurs peak electricity demand due to domestic, commercial and industrial heating and lighting requirements9. With climate change, winter cold waves subside while summer heat waves intensify10,11, shifting some of the demand from winter heating to summer cooling12,13,14. Apart from seasonality shifts, day-to-day activities influence electricity demand as well. Similar to many other industrialized countries, North American educational, industrial and commercial activity intensifies on weekdays (Monday through Friday) but abates on weekends, particularly on Sundays9. This weekly rhythm of socioeconomic activity can thus impact water retention and releases in regulated rivers15. These rapid, frequent and periodic flow fluctuations downstream of regulation points are commonly termed \u2018hydropeaking\u2019 events and are known to disrupt a range of ecohydrological processes16,17. Yet, the characteristics and trends in weekly hydropeaking cycles due to daily variation in electricity demands remain largely unknown. This is despite the general availability of discharge data at a daily time scale and the distinct weekly rhythm of socioeconomic activity including hydropower production, and hence water releases in regulated waterways, which impact ecohydrological processes.\n\nTo address that knowledge gap and a demand for global attention to hydropeaking rivers18, we assess here the prevalence of weekly hydropeaking cycles for 500 gauging sites along rivers of the USA and Canada spanning a wide range of basin characteristics, regulation, hydrological and climatic regimes. Specifically, we develop a scale-independent and dynamic weekly hydropeaking index (WHI) with both time and frequency domain terms, allowing quantification of weekly flow periodicity. As such, the WHI defines the prevalence and intensity of weekly periodicity in flows tied to hydropower production. We show that the WHI captures well the typical weekly rhythm observed in hydropeaking rivers, with low flows on weekends when hydropower demand wanes then high flows on weekdays when hydropower demand waxes. Application of the WHI to 1920-2019 time series of river discharge then provides evidence of vanishing weekly hydropeaking cycles in many regulated rivers of the USA and Canada with the 2010s comparable to the 1920s for hydropeaking prevalence. We propose that increased commercial and industrial activity on weekends, a shift towards other modes of energy production, policy changes altering water management practices, electrical grid interconnectivity and deregulation of electricity generation, plus a relatively wet decade in the 2010s across parts of the study area are likely contributing factors to waning weekly hydropeaking cycles. Thus this work is particularly relevant for long-term planning within the hydropower industry, power system operators and water resources managers.",
+            "section_image": []
+        },
+        {
+            "section_name": "Results",
+            "section_text": "The USA and Canada harbor abundant freshwater resources that include some of the world\u2019s largest rivers (by annual volumetric flows) including the Mississippi, St. Lawrence, Mackenzie, Ohio and Columbia rivers19. Many of these rivers and/or their tributaries have been impounded for hydropower generation, flood control, irrigation, potable water supply, navigation and recreation, leading to fragmented river networks and regulated flows4,6. Indeed, numerous dams have been built across the USA and Canada in the 20th and early 21st centuries2,3. Most dams in North America are operated for multi-purposes shaping seasonal and subseasonal patterns. Hydropower remains a principal component for sub-monthly variations along with flood control. Distinct weekly patterns mark hydropower production except perhaps at run-of-river facilities and those supplying industries continuously in operation such as aluminum smelters or pulp and paper mills8,9. As such, this study focuses on both regulated and unregulated waterways of the USA and Canada to explore the prevalence and intensity of weekly periodicity in discharge.\n\nAs defined, the WHI quantifies weekly periodicity in flows, with larger positive values indicating stronger weekly hydropeaking cycles and negative numbers its absence. The 1980\u20132019 mean, median, and standard deviation of WHI for the 500 sites reach 0.183, 0.056 and 1.121, respectively (Supplementary Table\u00a01). Thirty-eight sites attain a mean annual WHI\u2009\u2265\u20092.0 for 1980\u20132019 with another 64 sites achieving WHI\u2009\u2265\u20091.0. A list of sites with the top ten ranking WHI values reveals their wide regional distribution with foci in the Chattahoochee, Colorado, Etowah, Great Lakes-St. Lawrence, Nelson, Smith and Wallenpaupack drainage basins (Supplementary Table\u00a02), all of which are heavily dammed. The Smith River near Philpott claims the top WHI score of 3.783 while the Namakan River shows the lowest score of \u22123.168. Some highly regulated systems such as Manitoba\u2019s Burntwood River, which funnels water diverted from the Churchill River into the Nelson River, exhibit large negative WHI values (\u22121.884) as Notigi (the upstream point of regulation) is a control structure for a large reservoir operated in a longer term (e.g., seasonal) manner. Similarly, while several large dams impound the Missouri River, they are managed not only for hydropower production but also for flood control, irrigation, navigation and recreational values. As such, the four sites along the Missouri River used in this study exhibit an average WHI\u2009=\u2009\u22120.416 revealing an absence of significant weekly hydropeaking cycles.\n\nA map of the 1980\u20132019 average annual WHI values reveals that weekly hydropeaking rivers abound across the USA and Canada. Clusters of high WHI values emerge in the Alabama, Chattahoochee, Cumberland and Tennessee river basins of the southeastern USA, in waterways draining the Ozark Mountains, the Colorado River and in northern Ontario rivers draining into the Great Lakes (Fig.\u00a01). The Columbia River has several major points of regulation (WHI\u2009\u2265\u20091.5) from its headwaters in BC to its outlet in the Pacific Ocean. Highly hydropeaking sites (WHI\u2009\u2265\u20092.0) appear in both small (e.g., Alberta\u2019s Kananaskis River, A\u2009=\u2009899 km2) and large (Manitoba\u2019s Nelson River, A\u2009=\u20091.1\u2009\u00d7\u2009106 km2) systems. In contrast to their adjacent regulated rivers, free-flowing rivers of northern Canada, particularly those draining into Hudson Bay, exhibit large, negative WHI values. These unregulated rivers manifest strong annual cycles dominated by snowmelt-driven freshets and contain large natural storage capacity in the form of extensive lakes, ponds and wetlands. Free-flowing, pluvial rivers of the southeastern USA (e.g., the Choctawhatchee, Ogeechee, Pascagoula, Satilla and Suwannee rivers) also exhibit negative, albeit\u2009>\u2009\u22121.5, WHI scores. More than 20% of Canadian rivers exhibit WHI\u2009<\u2009\u22121.5 while this proportion reaches 0.6% in American waterways (Supplementary Table\u00a03). Nevertheless, both countries have a similar fraction of sites exhibiting WHI\u2009>\u20090.75 with 26.4% in the USA and 23.7% in Canada.\n\nCircle size corresponds to the two consecutive days with low flows beginning with the Saturday/Sunday (SS, largest symbols) combination and ending with the Friday/Saturday (FS, smallest symbols).\n\nWHI values diminish moving downstream from a point of regulation. For instance, WHI\u2009=\u20091.437 on the Peace River just downstream of BC\u2019s WAC Bennett and Peace Canyon dams where minimum flows arise on weekends; 400\u2009km downstream from the dams20, however, WHI declines to 0.912 at the community of Peace River in Alberta where minimum flows occur on Mondays/Tuesdays, indicating a 2-day delay in signal propagation. A cascade of dams and reservoirs can accentuate the hydropeaking signals along waterways (e.g., the Colorado, Columbia, and Tennessee rivers) or attenuate them (e.g., Ottawa River).\n\nSites with high values of WHI (\u2009\u2265\u20091.5) also show a preponderance of flow reductions on the weekends (Saturdays/Sundays) as identified by the larger symbols in Fig.\u00a01. Of the 63 sites with WHI\u2009\u2265\u20091.5, 56 experience the two consecutive days with low flows on weekends. In contrast, sites with negative WHI values show a range of low flow days with no distinct pattern emerging. No less than 31.0% of all sites used in this study exhibit low flows on Saturdays/Sundays, more than twice the expected value (Fig.\u00a02). This disproportionate amount of weekend low flows occurs mainly in hydropeaking rivers (WHI\u2009>\u20090). Weekday combinations show frequencies at, or lower than, the expected value with the Friday/Saturday sequence appearing at only 6.8% of sites. A Chi-Square test applied to the frequency of two consecutive low flow days reveals that the results differ significantly from the expected value of 0.143 (\u03c72\u2009=\u2009136.33, p\u2009<\u20092.2\u2009\u00d7\u200910\u221216, n\u2009=\u20097 with six degrees of freedom). The mean WHI equals 0.326 for 155 sites with low flows on weekends while it remains near zero or slightly negative for the six other two-day combinations. The distribution of mean WHI for the two-day combinations differs significantly from a uniform distribution based on a Chi-Square test (\u03c72\u2009=\u200912.286, p\u2009=\u20090.027 based on 10,000 replicates with n\u2009=\u20097).\n\nBlack bars denote the two consecutive days with low flows while red bars represent the WHI values for 500 sites across the USA and Canada, 1980\u20132019. Fractions of the two consecutive days with low flows are partitioned according to positive (solid) and negative (dotted) WHI values. The days of the week begin with the Saturday/Sunday (SS) combination and end with the Friday/Saturday (FS) combination. The horizontal black line denotes the expected value if the two-day low flows were distributed randomly while the horizontal red line marks the mean WHI across the 500 sites.\n\nThe temporal evolution of the mean and median WHI shows a rapid increase in hydropeaking intensity from the 1920s to the 1950s at which point they level off and fluctuate near zero (Fig.\u00a03). Starting in the 1990s, though, there is a gradual decline in both the mean and median WHI values with a return in the 2010s to statistics first seen in the 1930s (largely pre-regulation), a pattern observed both in the USA and Canada (not shown). The discharge-weighted WHIQ, whereby the sum of WHI times mean annual discharge are then normalized by annual discharge summed at all available sites, emphasizes the increasing volumes of regulated flows starting from the 1920s through the 1980s; however, WHIQ also declines markedly thereafter into the 21st century. In 1920, only 1.8% of available sites rank in the top decile of 1920\u20132019 WHI values (WHI\u2009\u2265\u20092.117). This fraction peaks at 18.9% of available sites in 1963 but thereafter diminishes consistently. In 2000, 66 or 13.8% of available sites score in the top decile of 1920\u20132019 WHI values but these counts fall precipitously to just 15 or 3.1% of the available sites by 2019, marking a 21st century declining pattern in weekly hydropeaking intensity. Trend analysis applied to the overall mean annual WHI reveals a statistically significant decline of \u22120.40 over 1980\u20132019 (Supplementary Fig.\u00a01). These temporal results, however, rely on the availability of discharge data, as the record length averages 79.7 years, ranging from a minimum of 24 years at one site to a full century at 69 sites (Supplementary Fig.\u00a02). The number of available sites increases steadily from 1920 into the early 1990s and peaks at 492 sites in 1985 and 1992 but then declines to 468 sites by 1996 thereafter averaging 483\u2009\u00b1\u20095 sites until 2019. Notable gaps appear in the discharge records starting in the 1990s, particularly for regulated rivers in Ontario and Qu\u00e9bec; however, adjusting the time series of mean annual WHI for unavailable sites reveals little difference in the overall pattern and trend of WHI during 1980\u20132019 (Supplementary Fig.\u00a01).\n\nThe (a) annual mean, median and discharge-weighted WHI values for up to 500 sites in the USA and Canada, the (b) number and (c) percentage of available sites ranking in the top decile of all WHI values, and the (d) annual maximum WHI values, 1920\u20132019.\n\nData availability also factors in the appraisal of the decadal evolution of hydropeaking intensity across the USA and Canada (Fig.\u00a04). Nevertheless, this shows the gradual inception of hydropeaking cycles during the 1920s and 1930s, particularly in the north-central, northeastern, and southeastern USA and in northern Ontario. The 1940s show an expansion of weekly hydropeaking rivers into the western USA including within the Colorado, Columbia and Sacramento river basins as the 1930s New Deal projects came online. The 1940s and 1950s mark an intensification of regulation in the Tennessee and Alabama river basins, the Ottawa Valley as well as rivers of northern Ontario draining to Lakes Superior and Huron. A pronounced expansion and amplification of the hydropeaking signal appears in the 1960s, particularly across the Great Lakes-St. Lawrence river basin in Ontario and Qu\u00e9bec. Some stabilization of the hydropeaking pattern marks the 1970s but a resurgence follows in the 1980s and 1990s when additional hydropeaking rivers emerge in western Canada. The 2000s retain a wide distribution of hydropeaking rivers across both countries; yet, by the 2010s, the number of highly hydropeaking rivers diminishes considerably, particularly in parts of the Great Lakes-St. Lawrence and Tennessee river basins.\n\nMaps are shown for (a) 1920\u20131929, (b) 1930\u20131939, (c) 1940\u20131949, (d) 1950\u20131959, (e) 1960\u20131969, (f) 1970\u20131979, (g) 1980\u20131989, (h) 1990\u20131999, (i) 2000\u20132009, and (j) 2010\u20132019. Circle size corresponds to the two consecutive days with low flows beginning with the Saturday/Sunday (SS, largest symbols) combination and ending with the Friday/Saturday (FS, smallest symbols). Results are shown only when ny\u2009\u2265\u20095 years in a given decade.\n\nThe decadal distribution of the 10 WHI bins (Fig.\u00a05a) further highlights the peak fraction of sites with WHI\u2009\u2265\u20091.5 attained in the 1960s (20.7%), with nearly matching minimum values in the 1920s (6.8%) and 2010s (7.8%). After the 1960s, there is a steady decline in the relative number of sites with low flows either on the Saturday/Sunday or Sunday/Monday combinations, indicating waning differences between weekday and weekend flows across the USA and Canada (Fig.\u00a05b).\n\nIn (a), WHI bins match those used in Fig.\u00a04 with a similar color palette (e.g., the maroon bars indicate WHI\u2009\u2265\u20093.0 starting at a zero cumulative percentage). In (b), the two-day combinations with low flows start on Friday/Saturday (FS) at a zero cumulative percentage (maroon bars) and end on Saturday/Sunday (SS) at 100% (black bars).\n\nThe temporal evolution of the annual maximum WHI value shows a rapid increase from ~3.0 in the 1920s to > 4.0 in the 1930s onward (Fig.\u00a03d). Annual peak WHI values > 4.0 are generally sustained for the remainder of the 20th century but then remain near 4.0 or below that threshold starting in 2003 until 2019. The peak WHI value each year over the study period is distributed among 21 sites, with the Smith River at Philpott capturing the top spot most often at 28 times and scoring the overall maximum WHI of 4.752 in 1955 (Supplementary Fig.\u00a03).\n\nFurther statistical analysis reveals an abundance of strong, negative WHI trends interspersed with positive ones for the 479 sites with ny\u2009\u2265\u200930 years over 1980\u20132019 (Fig.\u00a06). A total of 138 sites show locally statistically significant (p\u2009<\u20090.05) declines in WHI while 28 show locally statistically significant inclines. Of the 166 locally significant trends, 129 remain globally significant. Significant negative WHI trends abound in the southeastern and northeastern USA, the Great Lakes-St. Lawrence basin, and the Pacific Northwest while a cluster of positive trends arises in Qu\u00e9bec\u2019s Saguenay watershed. Clusters of negative WHI trends lie primarily within the Western, Northeastern and Southeastern Interconnects. While regulated rivers of Newfoundland show increasing WHI values, their unregulated counterparts show similar tendencies. Furthermore, in New Brunswick, the regulated Saint John River shows a decreasing trend in WHI while the proximal, unregulated Southwest Miramichi River shows an increasing trend. Sixty-nine percent of the locally significant WHI trends arise in hydropeaking rivers (WHI\u2009>\u20090) with fewer locally significant trends in non-hydropeaking rivers (WHI\u2009<\u20090; Supplementary Fig.\u00a04). Application of the Pettitt test21 reveals that two-thirds of the locally significant trends detected with the Mann-Kendall test also correspond to statistically significant break points in the WHI time series (Supplementary Data\u00a01).\n\nRed upward (blue downward) pointing triangles indicate positive (negative) trends. Trend magnitudes are proportional to the triangle sizes and green circles (pink outlines) indicate locally (globally) statistically significant trends (p\u2009<\u20090.05). Results are shown only when ny\u2009\u2265\u200930 years.\n\nThe commissioning and operation of hydroelectric facilities including dams and reservoirs markedly influences WHI evolution. For instance, development of hydropower dams by the Tennessee Valley Authority (TVA) in the first half of the 20th century across the Tennessee River Basin induces sharp increases in WHI, often from large negative to positive values (Supplementary Fig.\u00a05). In some cases, however, the WHI is already elevated at gauging sites, reflecting the presence of additional upstream points of regulation (e.g., the Hiwassee Dam on the Hiwassee River commissioned in 1940). Consistent with the general pattern observed across the USA and Canada, most sites operated by the TVA show an attenuation of WHI values in the 2010s. Where present, reservoir influence on WHI depends on its function (Supplementary Fig.\u00a06). When managed for hydropower production among other functions, the WHI typically stays positive with many surpassing scores of 2.0. At sites where reservoirs serve other purposes, the WHI remains substantially lower alternating between positive and negative values.\n\nWater management practices and climate variability, among other factors, yield significant interannual variation in hydropeaking intensity. For example, the Colorado River at Lees Ferry shows marked declines in WHI during high flow years (Supplementary Fig.\u00a07a). Indeed, heavy precipitation during strong El Ni\u00f1o events in the early 1980s induced high flows in the Colorado River including at Lees Ferry. Due to the unusually wet weather, the bypass tubes and spillway at Glen Canyon Dam were used to release additional water downstream, thereby moderating hydropeaking signals from 1983 to 198622. Similar declines in WHI appear in 1997 and 2011 when flows exceed the recent annual average. Computing the Pearson correlation coefficient between the 1980\u20132019 annual river discharge and the corresponding WHI yields 94 statistically significant negative correlations and only 19 statistically significant positive correlations (Supplementary Fig.\u00a07b). Thus high flows over extended periods attenuate weekly periodicity even in heavily regulated rivers such as the Colorado.\n\nThis analysis suggests that sustained wet periods may attenuate hydropeaking intensity while dry periods may accentuate it. Binned distributions of decadal standardized anomalies in river discharge reveal the contrasting dry 1930s vs. the wet 1970s, the latter coinciding with a suppression of hydropeaking across the USA and Canada (Supplementary Fig.\u00a08a). Yet, while the 2010s experienced relatively high flows, 7.8% of sites have WHI\u2009\u2265\u20091.5 whereas in the similarly wet 1990s, 17.4% of sites achieve WHI\u2009\u2265\u20091.5. Of 19 sites with large (> 1), positive standardized discharge anomalies during the 2010s, only four (the Betsiamites, La Grande and Nelson rivers plus Wallenpaupack Creek) have WHI\u2009>\u20091, which are likely more in response to enhanced diverted flows (excluding Wallenpaupack Creek) rather than high precipitation. While there are robust positive discharge anomalies in the north-central plains and northeastern USA and parts of central Canada in the 2010s, other regions with significant WHI declines exhibit near neutral or even negative discharge anomalies (Supplementary Fig.\u00a08b). Thus it is unlikely interdecadal climate variations alone account for recent WHI declines.\n\nApart from climate variations, changes in day-of-the-week flows may influence WHI trends. Sites with WHI\u2009>\u20090 generally observe greater dispersion of day-of-the-week flows although pluvial and intermittent rivers, particularly in the southern USA, also experience greater day-to-day flow variations (Supplementary Fig.\u00a09a). A trend analysis reveals significant declines in the dispersion of flows across the seven days of the week, concomitant with diminishing WHI values from 1980 to 2019 (Supplementary Fig.\u00a09b). As an example, an abrupt reduction in dispersion of day-of-the-week flows in Labrador\u2019s Churchill River appears in 1997 and is then sustained, suggesting factors other than climate variations are altering daily flows (Supplementary Fig.\u00a010).",
+            "section_image": [
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-021-27465-4/MediaObjects/41467_2021_27465_Fig1_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-021-27465-4/MediaObjects/41467_2021_27465_Fig2_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-021-27465-4/MediaObjects/41467_2021_27465_Fig3_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-021-27465-4/MediaObjects/41467_2021_27465_Fig4_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-021-27465-4/MediaObjects/41467_2021_27465_Fig5_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-021-27465-4/MediaObjects/41467_2021_27465_Fig6_HTML.png"
+            ]
+        },
+        {
+            "section_name": "Discussion",
+            "section_text": "The recent decline in weekly hydropeaking cycles in the USA and Canada emerges as a key finding in this study. Several possible factors may be contributing to this general pattern observed over the study area. Firstly, electricity demand, production and consumption may have shifted in recent years, thereby diminishing differences between weekdays vs. weekends. For instance, there has been a gradual shift towards more commercial (including e-commerce) and industrial activity on weekends that could alter the weekly discharge patterns in regulated rivers23,24. A shifting manufacturing sector, globalization, and lifestyle changes are all socioeconomic factors modifying electricity demand. Another possible factor is the development and expansion of other modes of energy production such as dispatchable combustion turbines and non-dispatchable solar and wind energy (Supplementary Fig.\u00a011). Solar and wind energy production activate during favourable weather conditions with hydropower otherwise matching the demand, which may disrupt the typical weekly pattern in regulated flows while allowing hydropower to offer new types of services such as capacity markets. Furthermore, the rapid increase in electricity production from non-hydro renewable sources coincides with the sharp decline of weekly hydropeaking intensity in the 2010s (Supplementary Fig.\u00a012).\n\nRegulatory bodies and changing governmental policies may also be altering how utilities manage regulated waterways. Indeed, there is renewed interest for environmental, ecological and cultural (e.g., from a First Nations or Indigenous perspective) flows in human-influenced systems, with emerging regulations and policies supporting their implementation25. For instance, regulatory changes in the operation of the Prickett hydroelectric facility from a peaking to run-of-river site to assist spawning lake sturgeon26 induced a significant WHI decline (of \u22120.216 decade\u22121) along the Sturgeon River in the upper peninsula of Michigan starting in the 1990s. Indeed, changes in operation away from peaking hydropower generating stations, whether mandated or voluntary, could influence hydropeaking patterns.\n\nThe increasing interconnectivity of the North American power grid, deregulation, and centralization of electricity dispatching may further contribute to a recent reduction of hydropeaking intensity. Finally, climate variations may also play a role in hydropower production as wet periods may require greater spillage of water from reservoirs thereby diminishing hydropeaking intensity. Alternatively, wet years may lead utilities to generate continuous baseload energy instead of peaking hydropower, inducing a similar effect. The relatively wet climate of the 2010s could account for part of the recent declines in WHI across Canada and the northern half of the conterminous USA. Thus a combination of factors including changing electricity demand patterns tied to lifestyle factors and socioeconomic activity, the emergence of alternative modes of energy production plus power grid interconnectivity, implementation of regulations and policies, and climate variations may be influencing the day-to-day hydrology of many regulated waterways across the USA and Canada.\n\nGiven the vast territory of the USA and Canada, their waterways often drain multiple jurisdictions including international transboundary watersheds (e.g., the Rio Grande, Great Lakes-St. Lawrence, Winnipeg and Columbia rivers). Regional water authorities, inter-jurisdictional water boards, federal, provincial, and state legislation, and international water treaties and commissions all affect how waterways are managed. Furthermore, synchronous inter-jurisdictional power grids (e.g., interconnections) can also affect hydropower generation and hence regulated flows, leading to distinct spatio-temporal patterns in hydropeaking intensity. Decadal maps of WHI values reveal the progression of weekly hydropeaking systems from the eastern and central USA to the Pacific Northwest in the 1960s when development in the Columbia River Basin expanded rapidly. The international Columbia River Treaty implemented in 1961 led to the construction of three major dams along the Columbia River (Duncan, Keenleyside and Mica Dams in Canada) plus another on the Kootenai River (Libby Dam in the USA)27. These dams and generating stations expanded the presence of hydropeaking cycles from the lower to the upper Columbia River Basin in the 1970s and 1980s (Fig.\u00a04). As such, regulation in the Canadian portion of the Columbia River Basin now leads to downstream propagation of hydropeaking into the northern USA where it is regenerated at multiple points of regulation including Grand Coulee Dam and the Dalles.\n\nAnother noticeable pattern in the decadal results is the WHI decline in many rivers of southern Qu\u00e9bec in the 1970s and 1980s. As the 5,428\u2009MW Churchill Falls generating station in Labrador came online in late 1971 (with hydropower sold mainly to the provincial utility Hydro-Qu\u00e9bec)28, followed a decade later by the 17,418\u2009MW James Bay Hydroelectric Complex in northern Qu\u00e9bec15, a northward shift in hydropower generation abated the weekly hydropeaking cycles in more southern waterways. Simultaneous reductions in WHI in the northeastern USA (e.g., Hudson and Connecticut Rivers) may also be tied to transboundary power grid interconnections and Hydro-Qu\u00e9bec\u2019s large export capacity (7,974\u2009MW in 201929). Similar to regional climate trends30, synchronous power grids thus have the capacity to shift the intensity of hydropeaking signals 1000s of kms away from points where hydropower is consumed, thereby creating hydropeaking teleconnections with potential for far-reaching social and ecohydrological effects.\n\nEcohydrological impacts of hydropeaking are site-specific and may include rapid changes in water temperature (i.e., \u2018thermo-peaking\u2019), increases in soil erosion and suspended matter, and habitat degradation, which affect ecosystems, reduce species abundance, and limit biodiversity (e.g., fish, riparian plants, macroinvertebrates)16,31,32. Across the USA and southern Canada, hydropeaking emerged relatively early in the 20th century with the proliferation of dams and flow regulation in these regions. Starting in the 1960s, hydropower infrastructure expanded northwards into regions previously devoid of any significant flow regulation and hydropeaking. This includes major waterways like BC\u2019s Peace River, Manitoba\u2019s Nelson River, Ontario\u2019s Moose and Abitibi rivers, and Qu\u00e9bec\u2019s La Grande Rivi\u00e8re. On these systems, major dams and reservoirs were built from the 1960s to early 1980s, vastly expanding the northern reach of hydropeaking rivers (Supplementary Fig.\u00a013). This shifted potential ecohydrological impacts of hydropeaking to areas also undergoing rapid climate change through Arctic amplification of global warming33. As such, sub-Arctic species of fish (e.g., brook trout, lake sturgeon, northern pike, and walleye), insects and riparian plants may now be exposed to the cumulative impacts of these environmental stressors17. Additionally, winter frazil ice production and ice jams may be precipitated and accentuated downstream of hydroelectric facilities with persistent hydropeaking signals such as in the Peace River20.\n\nDespite their recent northward expansion, weekly hydropeaking cycles are generally waning across the USA and southern Canada, suggesting a 21st century hydropeaking recovery in some of these river systems. Indeed, prior ecohydrological impacts of hydropeaking may be partially offset, benefiting local biota and ecosystem biodiversity34. For instance, recovery of lake sturgeon in the northern peninsula of Michigan demonstrates some of the benefits of shifting away from peaking hydropower operations26. This is particularly important as evidence is also mounting that hydropeaking influences aquatic species in rivers of Canada35,36,37,38. Other aspects of flow regulation, such as sub-daily flow fluctuations and associated ramping up and down cycles not investigated in this study, may negate this hydropeaking recovery16,17. Additional research is thus needed to explore hydropeaking cycles at other temporal scales to establish their site-specific ecohydrological impacts.\n\nThe proposed index to infer weekly hydropeaking signals provides a complementary metric to those developed in other studies5,39,40. Advantages of our approach include its scale independence, dynamic response, and relatively simple implementation. The WHI can be applied from small (<1\u2009\u00d7\u2009103\u2009km2) to large (>1\u2009\u00d7\u2009106\u2009km2) river basins with available daily discharge data (whether observed, reconstructed or simulated). The WHI responds to interannual variability in climate (e.g., wet/dry periods), changes in water management practices and policies, commissioning of new hydroelectric facilities or decommissioning of old ones, and other factors that affect flows. The use of daily discharge data also avoids the need for extensive databases on dams, reservoirs and other infrastructure that influence flows. Its possible implementation for short-term flow predictions emerges as another distinct advantage of the WHI. As an example, a running value of the WHI can be computed on the past year\u2019s daily flows and used to infer the possible deviations in daily flows over a given week based on recent historical patterns. Its computational simplicity, coded in our study in Fortran, allows processing of results for the 500 sites in <4\u2009min. As such, it is feasible to implement a version of the code for short-term flow predictions so long as up-to-date daily flow records remain available. It would also be relatively straightforward to adapt the code to explore sub-daily hydropeaking cycles9 if appropriate discharge data are available.\n\nOne challenge in implementing the WHI is access to daily discharge records. While considerable gauging stations exist in most of the USA and southern Canada, other waterways are not necessarily well monitored. A late 20th century decline in hydrometric stations due to budget restraints41 and the Water Survey of Canada\u2019s curtailment of data collection combined with stricter quality standards from third parties have exacerbated hydrological data accessibility. As well, private industry and government-owned corporations often record discharge at or near their hydroelectric facilities, but may consider these data as sensitive such that they are not released publicly or remain difficult to access. Thus, acquisition of daily discharge data in regulated systems, particularly as the number of small, private firms operating run-of-river hydroelectric facilities expands3, yields a distinct challenge in accessing flow data. Therefore, remote sensing42, data reconstructions (e.g., from statistical models or machine learning methods43) and numerical simulations that incorporate regulation44 are key in filling spatio-temporal gaps where and when in situ observations are lacking.\n\nAs hydropower generation and infrastructure development continue to expand across the USA and Canada, it is imperative to establish how water management practices affect downstream river flows and ecosystems. Common features in regulated rivers are discharge periodicities associated with hydropower production ebbs and flows including weekly cycles. In this study, a measure of this weekly rhythm in flows, the weekly hydropeaking index (WHI), is formulated and applied to 500 sites over parts of North America. Our analyses reveal that 29% of sites with at least three decades of available data during 1980\u20132019 exhibit locally statistically significant declines in WHI while only 6% show inclines. Moreover, the fraction of sites with WHI\u2009\u2265\u20091.5 dropped by half from the 2000s to the 2010s reverting to a value observed in the 1920s. Major watersheds observing significant declines in weekly hydropeaking include the Alabama, Columbia, Cumberland, Great Lakes-St. Lawrence, and upper Mississippi, which fall within the Eastern and Western Interconnects. Regional clusters of declining WHI highlight hydropower operations and river regulation governed at the watershed-, interconnect- and utility-scale.\n\nFactors possibly yielding vanishing weekly hydropeaking cycles include increased commercial and industrial activity on weekends, a shift towards other modes of energy production during peak demand hours or days, and policy changes altering water management practices including for cultural, ecological and environmental flows. This reduction in weekly hydropeaking also may benefit aquatic species, insects and riparian vegetation that otherwise are susceptible to rapid shifts in flows and water levels. Future efforts should therefore establish the ecohydrological implications of waning weekly hydropeaking cycles. The application of the WHI to other regions over the globe would provide broader perspectives on the commonality of this feature in regulated rivers. Lastly, detailed investigations at various spatial (e.g., watershed, interconnect, utility) and temporal (e.g., seasonal) scales should be undertaken to elucidate the role of governing agencies and hydroclimate on hydropeaking globally.",
+            "section_image": []
+        },
+        {
+            "section_name": "Methods",
+            "section_text": "A total of 500 sites across the USA and Canada ranging 190\u20121,805,222\u2009km2 in gauged area (A), 25\u201360\u00b0N in latitude, 54\u2013132\u00b0W in longitude, and 0.02\u2013268.28\u2009km3 in mean annual discharge are selected for this study (Supplementary Fig.\u00a014 and Supplementary Data\u00a02). A primary site selection criterion is daily discharge data availability for \u226524 years between 1920\u20132019, with \u226514 years during the focus period of 1980\u20132019. The chosen sites span a wide range of hydrological regimes from pluvial rivers in warmer climates (e.g., BC\u2019s Yakoun River) to nival and glacial systems at higher elevations or latitudes in cooler climates (e.g., BC\u2019s Lillooet River)45. Thus, the study area spans regions with little to no snowmelt where sub-annual scales govern temporal variability while others are mainly snowmelt-driven with predominant annual cycles46. The database also includes intermittent streams in warmer, drier climates such as California\u2019s Santa Ana River and Arizona\u2019s Little Colorado River. Regulated and unregulated rivers are selected (using guidance from Benke and Cushing19) to allow comparisons between sites. Some sites such as Lees Ferry on the Colorado River include extended records that cover pre- and post-regulation effects on flows.\n\nData and metadata (station ID, gauge coordinates, and gauged area) are extracted from various sources including publicly accessible databases maintained by federal, provincial and state agencies in addition to proprietary or unpublished data from private industry, government-owned utilities and international commissions. For most unregulated rivers, daily discharge data are sourced partly from the Water Survey of Canada\u2019s Hydrometric Database (HYDAT), the Centre d\u2019Expertise Hydrique du Qu\u00e9bec (CEHQ) and the United States Geological Survey (USGS). For regulated rivers, though, daily discharge data are not necessarily available from these sources or other public repositories as they are partially or entirely collected, quality controlled and archived by government-controlled utilities or private industry (see Supplementary Data\u00a02 and 3). This includes: Nalcor Energy for the Salmon and Exploits rivers plus the Churchill Falls (Labrador) Corporation Limited for the Churchill River at Churchill Falls Powerhouse in Newfoundland and Labrador; NB Power for the Saint John River in New Brunswick; Rio Tinto for the Kemano Powerhouse in BC and the Saguenay and P\u00e9ribonca rivers in Qu\u00e9bec; Hydro-Qu\u00e9bec for La Grande Rivi\u00e8re, Betsiamites, Gatineau, Manicouagan, des Outaouais, des Outardes and St-Maurice rivers; Evolugen by Brookfield Renewable for the Coulonge, Li\u00e8vre, and Noire rivers in Qu\u00e9bec and Mississagi and Aux Sables rivers in Ontario; Ontario Power Generation for the Abitibi, English, Kaministiquia, Madawaska, Mattagami (tributary to the Moose River), Montreal and Ottawa rivers; H2O Power for the Abitibi River; Manitoba Hydro for the Nelson and Winnipeg rivers; TransAlta for the North Saskatchewan and Kananaskis rivers; and BC Hydro for the Columbia River at Mica Dam. Additional data for gauges along the Rio Grande on the border between the USA and Mexico and the Pecos River are provided by the International Boundary and Water Commission. Data at 14 sites in the Tennessee River Basin and another site in the Cumberland River Basin are provided by the Tennessee Valley Authority. The United States Army Corps of Engineers (USACE) shared data for nine sites they manage in the Cumberland River Basin. Recent records of daily discharge from the US Bureau of Reclamation supplement those from the USGS for sites on the Colorado and upper Rio Grande rivers. Potential errors associated with discharge measurements and implications to our results are discussed in the\u00a0Supplementary Methods.\n\nThe overall study period spans 1 January 1920 to 31 December 2019 for which at least partial, extended (\u226524 years) records of daily discharge are available at all sites. Time series of daily streamflow (in m3 s\u22121) are constructed based on data availability for each site of interest (Supplementary Data\u00a02) and follows D\u00e9ry et al.47 in its approach. Daily discharge data sourced from the USGS, US Bureau of Reclamation, Tennessee Valley Authority, Nalcor Energy (Exploits River), USACE and NB Power are converted to metric units prior to analysis. For several waterways (e.g., the Nelson and Saguenay Rivers), data furthest downstream are first used, but when unavailable (prior to construction of dams and hydroelectric facilities), are replaced with those from the closest upstream gauging station while adjusting the data for the missing contributing area as necessary47,48. Gaps are in-filled with the mean daily discharge over the period of record; however, any calendar year with \u226510% missing records is excluded from analysis. Supplementary Data\u00a02 lists the percentage of in-filled data at each site (average: 0.02%, maximum: 0.58%) omitting years when \u226510% of the data remain unavailable. Uncertainty in the results associated with data homogeneity and the gap-filling strategy is evaluated and discussed in the\u00a0Supplementary Methods.\n\nVarious approaches are commonly used to explore flow alterations in regulated rivers including comparisons of hydrographs pre- and post-regulation9,49,50, trends in peak and/or low flows51 or of naturalized versus observed (regulated) flows52,53,54. A broader approach employs a set of multiple (up to 64) indicators of hydrologic alteration to quantify changes over the water year arising from regulation55,56,57. Another method combines hydrological data, reservoir information and a database of large dams in developing river regulation and fragmentation indices with a matrix of impact for application to all major global watersheds4,5. Apart from time domain analyses, Discrete Fourier Transforms or wavelet analyses offer additional insights on impacts of flow alterations from human interventions15,22,46,58. Consult Jumani et al.40 for a review of river regulation and fragmentation indices including their applications, advantages and limitations.\n\nWhile various approaches exist to infer hydrologic alterations from diversions, dam and reservoir operations including sub-daily hydropeaking cycles59,60, none focuses on the weekly timescale, a primary periodicity of socioeconomic activity. Therefore, we develop a WHI that combines time and frequency domain terms to quantify weekly periodicity in river discharge. The time domain term (TT, %) counts the number of weeks (Dw) in a given calendar year when two consecutive days exhibit flows lower than the corresponding weekly average (\\(\\overline{{{Q}_{1-7}}}\\)), followed by five sequential days above the corresponding weekly average:\n\nThis sequence of daily flows is chosen to emphasize the typical weekly rhythm observed in hydropeaking rivers: low flows on weekends when electricity demand wanes, followed by high flows on weekdays when electricity demand waxes9. A partial score of 0.25 is ascribed to sites where six consecutive days above the weekly average follow a single low flow day for that week. As some gauging sites lie downstream from points of regulation such that low flows are shifted later in the week rather than occurring on Saturdays and Sundays, we test all seven possible combinations of two consecutive days (e.g., Saturday/Sunday, Sunday/Monday, \u2026, Friday/Saturday) and select the one that maximizes WHI at each site over the period of record. This approach for the time domain term attenuates the effects of cyclical (rather than periodic) variations from synoptic-scale storm activity, which otherwise leads to marked weekly cycles in pluvial rivers46.\n\nAn application of Discrete Fourier Transforms to the daily discharge data provides the frequency domain term. Here we follow Wilks61 in partitioning the daily discharge time series into sine and cosine waves of amplitude Ck for harmonic k. Discrete Fourier Transforms are computed for each calendar year with the 52nd harmonic representing the weekly timescale of interest here. Then we compute the explained variance of the 52nd harmonic (TF):\n\nwhere n is the number of days in a given year (365 or 366 for a leap year), C52 is the amplitude of the 52nd harmonic, and sQ is the standard deviation in discharge.\n\nAfter expressing TT and TF as percentages, we take the base 10 logarithm of their product to obtain an annual WHI:\n\nin which B (= 10) is a coefficient chosen so that the median WHI\u2009\u2248\u20090 among all 500 sites. Annual WHI values range typically from about \u22124 to +4 (although WHI values have no theoretical upper or lower bounds), with large positive values indicating strong weekly periodicity attributed to flow regulation at hydropower stations. In contrast, rivers with robust annual cycles with flows dominated by potent snowmelt-driven freshets and/or large (natural) storage capacity within abundant lakes, ponds and wetlands exhibit large negative WHI values. The transition between negative to positive WHI values marks a shift from annual to weekly dominant time scales of variability in flows. The 1980\u20132019 mean daily flows (considering the day of the week) for the Namakan River (Minnesota/Ontario), St. Croix River (Maine/New Brunswick), and Smith River near Philpott (Virginia) illustrate the WHI ranging from the minimum, median, and maximum values (Supplementary Fig.\u00a015). WHI values remain site-specific and must be interpreted with care, particularly moving away (both upstream and downstream) from measurement sites with an intervening body of water, a confluence or another point of regulation altering hydropeaking intensity.\n\nWe first compute WHI time series at all 500 sites (Supplementary Data\u00a04) and develop a \u2018climatology\u2019 of index values for 1980\u20132019, with 14 years \u2264\u00a0ny\u2009\u2264\u200940 years depending on data availability at each site. Results for 1980\u20132019 are also tabulated in WHI bins of 0.75 across all sites, the USA and Canada. Summary statistics (mean, median, standard deviation, etc.) of the 1980\u20132019 WHI data are tabulated and their distribution tested for normality using the Shapiro-Wilk test. Similar climatological analyses are developed for each decade (1920s to 2010s) with results reported when ny\u2009\u2265\u20095 years at a given site. The Mann-Kendall test (MKT62,63) applied to all WHI time series with ny\u2009\u2265\u200930 years over 1980\u20132019 yields linear, monotonic trends in hydropeaking intensity, with p\u2009<\u20090.05 considered locally statistically significant. The field (or global) significance of the individual (or local) trend tests is assessed following Wilks61. The approach minimizes the false discovery rate (FDR) by first ranking p-values in ascending order for all trend tests with ny\u2009\u2265\u200930 years. Trends are then globally significant if p < pFDR depending on the distribution of sorted p values as:\n\nin which we set \u03b1global\u2009=\u20090.10. Trend analysis sensitivity to autocorrelation is tested in the\u00a0Supplementary Methods. As the MKT does not distinguish between gradual versus abrupt changes in a variable, we implement the Pettitt test21 while considering p\u2009<\u20090.05 as a break point in WHI time series. The year when the change point is identified along with the mean WHI prior to and after the break point years are tabulated.\n\nWe assess the 1920 to 2019 annual mean, median and maximum WHI across all sites with available data in a given year to track the overall evolution of hydropeaking intensity across the USA and Canada. We also count the annual number and percentage of sites that fall in the top decile of all 1920\u20132019 WHI scores. An additional metric reported is the discharge-weighted WHIQj computed each calendar year (index j) as:\n\nwhere Qi,j (km3 yr\u22121) denotes the annual discharge and i is the site index. This yields a relative measure of annual volumetric flows affected by weekly hydropeaking cycles rather than just the number of sites. For monotonic trend analysis, the MKT is applied to time series of overall mean annual WHI over the 1980\u20132019 focus period. The potential influence of missing data on the evolution of average WHI over 1980\u20132019 is assessed by substituting incomplete time series with each missing site\u2019s average WHI computed over the remainder of the focus period. This yields an adjusted mean annual WHI time series for a first order assessment of the influence of incomplete data.\n\nA histogram illustrates the distribution of two consecutive days when low flows emerge relative to the expected value of 1/7\u2009=\u20090.143 were these randomly distributed. Fractions of the seven possible two-day combinations are partitioned according to WHI \u22db 0. The histogram also includes the corresponding mean WHI across all rivers for a given two-day combination of low flows. A Chi-Square goodness-of-fit test64 verifies the hypothesis of whether the distribution of low flow days differs significantly from the expected value with threshold p = 0.05. Similarly, we test if the corresponding mean WHI values for the two-day pairs with low flows follow a uniform distribution using a Chi-Square test. The relationship between annual WHI values and mean annual flows over 1980\u20132019 is evaluated using Pearson\u2019s correlation coefficient with p\u2009<\u20090.05 considered statistically significant values. Next, we transform annual discharge time series to standardized anomalies over the period of record at each site (with <10% missing data in a calendar year). Decadal mean standardized anomalies for all available sites are then computed when ny\u2009\u2265\u20095 years in a given decade. These decadal average anomalies are binned in increments of 0.25 standard anomaly for comparison with WHI decadal distributions.\n\nThe influence of dams on the temporal evolution of WHI values is assessed using 14 hydroelectric facilities managed by the Tennessee Valley Authority (TVA;65 Supplementary Data\u00a03). Here, we take the year a project was completed as its commissioning year to establish the response of the WHI to flow regulation. Then, we report the influence of 14 multi-purpose reservoirs66 including those managed for hydropower production on the WHI computed for sites on downstream waterways.\n\nTo explore possible factors contributing to WHI trends we assess whether the dispersion of flows across the seven days of the week is changing over time. Here, we first compile total annual flows (in m3 s\u22121) for each of the seven days of the week, as well as the overall average, over each calendar year. Then, we quantify departures (as a percentage) for each day of the week relative to the annual mean. Next, we calculate standard deviations (\u03c3) in the percentage departures for the seven days of the week each year, creating \u03c3 time series for all 500 sites over 1980\u20132019. Finally, application of the MKT on the \u03c3 time series (when ny\u2009\u2265\u200930 years) yields 1980\u20132019 dispersion trends.",
+            "section_image": []
+        },
+        {
+            "section_name": "Data availability",
+            "section_text": "Data related to this article can be found in the\u00a0Supplementary Data files. Discharge data used in this study are available in the following publicly accessible databases: Centre d\u2019Expertise Hydrique du Qu\u00e9bec (http://www.cehq.gouv.qc.ca/hydrometrie/historique_donnees/info_validite.htm), US Bureau of Reclamation (https://data.usbr.gov/), United States Geological Survey (https://waterdata.usgs.gov/nwis), Water Survey of Canada\u2019s Hydrometric Database (https://wateroffice.ec.gc.ca), and the International Boundary and Water Commission (https://www.ibwc.gov/Water_Data/). For some regulated rivers, proprietary or unpublished discharge data can be requested from the following data providers: BC Hydro, Evolugen, H2O Power, Hydro-Qu\u00e9bec, International Boundary and Water Commission, Manitoba Hydro, Nalcor Energy, NB Power, Ontario Power Generation, Rio Tinto, Tennessee Valley Authority, TransAlta, and USACE (see Supplementary Data\u00a03). Source data are provided with this paper.",
+            "section_image": []
+        },
+        {
+            "section_name": "Code availability",
+            "section_text": "The Fortran code used in this study is available online with an explanation at https://doi.org/10.5281/zenodo.5646458.",
+            "section_image": []
+        },
+        {
+            "section_name": "References",
+            "section_text": "International Hydropower Association. 2020 Hydropower Status Report, https://hydropower-assets.s3.eu-west-2.amazonaws.com/publications-docs/2020_hydropower_status_report.pdf (2020).\n\nGraf, W. L. Dam nation: A geographic census of American dams and their large-scale hydrologic impacts. Water Resour. Res. 35, 1305\u20131311 (1999).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nCouto, T. B. A. & Olden, J. D. Global proliferation of small hydropower plants: science and policy. Front. Ecol. Environ. 16, 91\u2013100 (2018).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nDynesius, M. & Nilsson, C. Fragmentation and flow regulation of river systems in the northern third of the world. Science 266, 753\u2013762 (1994).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGrill, G. et al. An index-based framework for assessing patterns and trends in river fragmentation and flow regulation by global dams at multiple scales. Environ. Res. Lett. 10, 015001 (2015).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGrill, G. et al. Mapping the world\u2019s free-flowing rivers. Nature 569, 215\u2013219 (2019).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMideksa, T. K. & Kallbekken, S. The impact of climate change on the electricity markets: A review. Energ. Policy 38, 3579\u20133585 (2010).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWang, C., Grozev, G. & Seo, S. Decomposition and statistical analysis for regional electricity demand forecasting. Energy 41, 313\u2013325 (2012).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nRosenberg, D. M. et al. Large-scale impacts of hydroelectric development. Environ. Rev. 5, 27\u201354 (1997).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nContosta, A. R., Casson, N. J., Nelson, S. J. & Garlick, S. Defining frigid winter illuminates its loss across seasonally snow-covered areas of eastern North America. Environ. Res. Lett. 15, 034020 (2020).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nPerkins-Kirkpatrick, S. E. & Lewis, S. C. Increasing trends in regional heat waves. Nat. Commun. 11, 3357 (2020).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBarthendu, S. & Cohen, S. J. Impact of CO2-induced climate change on residential heating and cooling energy requirements in Ontario, Canada. Energ. Build. 10, 99\u2013108 (1987).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nFilion, Y. Climate change: Implications for Canadian water resources and hydropower production. Can. Water Resour. J. 25, 255\u2013269 (2000).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSchaeffer, R. et al. Energy sector vulnerability to climate change: A review. Energy 38, 1\u201312 (2012).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nD\u00e9ry, S. J., Stadnyk, T. A., MacDonald, M. K., Koenig, K. A. & Guay, C. Flow alteration impacts on Hudson Bay river discharge. Hydrol. Process. 32, 3576\u20133587 (2018).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGreimel, F. et al. Hydropeaking Impacts and Mitigation. in Riverine Ecosystem Management (eds Schmutz, S. & Sendzimir, J.). ch 5 (Springer, Cham, 2018). https://doi.org/10.1007/978-3-319-73250-3_5.\n\nSmokorowski, K. E. The ups and downs of hydropeaking: a Canadian perspective on the need for, and ecological costs of, peaking hydropower production. Hydrobiologia https://doi.org/10.1007/s10750-020-04480-y (2021).\n\nBattala, R. J. et al. Hydropeaked rivers need attention. Environ. Res. Lett. 16, 021001 (2021).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBenke, A. C. & Cushing, C. E. (eds) Rivers of North America (Elsevier Academic Press, Amsterdam, 2005).\n\nShe, Y., Hicks, F. & Andrishak, R. The role of hydro-peaking in freeze-up consolidation events on regulated rivers. Cold Reg. Sci. Technol. 73, 41\u201349 (2012).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nPettitt, A. N. A non-parametric approach to the change point problem. J. Appl. Stat. 38, 126\u2013135 (1979).\n\nArticle\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWhite, M. A., Schmidt, J. C. & Topping, D. J. Application of wavelet analysis for monitoring the hydrologic effects of dam operation: Glen Canyon Dam and the Colorado River at Lees Ferry, Arizona. River Res. Appl. 21, 551\u2013565 (2005).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHorner, N. C., Shehabi, A. & Azevedo, I. Known unknowns: Indirect energy effects of information and communication technology. Environ. Res. Lett. 11, 103001 (2016).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSkuterud, M. The impact of Sunday shopping on employment and hours of work in the retail industry: Evidence from Canada. Eur. Economic Rev. 49, 1953\u20131978 (2005).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nAnderson, E. P. et al. Understanding rivers and their social relations: a critical step to advance environmental water management. Wiley Interdiscip. Rev.: Water 6, e1381 (2019).\n\n\n                    Google Scholar\u00a0\n                \n\nAuer, N. A. Response of spawning lake sturgeons to change in hydroelectric facility operation. T. Am. Fish. Soc. 125, 66\u201377 (1996).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKarier, T. Economics of the Columbia River Treaty. Electr. J. 33, 106731 (2020).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMatbouli, Y. T., Hipel, K. W. & Kilgour, D. M. Strategic analysis of the Great Canadian hydroelectric power conflict. Energy Strateg. Rev. 4, 43\u201351 (2014).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHydro-Qu\u00e9bec. Hydro-Qu\u00e9bec: North America\u2019s leading provider of clean energy, https://www.hydroquebec.com/clean-energy-provider/, (2020).\n\nVoisin, N. et al. Impact of climate change on water availability and its propagation through the western U.S. power grid. Appl. Energy 276, 115467 (2020).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBejarano, M. D., Jansson, R. & Nilsson, C. The effects of hydropeaking on riverine plants: a review. Biol. Rev. 93, 658\u2013673 (2018).\n\nArticle\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMoreira, M. et al. Ecologically based criteria for hydropeaking mitigation: A review. Sci. Total Environ. 657, 1508\u20131522 (2019).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSerreze, M. C. & Francis, J. A. The Arctic amplification debate. Climatic Change 76, 241\u2013265 (2006).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nPoff, N. L. & Schmidt, J. C. How dams can go with the flow. Science 353, 1099\u20131100 (2016).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHarvey-Lavoie, S., Cooke, S. J., Gu\u00e9nard, G. & Boisclair, D. Differences in movements of northern pike inhabiting rivers with contrasting flow regimes. Ecohydrology 9, 1687\u20131699 (2016).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nJones, N. E. & Petreman, I. C. Environmental influences on fish migration in a hydropeaking river. River Res. Appl. 31, 1109\u20131118 (2015).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKelly, B., Smokorowski, K. E. E. & Power, M. Downstream effects of hydroelectric dam operation on thermal habitat use by Brook Trout (Salvelinus fontinalis) and Slimy Sculpin (Cottus cognatus). Ecol. Freshw. Fish. 26, 552\u2013562 (2017).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nTaylor, M. K. et al. Reach-scale movements of bull trout (Salvelinus confluentus) relative to hydropeaking operations in the Columbia River, Canada.\u00a0Ecohydrology\u00a07, 1079\u20131086 (2014).\n\n\n                    Google Scholar\u00a0\n                \n\nHaghighi, A. T., Marttila, H. & Klove, B. Development of a new index to assess river regime impacts after dam construction. Glob. Planet. Change 122, 186\u2013196 (2014).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nJumani, S. et al. River fragmentation and flow alteration metrics: a review of methods and directions for future research. Environ. Res. Lett. 15, 123009 (2020).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMlynowski, T. J., Hern\u00e1ndez-Henr\u00edquez, M. A. & D\u00e9ry, S. J. An evaluation of hydrometric monitoring across the Canadian pan-Arctic region, 1950-2008. Hydrol. Res. 42, 479\u2013490 (2011).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGleason, C. J. & Smith, L. C. Toward global mapping of river discharge using satellite images and at-many-stations hydraulic geometry. Proc. Natl Acad. Sci. USA 111, 4788\u20134791 (2014).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nTongal, H. & Booij, M. J. Simulation and forecasting of streamflows using machine learning models coupled with base flow separation. J. Hydrol. 564, 266\u2013282 (2018).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nTefs, A. A. G. et al. Simulating river regulation and reservoir performance in a continental-scale hydrologic model. Environ. Modell. Softw. 141, 105025 (2021).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nD\u00e9ry, S. J. et al. Detection of runoff timing changes in pluvial, nival and glacial rivers of western Canada. Water Resour. Res. 45, W04426 (2009).\n\nADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSmith, L. C., Turcotte, D. L. & Isacks, B. L. Stream flow characterization and feature detection using a discrete wavelet transform. Hydrol. Process. 12, 233\u2013249 (1998).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nD\u00e9ry, S. J., Stadnyk, T. A., MacDonald, M. K. & Gauli-Sharma, B. Recent trends and variability in river discharge across northern Canada. Hydrol. Earth Syst. Sci. 20, 4801\u20134818 (2016).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nD\u00e9ry, S. J., Stieglitz, M., McKenna, E. C. & Wood, E. F. Characteristics and trends of river discharge into Hudson, James, and Ungava Bays, 1964\u20132000. J. Clim. 18, 2540\u20132557 (2005).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nPeters, D. L. & Buttle, J. M. The effects of flow regulation and climatic variability on obstructed drainage and reverse flow contribution in a northern river-lake-delta complex, Mackenzie Basin headwaters. River Res. Appl. 26, 1065\u20131089 (2010).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWoo, M., Thorne, R., Szeto, K. & Yang, D. Streamflow hydrology in the boreal region under the influence of climate and human interference. Philos. T. R. Soc. B. 363, 2251\u20132260 (2008).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nAssani, A. A., Stichelbout, E., Roy, A. G. & Petit, F. Comparison of impacts of dams on the annual maximum flow characteristics in three regulated hydrologic regimes in Qu\u00e9bec (Canada). Hydrol. Process. 20, 3485\u20133501 (2006).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nNaik, P. K. & Jay, D. A. Distinguishing human and climate influences on the Columbia River: changes in mean flow and sediment transport. J. Hydrol. 404, 259\u2013277 (2011).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSt. Jacques, J.-M., Sauchyn, D. J. & Zhao, Y. Northern Rocky Mountain streamflow records: global warming trends, human impacts or natural variability? Geophys. Res. Lett. 37, L06407 (2010).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nYe, B., Yang, D. & Kane, D. L. Changes in Lena River streamflow hydrology: human impacts versus natural variations. Water Resour. Res. 39, 1200 (2003).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nRichter, B. D., Baumgartner, J. V., Powell, J. & Braun, D. P. A method for assessing hydrologic alteration within ecosystems. Conserv. Biol. 10, 1163\u20131174 (1996).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nTimpe, K. & Kaplan, D. The changing hydrology of a dammed Amazon. Sci. Adv. 3, e1700611 (2017).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nZhou, X., Huang, X., Zhao, H. & Ma, K. Development of a revised method for indicators of hydrologic alteration for analyzing the cumulative impacts of cascading reservoirs on flow regime. Hydrol. Earth Syst. Sci. 24, 4091\u20134107 (2020).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nTongal, H., Demirel, M. C. & Moradkhani, H. Analysis of dam-induced patterns on river flow dynamics. Hydrol. Sci. J. 62, 626\u2013641 (2017).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nCarolli, M. et al. A simple procedure for the assessment of hydropeaking flow alterations applied to several European streams. Aquat. Sci. 77, 639\u2013653 (2015).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nAshraf, F. B. H. et al. Changes in short term river flow regulation and hydropeaking in Nordic rivers. Sci. Rep. 8, 17232 (2018).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWilks, D. S. Statistical Methods in the Atmospheric Sciences (Elsevier, Amsterdam, 4th edition, 2019).\n\nMann, H. B. Nonparametric test against trend. Econometrica 13, 245\u2013259 (1945).\n\nArticle\u00a0\n    MathSciNet\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKendall, M. G. Rank Correlation Methods (Charles Griffin, London, 1975).\n\nMcCuen, R. H. Modeling Hydrologic Change: Statistical Methods (Lewis Publishers, Boca Raton, 2003).\n\nTennessee Valley Authority. Hydroelectric, https://www.tva.com/energy/our-power-system/hydroelectric, (2021).\n\nFerrazi, M., Woods, R. A. & Botter, G. Climatic signatures in regulated flow regimes across the Central and Eastern United States. J. Hydrol. Reg. Stud. 35, 100809 (2021).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nDownload references",
+            "section_image": []
+        },
+        {
+            "section_name": "Acknowledgements",
+            "section_text": "Thanks to the Water Survey of Canada and its provincial and territorial partners, the Centre d\u2019Expertise Hydrique du Qu\u00e9bec, USGS, BC Hydro, Evolugen by Brookfield Renewable, TransAlta, Manitoba Hydro, Ontario Power Generation, H2O Power, Rio Tinto, Hydro-Qu\u00e9bec, NB Power, Nalcor Energy, the Tennessee Valley Authority, the International Boundary and Water Commission, USACE and the US Bureau of Reclamation for providing hydrometric data. Thanks to Aseem Sharma (UNBC/NRCan) for preparing the spatial plots, Clyde McLean and Joanna Barnard (Nalcor Energy), Jim Samms (NB Power), Marie Broesky, Kevin Gawne, Kristina Koenig, Phil Slota, Kevin Sydor, Efrem Teklemariam, Mike Vieira and Shane Wruth (Manitoba Hydro), Matt MacDonald (Ontario Power Generation), Erik Richards and Marc Mantha (H2O Power), Samer Alghabra and Mokhtar Moujahid (Hydro-Qu\u00e9bec), Bruno Larouche and Richard Loubier (Rio Tinto), Michael Smilski (TransAlta), Jim Li, Debbie Rinvold and Stephanie Smith (BC Hydro), Adrian Cortez and Delbert Humberson (International Boundary and Water Commission), Kelly Withers and Matti Hanninen (Evolugen), and Robert Dillingham (USACE) for providing comments on this work and for additional data for regulated rivers, Dwayne Akerman, Amber Brown, Michel Desjardins, Matt Falcone, Samantha Hussey, Lyssa Maurer, Angus Pippy, Melanie Taylor, and Frank Weber (Water Survey of Canada) for sharing supplemental hydrometric data, and Huilin Gao (Texas A&M), John Zhu (Texas Water Development Board), Rajtantra Lilhare (UNBC), Julie Th\u00e9riault (UQAM) and Mike Vieira and Kristina Koenig (Manitoba Hydro) for logistical support. This research was supported by the Natural Sciences and Engineering Research Council of Canada, Manitoba Hydro, and partners through funding of the BaySys project.",
+            "section_image": []
+        },
+        {
+            "section_name": "Author information",
+            "section_text": "Department of Geography, Earth and Environmental Sciences, University of Northern British Columbia, 3333 University Way, Prince George, BC, V2N 4Z9, Canada\n\nStephen J. D\u00e9ry\u00a0&\u00a0Marco A. Hern\u00e1ndez-Henr\u00edquez\n\nDepartment of Geography, University of Calgary, Calgary, AB, T2N 1N4, Canada\n\nTricia A. Stadnyk\n\nDepartment of Civil Engineering, University of Victoria, Victoria, BC, V8W 2Y2, Canada\n\nTara J. Troy\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nS.J.D. designed the study, extracted hydrometric data and constructed time series of daily discharge for all rivers, formulated the weekly hydropeaking index, developed the codes, performed the statistical and computational analyses, and drafted line graphs with support from M.A.H.H., T.A.S., and T.J.T. S.J.D. wrote the manuscript with contributions from all co-authors and all contributed to manuscript refinement and revisions.\n\nCorrespondence to\n                Stephen J. D\u00e9ry.",
+            "section_image": []
+        },
+        {
+            "section_name": "Ethics declarations",
+            "section_text": "The authors declare no competing interests.",
+            "section_image": []
+        },
+        {
+            "section_name": "Peer review information",
+            "section_text": "Nature Communications thanks Huilin Gao, Nathalie Voisin and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available.",
+            "section_image": []
+        },
+        {
+            "section_name": "Additional information",
+            "section_text": "Publisher\u2019s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.",
+            "section_image": []
+        },
+        {
+            "section_name": "Rights and permissions",
+            "section_text": "Open Access  This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article\u2019s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article\u2019s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.\n\nReprints and permissions",
+            "section_image": []
+        },
+        {
+            "section_name": "About this article",
+            "section_text": "D\u00e9ry, S.J., Hern\u00e1ndez-Henr\u00edquez, M.A., Stadnyk, T.A. et al. Vanishing weekly hydropeaking cycles in American and Canadian rivers.\n                    Nat Commun 12, 7154 (2021). https://doi.org/10.1038/s41467-021-27465-4\n\nDownload citation\n\nReceived: 31 March 2021\n\nAccepted: 22 November 2021\n\nPublished: 09 December 2021\n\nVersion of record: 09 December 2021\n\nDOI: https://doi.org/10.1038/s41467-021-27465-4\n\nAnyone you share the following link with will be able to read this content:\n\nSorry, a shareable link is not currently available for this article.\n\n\n\n\n                            Provided by the Springer Nature SharedIt content-sharing initiative\n                        ",
+            "section_image": [
+                "https://data:image/svg+xml;base64,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"
+            ]
+        },
+        {
+            "section_name": "This article is cited by",
+            "section_text": "Communications Earth & Environment (2024)\n\nCommunications Earth & Environment (2024)\n\nnpj Urban Sustainability (2022)",
+            "section_image": []
+        },
+        {
+            "section_name": "Associated content",
+            "section_text": "Collection",
+            "section_image": []
+        }
+    ]
+}

03e6375d91ba2acc0d93fa9c97f15a5154400d7aed2f4249391cfd96e1971e8e/metadata.json

@@ -0,0 +1,163 @@
+{
+    "title": "Ultra-rapid and highly efficient enrichment of organic pollutants via magnetic mesoporous nanosponge for ultrasensitive nanosensors",
+    "pre_title": "Ultra-rapid and highly efficient enrichment of organic pollutants via magnetic nanoparticles/mesoporous nanosponge compounds for ultrasensitive nanosensors",
+    "journal": "Nature Communications",
+    "published": "25 November 2021",
+    "supplementary_0": [
+        {
+            "label": "Supplementary Information",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-021-27100-2/MediaObjects/41467_2021_27100_MOESM1_ESM.docx"
+        },
+        {
+            "label": "Peer Review File",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-021-27100-2/MediaObjects/41467_2021_27100_MOESM2_ESM.pdf"
+        }
+    ],
+    "supplementary_1": NaN,
+    "supplementary_2": NaN,
+    "source_data": [],
+    "code": [],
+    "subject": [
+        "Nanophotonics and plasmonics",
+        "Raman spectroscopy"
+    ],
+    "license": "http://creativecommons.org/licenses/by/4.0/",
+    "preprint_pdf": "https://www.researchsquare.com/article/rs-127668/v1.pdf?c=1637932050000",
+    "research_square_link": "https://www.researchsquare.com//article/rs-127668/v1",
+    "nature_pdf": "https://www.nature.com/articles/s41467-021-27100-2.pdf",
+    "preprint_posted": "06 Jan, 2021",
+    "research_square_content": [
+        {
+            "section_name": "Abstract",
+            "section_text": "Developing advanced sensing and detection technologies to effectively monitor organic micropollutants in water is under urgent demand in both scientific and industrial communities. Currently, owing to the ultrahigh sensitivity on the single-molecule level with highly informative spectra characteristics, SERS technique is regarded as the most direct and effective detection technique. However, some weakly adsorbed molecules, such as most of persistent organic pollutants, cannot exhibit strong SERS signals, which is a long-standing key challenge that has not been solved. Here, we show an enrichment-typed sensing strategy based on a powerful porous composite material, call mesoporous nanosponge. The nanosponge consists of magnetic nanoparticles immobilized porous \u03b2-cyclodextrin polymers, demonstrating remarkable capability of effective and fast removal of organic micropollutants, e.g. ~90% removal efficiency within ~1 min. With the anchoring of magnetic nanoparticles, the current new polymer adsorbent can be easily recycled from water and re-dispersed in ethanol so that the target molecules in the cavity of adsorbent is concentrated, with an enrichment factor up to ~103. By means of the current enrichment strategy, the limit of detection (LOD) of the typical organic pollutants can be significantly improved, i.e. increasing 2~3 orders of magnitude, compared with the detection without molecule enrichment protocol. Consequently, the current enrichment strategy is proved to be applicable in a variety of fields for portable and fast detection, such as Raman and fluorescent.Photonics/opticsOptics/LasersOptical Materials and DevicesnanosensorsSERS technique",
+            "section_image": []
+        },
+        {
+            "section_name": "Figures",
+            "section_text": "Figure 1Figure 2Figure 3Figure 4Figure 5",
+            "section_image": [
+                "https://assets-eu.researchsquare.com/files/rs-127668/v1/00cb65318e85e4ad15aa5798.jpg%3FmaxDims%3D150x150&w=256&q=75.png",
+                "https://assets-eu.researchsquare.com/files/rs-127668/v1/1b0a60fdd093f351e3c7345b.jpg%3FmaxDims%3D150x150&w=256&q=75.png",
+                "https://assets-eu.researchsquare.com/files/rs-127668/v1/f67253d177f9dfbc4bc5a11a.jpg%3FmaxDims%3D150x150&w=256&q=75.png",
+                "https://assets-eu.researchsquare.com/files/rs-127668/v1/ef1614b14d3fa9c3b5fe0c39.jpg%3FmaxDims%3D150x150&w=256&q=75.png",
+                "https://assets-eu.researchsquare.com/files/rs-127668/v1/dd97a2809188c77f1ca3ff18.jpg%3FmaxDims%3D150x150&w=256&q=75.png"
+            ]
+        },
+        {
+            "section_name": "Introduction",
+            "section_text": "The Stockholm Convention on Persistent Organic Pollutants (POPs) was endorsed by 131 nations in 2004 to eliminate most persistent bioaccumulative and toxic substances in the world.1 Organic micropollutants of ground and surface water resources, such as pesticides and plastic components, have aroused great concerns about potential negative effects on aquatic ecosystems and human health.2, 3 Therefore, parallel to the researches of adsorbent materials to remove organic pollutants from water, the ultrasensitive detection of organic pollutants is another crucial field, since the solubility of organic micropollutants in water is always at the trace level.4 Among diverse detection approaches, surface-enhanced Raman scattering (SERS), which achieved significant breakthroughs in 1997 and became the first vibrational spectroscopy technique that could provide delicate information on molecular fingerprints with potential single-molecule level of sensitivity,5-8 is regarded as the most simple, fast, flexible and portable detection technique.\nHowever, up to date, the superiority of single-molecule SERS in the detection of diverse molecules with intrinsic small cross-sections or low affinity for the plasmonic surface has not been into full play, particularly in real complex situation.9 As is well known, the SERS process involves complicated coupled three-body interactions among photons, molecules, and nanostructures.10, 11 Besides the interaction between light and nanostructures, the investigation on the interaction between molecule and plasmonic surface is of importance.12 On the one hand, SERS is an optical near-field effect.12-14 A high activity can be obtained only when the target molecule is very close to the plasmonic surface. On the other hand, most organic pollutants in water can not be effectively adsorbed onto a metallic surface because of their low affinity toward the metal.7 Therefore, recently, some strategies, such as selective enrichment and spatial localization of target molecules,15-17 are suggested to solve this long-standing challenge.\nIn this work, we propose a new sensing strategy based on the efficient enrichment and rapid separation of POPs by means of the magnetic nanoparticles immobilized porous \u03b2-CD polymer (MN-PCDP), called mesoporous nanosponge. The microporous \u03b2-cyclodextrin (\u03b2-CD) material, an inexpensive and renewable carbohydrate, which is featured by small pores and high surface areas,18, 19 was used in this work as an excellent adsorbent. In fact, microporous \u03b2-CD material has been widely studied because of outstanding adsorption efficiency through forming host-guest inclusion with many hydrophobic organic pollutants.20, 21 The magnetic nanoparticles are introduced into the MN-PCDP compounds to rapidly separate the adsorbent from water. The current strategy (the schematic description of the protocol is shown in Fig. 1) demonstrates several remarkable advantages. Firstly, as shown in Fig. 1a, when the MN-PCDP adsorbent (shown in Fig. 1b) is dispersed into water in the beaker, e.g. ~1000 ml, containing organic pollutants, ultra-rapid adsorption and magnetic separation can be accomplished, i.e., totally within ~ 1 min. Secondly, the adsorbed pollutant in MN-PCDP from water can be desorbed in ethanol with a volume of ~1 ml, for further analysis such as UV-vis, Raman and fluorescent spectroscopy. Thus, an ultra-high enrichment efficiency with an enrichment factor up to \u223c103 times can be obtained (Fig. 1c). With current enrichment strategy, the limit of detection (LOD) in a variety of sensing applications, such as SERS and fluorescent, can be lowered by 2~3 orders of magnitude. Furthermore, through the magnetic separation, the MN-PCDP mesoporous nanosponge can selectively adsorb the target organic pollutants, avoiding the disturbance of complex matrix. The current sensing strategy can be believed to be applicable to a wider range of sensing areas for an economical, simple, fast, flexible, and portable detection.",
+            "section_image": []
+        },
+        {
+            "section_name": "Results",
+            "section_text": "Synthesis and characterization. The MN-PCDP was prepared by cross-linking polymerization of \u03b2-CD and cross-linking agent (tetrafluoroterephthalonitrile (TFT)), with magnetic nanoparticles (Fe3O4) in one-step solvothermal reaction. Fig. 2a-c show the transmission electron microscope (TEM) images of magnetic nanoparticles (MN, Fe3O4), porous \u03b2-CD polymer (PCDP) and MN-PCDP, respectively. As shown in Fig. 2a, the synthesized MN exhibits regular spheres with good dispersibility and uniform size (average size ~200 nm). The Fourier transform-infrared spectroscopy (FT-IR) spectrum of MN is displayed in Fig. 2d. The absorption bands at 1652 cm-1 and 1396 cm-1 of the MN can be associated with carboxylate group22 and that also appear in the MN-PCDP. Fig. 2b and Supplementary Fig. 1 exhibit that the PCDP is porous network structure. After the immobilization of MN, as shown in Fig. 2c, the porous network structure of MN-PCDP is not disrupted. The FT-IR spectrum of the MN-PCDP not only obviously combines the characteristic peaks of the TFT and the \u03b2-CD but also displays a new peak at 1265 cm-1 in relation to the newly formed C-F group, implying that the \u03b2-CD has been crosslinked with TFT.23, 24 Fig. 2e indicates that the Brunauer-Emmett-Teller surface areas (SBET) of MN-PCDP is about 66 m2 g-1. The pores with diameter of 1.7-3.0 nm comprise the majority of the free volume of MN-PCDP and its average pore diameter is 2.12 nm.\nAdsorption of MN-PCDP nanosponges. The high surface area and permanent porosity of MN-PCDP mesoporous nanosponge enable the rapid removal of organic micropollutants from water.25 As shown in Supplementary Fig. 2, the PCDP and MN-PCDP displays the same properties in time-dependent adsorptions of bisphenol A (BPA), illustrating the immobilization of magnetic nanoparticles has no remarkable influence on the adsorption performance of PCDP. The time-dependent adsorptions of various organic micropollutants adsorbed by MN-PCDP, including plastic components, pesticide and aromatic model compounds (Fig. 3a), are shown in Fig. 3b, Supplementary Fig. 3 and Supplementary Table S1. The removal rate of the above organic micropollutants is very fast, which tends to be constant within 1 min. The removal efficiencies of BPA, parathion, carbendazim and 2-naphthol (2-NO) are more than 80% in 30 sec, which is much higher than the Norit ROW 0.8 supra extruded activated carbon (NAC) as presented in Fig. 3c, Supplementary Fig. 4-5 and Supplementary Table S2. We further probe the readily accessible binding sites of MN-PCDP by determining the flow-through uptake of different organic micropollutants. In these experiments, the adsorbent (~5 mg) was trapped as a thin layer on a 0.22 \u03bcm syringe filter, and aqueous organic pollutants (5 ml, 0.1 mM) passed rapidly through the filter at a flow rate of 10 ml min-1 (Supplementary Fig. 6). Under these conditions, for example, 76% of the BPA is removed from the solution, corresponding to more than 84% of its equilibrium adsorption, confirming that the host-guest interaction plays a major role in the filtration process by syringe.26 The superior performance of MN-PCDP can be further indicated that its \u03b2-CD moieties are easily accessed by most of organic micropollutants. Furthermore, these molecules are rapidly trapped. In addition, the influence of the concentrations of adsorbent on the BPA adsorption efficiency is studied as shown in Fig. 4b, Supplementary Fig. 7 and Supplementary Table S3. When the concentration of adsorbent increases from 0.1 mg L-1 to 1.0 mg L-1, the adsorption efficiency of BPA is enhanced from 25.12% to 87.09% within 1 min and from 35.07% to 89.82% within 10 min.\nDesorption and enrichment of MN-PCDP nanosponges. As we all know, organic micropollutants exhibit good solubility in organic solution, such as ethanol and methanol.18 Hence, after adsorption process, we utilized ethanol to desorb the organic micropollutants from MN-PCDP mesoporous nanosponges, and then obtained the enriched pollutant solution through magnetic nanoparticle separation. In order to obtain higher concentration of desorbed micropollutant solution, in this work, we chose 1 mL ethanol to desorb organic micropollutants adsorbed in 100 mg MN-PCDP adsorbent. As shown in Supplementary Fig. 8, using current enrichment processes, the concentration of BPA can be increased to 88.5 times of its initial concentration with a recipe of 100 mL organic pollutant (BPA) solution and 100 mg MN-PCDP adsorbent. This result reveals that more than 98% of the adsorbed organic micropollutants are desorbed into the ethanol solution. That is to say, for 100 mL organic pollutant solution, we realize ~102 times enrichment of target molecule. As the amount of adsorbent increases, the adsorption efficiency tends to reach equilibrium. Considering the cost increase of sample preparation and the operation in the desorption process (with 1 mL ethanol) resulting from the increase of adsorbent dosage, 100 mg of adsorbent is selected as the amount of material for subsequent experiments.\nIn order to further improve the enrichment effect of 100 mg adsorbent in total 1000 mL organic micropollutants, herein, we attempted three methods, including 100 mL\u00d710 times, 250 mL\u00d74 times and 500 mL\u00d72 times, to optimize the adsorption and desorption processes. Importantly, the adsorbent can be simply separated by magnet in every adsorption cycle, and desorbed in ethanol in the last adsorption cycle. As shown in Fig. 4c, Supplementary Fig. 9 and Supplementary Table S4, with adsorption times increasing, the removal efficiencies of these three methods gradually decrease. The removal efficiencies of these three methods (100 mL\u00d710 times, 250 mL\u00d74 times and 500 mL\u00d72 times) are 50.78%, 62.58% and 41.22%, respectively. Meanwhile, the enrichment efficiencies of these three methods are 485, 605 and 396 times of the initial concentration (Fig. 4d), respectively. Therefore, we achieve over 600 times\u2019 enrichment of organic pollutants (with 1000 mL of initial micropollutants) in the optimized adsorption and desorption processes. Here, the selected parameters (100 mg adsorbent in 250 mL organic solution for 4 cycle times) were used for the succedent experiments. Meanwhile, it is also worth pointing out that the separation process by magnet is very fast and facile, such as 90 sec for 250 mL solution (Fig. 4a and Supplementary Fig. 10b), 60 sec for 100 mL (Supplementary Fig. 10a), and 150 sec for in 500 mL (Supplementary Fig. 10c), fully meeting the requirement of immediate-pretreatment detection application.\nSERS and fluorescence measurement of pollutants based on the enrichment of MN-PCDP nanosponges. Many POPs are mutagenic, carcinogenic and not degradable by direct biological treatment, some of which damage nerve, endocrine systems of human body, and the ecological balance due to their toxicity in nature.27 Based on the above experimental researches, the MN-PCDP nanosponges were used to absorb organic micropollutants and then were collected from water (Fig. 5a), so as to realize the rapid removal and enrichment of POPs. Moreover, we evaluated fluorescence and enhanced Raman spectra of POPs (carbendazim and BPA) to demonstrate the enrichment effect of current enrichment strategy. The enhanced Raman spectra of carbendazim molecular with ~55 nm Au nanoparticles (Supplementary Fig. 11) were measured under 785 nm laser. As shown in Figure 5b-c, without the enrichment process, the LOD of SERS for carbendazim is 1 nM, but after the enrichment using MN-PCDP adsorbent, this value reaches to ~5 pM, which shows an increase of 102~103. In Supplementary Fig. 12, for BPA molecule, with the help of adsorbent, the LOD of fluorescence is also greatly improved. In Fig. 5d-e and Supplementary Fig. 13, based on the current enrichment sensing strategy, the LODs of fluorescence detections for the pure solution of carbendazim and BPA are lower by 2~3 orders of magnitude. In this study, the enrichment strategy based on the adsorption and desorption processes of MN-PCDP adsorbent may significantly increase the sensitivity of plasmonic sensors, compared with the LOD for similar molecules,28-30 illustrating its wide applicability.\nOwing to the excellent enrichment and easily-separated features, the current strategy was believed that the mesoporous nanosponges could be served as a preprocessing for direct, rapid and ultrasensitive detection of contaminants in complex situations. After adsorption process, the MN-PCDP adsorbent was easily collected on the wall of beaker (Fig. 5a) with a magnet, avoiding the interference of complex matrix, such as mud and microorganism. Fig. 5f reveals that both the characteristic peaks of BPA (830 and 1179 cm-1) and carbendazim (1008, 1244 and 1263 cm-1) evidently appear in the Raman spectrum of mixture solution, including 1 \u00b5M BPA and 10 nM carbendazim. Furthermore, the MN-PCDP demonstrates a superior reusability as shown in Supplementary Fig. 14. Six consecutive BPA adsorption/desorption cycles were performed and the regenerated MN-PCDP exhibited almost no decrease (90.2% to 87.5%) in performance compared to the as-synthesized polymer.",
+            "section_image": []
+        },
+        {
+            "section_name": "Discussion",
+            "section_text": "In summary, we have developed a robust and rapid sensing strategy based on the MN-PCDP mesoporous nanoponges to capture and enrich organic pollutants from water. In this strategy, the MN-PCDP adsorbent exhibits excellent adsorption capacity for various kinds of pollutants owing to the unique cavity structures. Moreover, the adsorbed pollutant in MN-PCDP can be desorbed in ethanol with a very fast and facile operation. In SERS detection of organic pollutants, i.e. carbendazim and BPA, in this work, the current sensing strategy may significantly increase the sensitivity of plasmonic sensors with 2~3 orders of magnitude. Therefore, the current robust sensing strategy with the ultra-rapid and highly efficient sample pretreatment and molecule enrichment is believed to be applicable to a wider range of sensing devices, such as fluorescent, Raman and infrared spectroscopes for a cost-effective, simple, fast, flexible and portable detection.",
+            "section_image": []
+        },
+        {
+            "section_name": "Declarations",
+            "section_text": "Data availability\nThe data that support the findings of this study are available within the paper and its Supplementary Information or from the corresponding authors on reasonable request.\nAcknowledgements\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 \nThis work was supported by the programs supported by the National Natural Science Foundation of China (No. 21675122, 21874104, 22074115), the Key Research Program in Shaanxi (2017NY-114), Basic Public Welfare Research Project of Zhejiang Province (No. LY20E010007), and Natural Science Foundation of Shaanxi Province (No. 2019JLP-19), the World-Class Universities (Disciplines) and the Characteristic Development Guidance Funds for the Central Universities.\nAuthor contributions\nL.L.Z. synthesized the materials, carried out the characterizations and performance, analyzed the data, and wrote the manuscript. R. H., H. N., Y. Z. D. contributed in part of the TEM, Raman and fluorescence characterizations. H.J.Y. and J.X.F., supervised the project, designed the experiments, contributed in discussions, comments and writing of manuscript. All authors discussed the results and commented on the manuscript.\nAdditional information\nSupplementary Information accompanies this paper at\ncompeting of Interest: The authors declare no competing of interest.\nReprints and permission information is available online at\nJournal peer review information:\nPublisher\u2019s note: Springer Nature remains neutral with regard to jurisdictional claims in\npublished maps and institutional affiliations.",
+            "section_image": []
+        },
+        {
+            "section_name": "References",
+            "section_text": "1\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 Kelly, B. C., Ikonomou, M. G., Blair, J. D., Morin, A. E. & F. A. P. C. Gobas. Food web-specific biomagnification of persistent organic pollutants. Science 317, 236-239 (2007).\n2\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 Schwarzenbach, R. P. et al. The challenge of micropollutants in aquatic systems. Science 313, 1072-1077, doi:10.1126/science.1127291 (2006).\n3\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 Jones, K. C. & Voogt, P. De. Persistent organic pollutants (POPs): state of the science. Environ. Pollut. 100, 209-221 (1999).\n4\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 Pi, Y. et al. Adsorptive and photocatalytic removal of Persistent Organic Pollutants (POPs) in water by metal-organic frameworks (MOFs). Chem. Eng. J 337, 351-371, doi:10.1016/j.cej.2017.12.092 (2018).\n5\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 Xu, H., Bjerneld, E. J., K\u00e4ll, M. & B\u00f6rjesson, L. Spectroscopy of Single Hemoglobin Molecules by Surface Enhanced Raman Scattering. Phys. Rev. Lett. 83, 4357-4360, doi:10.1103/PhysRevLett.83.4357 (1999).\n6\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 Xu, H., Aizpurua, J., Kall, M. & Apell, P. Electromagnetic contributions to single-molecule sensitivity in surface-enhanced raman scattering. Phys. Rev. E 62, 4318-4324, doi:10.1103/physreve.62.4318 (2000).\n7\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 Pieczonka, N. P. & Aroca, R. F. Single molecule analysis by surfaced-enhanced Raman scattering. Chem. Soc. Rev. 37, 946-954, doi:10.1039/b709739p (2008).\n8\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 Li, J. F. et al. Shell-isolated nanoparticle-enhanced Raman spectroscopy. Nature 464, 392-395, doi:10.1038/nature08907 (2010).\n9\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 Panneerselvam, R. et al. Surface-enhanced Raman spectroscopy: bottlenecks and future directions. Chem. Commun. 54, 10-25, doi:10.1039/c7cc05979e (2017).\n10\u00a0\u00a0\u00a0\u00a0\u00a0 Otto, A., Mrozek, I., Grabhorn, H. & Akemann, W. J. J. o. P. C. M. Surface-enhanced Raman scattering. J. Phys-Condens. Matter. 4, 1143-1212 (1992).\n11\u00a0\u00a0\u00a0\u00a0\u00a0 Schatz & George, C. Theoretical studies of surface enhanced Raman scattering. Acc. Chem. Res 17, 370-376 (1984).\n12\u00a0\u00a0\u00a0\u00a0\u00a0 Schlucker, S. Surface-enhanced Raman spectroscopy: concepts and chemical applications. Angew Chem Int Ed Engl 53, 4756-4795, doi:10.1002/anie.201205748 (2014).\n13\u00a0\u00a0\u00a0\u00a0\u00a0 Moskovits, M. Surface-enhanced Raman spectroscopy: a brief retrospective. J. Raman Spectrosc. 36, 485-496 (2005).\n14\u00a0\u00a0\u00a0\u00a0\u00a0 Zeman, E. J. & Schatz, G. C. An accurate electromagnetic theory study of surface enhancement factors for silver, gold, copper, lithium, sodium, aluminum, gallium, indium, zinc, and cadmium. J. Phys. Chem. 91, 634-643 (1987).\n15\u00a0\u00a0\u00a0\u00a0\u00a0 Zhang, D. et al. Buoyant particulate strategy for few-to-single particle-based plasmonic enhanced nanosensors. Nat. Commun. 11, 2603, doi:10.1038/s41467-020-16329-y (2020).\n16\u00a0\u00a0\u00a0\u00a0\u00a0 Hao, R., You, H., Zhu, J., Chen, T. & Fang, J. \"Burning Lamp\"-like Robust Molecular Enrichment for Ultrasensitive Plasmonic Nanosensors. ACS Sens. 5, 781-788, doi:10.1021/acssensors.9b02423 (2020).\n17\u00a0\u00a0\u00a0\u00a0\u00a0 De Angelis, F. et al. Breaking the diffusion limit with super-hydrophobic delivery of molecules to plasmonic nanofocusing SERS structures. Nat. Photonics 5, 682-687, doi:10.1038/nphoton.2011.222 (2011).\n18\u00a0\u00a0\u00a0\u00a0\u00a0 Alsbaiee, A. et al. Rapid removal of organic micropollutants from water by a porous beta-cyclodextrin polymer. Nature 529, 190-194, doi:10.1038/nature16185 (2016).\n19\u00a0\u00a0\u00a0\u00a0\u00a0 Xiao, L. et al. Beta-Cyclodextrin Polymer Network Sequesters Perfluorooctanoic Acid at Environmentally Relevant Concentrations. J. Am. Chem. Soc. 139, 7689-7692, doi:10.1021/jacs.7b02381 (2017).\n20\u00a0\u00a0\u00a0\u00a0\u00a0 Liu, X. et al. A magnetic graphene hybrid functionalized with beta-cyclodextrins for fast and efficient removal of organic dyes. J. Mater. Chem. A 2, doi:10.1039/c4ta00753k (2014).\n21\u00a0\u00a0\u00a0\u00a0\u00a0 Huang, D. et al. Fabrication of water-compatible molecularly imprinted polymer based on \u03b2-cyclodextrin modified magnetic chitosan and its application for selective removal of bisphenol A from aqueous solution. J. Taiwan Inst. Chem. E 77, 113-121, doi:10.1016/j.jtice.2017.04.030 (2017).\n22\u00a0\u00a0\u00a0\u00a0\u00a0 Liu, J. et al. Highly water-dispersible biocompatible magnetite particles with low cytotoxicity stabilized by citrate groups. Angew. Chem. Int. Ed. 48, 5875-5879, doi:10.1002/anie.200901566 (2009).\n23\u00a0\u00a0\u00a0\u00a0\u00a0 Jiang, H.-L. et al. A novel crosslinked \u03b2-cyclodextrin-based polymer for removing methylene blue from water with high efficiency. Colloid. Surfaces. A 560, 59-68, doi:10.1016/j.colsurfa.2018.10.004 (2019).\n24\u00a0\u00a0\u00a0\u00a0\u00a0 Huang, D., Zhang, Y. & Zhang, H. A novel synthesis of ethyl carbonate derivatives of beta-cyclodextrin. Carbohyd. Res. 370, 82-85, doi:10.1016/j.carres.2013.01.022 (2013).\n25\u00a0\u00a0\u00a0\u00a0\u00a0 Huang, Q., Chai, K., Zhou, L. & Ji, H. A phenyl-rich \u03b2-cyclodextrin porous crosslinked polymer for efficient removal of aromatic pollutants: Insight into adsorption performance and mechanism. Chem. Eng. J 387, doi:10.1016/j.cej.2020.124020 (2020).\n26\u00a0\u00a0\u00a0\u00a0\u00a0 Wang, Z., Zhang, P., Hu, F., Zhao, Y. & Zhu, L. A crosslinked beta-cyclodextrin polymer used for rapid removal of a broad-spectrum of organic micropollutants from water. Carbohyd. Polym. 177, 224-231, doi:10.1016/j.carbpol.2017.08.059 (2017).\n27\u00a0\u00a0\u00a0\u00a0\u00a0 Rusiecki, J. A. et al. Global DNA hypomethylation is associated with high serum-persistent organic pollutants in Greenlandic Inuit. Environ. Health. Persp. 116, 1547-1552, doi:10.1289/ehp.11338 (2008).\n28\u00a0\u00a0\u00a0\u00a0\u00a0 Chen, X. et al. Detection and quantification of carbendazim in Oolong tea by surface-enhanced Raman spectroscopy and gold nanoparticle substrates. Food Chem 293, 271-277, doi:10.1016/j.foodchem.2019.04.085 (2019).\n29\u00a0\u00a0\u00a0\u00a0\u00a0 Zhu, X. et al. A novel graphene-like titanium carbide MXene/Au\u2013Ag nanoshuttles bifunctional nanosensor for electrochemical and SERS intelligent analysis of ultra-trace carbendazim coupled with machine learning. Ceram. Int. doi:10.1016/j.ceramint.2020.08.121 (2020).\n30\u00a0\u00a0\u00a0\u00a0\u00a0 Zhai, Y. et al. Metal-organic-frameworks-enforced surface enhanced Raman scattering chip for elevating detection sensitivity of carbendazim in seawater. Sensors Actuat. B-Chem. 326, doi:10.1016/j.snb.2020.128852 (2021).",
+            "section_image": []
+        },
+        {
+            "section_name": "Additional Declarations",
+            "section_text": "There is NO Competing Interest.",
+            "section_image": []
+        },
+        {
+            "section_name": "Supplementary Files",
+            "section_text": "SupportingInformationNC.docx",
+            "section_image": []
+        }
+    ],
+    "nature_content": [
+        {
+            "section_name": "Abstract",
+            "section_text": "Currently, owing to the single-molecule-level sensitivity and highly informative spectroscopic characteristics, surface-enhanced Raman scattering (SERS) is regarded as the most direct and effective detection technique. However, SERS still faces several challenges in its practical applications, such as the complex matrix interferences, and low sensitivity to the molecules of intrinsic small cross-sections or weak affinity to the surface of metals. Here, we show an enrichment-typed sensing strategy with both excellent selectivity and ultrahigh detection sensitivity based on a powerful porous composite material, called mesoporous nanosponge. The nanosponge consists of porous \u03b2-cyclodextrin polymers immobilized with magnetic NPs, demonstrating remarkable capability of effective and fast removal of organic micropollutants, e.g., ~90% removal efficiency within ~1\u2009min, and an enrichment factor up to ~103. By means of this current enrichment strategy, the limit of detection for typical organic pollutants can be significantly improved by 2~3 orders of magnitude. Consequently, the current enrichment strategy is proved to be applicable in a variety of fields for portable and fast detection, such as Raman and fluorescent sensing.",
+            "section_image": []
+        },
+        {
+            "section_name": "Introduction",
+            "section_text": "The Stockholm Convention on Persistent Organic Pollutants (POPs) was endorsed by 131 nations in 2004 to eliminate the most persistent bioaccumulative and toxic substances in the world1. Organic micropollutants of ground and surface water resources, such as pesticides and plastic components, have aroused great concerns about potential negative effects on aquatic ecosystems and human health2,3. Therefore, parallel to the researches of adsorbent materials to remove organic pollutants from water, the ultrasensitive detection of organic pollutants is another crucial field, since the solubility of organic micropollutants in water is always at the trace level4. Among diverse detection approaches, surface-enhanced Raman spectroscopy (SERS), achieved breakthroughs in 1997 and became the first vibrational spectroscopy technique that could provide delicate information on molecular fingerprints with a potential of single-molecule sensitivity5,6,7,8,9, thus regarded as the most simple, fast, flexible, and portable detection technique.\n\nHowever, it seems surprising that after fifty decades, SERS has not yet been widely used in practical applications10. This is owing to the fact that, besides the stability of SERS substrates and reproducibility of spot-to-spot, SERS still faces two major bottlenecks in the commercial market. The first is the low detection sensitivity to the molecules of intrinsic small cross-sections or weak affinity to metal surface. As we know, SERS is an optical near-field effect11,12,13. A high activity can be obtained only when the target molecule is very close to the plasmonic surface. Therefore, researchers exploited diverse approaches to capture the target molecules onto the metal surface by means of antibodies, aptamers, ion liangs, et al.14,15. However, the ultra-rapid capture to meet on-site and portable detection remains a challenge. The second is the interference from the complex matrices16. In real-sample detection, most organic pollutants in water cannot be effectively adsorbed onto the metallic surface because of their low affinity toward the metal, hence the metal surface is usually inactivated due to unspecific adsorption by the interference from matrix molecules in the surrounding environment. Thus, for commercial applications, an ultra-fast and effective pretreatment is of importance to eliminate the most matrix interference. Therefore, recently, some strategies, e.g., selective separation, concentration, enrichment from complex matrix, and spatial localization of target molecules17,18,19, are suggested to solve this long-standing challenge.\n\nIn this work, we propose a new sensing strategy in rapid separation and highly efficient enrichment of POPs from complicated real-sample matrices by means of the magnetic NPs immobilized porous \u03b2-CD polymer (MN-PCDP), called mesoporous nanosponge. The current strategy (the schematic description of the protocol is shown in Fig.\u00a01) demonstrates several remarkable advantages. Firstly, specific and selective absorption and separation of target molecules eliminate the matrix interference. When the MN-PCDP adsorbent is dispersed into the water in the beaker containing organic pollutants and impurities (shown in Fig.\u00a01a, b), specific and selective adsorption of target molecules can be achieved. In fact, microporous \u03b2-CD material has been widely studied because of its outstanding adsorption efficiency through forming host-guest inclusion with many hydrophobic organic pollutants20,21. The magnetic NPs are introduced into the MN-PCDP compounds to rapidly separate the adsorbent from water, thus eliminates the interference of unspecific adsorption from the complex matrix in the real-samples. Secondly, highly efficient concentration and enrichment capability thus ultra-high detection sensitivity can be obtained. The MN-PCDP adsorbent demonstrates a remarkable removal efficiency on organic micropollutants, e.g., ~90% (relative standard deviation, RSD\u2009<\u20091%). Meanwhile, the adsorbed pollutants from initial water of ~1000\u2009ml can be desorbed in ethanol with a volume of ~1\u2009ml (Fig.\u00a01c), for further analysis such as UV-vis, Raman, and fluorescent spectroscopy. Thus, an ultra-high enrichment efficiency with an enrichment factor up to \u223c103 times can be obtained (RSD\u2009<\u20095%), and the limit of detection (LOD) in a variety of sensing applications can be lowered by 2\u22123 orders of magnitude. Thirdly, ultra-quick enrichment processes thus on-site portable detection can be realized. In the current strategy, ultra-fast adsorption, magnetic separation, and desorption can be accomplished, i.e., totally within 2\u22123\u2009min. Thus, the current sensing strategy can be believed to be applicable to a wider range of sensing areas for an economical, simple, fast, flexible, and portable detection.\n\na Adsorption and c desorption processes using magnetic NPs immobilized porous \u03b2-CD polymer (MN-PCDP) with ~1000 times enrichment. b Optical photograph and TEM imagine of MN-PCDP.",
+            "section_image": [
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-021-27100-2/MediaObjects/41467_2021_27100_Fig1_HTML.png"
+            ]
+        },
+        {
+            "section_name": "Results",
+            "section_text": "The microporous MN-PCDP material, an inexpensive and renewable carbohydrate, which is featured by small pores and high surface areas, is used in this work as an excellent adsorbent. The MN-PCDP is prepared by cross-linking polymerization of \u03b2-CD and cross-linking agent (tetrafluoroterephthalonitrile (TFT)), with magnetic NPs (Fe3O4) in one-step solvothermal reaction. Supplementary Fig.\u00a01a\u2212c shows the transmission electron microscope (TEM) images of magnetic NPs (MN, Fe3O4), porous \u03b2-CD polymer (PCDP), and MN-PCDP, respectively. As shown in Supplementary Fig.\u00a01a, the synthesized MN exhibits regular spheres with good dispersibility and uniform size (average size ~200\u2009nm). Supplementary Fig.\u00a01b exhibits that the PCDP is a porous network structure. After the immobilization of MN, as shown in Supplementary Fig.\u00a01c, the porous network structure of MN-PCDP is not disrupted. The Fourier transform-infrared spectroscopy (FT-IR) spectra of MN, \u03b2-CD, TFT, PCDP, MN-PCDP are displayed in Supplementary Fig.\u00a01d. The absorption bands at 1652 and 1396\u2009cm\u22121 of the MN can be associated with carboxylate group22. The FT-IR spectrum of the MN-PCDP obviously combines the characteristic peaks of the TFT and the \u03b2-CD. The signal intensity of absorption peak at 1265\u2009cm\u22121 in relation to C-F stretching vibration is weaker than that in TFT owing to the partial replacement of F23,24, implying that the \u03b2-CD has been crosslinked with TFT (Supplementary Fig.\u00a02). Supplementary Figure\u00a01e indicates that the Brunauer\u2212Emmett\u2212Teller surface areas (SBET) of MN-PCDP is about 66 m2 g\u22121. The pores with diameter of 1.7\u22123.0\u2009nm comprise the majority of the free volume of MN-PCDP and its average pore diameter is ~2.12\u2009nm.\n\nThe high surface area and permanent porosity of MN-PCDP mesoporous nanosponge enable the rapid removal of organic micropollutants from water25. As shown in Supplementary Fig.\u00a03, the PCDP and MN-PCDP display almost the same properties in time-dependent adsorptions of bisphenol A (BPA), revealing the immobilization of magnetic NPs has no remarkable influence on the adsorption performance of PCDP. The time-dependent adsorptions of various organic micropollutants adsorbed by MN-PCDP, including plastic components, pesticide, and aromatic model compounds (Fig.\u00a02a), are shown in Fig.\u00a02b, Supplementary Fig.\u00a04, and Supplementary Table\u00a01. Each organic micropollutant is rapidly removed, reaching ~95% of its equilibrium uptake in 10 s20. The removal efficiencies of BPA, parathion, carbendazim, and 2-naphthol (2-NO) by MN-PCDP are more than 80% in 30\u2009s, which is much higher than the Norit ROW 0.8 supra extruded activated carbon (NAC) as presented in Fig.\u00a02c, Supplementary Fig.\u00a05, 6 and Supplementary Table\u00a02. In this work, different cross-linking agent, e.g., epichlorohydrin (EPI) is compared (Supplementary Fig.\u00a07). As shown in Supplementary Fig.\u00a08, the removal efficiency of BPA by MNEPI-CDP in 1\u2009min is 19.5%, which is much lower than MN-PCDP. We further probe the readily accessible binding sites of MN-PCDP by determining the flow-through uptake of different organic micropollutants. In these experiments, the adsorbent (~5\u2009mg) is trapped as a thin layer on a 0.22 \u03bcm syringe filter, and aqueous organic pollutants (5\u2009mL, 0.1\u2009mM) passed rapidly through the filter at a flow rate of 10\u2009ml\u2009min\u22121 (Supplementary Fig.\u00a09). Under these conditions, for example, 76% of the BPA is removed from the solution, corresponding to more than 84% of its equilibrium adsorption, confirming that the host\u2212guest interaction plays a major role in the filtration process by syringe26.\n\na Structures of each tested organic pollutant. b Time-dependent adsorption of each pollutant (0.1\u2009mM) by MN-PCDP (1\u2009mg\u2009mL\u22121). c Percentage removal efficiency of each pollutant obtained by stirring NAC (blue), stirring MN-PCDP (red), and rapidly flowing through a thin MN-PCDP layer (green). The data are reported as the average uptake of triplicate experiments. Error bars mean standard deviations.\n\nAs is known, the hydroxyl groups of \u03b2-CD are located at the outer surface of the molecule, that is, primary hydroxyls at the narrow side and secondary hydroxyls at the wider side, which makes \u03b2-CD water\u2212soluble but simultaneously generates an inner cavity that is relatively hydrophobic27. Because of their hydrophobic interior cavity, \u03b2-CD can either partially or entirely accommodate suitably sized lipophilic low molecular weight molecules or even polymers28. The superior performance of MN-PCDP can be helpful to that its \u03b2-CD moieties are easily accessed by most of the organic micropollutants, and these molecules can be rapidly trapped in the cavity of \u03b2-CD. For example, MN-PCDP exhibits a remarkable adsorption capability and selectivity for most aromatics and some chain compounds, as shown in Fig.\u00a02b and Supplementary Figs.\u00a010, 11. Furthermore, by means of particular treatments such as changing pH value of solution29, the adsorption feature of molecules can be tuned. Thus, the MN-PCDP mesoporous nanosponge will display a wide applicability and selectivity in a variety of molecules.\n\nFurthermore, the influence of the concentrations of adsorbent on the adsorption efficiency of BPA is studied and shown in Fig.\u00a03b, Supplementary Fig.\u00a012, and Supplementary Table\u00a03. When the concentration increases from 0.1 to 1.0\u2009mg\u2009L\u22121, the adsorption efficiency of BPA is enhanced from 25.12 to 87.09% within 1\u2009min and from 35.07 to 89.82% within 10\u2009min.\n\na Optical photographs of MN-PCDP separation process by a magnet in continuous time. b Time-dependent adsorption of BPA (0.1\u2009mM) using MN-PCDP with different dosage (0.1, 0.25, 0.5, 0.75, and 1\u2009mg\u2009L\u22121). c Removal efficiency of BPA (0.01\u2009mM) using MN-PCDP (100\u2009mg) in three methods (100\u2009mL for 10 times, 250\u2009mL for 4 times and 500\u2009mL for 2 times). d Average removal (black) and enrichment (red) efficiency of the three methods in (c).\n\nAs we all know, organic micropollutants exhibit good solubility in organic solution, such as ethanol and methanol20. Hence, after adsorption process, we may separate the MN-PCDP mesoporous nanosponges from the solution quickly by the magnet, and utilize ethanol to desorb the organic micropollutants from the adsorbent, thus obtain the enriched pollutant solution. In order to obtain higher concentration of desorbed micropollutant solution, in this work, we chose 1\u2009mL ethanol to desorb organic micropollutants adsorbed in MN-PCDP adsorbent. As shown in Supplementary Fig.\u00a013, the concentration of BPA can be increased to 88.5 times of its initial concentration with a recipe of 100\u2009mL organic pollutant (BPA) solution and 100\u2009mg MN-PCDP adsorbent. This result reveals that more than 98% of the adsorbed organic micropollutants are desorbed into the ethanol solution. As discussed in Fig.\u00a03b, with the concentration of adsorbent increases, the adsorption efficiency tends to reach equilibrium. Considering the cost increase of sample preparation and dosage of adsorbent in the desorption process (with 1\u2009mL ethanol), 20\u2212100\u2009mg/100\u2009mL of adsorbent is selected as the adsorbent concentration in subsequent experiments.\n\nIn order to further improve the enrichment efficiency of 100\u2009mg adsorbent in total 1000\u2009mL organic micropollutants, herein, we attempted three methods during the adsorption and desorption processes, including 100\u2009mL \u00d7 10 times, 250\u2009mL \u00d7 4 times, and 500\u2009mL \u00d7 2 times. Importantly, the adsorbent can be simply separated by a magnet in every adsorption cycle, then desorbed in ethanol. As shown in Fig.\u00a03c, Supplementary Fig.\u00a014, and Supplementary Table\u00a04, as the recycling adsorption times increase, the removal efficiencies of these three methods gradually decrease. The average removal efficiencies in methods of 100\u2009mL \u00d7 10 times, 250\u2009mL \u00d7 4 times, and 500\u2009mL \u00d7 2 times are 50.98, 62.58, and 41.22%, respectively. These results represent an enrichment capability of 485, 605, and 396 times of the initial concentration (Fig.\u00a03d), respectively. Thus, we achieve a notable enrichment factor of above 600 times of initial organic pollutants via 1000\u2009mL initial solution by means of the optimization of adsorption and desorption processes. Here, the optimized parameters, i.e., 100\u2009mg adsorbent in 250\u2009mL \u00d7 4 cycle times, are used for the succedent experiments. Meanwhile, it is also worth pointing out that the separation process by a magnet is very fast and facile, such as 100\u2009mL with 60\u2009s (Supplementary Fig.\u00a015a), 250\u2009mL with 90\u2009s (Fig.\u00a03a and Supplementary Fig.\u00a015b), and 500\u2009mL with 150\u2009s (Supplementary Fig.\u00a015c). Therefore, the current ultra-fast enrichment protocol may fully meet the requirement of on-site and portable detection applications.\n\nIn order to evaluate the advantage of the current enrichment protocol on the detection sensitivity, fluorescence, and SERS spectra of five POPs molecules, e.g., BPA, carbendazim, tetramethyl thiuram disulfide (TMTD, thiram), diquat, and anthracene, are measured using MN-PCDP nanosponges as adsorbent. The hydrophobic slippery SERS platform30 with ~55\u2009nm Au NPs (Supplementary Fig.\u00a016) is adopted for the measurement of SERS spectra. As shown in Fig.\u00a04a, b, without the enrichment process, the LOD of SERS for TMTD is around 1 pM. However, after the enrichment using MN-PCDP adsorbent, this value reaches to ~5 fM, showing an increase of 102\u2212103. Meanwhile, based on this enrichment protocol, the LOD of BPA, carbendazim, diquat and\u00a0anthracene are up to 0.1\u2009nM, 5 pM, 1 pM, and 1\u2009nM (Supplementary Figs.\u00a017\u221220), which are much lower than most of the magnetic SERS-based sensors (Table\u00a01). Furthermore, multiple adsorption and desorption experiments by MN-PCDP for the above five organic molecules are implemented to illustrate the reproducibility of this adsorbent. In Supplementary Figs.\u00a021\u221223 and Supplementary Tables\u00a05, 6, the removal efficiencies and enrichment efficiencies of MN-PCDP adsorbent are excellent for target molecules with RSD less than 1 and 5%, respectively. The Raman detectable reproducibility of TMTD by hydrophobic slippery SERS platform is shown in Supplementary Figs.\u00a024\u221230. In Fig.\u00a04c, the Raman signals of TMTD characteristic peaks are acquired with 100% detection probability in 0.5 pM, ~65% in 50 fM, and ~10% in 5 fM. In addition, the solution-based aggregation approach, a simplest and effective way in commercial detection platforms at present, is adopted to clarify the consistency of SERS signal. As shown in Fig.\u00a04d and Supplementary Figs.\u00a031, 32, the SERS signals display superior spectral reproducibility and uniformity with a 100% detection probability and RSD value of ~5%, even at TMTD concentration of 10\u221212\u2009M. In Fig.\u00a04e, f and Supplementary Fig.\u00a033, using the current enrichment-typed sensing strategy, the LODs of fluorescence detections for the concentrated and enriched molecules of carbendazim and BPA are also enhanced by 2\u22123 orders of magnitude. In this study, the enrichment protocol based on the adsorption and desorption processes of MN-PCDP adsorbent may significantly increase the sensitivity of plasmonic sensors, compared with the LOD for the same molecules with different SERS and fluorescence detection protocols (Table\u00a01)31,32,33. Thus, the current strategy has wider applicability for mass spectrometry, chromatography, and other detection protocols.\n\nEnhanced Raman spectra of TMTD a before and b after enrichment process of MN-PCDP. c Probability of SERS signals with different concentrations of TMTD after enrichment process of MN-PCDP by hydrophobic slippery SERS platform. The inserted schematic diagram shows hydrophobic slippery SERS platform. The data are reported as the average enrichment of five experiments. Error bars mean standard deviations. Shaded areas mean the characteristic peaks (555, 1138, and 1370\u2009cm\u22121) of TMTD molecule. d Detection probability (red) and intensity of SERS signals (black) with different concentrations of TMTD after enrichment process of MN-PCDP by aggregating approach with Au colloid. The inserted schematic diagram shows aggregating approach with Au colloid. The data are reported as the average enrichment of five experiments. Error bars mean standard deviations. Fluorescence spectra of carbendazim e before and f after the enrichment of MN-PCDP.\n\nBased on the distinguishing and selective absorption capacity for different molecules (Fig.\u00a02 and Supplementary Figs.\u00a010, 11)29,34, the mesoporous nanosponge is expected to be used in the separation of interested molecules from the mixed systems. As shown in Fig.\u00a05b and Supplementary Fig.\u00a034, the malachite green (MG) and sunset yellow (SY) molecules are firstly mixed together and became to be the mixture solution (MG\u2009+\u2009SY) (Fig.\u00a05b-i-ii). Then, after adsorption and separation processes using MN-PCDP, clearly, there is only SY left in the mixture solution (Fig.\u00a05b-iii). Similarly, with a desorption process, the MG molecule is also successfully separated (Fig.\u00a05b-iv). Following, two different pesticide molecules, namely chlormequat chloride (CCC) and TMTD, are used to further evaluate the selective separation using SERS detection (Fig.\u00a05c-i-iii). Obviously, once TMTD is captured and separated, the SERS signals of CCC in the mixture solution significantly increases (Fig.\u00a05c-iv), indicating that most TMTD molecules have been adsorbed. Moreover, the single Raman (Fig.\u00a05c-v) and UV\u2212Vis spectra (Supplementary Fig.\u00a035) from TMTD molecules can be observed, illustrating only TMTD is effectively selected and separated by MN-PCDP.\n\na Optical photographs about the enrichment process of MN-PCDP in mud water. b UV\u2212Vis spectra of SY (i, 0.01\u2009mM), MG (i, 0.01\u2009mM) and SY/MG (ii, both with 0.01\u2009mM) mixture solution. iii The filtration mixture solution after the absorption of MN-PCDP for MG. iv The desorption solution of captured molecules (MG) redispersed in ethanol from the cavity of MN-PCDP adsorbent. c Raman spectra of CCC (i, 0.1\u2009mM), TMTD (ii, 1\u2009\u00b5M) and iii CCC (0.1\u2009mM) /TMTD (1\u2009\u00b5M) mixture solution before and after iv absorption, v desorption. d UV\u2212Vis spectra about selective enrichment of organic pollutant molecules (carbendazim, 0.01\u2009mM) from practical samples. e Raman spectra about selective enrichment of organic pollutant molecules (carbendazim, 1\u2009nM) from practical samples. f Raman spectrum of mixture (BPA and carbendazim) after the enrichment process of MN-PCDP in real samples. Shaded areas mean the characteristic peaks of BPA (green, 830 and 1179\u2009cm\u22121) and carbendazim (yellow, 1004, 1222, and 1263\u2009cm\u22121) molecules.\n\nAnother advantage of the current enrichment protocol is that the interference of complex matrix can be effectively eliminated in the detection of real-sample system35,36,37. After the adsorption process, the MN-PCDP adsorbent can be easily separated by a magnet from a complicated environment containing, e.g., mud and microorganism (Fig.\u00a05a). Figure\u00a05d exhibits the UV\u2212Vis spectra of carbendazim at a concentration of 0.01\u2009mM in the filtered soil solution. It can be observed that the characteristic absorption peaks of carbendazim are very weak even after filtration. However, using the current selective separation and enrichment protocol, the absorption peaks intensity of carbendazim (Fig.\u00a05d), BPA and MG (Supplementary Fig.\u00a036a, b) are respectively increased 463, 550, and 516 times comparing with pure solution. Similarly, as shown in Fig.\u00a05e, when the concentration of detection molecules goes down to 1\u2009nM, the SERS signals of carbendazim molecule in practical soil solution samples even after filtration are almost unobservable, whereas they can be easily detected after selective enrichment by MN-PCDP. The intensities of Raman characteristic peak from BPA and MG (Supplementary Fig.\u00a037a, b) in complex matrix are also greatly improved.\n\nIn addition, in the real environment, more than one molecule is always studied, thus it is very important to realize the simultaneous detection of multiple molecules, especially in the existence of matrix interference. Figure\u00a05f reveals that both the characteristic peaks of BPA (830 and 1179\u2009cm\u22121) and carbendazim (1004, 1222, and 1263\u2009cm\u22121) evidently appear in the Raman spectra of mixture solution, e.g., 1\u2009\u00b5M BPA and 10\u2009nM carbendazim, indicating the great absorption and detection capability for different molecules at the same time. Furthermore, the MN-PCDP demonstrates a superior reusability. As shown in Supplementary Fig.\u00a038, six consecutive BPA adsorption/desorption cycles are performed and the regenerated MN-PCDP exhibited almost inappreciable decrease (90.2\u221284.3%) in performance compared to the as-synthesized polymer.",
+            "section_image": [
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-021-27100-2/MediaObjects/41467_2021_27100_Fig2_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-021-27100-2/MediaObjects/41467_2021_27100_Fig3_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-021-27100-2/MediaObjects/41467_2021_27100_Fig4_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-021-27100-2/MediaObjects/41467_2021_27100_Fig5_HTML.png"
+            ]
+        },
+        {
+            "section_name": "Discussion",
+            "section_text": "In summary, we have developed a robust and efficient sensing strategy based on the MN-PCDP mesoporous nanoponges to capture and enrich organic pollutants from water. In this strategy, the MN-PCDP adsorbent exhibits excellent selective adsorption and enrichment capacity for various kinds of pollutants, eliminating the interference of complex matrix in the real-sample environments. Meanwhile, in diverse detection protocols of organic pollutants, e.g., UV\u2212Vis, Raman, and fluorescent, the current sensing strategy significantly may increase the sensitivity with 2\u22123 orders of magnitude. Moreover, using the immobilization of magnetic NPs, the adsorption, separation, and enrichment processes by MN-PCDP can be completed within 2\u22123\u2009min. Therefore, the current robust sensing strategy with the ultra-rapid, selective, and highly efficient molecule enrichment capability is believed to be applicable to a wider range of sensing devices for a cost-effective, simple, fast, flexible and portable detection. In the future, single-particle-MN-PCDP combining with Au NPs (SERS substrate) could dramatically lower the detection limit and enables higher spatial and temporal resolution38,39,40,41,42, thus build single NP sensor to improve detection sensitivity (Supplementary Fig.\u00a039).",
+            "section_image": []
+        },
+        {
+            "section_name": "Methods",
+            "section_text": "The carboxyl-functionalized magnetite NPs (Fe3O4) with highly water-dispersibility were synthesized by a modified solvothermal reaction approach22. Typically, FeCl3\u00b76H2O (1.08\u2009g, 4.0\u2009mmol) and trisodium citrate (0.20\u2009g, 0.68\u2009mmol) were dissolved in ethylene glycol (20\u2009mL) with stirring at 500\u2009rpm. Afterward, sodium acetate trihydrate (2.0\u2009g, 15\u2009mmol) was added and the mixture was stirred for 30\u2009min. Then, the mixture was sealed in a Teflon-lined stainless-steel autoclave (50\u2009mL). The autoclave was heated at 200\u2009\u00b0C for 12\u2009h, and then allowed to cool to room temperature. The black products were washed with ethanol and ultrapure water for several times. Finally, the carboxyl-functionalized magnetite NPs (Fe3O4) were separated by magnet, re-dispersed in ethanol, and dried in vacuum drying oven at 30\u2009\u00b0C.\n\nThe MN-PCDP composites were then prepared by modification of nucleophilic aromatic substitution method of hydroxyl groups of \u03b2-CD20. A dried 100\u2009mL Shrek reaction vial with a magnetic stir bar was charged with \u03b2-CD (0.82\u2009g, 0.724\u2009mmol), TFT (0.40\u2009g, 1.03\u2009mmol), and K2CO3 (1.28\u2009g, 9.28\u2009mmol) and dried Fe3O4 (0.041\u2009g). The vial was flushed with N2 gas for 10\u2009min, then an anhydrous THF/DMF mixture (9:1\u2009v/v, 40\u2009mL) was added and the vial was purged with N2 for an additional 5\u2009min. After that, the N2 inlet was removed. The mixture was stirred at 500\u2009rpm and refluxed at 85\u2009\u00b0C for 36\u2009h under nitrogen protection. The brown suspension was cooled to room temperature and magnetically separated the supernatant by magnet. The precipitate was washed twice with an appropriate amount of distilled water, THF, ethanol, and CH2Cl2, respectively. The final precipitate was vacuum dried at 77\u2009K in a liquid nitrogen bath for 24\u2009h and then the magnetic NPs immobilized porous \u03b2-CD polymer (MN-PCDP) was obtained.\n\nIn studies, the dried polymer (MN-PCDP, 20\u2009mg) was initially washed with H2O for 2 times and then separated by a magnet. Adsorption kinetic studies for different pollutants were performed in 30\u2009mL scintillation vials with 20\u2009mL organic pollutant solution and 20\u2009mg adsorbent, at ambient temperature on a hot plate at 25\u2009\u00b0C. Then the sample was shaken at 250\u2009rpm until the adsorption equilibrium was reached. The mixture was immediately stirred and 1\u2009mL aliquots of the suspension were taken at certain intervals via syringe and filtered immediately by a 0.22 \u03bcm PTFE membrane filter. The residual concentration of the pollutant in each sample was determined by UV\u2013vis spectroscopy.\n\nThe removal efficiency of pollutant removal by the adsorbent was determined by the following equation:\n\nwhere C0 and Ct are the initial and residual concentration of pollutant in the stock solution and filtrate, respectively.\n\nIndividual pollutants were at high concentrations (mM). 5.0\u2009mg of the MN-PCDP adsorbent was washed with deionized H2O for 2 times, then the precipitate was pushed by a syringe through a 0.22 \u03bcm PTFE membrane filter to form a thin layer of the adsorbent on the filter membrane. 5\u2009mL of the pollutant stock solution was then pushed through the adsorbent in ~30\u2009s (10\u2009mL\u2009min\u22121 flow rate). The filtrate was then measured by UV\u2013Vis spectroscopy to determine the pollutant removal efficiency.\n\n100.0\u2009mg of the adsorbent was washed with deionized H2O for 2 times, and then added to the organic pollutant stock solution (0.01\u2009mM) with determined volume (100, 250, and 500\u2009mL). The mixture was shaken at 250\u2009rpm for 1\u2009min at 25\u2009\u00b0C. After separating the supernatant and the adsorbent by an external magnet, the supernatant was filtered through a 0.22 \u03bcm filter membrane and determined by UV\u2212Vis spectroscopy. Meanwhile, the precipitate was evaporated to dryness with a gentle nitrogen stream, then the residue was dissolved in 1\u2009mL of ethanol to desorb the adsorbed organic pollutant. The desorption solution was measured by UV\u2212Vis spectroscopy and compared with the initial concentration of pollutant in the stock solution.\n\nThe enrichment efficiency of pollutant adsorbed by the adsorbent was determined by the following equation:\n\nWhere C0 and C are the initial and desorbed solution concentration of pollutant, respectively.\n\nThe fluorescence spectra of pure solution were directly measured by a fluorescence spectrophotometer.\n\nThe Au NPs with different sizes in diameter were synthesized based on a modified citrate reduction approach. The growth process of gold NPs with different sizes included three steps. For step 1, 100\u2009mL of ultrapure water was added into a conical flask and heated to boiling. Then, 4\u2009ml of 1\u2009wt% sodium citrate (SC) solution was injected immediately, and 3.2\u2009mL of 10\u2009mM HAuCl4 was added after 3\u2009min. Kept the reaction for 25\u2009min and made it natural cooling, then the Au seeds were obtained. For step 2, 80\u2009mL of ultrapure water and 20\u2009mL of Au seeds were mixed into the conical flask and heated to boiling. Then, 2\u2009mL of 1\u2009wt% sodium citrate solution was injected immediately, and 0.2\u2009mL of HAuCl4 was added 3\u2009min later. Then additional 0.2\u2009mL \u00d7 9 dosage of HAuCl4 was injected every 8\u2009min. After the last precursor was added, the reaction was kept for 25\u2009min, and Au NPs-25 nm were obtained. For step 3, Au NPs prepared in step 2 were used as the seed solution, and the growth process was repeated as growth steps 2, and then Au NPs-55 nm were obtained in this step.\n\nSERS measurement is based on the hydrophobic slippery surface17. Concentrated molecules and Au NPs were prepared on a hydrophobic slippery Teflon membrane as follows: First, a Teflon membrane was attached on a flat glass slide (5\u2009cm \u00d7 5\u2009cm) by using a double-sided adhesive. Then, 0.5\u2009mL of perfluorinated fluid was dispersed by spin coating. The low speed was 300\u2009rpm for 30\u2009s, and the high speed was 1500\u2009rpm for 1\u2009min. After the excess lubricating liquid was removed by centrifugal force, and the infused membrane was heated for 30\u2009min. Lastly, 50\u2009\u03bcL of probe molecules and 10\u2009\u03bcL of Au colloids were simultaneously dropped onto the slippery surface. During drying, the contact line shrunk because of the low friction of the lubricated Teflon surface. As a result, the initial droplet could be concentrated into a small area less than 0.5\u2009mm in diameter.",
+            "section_image": []
+        },
+        {
+            "section_name": "Data availability",
+            "section_text": "The data that support the findings of this study are available within the paper and its Supplementary Information or from the corresponding authors on reasonable request.",
+            "section_image": []
+        },
+        {
+            "section_name": "References",
+            "section_text": "Kelly, B. C., Ikonomou, M. G., Blair, J. D., Morin, A. E. & Gobas, F. A. P. C. Food web-specific biomagnification of persistent organic pollutants. Science 317, 236\u2013239 (2007).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSchwarzenbach, R. P. et al. The challenge of micropollutants in aquatic systems. Science 313, 1072\u20131077 (2006).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nJones, K. C. & Voogt, P. De. Persistent organic pollutants (POPs): state of the science. Environ. Pollut. 100, 209\u2013221 (1999).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nPi, Y. H. et al. Adsorptive and photocatalytic removal of persistent organic pollutants (POPs) in water by metal-organic frameworks (MOFs). Chem. Eng. J. 337, 351\u2013371 (2018).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nXu, H. X., Bjerneld, E. J., K\u00e4ll, M. & B\u00f6rjesson, L. Spectroscopy of single hemoglobin molecules by surface enhanced Raman scattering. Phys. Rev. Lett. 83, 4357\u20134360 (1999).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKneipp, K. ; et al. Single molecule detection using surface-enhanced Raman scattering (SERS). Phys. Rev. Lett. 78, 1667\u20131670 (1997).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nNie, S. M. & Emory, S. R. Probing single molecules and single nanoparticles by surface-enahnced Raman scattering. Science 275, 1102\u20131106 (1997).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nXu, H. X., Aizpurua, J., Kall, M. & Apell, P. Electromagnetic contributions to single-molecule sensitivity in surface-enhanced Raman scattering. Phys. Rev. E 62, 4318 (2000).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLi, J. F. et al. Shell-isolated nanoparticle-enhanced Raman spectroscopy. Nature 464, 392\u2013395 (2010).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nPanneerselvam, R. et al. Surface-enhanced Raman spectroscopy: bottlenecks and future directions. Chem. Commun. 54, 10\u201325 (2017).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nOtto, A., Mrozek, I., Grabhorn, H. & Akemann, W. Surface-enhanced Raman scattering. J. Phys.-Condens. Matter 4, 1143\u20131212 (1992).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSchatz & George, C. Theoretical studies of surface enhanced Raman scattering. Acc. Chem. Res. 17, 370\u2013376 (1984).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSchlucker, S. Surface-enhanced Raman spectroscopy: concepts and chemical applications. Angew. Chem. Int. Ed. Engl. 53, 4756\u20134795 (2014).\n\nArticle\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMoskovits, M. Surface-enhanced Raman spectroscopy: a brief retrospective. J. Raman Spectrosc. 36, 485\u2013496 (2005).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nZeman, E. J. & Schatz, G. C. An accurate electromagnetic theory study of surface enhancement factors for silver, gold, copper, lithium, sodium, aluminum, gallium, indium, zinc, and cadmium. J. Phys. Chem. 91, 634\u2013643 (1987).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGuo, X. T. et al. Molecular imprinting-based surface-enhanced Raman scattering sensors. ACS Sens. 5, 601\u2013619 (2020).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nZhang, D. J. et al. Buoyant particulate strategy for few-to-single particle-based plasmonic enhanced nanosensors. Nat. Commun. 11, 2603 (2020).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHao, R., You, H. J., Zhu, J., Chen, T. & Fang, J. X. \u201cBurning Lamp\u201d-like robust molecular enrichment for ultrasensitive plasmonic nanosensors. ACS Sens. 5, 781\u2013788 (2020).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nDe Angelis, F. et al. Breaking the diffusion limit with super-hydrophobic delivery of molecules to plasmonic nanofocusing SERS structures. Nat. Photonics 5, 682\u2013687 (2011).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nAlsbaiee, A. et al. Rapid removal of organic micropollutants from water by a porous beta-cyclodextrin polymer. Nature 529, 190\u2013194 (2016).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nXiao, L. L. et al. Beta-cyclodextrin polymer network sequesters perfluorooctanoic acid at environmentally relevant concentrations. J. Am. Chem. Soc. 139, 7689\u20137692 (2017).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLiu, J. et al. Highly water-dispersible biocompatible magnetite particles with low cytotoxicity stabilized by citrate groups. Angew. Chem. Int. Ed. 48, 5875\u20135879 (2009).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nJiang, H. L. et al. A novel crosslinked \u03b2-cyclodextrin-based polymer for removing methylene blue from water with high efficiency. Colloid Surf. A 560, 59\u201368 (2019).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nXu, G. Z. et al. Simple synthesis of a swellable porous \u03b2-cyclodextrin-based polymer in the aqueous phase for the rapid removal of organic micro-pollutants from water. Green. Chem. 21, 6062\u20136072 (2019).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHuang, Q. Y., Chai, K. G., Zhou, L. Q. & Ji, H. B. A phenyl-rich \u03b2-cyclodextrin porous crosslinked polymer for efficient removal of aromatic pollutants: insight into adsorption performance and mechanism. Chem. Eng. J. https://doi.org/10.1016/j.cej.2020.124020 (2020).\n\nWang, Z. H., Zhang, P. B., Hu, F., Zhao, Y. F. & Zhu, L. P. A crosslinked beta-cyclodextrin polymer used for rapid removal of a broad-spectrum of organic micropollutants from water. Carbohyd. Polym. 177, 224\u2013231 (2017).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nManakker, F. V. D., Vermonden, T., Nostrum, C. F. V. & Hennink, W. E. Cyclodextrin-based polymeric materials: synthesis, properties, and pharmaceutical/biomedical applications. Biomacromolecules 10, 3157\u20133175 (2009).\n\nArticle\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWankar, J. et al. Recent advances in host-guest self\u2010assembled cyclodextrin carriers: implications for responsive drug delivery and biomedical engineering. Adv. Funct. Mater. 30, 1909049 (2020).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nZhao, R., Wang, Y., Li, X., Sun, B. & Wang, C. Synthesis of \u03b2-cyclodextrin-based electrospun nanofiber membranes for highly efficient adsorption and separation of methylene blue. ACS Appl. Mater. Interfaces 7, 26649\u201326657 (2015).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nZhang, D. J. et al. Hydrophobic slippery surface-based surface-enhanced raman spectroscopy platform for ultrasensitive detection in food safety applications. Anal. Chem. 91, 4687\u20134695 (2019).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nChen, X. et al. Detection and quantification of carbendazim in Oolong tea by surface-enhanced Raman spectroscopy and gold nanoparticle substrates. Food Chem. 293, 271\u2013277 (2019).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nZhu, X. Y. et al. A novel graphene-like titanium carbide MXene/Au\u2013Ag nanoshuttles bifunctional nanosensor for electrochemical and SERS intelligent analysis of ultra-trace carbendazim coupled with machine learning. Ceram. Int. https://doi.org/10.1016/j.ceramint.2020.08.121 (2020).\n\nZhai, Y. et al. Metal-organic-frameworks-enforced surface enhanced Raman scattering chip for elevating detection sensitivity of carbendazim in seawater. Sensors Actuat. B-Chem. https://doi.org/10.1016/j.snb.2020.128852 (2021).\n\nJiang, H. L. et al. Selective adsorption of anionic dyes from aqueous solution by a novel \u03b2-cyclodextrin-based polymer. J. Mol. Struct. 1203, 127373 (2019).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSong, D. et al. A label-free SERRS-based nanosensor for ultrasensitive detection of mercury ions in drinking water and wastewater effluent. Anal. Methods 9, 154\u2013162 (2017).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBao, Z. Y., Dai, J., Lei, D. Y. & Wu, Y. Maximizing surface-enhanced Raman scattering sensitivity of surfactant-free Ag\u2212Fe3O4 nanocomposites through optimization of silver nanoparticle density and magnetic self-assembly. J. Appl. Phys. 114, 124305 (2013).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBao, Z. Y. et al. Quantitative SERS detection of low-concentration aromatic polychlorinated biphenyl-77 and 2,4,6-trinitrotoluene. J. Hazard. Mater. 280, 706\u2013712 (2014).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLi, Q. et al. Detection of single proteins with a general nanopore sensor. ACS Sens. 4, 1185\u20131189 (2019).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLu, S. M. et al. Electrochemical sensing at a confined space. Anal. Chem. 92, 5621\u20135644 (2020).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nYing, Y. L. et al. Electrochemical confinement effects for innovating new nanopore sensing mechanisms. Small Methods 2, 1700390 (2018).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nShi, L. et al. Plasmon resonance scattering spectroscopy at the single nanoparticle level: real-time monitoring of a click reaction. Angew. Chem. Int. Ed. 52, 6011\u20136014 (2013).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGao, Q. et al. Highly sensitive impedimetric sensing of DNA hybridization based on the target DNA-induced displacement of gold nanoparticles attached to ssDNA probe. Electrochem. Commun. 13, 335\u2013337 (2013).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHe, Q. et al. Fabrication of Fe3O4@SiO2@Ag magnetic-plasmonic nanospindles as highly efficient SERS active substrates for label-free detection of pesticides. N. J. Chem. 41, 1582\u20131590 (2017).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nTang, S. Y. et al. Efficient enrichment and self-assembly of hybrid nanoparticles into removable and magnetic SERS substrates for sensitive detection of environmental pollutants. ACS Appl. Mater. Interfaces 9, 7472\u20137480 (2017).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWang, C. W. et al. Seed-mediated synthesis of high-performance silver-coated magnetic nanoparticles and their use as effective SERS substrates. Colloids Surf. A: Physicochem. Eng. Asp. 506, 393\u2013401 (2016).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSun, M. et al. Cube-like Fe3O4@SiO2@Au@Ag magnetic nanoparticles: a highly efficient SERS substrate for pesticide detection. Nanotechnology 29, 165302 (2018).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWang, C. W. et al. Sonochemical synthesis of highly branched flower-like Fe3O4@SiO2@Ag microcomposites and their application as versatile SERS substrates. Nanoscale 8, 19816\u201319828 (2016).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWang, C. W. et al. Polyethylenimine-interlayered core-shell-satellite 3D magnetic microspheres as versatile SERS substrates. Nanoscale 7, 18694\u201318707 (2015).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nYang, T. X. et al. Au dotted magnetic network nanostructure and its application for on-site monitoring femtomolar level pesticide. Small 10, 1325\u20131331 (2014).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLi, Q. et al. A gold/Fe3O4 nanocomposite for use in a surface plasmon resonance immunosensor for carbendazim. Microchim. Acta 186, 313 (2019).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKadasala, N. R. & Wei, A. Trace detection of tetrabromobisphenol A by SERS with DMAP-modified magnetic gold nanoclusters. Nanoscale 7, 10931\u201310935 (2015).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nXiao, R., Wang, W. C., Zhu, A. N. & Long, F. Single functional magnetic-bead as universal biosensing platform for trace analyte detection using SERS-nanobioprobe. Biosens. Bioelectron. 79, 661\u2013668 (2016).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nDu, J. J. & Jing, C. Y. Preparation of thiol modified Fe3O4@Ag magnetic SERS probe for PAHs detection and identification. J. Phys. Chem. C 115, 17829\u201317835 (2011).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nDownload references",
+            "section_image": []
+        },
+        {
+            "section_name": "Acknowledgements",
+            "section_text": "This work was supported by the programs supported by the National Natural Science Foundation of China (No. 21675122, 21874104, 22074115 awarded to J.X.F. and 62022001 awarded to D.Y.L.), the Key Research Program in Shaanxi (2017NY-114 awarded to J.X.F.), and Natural Science Foundation of Shaanxi Province (No. 2019JLP-19 awarded to H.J.Y.), the World-Class Universities (Disciplines, awarded to J.X.F.) and the Characteristic Development Guidance Funds for the Central Universities (awarded to J.X.F.). The characterizations of materials are supported by the Instrument Analysis Center of Xi\u2019an Jiaotong University.",
+            "section_image": []
+        },
+        {
+            "section_name": "Author information",
+            "section_text": "These authors contributed equally: Lingling Zhang, Yu Guo.\n\nKey Laboratory of Physical Electronics and Devices of Ministry of Education, School of Electronic Science and Engineering, Faculty of Electronic and Information Engineering, Xi\u2019an Jiaotong University, Xi\u2019an, Shaanxi, 710049, China\n\nLingling Zhang,\u00a0Yu Guo,\u00a0Rui Hao,\u00a0Yafei Shi\u00a0&\u00a0Jixiang Fang\n\nKey Laboratory of Biomedical Information Engineering of Ministry of Education, School of Life Science and Technology, Xi\u2019an Jiaotong University, Xi\u2019an, Shaanxi, 710049, China\n\nHongjun You\u00a0&\u00a0Jixiang Fang\n\nSchool of Microelectronics, Faculty of Electronic and Information Engineering, Xi\u2019an Jiaotong University, Xi\u2019an Jiaotong University, Xi\u2019an, Shaanxi, 710049, China\n\nHu Nan\u00a0&\u00a0Yanzhu Dai\n\nDepartment of Materials Science and Engineering, City University of Hong Kong, 83 Tat Chee Avenue, Kowloon, 999077, Hong Kong, China\n\nDanjun Liu\u00a0&\u00a0Dangyuan Lei\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nL.L.Z. synthesized the materials, carried out the characterizations and performance, analyzed the data, and wrote the draft of the manuscript. Y.G., R.H., Y.F.S., H.N., Y.Z.D., D.J.L. and D.Y.L. contributed in part of the TEM, Raman, and fluorescence characterizations. J.X.F. supervised the project, designed the experiments, contributed in discussions, comments, and writing of the manuscript. H.J.Y. designed the partial experiments, contributed in discussions and comments. All authors discussed the results and commented on the manuscript.\n\nCorrespondence to\n                Jixiang Fang.",
+            "section_image": []
+        },
+        {
+            "section_name": "Ethics declarations",
+            "section_text": "The authors declare no competing of interest.",
+            "section_image": []
+        },
+        {
+            "section_name": "Additional information",
+            "section_text": "Peer review information\u2009Nature Communications thanks the anonymous reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available.\n\nPublisher\u2019s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.",
+            "section_image": []
+        },
+        {
+            "section_name": "Rights and permissions",
+            "section_text": "Open Access  This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article\u2019s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article\u2019s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.\n\nReprints and permissions",
+            "section_image": []
+        },
+        {
+            "section_name": "About this article",
+            "section_text": "Zhang, L., Guo, Y., Hao, R. et al. Ultra-rapid and highly efficient enrichment of organic pollutants via magnetic mesoporous nanosponge for ultrasensitive nanosensors.\n                    Nat Commun 12, 6849 (2021). https://doi.org/10.1038/s41467-021-27100-2\n\nDownload citation\n\nReceived: 13 December 2020\n\nAccepted: 19 October 2021\n\nPublished: 25 November 2021\n\nVersion of record: 25 November 2021\n\nDOI: https://doi.org/10.1038/s41467-021-27100-2\n\nAnyone you share the following link with will be able to read this content:\n\nSorry, a shareable link is not currently available for this article.\n\n\n\n\n                            Provided by the Springer Nature SharedIt content-sharing initiative\n                        ",
+            "section_image": [
+                "https://data:image/svg+xml;base64,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"
+            ]
+        },
+        {
+            "section_name": "This article is cited by",
+            "section_text": "Nature Communications (2025)\n\nJournal of Inorganic and Organometallic Polymers and Materials (2025)\n\nNature Communications (2022)",
+            "section_image": []
+        }
+    ]
+}

0417a97ecc0d13717181d9623733f3148d5ba9e7ac583590ea0cf0862739b60a/metadata.json

@@ -0,0 +1,154 @@
+{
+    "title": "Hyperuniformity and phase enrichment in vortex and rotor assemblies",
+    "pre_title": "Hyperuniformity and phase enrichment in vortex and rotor assemblies",
+    "journal": "Nature Communications",
+    "published": "10 February 2022",
+    "supplementary_0": [
+        {
+            "label": "Supplementary Information",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-022-28375-9/MediaObjects/41467_2022_28375_MOESM1_ESM.pdf"
+        },
+        {
+            "label": "Peer Review File",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-022-28375-9/MediaObjects/41467_2022_28375_MOESM2_ESM.pdf"
+        }
+    ],
+    "supplementary_1": NaN,
+    "supplementary_2": NaN,
+    "source_data": [
+        "https://github.com/dbstein/rotor_hyperuniformity",
+        "/articles/s41467-022-28375-9#ref-CR64"
+    ],
+    "code": [
+        "https://github.com/dbstein/rotor_hyperuniformity",
+        "/articles/s41467-022-28375-9#ref-CR64"
+    ],
+    "subject": [
+        "Fluid dynamics",
+        "Membrane biophysics",
+        "Self-assembly",
+        "Statistical physics, thermodynamics and nonlinear dynamics"
+    ],
+    "license": "http://creativecommons.org/licenses/by/4.0/",
+    "preprint_pdf": "https://www.researchsquare.com/article/rs-385285/v1.pdf?c=1644511741000",
+    "research_square_link": "https://www.researchsquare.com//article/rs-385285/v1",
+    "nature_pdf": "https://www.nature.com/articles/s41467-022-28375-9.pdf",
+    "preprint_posted": "26 Apr, 2021",
+    "research_square_content": [
+        {
+            "section_name": "Abstract",
+            "section_text": "Ensembles of particles rotating in a two-dimensional fluid can exhibit chaotic dynamics yet develop signatures of hidden order. Such \u201crotors\u201d are found in the natural world spanning vastly disparate length scales \u2014 from the rotor proteins in cellular membranes to models of atmospheric dynamics. Here we show that an initially random distribution of either ideal vortices in an inviscid fluid, or driven rotors in a viscous membrane, spontaneously self assembles. Despite arising from drastically different physics, these systems share a Hamiltonian structure that sets geometrical conservation laws resulting in distinct structural states. We find that the rotationally invariant interactions isotropically suppress long wavelength fluctuations \u2014 a hallmark of a disordered hyperuniform material. With increasing area fraction, the system orders into a hexagonal lattice. In mixtures of two co-rotating populations, the stronger population will gain order from the other and both will become phase enriched. Finally, we show that classical 2D point vortex systems arise as exact limits of the experimentally accessible microscopic membrane rotors, yielding a new system through which to study topological defects.Plasma and FluidsThermodynamics and statistical mechanicsBiophysicsParticle EnsemblesTwo-dimensional FluidSpontaneous Self-assemblyHamiltonian StructureTopological Defects",
+            "section_image": []
+        },
+        {
+            "section_name": "Figures",
+            "section_text": "Figure 1Figure 2Figure 3Figure 4",
+            "section_image": [
+                "https://assets-eu.researchsquare.com/files/rs-85285/v1/25dfe053265f6e1d42c3d03a.png",
+                "https://assets-eu.researchsquare.com/files/rs-85285/v1/2fde83f8a67c8aaf66ace71c.png",
+                "https://assets-eu.researchsquare.com/files/rs-85285/v1/d0dee7f814d3e81a142f4006.png",
+                "https://assets-eu.researchsquare.com/files/rs-85285/v1/5fbe8539260cab605a265869.png"
+            ]
+        },
+        {
+            "section_name": "Additional Declarations",
+            "section_text": "There is NO Competing Interest.",
+            "section_image": []
+        },
+        {
+            "section_name": "Supplementary Files",
+            "section_text": "SI.pdfSupplementary Information",
+            "section_image": []
+        }
+    ],
+    "nature_content": [
+        {
+            "section_name": "Abstract",
+            "section_text": "Ensembles of particles rotating in a two-dimensional fluid can exhibit chaotic dynamics yet develop signatures of hidden order. Such rotors are found in the natural world spanning vastly disparate length scales \u2014 from the rotor proteins in cellular membranes to models of atmospheric dynamics. Here we show that an initially random distribution of either driven rotors in a viscous membrane, or ideal vortices with minute perturbations, spontaneously self assemble into a distinct arrangement. Despite arising from drastically different physics, these systems share a Hamiltonian structure that sets geometrical conservation laws resulting in prominent structural states. We find that the rotationally invariant interactions isotropically suppress long-wavelength fluctuations \u2014 a hallmark of a disordered hyperuniform material. With increasing area fraction, the system orders into a hexagonal lattice. In mixtures of two co-rotating populations, the stronger population will gain order from the other and both will become phase enriched. Finally, we show that classical 2D point vortex systems arise as exact limits of the experimentally accessible microscopic membrane rotors, yielding a new system through which to study topological defects.",
+            "section_image": []
+        },
+        {
+            "section_name": "Introduction",
+            "section_text": "Two-dimensional (or nearly so) fluid flows show rich and complex vortical dynamics. These can arise from flow interactions with boundaries1,2, the inverse cascades of 2D turbulence3,4,5, from Coriolis force-dominated atmospheric flows6, and from quantization effects in superfluid He-II7,8. Point vortices have long been staples for the modeling of such inertially dominated inviscid flows. Kirchoff9 was the first to describe point vortices using a Hamiltonian framework. His work was extended by many others10,11,12,13, notably, Onsager14 in his statistical mechanics treatment of 2D turbulence as clouds of point vortices.\n\nRemarkably, structurally identical Hamiltonian and moment constraints can arise in the microscopic, viscously dominated realm from a strict balance of dissipation with drive on immersed rotating objects. These objects include models of interacting transmembrane ATP-synthase rotor proteins15,16,17, and the planar interactions of rotors\u2014microscopic particles driven to rotate by an external torque18,19. We refer to such systems as BDD systems, as in balanced drive and dissipation. In modeling rotational BDD systems, other physical effects may also come into play, such as steric interactions, that can yield interesting complexities17. Assemblies of interacting, driven-to-rotate particles have become an area of intensifying interest in the active matter community18,19,20,21,22,23,24,25,26.\n\nHere, we study both a BDD system of rotating microscopic particles\u2014membrane rotors\u2014immersed in a flat membrane, and point vortices which are a particular limit of this BDD system. For both, symmetries in the Hamiltonian, \\({{{{{{{\\mathcal{H}}}}}}}}\\), lead to conservation laws that are geometrical in nature, bounding the proximity and distribution of particles. We derive a connection between the Hamiltonian and the structure factor, S(q) (where q is the wavevector), which can be used to place bounds on spatial correlations,\n\nwhere \\(\\widetilde{{{\\Psi }}}\\) is the Fourier transform of the stream function. In the case of point vortices, \\(\\widetilde{{{\\Psi }}}({{{{{{{\\bf{q}}}}}}}})=1/{q}^{2}\\), where q\u2009=\u2009|q|. As we show, Eq. (1) argues that this system should tend towards hyperuniformity. That is, the long-wavelength configuration at steady state is characterized by an isotropically vanishing structure factor, S(q\u2009\u2192\u20090)\u2009\u2192\u20090, leading to an isotropic band-gap27,28,29. To investigate this prediction, we numerically simulate assemblies of both BDD and point vortices and observe (see Fig.\u00a01): (i) hyperuniformity for BDD systems; (ii) evidence that point vortex systems can become hyperuniform depending on how they are perturbed; (iii) phase enrichment (in both cases); and (iv) crystallization (for BDD). Our observations lead us to conclude that rotational dynamics provide a mechanism for the self-assembly of particles into a disordered hyperuniform 2D material.\n\n(A) Hyperuniformity for Euler point vortices, (B)\u00a0Hyperuniformity for\u00a0QG rotors/surface rotors, (C) Phase enrichment induced by circulation differences where green (black) represents vortices of high (low) circulation, and (D) Crystallization arising from hydrosteric interactions. The insets of (A), (B), and (C) show the structure factor, S(q). In (A) and (B), S(q) decays to zero at small q, indicating that the distribution is hyperuniform. In (C), the structure factor shows the six distinct peaks of a hexagonal lattice.",
+            "section_image": [
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-022-28375-9/MediaObjects/41467_2022_28375_Fig1_HTML.png"
+            ]
+        },
+        {
+            "section_name": "Results",
+            "section_text": "We begin by introducing a single vortex in an ideal inviscid fluid. We then describe the flow generated by a point rotor in a viscous membrane and show that the two flows are identical in a biologically relevant limit. We use this equivalence and apply known tools from the study of ideal vortices on both systems. Namely, the linearity of the equations enables extending the result of a single vortex to the flow generated by an ensemble of vortices, which, in turn, could also be described by a Hamiltonian.\n\nAn ideal point vortex is given by a singular vorticity, \u03c9\u2009=\u2009\u2207\u2009\u00d7\u2009v\u2009=\u2009\u03b4(r). A 2D incompressible fluid can be described using a stream function \u03a8 such that the velocity, v, is given by v\u2009=\u2009\u2202\u22a5\u03a8. This equation, combined with the equation above gives, \\({{\\Psi }}=-\\frac{1}{2\\pi }\\log r\\) (ref. 12). The flow, v(r), therefore, scales as 1/r, where r\u2009=\u2009\u2223r\u2223.\n\nWe switch now to a point rotor in a viscous membrane, driven by an external torque \u03c4 (see Fig.\u00a02A for a schematic representation). Following Saffman and Delbr\u00fcck\u2019s seminal work30, and others that followed15,16,31,32,33, we assume that the membrane is incompressible (\u2207\u2009\u22c5\u2009v\u2009=\u20090), and that inertia is negligible. Under these assumptions, the Stokes momentum conservation equation for the membrane reads,\n\nwhere v is the 2D velocity in the plane of the membrane, u\u00b1 is the 3D flow in the outer fluids, \u03b72D is the 2D viscosity, and \u03b73D is the viscosity of the outer fluids. The second term on the right-hand side is the surface shear stress of the outer fluids, and the third term is the force due to a rotating point object. There is no pressure contribution when the motion is purely rotational. This equation is coupled to the equations of the outer fluids. It is easy to solve the above equations using a 2D Fourier transform (\\(\\widetilde{F}({{{{{{{\\bf{q}}}}}}}})=\\int\\nolimits_{-\\infty }^{\\infty }\\int\\nolimits_{-\\infty }^{\\infty }F({{{{{{{\\bf{r}}}}}}}}){e}^{-i{{{{{{{\\bf{q}}}}}}}}\\cdot {{{{{{{\\bf{r}}}}}}}}}{d}^{2}r\\)), giving:\n\nwhere \u0393\u2009=\u2009\u03c4/\u03b72D, and \u03bb\u2009=\u2009\u03b72D/2\u03b73D is the Saffman Delbr\u00fcck length. At small distances (r\u2009\u2009\u226a\u2009\u2009\u03bb), momentum travels in the plane of the membrane. At large distances (r\u2009\u226b\u2009\u03bb), momentum travels through the outer fluid as well34,35. In real space, \u03a8(r)\u2009=\u20091/4(H0(r/\u03bb)\u2009\u2212\u2009Y0(r/\u03bb)), where H0 and Y0 are zeroth-order Struve function and Bessel function of the second kind, respectively.\n\nA A representation of a membrane rotor\u2014a disk rotating due to a torque \u03c4 in the plane of the membrane. B The velocity field due to a membrane rotor (solid line) which scales as a point vortex v\u2009~\u20091/r at small distances (dotted), r/\u03bb\u2009\u2009\u226a\u2009\u20091, transitioning to a QG behavior at large distances v\u2009~\u20091/r2 (dashed). C Contour dynamics of an ellipse with radii ratios rl/rs\u2009\u2264\u20093, where rl (rs) is the major (minor) axis. Starting from the same contour, the dynamics differ according to the radius relative to the SD length. Blue is in the limit rl\u2009\u2009\u226a\u2009\u2009\u03bb. In this limit, the ellipse is rotating as a rigid body, as predicted by Kelvin65 for an elliptic patch in an Euler fluid. Black is in the limit rl\u2009\u226b\u2009\u03bb, no longer conserving its shape since the large distance flow is in the quasigeostrophic regime. D 200 point membrane rotors, blue is the initial random configuration, black is the final configuration. Solid line shows typical trajectory of an individual vortex. Note that the area did not change considerably since the system of vortices is self-bounding.\n\nIn the limit of small distances, r\u2009\u2009\u226a\u2009\u2009\u03bb, the stream function is, \\({{\\Psi }}\\approx -\\frac{1}{2\\pi }\\log r\\), i.e., exactly the same as for an ideal point vortex. In the opposite limit, r\u2009\u226b\u2009\u03bb, the stream function becomes \\({{\\Psi }}=\\frac{1}{2\\pi r}\\) as in quasigeostrophic (QG) flows\u2014atmospheric or oceanic flows coming from gradients in pressure coupled to the Coriolis force36, or driven rotors on the surface of a fluid22. A membrane rotor, therefore, transitions from a point vortex for Euler at small distances to that of QG flow at large distances. Thus, the velocity diverges (decays) as 1/r (1/r2) in the limit of small (large) distances (see Fig.\u00a02B, C). For simplicity, we work primarily in the limit of small distances, r\u2009\u2009\u226a\u2009\u2009\u03bb, since in this limit the dynamics in a membrane converge with those of point vortices (many results still apply to the more general case as shown in the Supplementary Figs.\u00a05 and 6). In what follows, we will use the term point vortices when there are only hydrodynamic interactions and the term rotors when the particles have steric interactions in addition to hydrodynamic ones.\n\nThe dynamics of N point vortices are dictated by Hamilton\u2019s equations,\n\nwhere \\({\\partial }_{i}^{\\perp }=(\\partial {y}_{i},-\\partial {x}_{i})\\), vi is the velocity of the ith vortex, and \u0393i is the circulation (proportional to the magnitude of the torque for rotors, \u0393i\u2009=\u2009\u03c4i/\u03b72D). The Hamiltonian depends on the conjugate variables ri\u2009=\u2009(xi,\u2009yi), [normalized by the circulation \\(\\sqrt{| {{{\\Gamma }}}_{i}| }\\,{{{{{{{\\rm{sgn}}}}}}}}({{{\\Gamma }}}_{i})\\)], i.e., the positions of the vortices12. The symmetries of the Hamiltonian correspond to conservation laws37. In this case, we have symmetries with respect to translation in time, space, and rotation, corresponding to the conservation of the Hamiltonian itself, and of the first and second moments of vorticity, L\u2009=\u2009\u2211i\u0393iri\u2009=\u20090 wlog, and \\(M={\\sum }_{i,j}{{{\\Gamma }}}_{i}{r}_{i}^{2}\\). The conservation of L and M are analogous to the conservation of the center of mass and to the moment of inertia, with sums weighted by circulation instead of mass. From the conservation laws we can deduce that the initial area cannot change dramatically. Particles cannot drift to infinity since the second moment is fixed, nor can they collapse to a point since the Hamiltonian is conserved. These properties are readily observed in simulations. Figure\u00a02D shows typical trajectories of 200 membrane rotors. The initial distribution is random in a predefined finite area, and the dynamics are chaotic38. The final configuration occupies nearly the same region of space as the initial configuration does, and the conservation laws hold to high precision in our simulations, as detailed in \u201cMethods\u201d. This self confining property of vortex dynamics has further consequences, as we now show.\n\nHyperuniformity is the suppression of density-density fluctuations at small wavenumbers (or correspondingly, at large distances)39,40,41. Disordered hyperuniformity can emerge due to short-ranged interactions such as those that arise in sheared suspensions42,43,44, jammed materials45, and for spinning particles46. Here, we will show hyperuniformity emerging from long-ranged interactions, similar to its emergence in sedimentation of irregular objects47. A good way to characterize hyperuniformity is the structure factor, defined as \\(S({{{{{{{\\bf{q}}}}}}}})={N}^{-1}| \\widetilde{\\rho }({{{{{{{\\bf{q}}}}}}}}){| }^{2}\\), where \u03c1(r)\u2009=\u2009\u2211i\u03b4(r\u2009\u2212\u2009ri) is the coarse-grained density. In a hyperuniform material, S(q) goes to zero as a power law at small wavenumbers. We present an argument that a density of point vortices should be hyperuniform due to the conservation of the Hamiltonian. For a density of rotors, the Hamiltonian is given by \\({{{{{{{\\mathcal{H}}}}}}}}[\\rho ({{{{{{{\\bf{r}}}}}}}})] \\sim \\frac{{{{\\Gamma }}}^{2}}{2}\\int {{{{{{{\\bf{dr}}}}}}}}\\int {{{{{{{\\bf{dr}}}}}}}}^{\\prime} \\rho ({{{{{{{\\bf{r}}}}}}}})\\rho ({{{{{{{\\bf{r}}}}}}}}^{\\prime} ){{\\Psi }}(| {{{{{{{\\bf{r}}}}}}}}-{{{{{{{\\bf{r}}}}}}}}^{\\prime} | ).\\) Using the convolution theorem, we find a general relation between the Hamiltonian and the structure factor given by Eq. (1). In the case of point vortices, \\(\\widetilde{{{\\Psi }}}({{{{{{{\\bf{q}}}}}}}})=1/{q}^{2}\\), which gives\n\nFor the integral of Eq. (5) to converge in 2D, S(q)\u2009~\u2009q\u03b1 near the origin, and we must have \u03b1\u2009>\u20090. In other words, an ensemble of point vortices should be hyperuniform. Figure\u00a03B shows an apparent \u03b1\u2009~\u20091.3 scaling for point vortices, consistent with the above argument.\n\nA Snapshots of 10,000 point vortices initially (left) and at steady state (right). Insets show the structure factor, S(q) with a distinct cavity at steady state. B Angular average of the structure factor shown in (A), in a log-log scale with solid line showing a q1.3 scaling. Error bars are standard deviation over ten well-separated timesteps. Inset shows the structure factor of the rotors shown in (C) with increasing hue corresponding to increased concentration \u03d5\u2009=\u2009(0.14,\u20090.24,\u20090.37,\u20090.54). Solid line is the same \u03b1\u2009~\u20091.3 scaling. C Steady-state configurations of 2000 membrane rotors with the corresponding structure factors, showing a transition from disordered hyperuniformity to a hexagonal lattice. Particles are colored according to their local bond orientation parameter \u03c86. For particle j, \\({\\psi }_{6}^{j}={\\sum }_{i}{e}^{6i{\\theta }_{ij}}\\), with the sum taken over nearest neighbors as found by a Voronoi diagram. The table gives ensemble-averaged values, \\({N}^{-1}{\\sum }_{j}{\\psi }_{6}^{j}\\). D A plot of the relative deviation for each particle, with the relative deviation of particle i defined by how far it is displaced from its position at the previous cycle, i.e., \u2223ri(t\u2009+\u2009tcyc)\u2009\u2212\u2009ri(t)\u2223. The cycle time is calculated at steady state as the average time it takes the system to rotate by 2\u03c0. Particles are colored by their relative deviation, from blue to yellow with increasing deviation. The plot at the bottom shows the strobed position of four particles during a time interval of \u0394t\u2009\u2248\u2009115 cycles; the particles in the strobed frame move along Brownian-like trajectories.\n\nUsing simulations, we show that a set of N vortices, uniformly distributed within a radius R, evolves to a disordered steady state with a hidden order visible to the naked eye (compare Fig.\u00a03A left and right). We quantitatively characterize the system in steady state in three ways: (1) The structure factor: At steady-state S(q) shows a distinct cavity, at q\u2009\u2248\u20090, S(q)\u2009\u2192\u20090, for both points vortices (Fig.\u00a03A) and rotors (Fig.\u00a03C). (2) Perturbations: We demonstrate that hyperuniformity is robust under different perturbations, be it in the form of numerical errors, repulsive interactions, or impurities (in the next section). For point vortices, the steady state appears later and later as the timestep is decreased (see Supplementary Fig.\u00a01), suggesting that perturbations are necessary for convergence, here very small but persistent time-stepping errors48. We suspect that perturbations that break the rotational symmetry of the Hamiltonian are required, as testing a smaller ensemble of 1000 point vortices with a symplectic integration scheme, based on the exact solution of pair interactions49, showed little sign of hyperuniformity (see Supplementary Fig.\u00a02 and \u201cMethods\u201d). The observed relaxation toward hyperuniformity is consistent with the critical slowing down reported for other systems40. Adding steric interactions, hyperuniformity appears on a timescale that is independent of the timestep. Moreover, with steric interactions, as the area fraction \u03d5 of the particles is increased, the system transitions from disordered hyperuniform, to an ordered hyperuniform hexagonal lattice at \u03d5\u2009~\u20090.5, as can be seen in Fig.\u00a03C (as a sanity check we show in Supplementary Fig.\u00a03 that a confined, rotationally sheared suspension does not become hyperuniform). The inset of Fig.\u00a03B shows the averaged structure factor where at intermediate area fractions we see Percus\u2013Yevick type features for the structure factor of disks50. (3)The relative deviation: We observe that at late times the ensemble of point vortices rotates almost as a rigid body and each particle nearly goes back to its position at the previous cycle (see Supplementary Fig.\u00a04). The system may seem to have reached an absorbing state, but note that the relative deviation (as defined in Fig.\u00a03D) only measures changes over a single cycle. The motion of vortices over many cycles is still chaotic. Figure\u00a03D shows the trajectory of a few vortices over ~115 cycles, showing Brownian-like dynamics. Similar results were obtained for membrane rotors at an area fraction of \u03d5\u2009=\u20090.1. For sufficiently large area fractions, the system crystallizes and the ensemble rotates as a rigid body where relative deviations are close to zero over many cycles.\n\nWe now show that for mixed populations of fast and slow rotating particles, there is phase enrichment of both populations and hyperuniformity of the fast ones. Consider a mixture of two equally numbered populations (\u03c1l\u2009=\u2009\u03c1h at t\u2009=\u20090) initially placed within the same radius R. \u03c1l rotates slowly with \u0393l\u2009\u2009\u226a\u2009\u2009\u0393h, where \u0393h is the circulation of the second population. Figure\u00a04A shows long-time simulation results for 10,000 point vortices. The two populations behave very differently. The fast vortices remain in a disk of only slightly smaller size than their initial area (Fig.\u00a04B). The slow particle distribution shows a significant expansion. In addition, there is a striking difference when comparing the independently computed structure factors of these two populations, the fast vortices are hyperuniform with S(q)\u2009~\u2009q1.4, whereas the slow ones show no signs of hyperuniformity (Fig.\u00a04C). This difference is dramatic enough to be visible in a cursory examination of the separate distributions; see Fig.\u00a04A.\n\nA Steady-state configuration for ten thousand point vortices of a circulation ratio \u03b3\u2009=\u2009\u0393h/\u0393l\u2009=\u2009128. Each inset shows a close-up view of one of the populations within the same physical region. B Density of the configuration in (A), \u03c1(r), averaged over angle as a function of distance from the center. Note how density fluctuations are suppressed for the high-circulation vortices, as is more clearly observed by the averaged structure factor, S(q), in (C), where the dashed green line shows a\u2009~\u2009q1.4 power law. Error bars are standard deviation over ten well-separated timesteps shown as transparent dots on top of the average result. D The second moment for N\u2009=\u200910,000 vortices. Plotted separately for the high (in green) and low (in black) vortices at steady state as a function of \u03b3 (i.e., increasing \u0393h). E LOSSLESS compression for the two populations showing an increase (decrease) in file size (an estimate of entropy) for the low (high) circulation vortices over a couple of cycles. In blue is the file size for the total system. Solid line is a moving average, time is normalized by an average cycle time tc.\n\nA heuristic model sheds light on this phenomenon (see Supplementary Note\u00a01). Above, each vortex population starts with uniform density within the disk of radius R. Consider as a candidate long-time solution one where each population remains of uniform density, \u03c1l and \u03c1h respectively, and confined within concentric circles of radii Rl and Rh. The circulations of each population, \u0393l and \u0393h, and the system Hamiltonian H and second moment M are fixed by the initial configuration, which restricts the possible values of Rl and Rh. There are two possible solutions. In the first, Rl\u2009=\u2009Rh\u2009=\u2009R. In the second, the radius of the fast vortices slightly decreases to Rh, allowing the slow vortices to expand to a larger radius Rl given by \\({R}_{l}^{2}=(\\gamma +1){R}^{2}-{R}_{h}^{2}\\gamma\\), where \u03b3\u2009=\u2009\u0393h/\u0393l (see Fig.\u00a04D). For large \u03b3, we find that Rh\u2009\u2243\u2009R(1\u2009\u2212\u2009\u03b2/\u03b3), where \u03b2 is a positive prefactor of order 1. The slow vortices radius asymptotes to \\({R}_{l}=R\\sqrt{1+2\\beta }+O(1/\\gamma )\\). The numerical results indicate that the outer radius indeed asymptotes to a finite constant as \u03b3\u2009\u2192\u2009\u221e (see Fig.\u00a04D and Supplementary Figs.\u00a06\u20138).\n\nWhy are solutions with two different radii those observed in our simulations? Such a solution is favored entropically as it maximizes the number of available states. At large \u03b3, the main entropical contribution is volumetric, \\({{\\Delta }}{{{{{{{{\\mathcal{S}}}}}}}}}_{{{{{{{{\\rm{volume}}}}}}}}}=2N\\log ({R}_{{{{{{{{\\rm{final}}}}}}}}}/{R}_{{{{{{{{\\rm{initial}}}}}}}}})\\). Since the high-circulation vortices hardly change radius, Rh \\(\\mathop{\\longrightarrow}\\limits^{\\gamma{\\to}\\infty}\\) R, the change in entropy is coming mainly from the expansion of the low circulation vortices and is given by \\({{\\Delta }}{{{{{{{\\mathcal{{S}}}}}}}_{{{{{{{{\\rm{total}}}}}}}}}}} \\sim N\\log (1+2\\beta ) \\; > \\; 0\\). Coupling the two populations allows one population to expand where before it was bounded. The situation is analogous to depletion interactions, where the net entropy of a system increases by condensing the large particles allowing for the small particles to explore a larger volume51. As Onsager first suggested14, in a bound system, configurational entropy must have a maximum as a function of energy above which demixing of two populations can be observed.\n\nA simple way to estimate the entropy in a system is by using LOSSLESS compression, as suggested by refs. 52,53. Compressing plots of particle positions in a system of 10,000 point vortices with circulation ratio \u0393h/\u0393l\u2009=\u2009128 shows an increase in file size for \u03c1l and a decrease for \u03c1h, while the combined system is increasing, see Fig.\u00a04E.",
+            "section_image": [
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-022-28375-9/MediaObjects/41467_2022_28375_Fig2_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-022-28375-9/MediaObjects/41467_2022_28375_Fig3_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-022-28375-9/MediaObjects/41467_2022_28375_Fig4_HTML.png"
+            ]
+        },
+        {
+            "section_name": "Discussion",
+            "section_text": "We have shown that driven rotors in a membrane or a soap film, like point vortices in an ideal 2D fluid, have geometrical conservation laws which limit their distribution. These conservation laws suggest different possible structural states such as hyperuniformity and phase enrichment. We suspect that a completely pure system of point vortices may never reach hyperuniformity due to a dynamical bottleneck, but have shown that hyperuniformity is robust to two forms of perturbations, whether arising due to numerical errors or steric interactions. For rotors with steric interactions, the unbounded ensemble crystallizes into a hexagonal lattice when the area fraction \u03d5\u2009\u2273\u20090.5 (see also ref. 17). We have limited the discussion to membrane rotors and vortices, but the results for hyperuniformity and phase enrichment hold for other settings in which particles are restricted to a 2D plane, e.g., rotors at the surface of a fluid (see Supplementary Figs.\u00a05 and 6). In fact, while this paper was under review, hyperuniformity was reported in populations of algae swimming in right circles at an interface54.\n\nWhat is especially interesting about our particular BDD system is its potential for experimental realizability, its moment and Hamiltonian structure, and that its near-field interactions (i.e., below the Saffman\u2013Delbruck length) are identical to those of Euler point vortices. Further, the far-field interactions of membrane rotors are identical to those of point vortices of the semi-quasigeostrophic equations36,55,56 used to model atmospheric flows. Thus, to observe the interesting dynamical features we describe, one does not need to go to the atmospheric scale, or cool a fluid to near-zero temperature. In principle, one can simply observe microscopic particles on a soap film, in smectic films, a membrane, or even at the surface of a fluid19,22,57,58.",
+            "section_image": []
+        },
+        {
+            "section_name": "Methods",
+            "section_text": "Simulations were performed in Python. Random initial configurations within the unit disk were found by rejection sampling (points in the unit rectangle were sampled uniformly, transformed to the rectangle [\u22121,\u20091]2, and those whose radius exceeded the target radius were discarded). The initial Hamiltonian H0 and second moment M0 are computed at t\u2009=\u20090, and the relative errors \u03f5H(t)\u2009=\u2009\u2223Ht\u2009\u2212\u2009H0\u2223/H0 and \u03f5M(t)\u2009=\u2009\u2223Mt\u2009\u2212\u2009M0\u2223/M0 are monitored as a measure of fidelity. For simulations of rotors (i.e., with steric repulsion), a 5th order explicit Runge\u2013Kutta method based on the Dormand\u2013Prince scheme59 with a fixed timestep size of \u03b4t\u2009=\u200910\u22127 was used. Long integration times were required for simulations of point vortices, and for these simulations an explicit eighth-order adaptive method based on the Dormand\u2013Prince scheme60,61 was used, with both relative and absolute tolerances set to 10\u22126. The specific implementation of the scheme used was the DOP853 method of scipy.integrate62. For simulations of 10,000 point vortices with \u0393\u2009=\u20092\u03c0, \u03f5H(t)\u2009<\u20091.5\u2009\u00d7\u200910\u22123 and \u03f5M(t)\u2009<\u20094\u2009\u00d7\u200910\u22125 up to t\u2009\u2248\u20091.6\u2009\u00d7\u2009104 cycles, while for simulations with 5000 vortices with \u0393\u2009=\u20092\u03c0 and 5000 vortices with \u0393\u2009=\u2009256\u03c0, \u03f5H(t)\u2009<\u20090.1, and \u03f5M(t)\u2009<\u20094\u2009\u00d7\u200910\u22125 up to t\u2009\u2248\u2009105 cycles. Time is normalized by the average cycle time, tc\u2009\u2248\u20094\u03c02R2/\u2211i\u0393i, where R is the initial radius. We tested running these simulations for 1000 particles with symplectic time integration based on an exact solution of two point vortices49. Simulations were run up to t\u2009~\u20091.4\u2009\u00d7\u2009105 cycles. Due to numerical constraints, we did not run larger ensembles or longer times. At these times, we did not observe clear signatures of hyperuniformity, though there was an indication of a slight decrease in S(q) for low q. Beginning the simulation with a hyperuniform state, and running the symplectic integration over 5\u2009\u00d7\u2009104 cycles preserved hyperuniformity (see Supplementary Fig.\u00a01).\n\nSteric interactions were taken as the repulsive part of a harmonic potential, i.e., for two particles whose centers are distance ri apart, F\u2009=\u2009\u2212\u2009krij if rij\u2009<\u20092a and zero otherwise. The use of a harmonic potential, rather than a sharp step function for hard core particles, provided improved numerical stability and convergence. A large k value was chosen to ensure no overlap between particles, k\u2009=\u2009106, for particles of radius a\u2009=\u20090.01.\n\nTo accurately compute the structure factor S(q) we use a type-1 non-uniform fast-Fourier transform63. Explicitly, points are restricted to a windowing region that is confined entirely within the unit disk. The frequencies \\(\\widetilde{\\rho }({{{{{{{\\bf{q}}}}}}}})\\) are computed for the first 512 modes in each direction, and the average value (i.e., \\(\\widetilde{\\rho }(0)\\)) is set to 0. This results in structure factors in the plane, such as those shown in Fig.\u00a03. Except in those cases where crystallization occurs, these structure factors are azimuthally isotropic. To summarize this information, the angular average over the structure factor was calculated by slicing the result to 1000 equal bins between \\({q}_{\\min }\\) and \\({q}_{\\max }\\) and taking the mean of the results that fell within each slice.\n\nA plot of the positions of the point vortices was compressed using PNG with AGG backend. Each vortex was plotted by a single pixel. The total size of the plots was kept fixed in time. The figure size was chosen to minimize overlap between neighboring vortices but maintaining a computationally accessible file size.",
+            "section_image": []
+        },
+        {
+            "section_name": "Data availability",
+            "section_text": "All data that support the findings of this study can be reproduced using the code available on https://github.com/dbstein/rotor_hyperuniformity64. Data files are also available upon request.",
+            "section_image": []
+        },
+        {
+            "section_name": "Code availability",
+            "section_text": "The code for generating and analyzing the data presented in this paper is available at https://github.com/dbstein/rotor_hyperuniformity64.",
+            "section_image": []
+        },
+        {
+            "section_name": "References",
+            "section_text": "King, R. A review of vortex shedding research and its application. Ocean Eng. 4, 141\u2013171 (1977).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nShelley, M. J. & Zhang, J. Flapping and bending bodies interacting with fluid flows. Annu. Rev. Fluid Mech. 43, 449\u2013465 (2011).\n\nArticle\u00a0\n    ADS\u00a0\n    MathSciNet\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nFj\u00f8rtoft, R. On the changes in the spectral distribution of kinetic energy for two-dimensional, nondivergent flow. Tellus 5, 225\u2013230 (1953).\n\nArticle\u00a0\n    ADS\u00a0\n    MathSciNet\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKraichnan, R. H. Inertial ranges in two-dimensional turbulence. Phys. Fluids 10, 1417\u20131423 (1967).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBernard, D., Boffetta, G., Celani, A. & Falkovich, G. Conformal invariance in two-dimensional turbulence. Nat. Phys. 2, 124\u2013128 (2006).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBehringer, R. P., Meyers, S. D. & Swinney, H. L. Chaos and mixing in a geostrophic flow. Phys. Fluids A: Fluid Dyn. 3, 1243\u20131249 (1991).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nAbrikosov, A. A. The magnetic properties of superconducting alloys. J. Phys. Chem. Solids 2, 199\u2013208 (1957).\n\nMatthews, M. R. et al. Vortices in a Bose-Einstein Condensate. Phys. Rev. Lett. 83, 2498\u20132501 (1999).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKirchhoff, G. Vorlesungen \u00fcber Mathematische Physik: Mechanik (BG Teubner, 1891).\n\nAref, H. Integrable, chaotic, and turbulent vortex motion in two-dimensional flows. Annu. Rev. Fluid Mech. 15, 345\u2013389 (1983).\n\nArticle\u00a0\n    ADS\u00a0\n    MathSciNet\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLin, C. C. On the motion of vortices in two dimensions: i. existence of the Kirchhoff-Routh function. Proc. Natl Acad. Sci. USA 27, 570\u2013575 (1941).\n\nArticle\u00a0\n    ADS\u00a0\n    MathSciNet\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nNewton, P. K. The N-vortex problem, Vol. 145. in Applied Mathematical Sciences (eds Jerrold E. Marsden & Sirovich L.) 1\u2013114 (Springer, 2001).\n\nBogatskiy, A. & Wiegmann, P. Edge wave and boundary layer of vortex matter. Phys. Rev. Lett. 122, 214505 (2019).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nOnsager, L. Statistical hydrodynamics. Il Nuovo Cim. 6, 279\u2013287 (1949).\n\nArticle\u00a0\n    ADS\u00a0\n    MathSciNet\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLenz, P., Joanny, J.-F., J\u00fclicher, F. & Prost, J. Membranes with rotating motors. Phys. Rev. Lett. 91, 108104 (2003).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLenz, P., Joanny, J.-F., J\u00fclicher, F. & Prost, J. Membranes with rotating motors: microvortex assemblies. Eur. Phys. J. E 13, 379\u2013390 (2004).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nOppenheimer, N., Stein, D. B. & Shelley, M. J. Rotating membrane inclusions crystallize through hydrodynamic and steric interactions. Phys. Rev. Lett. 123, 148101 (2019).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGrzybowski, B. A., Stone, H. A. & Whitesides, G. M. Dynamic self-assembly of magnetized, millimetre-sized objects rotating at a liquid-air interface. Nature 405, 1033\u20131036 (2000).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSoni, V. et al. The odd free surface flows of a colloidal chiral fluid. Nat. Phys. 15, 1184\u20131194 (2019).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nNguyen, N. H. P., Klotsa, D., Engel, M. & Glotzer, S. C. Emergent collective phenomena in a mixture of hard shapes through active rotation. Phys. Rev. Lett. 112, 075701 (2014).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLushi, E., Wioland, H. & Goldstein, R. E. Fluid flows created by swimming bacteria drive self-organization in confined suspensions. Proc. Natl Acad. Sci. USA 111, 9733\u20139738 (2014).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nYeo, K., Lushi, E. & Vlahovska, P. M. Collective dynamics in a binary mixture of hydrodynamically coupled microrotors. Phys. Rev. Lett. 114, 188301 (2015).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLushi, E. & Vlahovska, P. M. Periodic and chaotic orbits of plane-confined micro-rotors in creeping flows. J. Nonlinear Sci. 25, 1111\u20131123 (2015).\n\nArticle\u00a0\n    ADS\u00a0\n    MathSciNet\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGoto, Y. & Tanaka, H. Purely hydrodynamic ordering of rotating disks at a finite Reynolds number. Nat. Commun. 6, 5994 (2015).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nShen, Z. & Lintuvuori, J. S. Two-phase crystallization in a carpet of inertial spinners. Phys. Rev. Lett. 125, 228002 (2020).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBililign, E. S. et al. Motile dislocations knead odd crystals into whorls. Nat. Phys. 1\u20137 (2021).\n\nJohn, S. Strong localization of photons in certain disordered dielectric superlattices. Phys. Rev. Lett. 58, 2486 (1987).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nYablonovitch, E. Inhibited spontaneous emission in solid-state physics and electronics. Phys. Rev. Lett. 58, 2059\u20132062 (1987).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMan, W. et al. Isotropic band gaps and freeform waveguides observed in hyperuniform disordered photonic solids. Proc. Natl Acad. Sci. USA 110, 15886\u201315891 (2013).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSaffman, P. G. & Delbruck, M. Brownian motion in biological membranes. Proc. Natl Acad. Sci. USA 72, 3111\u20133113 (1975).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLevine, A. J., Liverpool, T. B. & MacKintosh, F. C. Mobility of extended bodies in viscous films and membranes. Phys. Rev. E 69, 021503 (2004).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSeki, K., Mogre, S. & Komura, S. Diffusion coefficients in leaflets of bilayer membranes. Phys. Rev. E 89, 022713 (2014).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nCamley, B. A. & Brown, F. L. H. Dynamic simulations of multicomponent lipid membranes over long length and time scales. Phys. Rev. Lett. 105, 148102 (2010).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nOppenheimer, N. & Diamant, H. Correlated diffusion of membrane proteins and their effect on membrane viscosity. Biophysical J. 96, 3041\u20133049 (2009).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nOppenheimer, N. & Stone, H. A. Effect of hydrodynamic interactions on reaction rates in membranes. Biophysical J. 113, 440\u2013447 (2017).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHeld, I. M., Pierrehubert, R. T., Garner, S. T. & Swanson, K. L. Surface quasi-geostrophic dynamics. J. Fluid Mech. 282, 1\u201320 (1995).\n\nArticle\u00a0\n    ADS\u00a0\n    MathSciNet\u00a0\n    CAS\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nNoether, E. Invariant variation problems. Transp. Theory Stat. Phys. 1, 186\u2013207 (1971).\n\nAref, H. & Pomphrey, N. Integrable and chaotic motions of four vortices I. The case of identical vortices. Proc. Natl Acad. Sci. USA 380, 359\u2013387 (1982).\n\nMathSciNet\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nTorquato, S. Hyperuniformity and its generalizations. Phys. Rev. E 94, 022122 (2016).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHexner, D. & Levine, D. Hyperuniformity of critical absorbing states. Phys. Rev. Lett. 114, 110602 (2015).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nAriel, G. & Diamant, H. Inferring entropy from structure. Phys. Rev. E 102, 022110 (2020).\n\nArticle\u00a0\n    ADS\u00a0\n    MathSciNet\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWeijs, J. H., Jeanneret, R., Dreyfus, R. & Bartolo, D. Emergent hyperuniformity in periodically driven emulsions. Phys. Rev. Lett. 115, 108301 (2015).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWilken, S., Guerra, R. E., Pine, D. J. & Chaikin, P. M. Hyperuniform structures formed by shearing colloidal suspensions. Phys. Rev. Lett. 125, 148001 (2020).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWang, J., Schwarz, J. M. & Paulsen, J. D. Hyperuniformity with no fine tuning in sheared sedimenting suspensions. Nat. Commun. 9, 2836 (2018).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nTorquato, S. Hyperuniform states of matter. Phys. Rep. 745, 1\u201395 (2018).\n\nArticle\u00a0\n    ADS\u00a0\n    MathSciNet\u00a0\n    CAS\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLei, Q.-L. & Ni, R. Hydrodynamics of random-organizing hyperuniform fluids. Proc. Natl Acad. Sci. USA 116, 22983\u201322989 (2019).\n\nArticle\u00a0\n    ADS\u00a0\n    MathSciNet\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGoldfriend, T., Diamant, H. & Witten, T. A. Screening, hyperuniformity, and instability in the sedimentation of irregular objects. Phys. Rev. Lett. 118, 158005 (2017).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nDai, W. & Shelley, M. J. A numerical study of the effect of surface tension and noise on an expanding Hele-Shaw bubble. Phys. Fluids A: Fluid Dyn. 5, 2131\u20132146 (1993).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nZhang, M. Q. & Qin, M. Z. Explicit symplectic schemes to solve vortex systems. Computers Math. Appl. 26, 51\u201356 (1993).\n\nArticle\u00a0\n    MathSciNet\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nPercus, J. K. & Yevick, G. J. Analysis of classical statistical mechanics by means of collective coordinates. Phys. Rev. 110, 1\u201313 (1958).\n\nArticle\u00a0\n    ADS\u00a0\n    MathSciNet\u00a0\n    CAS\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKardar, M. Statistical Physics of Particles (Cambridge University Press, 2007).\n\nMartiniani, S., Chaikin, P. M. & Levine, D. Quantifying hidden order out of equilibrium. Phys. Rev. X 9, 011031 (2019).\n\nCAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nAvinery, R., Kornreich, M. & Beck, R. Universal and accessible entropy estimation using a compression algorithm. Phys. Rev. Lett. 123, 178102 (2019).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHuang, M., Hu, W., Yang, S., Liu, Q. X. & Zhang, H. P. Circular swimming motility and disordered hyperuniform state in an algae system. Proc. Natl Acad. Sci. USA 118, 1\u20138 (2021).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nFalkovich, G. Symmetries of the turbulent state. J. Phys. A: Math. Theor. 42, 123001 (2009).\n\nArticle\u00a0\n    ADS\u00a0\n    MathSciNet\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nC\u00f3rdoba, D., Fefferman, C. & Rodrigo, J. L. Almost sharp fronts for the surface quasi-geostrophic equation. Proc. Natl Acad. Sci. USA 101, 2687\u20132691 (2004).\n\nArticle\u00a0\n    ADS\u00a0\n    MathSciNet\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nNguyen, Z. H. et al. Crossover between 2D and 3D fluid dynamics in the diffusion of islands in ultrathin freely suspended smectic films. Phys. Rev. Lett. 105, 268304 (2010).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLumay, G., Obara, N., Weyer, F. & Vandewalle, N. Self-assembled magnetocapillary swimmers. Soft Matter 9, 2420 (2013).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nDormand, J. R. & Prince, P. J. A family of embedded Runge-Kutta formulae. J. Comput. Appl. Math. 6, 19\u201326 (1980).\n\nArticle\u00a0\n    MathSciNet\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nPrince, P. J. & Dormand, J. R. High order embedded Runge-Kutta formulae. J. Comput. Appl. Math. 7, 67\u201375 (1981).\n\nArticle\u00a0\n    MathSciNet\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHairer, E., N\u00f8rsett, S. P. & Wanner, G. Solving Ordinary Differential Equations I (2nd Revised. Ed.): Nonstiff Problems (Springer-Verlag, 1993).\n\nVirtanen, P. et al. SciPy 1.0: fundamental algorithms for scientific computing in Python. Nat. Methods 17, 261\u2013272 (2020).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBarnett, A. H., Magland, J. & af Klinteberg, L. A parallel nonuniform fast Fourier transform library based on an \u201cexponential of semicircle\u201d kernel. SIAM J. Sci. Comput. 41, C479\u2013C504 (2019).\n\nArticle\u00a0\n    MathSciNet\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nOppenheimer, N., Stein, D. B., Ben Zion, M. Y. & Shelley, M. J. Hyperuniformity and phase enrichment in vortex and rotor assemblies. https://doi.org/10.5281/zenodo.5815476 (2022).\n\nThomson, W. Vibrations of a columnar vortex. Lond., Edinb., Dublin Philos. Mag. J. Sci. 10, 155\u2013168 (1880).\n\nArticle\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nDownload references",
+            "section_image": []
+        },
+        {
+            "section_name": "Acknowledgements",
+            "section_text": "We thank Haim Diamant for insightful discussions regarding the emergence of hyperuniformity from the conservation laws, to Martin Lenz for suggesting a simple heuristic model of the phase enrichment, and to Enkeleida Lushi. N.O. acknowledges support by the Israel Science Foundation (grant No. 1752/20). M.J.S. acknowledges support by the National Science Foundation under Awards Nos. DMR-1420073 (NYU MRSEC), DMS-1620331, and DMR-2004469.",
+            "section_image": []
+        },
+        {
+            "section_name": "Author information",
+            "section_text": "School of Physics and Astronomy, and the Center for Physics and Chemistry of Living Systems, Tel Aviv University, Tel Aviv, 6997801, Israel\n\nNaomi Oppenheimer\n\nCenter for Computational Biology, Flatiron Institute, New York, NY, 10010, USA\n\nDavid B. Stein\u00a0&\u00a0Michael J. Shelley\n\nLaboratoire Gulliver, UMR CNRS 7083, ESPCI Paris, PSL Research University, 75005, Paris, France\n\nMatan Yah Ben Zion\n\nCourant Institute, New York University, New York, NY, 10012, USA\n\nMichael J. Shelley\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nM.J.S. and N.O. initiated the research. N.O., D.B.S., and M.Y.B.Z. wrote initial versions of the code. Final versions of the code were written by D.B.S.. N.O. and D.B.S. analyzed the data. N.O. developed theory. All authors conceptualized, designed, reviewed the work, suggested ideas, and wrote the manuscript.\n\nCorrespondence to\n                Naomi Oppenheimer or Michael J. Shelley.",
+            "section_image": []
+        },
+        {
+            "section_name": "Ethics declarations",
+            "section_text": "The authors declare no competing interests.",
+            "section_image": []
+        },
+        {
+            "section_name": "Peer review",
+            "section_text": "Nature Communications thanks Juho Lintuvuori, and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available.",
+            "section_image": []
+        },
+        {
+            "section_name": "Additional information",
+            "section_text": "Publisher\u2019s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.",
+            "section_image": []
+        },
+        {
+            "section_name": "Rights and permissions",
+            "section_text": "Open Access  This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article\u2019s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article\u2019s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.\n\nReprints and permissions",
+            "section_image": []
+        },
+        {
+            "section_name": "About this article",
+            "section_text": "Oppenheimer, N., Stein, D.B., Zion, M.Y.B. et al. Hyperuniformity and phase enrichment in vortex and rotor assemblies.\n                    Nat Commun 13, 804 (2022). https://doi.org/10.1038/s41467-022-28375-9\n\nDownload citation\n\nReceived: 13 April 2021\n\nAccepted: 17 January 2022\n\nPublished: 10 February 2022\n\nVersion of record: 10 February 2022\n\nDOI: https://doi.org/10.1038/s41467-022-28375-9\n\nAnyone you share the following link with will be able to read this content:\n\nSorry, a shareable link is not currently available for this article.\n\n\n\n\n                            Provided by the Springer Nature SharedIt content-sharing initiative\n                        ",
+            "section_image": [
+                "https://data:image/svg+xml;base64,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"
+            ]
+        },
+        {
+            "section_name": "This article is cited by",
+            "section_text": "Nature Communications (2023)",
+            "section_image": []
+        }
+    ]
+}

042a95bb963a3fded623677d176d398a6df7372697fadd494fd320558dfef397/metadata.json

@@ -0,0 +1,166 @@
+{
+    "title": "Mapping lesion-specific response and progression dynamics and inter-organ variability in metastatic colorectal cancer",
+    "pre_title": "Mapping Intrapatient Response Heterogeneity and Lesion-specific Relapse Dynamics in Metastatic Colorectal Cancer",
+    "journal": "Nature Communications",
+    "published": "26 January 2023",
+    "supplementary_0": [
+        {
+            "label": "Supplementary Information",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-023-36121-y/MediaObjects/41467_2023_36121_MOESM1_ESM.pdf"
+        },
+        {
+            "label": "Peer review file",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-023-36121-y/MediaObjects/41467_2023_36121_MOESM2_ESM.pdf"
+        },
+        {
+            "label": "Description to Additional Supplementary Information",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-023-36121-y/MediaObjects/41467_2023_36121_MOESM3_ESM.pdf"
+        },
+        {
+            "label": "Supplementary Data 1",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-023-36121-y/MediaObjects/41467_2023_36121_MOESM4_ESM.xlsx"
+        },
+        {
+            "label": "Reporting Summary",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-023-36121-y/MediaObjects/41467_2023_36121_MOESM5_ESM.pdf"
+        }
+    ],
+    "supplementary_1": [
+        {
+            "label": "Source Data",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-023-36121-y/MediaObjects/41467_2023_36121_MOESM6_ESM.xlsx"
+        }
+    ],
+    "supplementary_2": NaN,
+    "source_data": [
+        "https://data.projectdatasphere.org/projectdatasphere/html/access",
+        "/articles/s41467-023-36121-y#Sec19"
+    ],
+    "code": [
+        "https://github.com/zhoujw14/Mapping-Metastasis.git"
+    ],
+    "subject": [
+        "Cancer therapeutic resistance",
+        "Colorectal cancer",
+        "Machine learning",
+        "Metastasis",
+        "Population dynamics"
+    ],
+    "license": "http://creativecommons.org/licenses/by/4.0/",
+    "preprint_pdf": "https://www.researchsquare.com/article/rs-1447896/v1.pdf?c=1648488855000",
+    "research_square_link": "https://www.researchsquare.com//article/rs-1447896/v1",
+    "nature_pdf": "https://www.nature.com/articles/s41467-023-36121-y.pdf",
+    "preprint_posted": "28 Mar, 2022",
+    "research_square_content": [
+        {
+            "section_name": "Abstract",
+            "section_text": "Achieving systemic tumor control across metastases is vital for long-term patient survival but remains intractable in many patients. High intrapatient heterogeneity persists, conferring many dissociated responses across metastatic lesions. Most studies of metastatic disease focus on tumor molecular and cellular features, which are crucial to elucidating the mechanisms underlying intrapatient heterogeneity. However, our understanding of intrapatient heterogeneity on the macroscopic level, such as lesion dynamics in growth, response, and relapse during treatment, remains rudimentary. This study investigated intrapatient heterogeneity through analyzing 116,542 observations of 40,612 lesions in 4,308 metastatic colorectal cancer (mCRC) patients. Despite significant differences in their response and relapse dynamics, metastatic lesions converged on four phenotypes that varied with anatomical site. Importantly, we found that organ-level relapse sequence was closely associated with patient survival, and that patients with the first relapses in the liver often had worse survival. In conclusion, our study provides insights into intrapatient response heterogeneity in mCRC and creates impetus for metastasis-specific therapeutics.",
+            "section_image": []
+        },
+        {
+            "section_name": "Additional Declarations",
+            "section_text": "There is NO Competing Interest.",
+            "section_image": []
+        },
+        {
+            "section_name": "Supplementary Files",
+            "section_text": "Supplementary3.13.pdfMapping Intrapatient Response Heterogeneity and Lesion-specific Relapse Dynamics in Metastatic Colorectal CancerClinicalTrialInformationNCOMMS2209853.xlsxClinical trials included in the analysis",
+            "section_image": []
+        }
+    ],
+    "nature_content": [
+        {
+            "section_name": "Abstract",
+            "section_text": "Achieving systemic tumor control across metastases is vital for long-term patient survival but remains intractable in many patients. High lesion-level response heterogeneity persists, conferring many dissociated responses across metastatic lesions. Most studies of metastatic disease focus on tumor molecular and cellular features, which are crucial to elucidating the mechanisms underlying lesion-level variability. However, our understanding of lesion-specific heterogeneity on the macroscopic level, such as lesion dynamics in growth, response, and progression during treatment, remains rudimentary. This study investigates lesion-specific response heterogeneity through analyzing 116,542 observations of 40,612 lesions in 4,308 metastatic colorectal cancer (mCRC) patients. Despite significant differences in their response and progression dynamics, metastatic lesions converge on four phenotypes that vary with anatomical site. Importantly, we find that organ-level progression sequence is closely associated with patient long-term survival, and that patients with the first lesion progression in the liver often have worse survival. In conclusion, our study provides insights into lesion-specific response and progression heterogeneity in mCRC and creates impetus for metastasis-specific therapeutics.",
+            "section_image": []
+        },
+        {
+            "section_name": "Introduction",
+            "section_text": "Metastasis is the leading cause of cancer mortality1. Unfortunately, antitumor therapies are still designed mostly based on the biology of primary tumors, with little consideration of metastases2,3. Achieving systemic tumor control across metastases is critical for long-term survival but remains intractable in many patients. Some metastases respond highly to treatment while others do not at all, resulting in many dissociated and heterogeneous responses within patients4,5,6,7. Lesion-level response and progression heterogeneity are common in many cancer types, but our understanding of such lesion-level response heterogeneity and its relevance to prognosis remains rudimentary.\n\nMost investigations of metastatic heterogeneity focus on tumor genetic mutations, clonal compositions, or transcriptomics8,9,10. These molecular and cellular characterizations are critical to elucidating the underlying mechanisms of lesion response heterogeneity11,12. However, it is equivalently critical to study metastatic heterogeneity on the macroscopic level, such as distinct lesion dynamics in growth, response, and recurrence during treatment, as well as their potential phenotypic convergence anatomically. These phenotypes would complement molecular and cellular analyses for a holistic view of lesion-specific heterogeneity. The local microenvironment selects tumor phenotypes in response to treatment, leading to heterogeneity across anatomically distinct lesions in terms of response and progression dynamics13,14. Characterizing their phenotypic differences (divergence) or similarities (convergence) could yield insights into tumor ecological features and systemic resistance. The spatiotemporal patterns of response and progression at the lesion or organ level could not only be informative to prognosis, but also could enrich our knowledge of metastasis-to-metastasis interactions and the systemic consequence of regional progression. This study sought to investigate spatiotemporal response heterogeneity through mapping lesion-specific response and progression dynamics in metastatic CRC (mCRC).\n\nColorectal cancer (CRC) is the third leading cause of cancer-related death15. About 20% of CRC patients have distant metastases at diagnosis; the five-year relative survival rate is only 14% for these patients16,17. Lesion-specific response heterogeneity is common in CRC patients treated with either standard chemotherapy alone or in combination with targeted therapy18. We, along with others, have found that high response heterogeneity is associated with worse survival18,19,20,21. Importantly, we also found favorable responses in liver metastases predicted longer patient survival, compared to lesions in the lungs and lymph nodes (LN)18. Characterizing lesion-specific response heterogeneity in mCRC is therefore valuable for prognosis and therapies.\n\nIn this study, to map the lesion-level response and progression patterns in mCRC, we first apply a mathematical model to capture tumor growth dynamics in 4,308 mCRC patients. Next, individual lesion-specific response and progression probabilities are mapped to predict their phenotypic divergence and convergence across anatomical sites. Last, we apply a machine learning approach to analyze the progression sequence across organs and its relevance to long-term patient survival. The spatiotemporal patterns of response and progression at the organ level could infer tumor evolution in space and time, affording more biological hypotheses. Our study provides insights into lesion-specific phenotypic heterogeneity in mCRC and yields substantial implications for designing metastasis-specific therapeutics.",
+            "section_image": []
+        },
+        {
+            "section_name": "Results",
+            "section_text": "To evaluate metastatic response and progression dynamics in mCRC, we collected longitudinal radiographical measurements of metastatic lesions in colorectal cancer (CRC) patients from Project Data Sphere. In total, 4308 patients with 40,612 lesions from eight Phase III trials were included. The inclusion and exclusion criteria are presented in Fig.\u00a01a. The distribution of lesion number across organs is shown in Fig.\u00a01b. The total target lesions were 19,180 with 94,174 radiographic measurements by CT scan, and there were 18,594 nontarget lesions and 2838 new lesions that had records of response status, appearance time, and anatomical site, and these lesions did not have longitudinal radiological measurements. Additional information including patients\u2019 demographic and clinical characteristics (e.g., age, gender, race, body mass index [BMI], tumor type, treatment history, RECIST response, and KRAS status), progression-free survival (PFS) and overall survival (OS) are reported in Table\u00a01. We also included the tumor longitudinal measurements in a head and neck squamous cell carcinomas (mHNSCC) trial for an external validation. The data was also from Project Data Sphere with similar inclusion/exclusion criteria as in the CRC data.\n\na CONSORT diagram of metastatic colorectal cancer data inclusion and exclusion criteria. b The number of all lesions (target, non-target, and new) and target lesions across organs. GR Genitourinary and Reproductive, CNS Central nervous system, GI Gastrointestinal tract, LN Lymph nodes. Source data are provided as a Source Data file.\n\nThe tumor growth dynamics of 19,180 target lesions with 94,174 radiographical measurements were fitted using a tumor growth model22. There are three dynamic parameters in the model: the regression (tumor-killing) rate Kd, the fraction of non-responding (or resistant) cells F, and the progression (or regrowth) rate Kg (Fig.\u00a02a). The model was optimized using a nonlinear mixed effect (NLME) modeling approach, which allows the estimation of three dynamic parameters at the individual level and their inter-lesion variance within the population. Overall, the model adequately recapitulated the longitudinal profiles of tumor radiographic measurements for each lesion. The goodness-of-fit and model visual predictive check plots, as well as representative individual fittings, show good model predictive performance (Supplementary Fig.\u00a01).\n\na Schematic plot of tumor growth model. b Box plots of model parameters Kd, F and Kg across organs. Significance was calculated using Kruskal-Wallis tests. The box extends from the 25th to 75th percentiles and the line in the middle is plotted as the median. The whiskers are drawn down to the 10th percentile and up to the 90th percentile. Points below and above the whiskers represent individual lesions. c The correlations between model parameters. d The correlations between model parameters and tumor baseline volume. The size of the dots represents lesion number (reported in b). The dashed lines with gray area are the linear regression with 95% confidence interval. The correlation coefficients and p-values were calculated using two-tailed Pearson correlation tests. Source data are provided as a Source Data file.\n\nPopulation estimates and inter-lesion variances in tumor dynamic parameters are summarized in Supplementary Table\u00a01. The parameters for individual lesions significantly differed across organs (p\u2009<\u20090.0001, Fig.\u00a02b). Among all metastases, lesions in the bone exhibited the lowest tumor shrinkage (1-F), while lesions in the genitourinary and reproductive (GR) system had the fastest progression rates (Kg), and kidney lesions showed the lowest regression rates (Kd). Among three most abundant metastatic sites (liver, lung, and LN), lesions in the liver showed the highest tumor shrinkage but the fastest progression rates, suggesting the unique response feature of liver lesions.\n\nHigher fractions of treatment-resistant cell (F) is associated with slower rates of regression (Kd, r\u2009=\u2009\u22120.69, p\u2009=\u20090.0014) and faster rates of progression (Kg, r\u2009=\u20090.53, p\u2009=\u20090.03, Fig.\u00a02c). Progression rates seemed to be independent of regression rates (Fig.\u00a02c). Remarkably, no significant correlations were observed between baseline tumor burden and all tumor dynamic parameters (Fig.\u00a02d). Large tumor burden, on the individual lesion level, did not necessarily confer slow regression rates, high treatment-resistant fractions, or slow progression rates, implying that tumor burden at baseline is not a robust prognostic factor in mCRC23,24,25. Notably, metastatic lesions under antibody targeted therapy (bevacizumab and/or panitumumab) plus chemotherapy (FOLFOX or FOLFIRI), compared to standard chemotherapy alone, showed significantly deeper response (effect size\u2009=\u20090.43) and lower progression rates (effect size\u2009=\u20090.26), but had a moderate effect on tumor regression rates (effect size\u2009=\u20090.06, Supplementary Fig.\u00a02).\n\nThe tumor growth model predicted the longitudinal profiles of response and progression for each target lesion. Time to response and progression were then derived as the duration from the start of treatment to the time of response or progression per RECIST v1.126, respectively. We compared our model-predicted lesion response and progression rates with patient response status per RECIST 1.1. Complete (CR) or partial responders (PR) had shorter time to response and more extended duration before progression than patients with stable disease (SD). (Supplementary Fig.\u00a03).\n\nWe integrated the time to response for both target and non-target lesions and the time to progression for all lesions, including the new ones, into random effect Cox proportional models27. The Cox model predicted the relative probabilities of lesion response or progression at the organ level. We tested the covariate effects of treatment, age, BMI, gender, race, surgical history, and line of therapy (first or second line) in the random effect Cox proportional model. The covariate effects were summarized in Supplementary Fig.\u00a04a and Supplementary Fig.\u00a04b. Significant covariates were included in the final model. Of note, treatment, surgical history, and line of therapy were included as significant covariates in the progression model, and treatment, race, and line of therapy were included in the response model.\n\nWith these covariate effects, we could more accurately estimate organ-intrinsic response and progression characteristics. The hazard ratios for the response and progression across organs are shown in Fig.\u00a03a and Fig.\u00a03b. With abdominal lesions as the reference, metastatic lesions in the liver were most likely to respond to treatments, whereas lesions in the brain/central nervous system (CNS) were least likely (Fig.\u00a03a). Lesions in the gastrointestinal (GI) system, skin, and bone were significantly less likely to respond than abdominal lesions. Lesions in the spleen, lung, and peritoneum showed comparable responses. The probability of progression also differed greatly across anatomical sites (Fig.\u00a03b). The metastatic lesions with the highest likelihood of progression were those in the brain/CNS, GR system, and liver, while lesions in the GI system, and regional and distal LNs were least likely.\n\na, b Data are presented as the hazard ratio estimates with 95% confidence interval by organs on lesion response and progression in colorectal cancer patients. c, d are the anatomical charts of organ-specific response and progression hazard ratios in metastatic colorectal cancer (mCRC) and metastatic head and neck squamous cell carcinomas (mHNSCC). e, f Data are presented as the hazard ratio estimates with 95% confidence interval in response and progression by organs stratified on treatments in mCRC. P-values in a, b, e, and f were calculated by two-sided likelihood ratio tests. TAR\u2009+\u2009Chemo, antibody targeted therapies (bevacizumab or panitumumab) plus chemotherapy; Chemo Alone, chemotherapy alone. Source data are provided as a Source Data file.\n\nWe then integrated organ-specific response and progression probabilities to investigate their potential phenotypic convergence across anatomical sites. As in Fig.\u00a03c, an anatomical chart of organ-specific response and progression probabilities was created based on their relative hazards in the Cox model. Four types of phenotypic features emerge in CRC-metastatic organs defined by their associated lesions\u2019 likelihood of response and progression. Notably, bone and brain lesions had low response and high progression probabilities (low-high phenotype), while liver lesions had high probabilities of both response and progression (high-high phenotype). Patients with these metastases, particularly those with low-high phenotype, had much worse survival outcomes than those with other phenotypes (OS median 378 days vs. 561 days, p\u2009<\u20090.0001, Supplementary Fig.\u00a05a). On the other side, metastatic lesions in the lung and LN showed high response and low progression probabilities (high-low phenotypes). Patients who have metastases in high-low phenotype organs only tend to have a better prognosis than patients with other phenotypic metastases do (OS median 770 days vs. 524 days, p\u2009<\u20090.0001, Supplementary Fig.\u00a05b).\n\nInterestingly, most metastatic lesions with high progression probabilities tend to occur in organs known to have immunosuppressive microenvironments, such as the liver, bone, and brain/CNS28,29,30,31. To show the anatomical pattern of lesion response and progression is beyond tumor biology and is more closely related to tissue microenvironments, we performed a validation analysis in a biologically distinct cancer type, head and neck squamous cell carcinomas (mHNSCC), to see whether a similar anatomical chart exists (Fig.\u00a03d). A total of 393 patients with 1892 lesions were analyzed, including eleven metastatic organs (Supplementary Fig.\u00a06a, b). Patients\u2019 demographics are reported in Supplementary Table\u00a02. We built random effect Cox proportional models to estimate hazard ratios across organs in mHNSCC, as we did in mCRC. Treatment, age, radiation history, and body surface area (BSA) were included as covariates in the progression model. Treatment and race were considered as covariates for the response model. The organ-specific hazard ratios for lesion progression and response were shown in Supplementary Fig.\u00a06c, d. In mHNSCC, metastases in the liver, and brain also showed high progression potential, in line with what we observed in mCRC. Metastatic lesions in the LNs exhibit a high-low phenotype, consistent with mCRC. Similar anatomical charts across cancer types suggest that organ-intrinsic microenvironmental factors, such as the local physical and immunological components, could be key modulators to the mechanisms underlying the probabilities of tumor response and progression. However, further investigations are warranted.\n\nIn mCRC, treatment effects on organ-specific responses were also investigated. For simplicity, treatments were divided into two groups, chemotherapy alone and in combination with antibody targeted therapy. The combined antibody targeted therapies are either panitumumab or bevacizumab, or both. Surprisingly, combination with the antibody targeted therapies did not significantly influence organ-specific response probabilities (Fig.\u00a03e), suggesting limited direct cytotoxic effects of antibody-based targeted therapies. Notably, the primary therapeutic benefit of antibody targeted therapies was to decrease lesion progression (Fig.\u00a03f). Progression hazards significantly decreased in most metastatic organs except for the skin, brain/CNS, spleen, and kidney. Taken together, antibody-targeted therapies showed effect primarily on decreasing lesion progression and had limited influence on the lesion response probability. Interestingly, cytotoxic chemotherapies did not seem to influence lesion progression patterns, and high-progression organs in Fig.\u00a03c also had remained to have high progression probability during cytotoxic chemotherapies (Fig.\u00a03f), reiterating a critical role for local tissue environments in long-term tumor control.\n\nWe built a k-means unsupervised clustering model32 to cluster patients based on their organ-level lesion progression sequence and investigate their relevance to patient survival. Elbow sum of square33 (Supplementary Fig.\u00a07a) and Silhouette score34 (Supplementary Fig.\u00a07b) were calculated to determine the optimal k in the final classification. Akaike information criterion (AIC) and Bayesian information criterion (BIC)35 were also applied to find optimal k (Supplementary Fig.\u00a07e). K\u2009=\u20094, 5, and 6 showed similar performance in the model evaluation metrics. The patient survival profiles were also compared using concordance, and the select k\u2009=\u20095 showed the finest separation of patient survival, resulting in distinct features of progression for each group. Five groups of patients were thus identified with distinct patterns of organ-specific progression sequences and were stratified by progressive organ number and first-progressive organ: Mono-Organ (n\u2009=\u20091425), Hetero-Organ (n\u2009=\u2009801), Lung-First (n\u2009=\u2009577), Liver-First (n\u2009=\u20091194), and the Other-First (n\u2009=\u2009888) groups. The clinical demographics and baseline information of each group are summarized in Supplementary Table\u00a03.\n\nOrgan-level progression sequence is significantly correlated with long-term patient survival (p\u2009<\u20090.0001, Fig.\u00a04b). As expected, patients with multiple organ progression had worse survival than patients with only one organ progression (OS median Hetero-Organ 385 days vs. Mono-Organ 653 days). Remarkably, despite comparable number of baseline metastases, patients whose first progression was in the liver had a much worse prognosis than those whose first progression was in lungs or other sites (OS median Liver-First 450 days vs. Lung-First 679 days vs. Other-First 581 days, Fig.\u00a04b and Supplementary Fig.\u00a08). This is consistent with earlier observations (Fig.\u00a03c) that lesions in the lung had high-low phenotype that is often associated with good patient prognosis. Patients with tumor progression first in the liver had faster subsequent progression than patients whose progression occurred in lungs or other sites, suggesting that progressive lesions in the liver may have high systemic consequences (p\u2009<\u20090.0001, Fig.\u00a04c). It also aligns with our previous finding that the response of liver lesions to treatments strongly predicted patient survival18.\n\na Patients were clustered into five groups based on their lesion progression sequence. The column labels are the progression sequence. Color of the heatmap represents the log10 scale of patient number (all plus one to avoid zero values). b Kaplan-Meier curves of clustered patients overall survival. c Box plots of the first lesion progression time (1st), time between first and second progression (2nd-1st), time between second and third progression (3rd-2nd), time between third and fourth progression (4th-3rd), and the average progression time in Lung-First (n\u2009=\u2009577), Other-First (n\u2009=\u2009639), and Liver-First (n\u2009=\u2009930). The box extends from the 25th to 75th percentiles and the line in the middle is plotted as the median. The whiskers are drawn down to the 10th percentile and up to the 90th percentile. Points below and above the whiskers represent individual lesions. P-values in c were calculated by two-sided Dunn\u2019s multiple comparisons. Source data are provided as a Source Data file.\n\nNext, we performed k-means unsupervised clustering in the Hetero-Organ group to further investigate progression patterns in patients with extensive metastases progression. Four groups of patients were optimally clustered (Supplementary Fig.\u00a07c, d, f), and one distinctive feature among these clusters was the progression order of liver lesions (Supplementary Fig.\u00a09a). Despite similar baseline metastases, patients with first or second progression occurring in the liver had worse survival than those with early progression occurring in other organs (Supplementary Fig.\u00a09b), but two groups showed no significant difference in subsequent time to progression (Supplementary Fig.\u00a09c). This observation further underlines the importance and systemic consequence of liver lesions to tumor response and resistance.\n\nWe compared the progression sequence in patients under different treatments (chemotherapy alone vs. combination with antibody targeted therapy). In patients with Liver-First, Lung-First or Other-First progression patterns, antibody targeted therapies significantly improved patient overall survival (p\u2009<\u20090.0001, Fig.\u00a05a). However, neither the proportion of patients with each progression cluster (Fig.\u00a05b) nor the sequence of progression across metastatic organs were significantly changed by antibody targeted therapies (Fig.\u00a05c\u2013e). Tumor progression in the GR and pancreas occurred slightly earlier in antibody targeted therapy, which did not seem to translate meaningful difference in patient survival. Despite similar sequences, patients under antibody targeted therapies had significantly slower first and second progression, even without significant difference in the third or later progression (Fig.\u00a05f, g). The average time to progression were much longer in combination therapy compared to chemotherapy alone.\n\na Lung-First, Other-First and Liver-First patients overall survival stratified by treatments. b Lung-First, Other-First and Liver-First patient proportions by treatments. c\u2013e are patient progression sequences stratified by treatments. f\u2013h are the box plots of the first lesion progression time (1st), time between first and second progression (2nd-1st), time between second and third progression (3rd-2nd), time between third and fourth progression (4th-3rd), and the average progression time by treatments of the groups in c\u2013e. N\u2009=\u2009307/n\u2009=\u2009440/n\u2009=\u2009335 patients from TAR\u2009+\u2009Chemo and n\u2009=\u2009270/n\u2009=\u2009490/n\u2009=\u2009304 patients from Chemo Alone were included in f\u2013h. The box extends from the 25th to 75th percentiles and the line in the middle is plotted as the median. The whiskers are drawn down to the 10th percentile and up to the 90th percentile. Points below and above the whiskers represent individual lesions. P-values in f\u2013h were calculated by two-sided Kruskal-Wallis tests. TAR\u2009+\u2009Chemo, antibody-targeted therapies (bevacizumab or panitumumab) plus chemotherapy; Chemo Alone Chemotherapy alone. Source data are provided as a Source Data file.\n\nIn patients with the Hetero-Organ pattern, antibody targeted therapies did not meaningfully improve overall survival (Supplementary Fig.\u00a010a) compared to chemotherapy alone, and the proportions of patients in each subcluster were similar between the two treatment groups (Supplementary Fig.\u00a010b). Patients\u2019 progression patterns and lesion time to progression were largely comparable, especially for those who had early liver lesion progression (Supplementary Fig.\u00a010c\u2013h). Similarly, antibody targeted therapies did not influence lesion progression sequence. Overall, the primary therapeutic benefit of antibody targeted therapies was to delay progression in patients with few (<\u20094) metastatic organs, but not in those with broad metastases.\n\nIn order to predict patient progression sequence at the time of diagnosis, we built a gradient boosting model using patient baseline characteristics and metastases profiles36. The model parameters are in Supplementary Table\u00a04. The area under the receiver operating characteristic (ROC) curve of the testing data was 0.91, which indicated fair performance (Supplementary Fig.\u00a011a). The model could predict Mono-Organ and Hetero-Organ groups better than Lung-First, Liver-First, and Other-First groups with higher area under the ROC curve. This indicates that more follow-up information about tumor early response is imperative to predict the progression sequences of the latter three groups (Supplementary Fig.\u00a011b).",
+            "section_image": [
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-023-36121-y/MediaObjects/41467_2023_36121_Fig1_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-023-36121-y/MediaObjects/41467_2023_36121_Fig2_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-023-36121-y/MediaObjects/41467_2023_36121_Fig3_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-023-36121-y/MediaObjects/41467_2023_36121_Fig4_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-023-36121-y/MediaObjects/41467_2023_36121_Fig5_HTML.png"
+            ]
+        },
+        {
+            "section_name": "Discussion",
+            "section_text": "Metastasis is responsible for the majority of cancer-related mortality. Unfortunately, systemic tumor control across metastases remains intractable in many patients. This study investigated inter-lesion heterogeneity by analyzing response dynamics of 40,612 lesions to multiple types of treatment in 4308 mCRC patients. Without molecular characterizations of metastases, we focused on the phenotypic features associated with lesion response and progression dynamics as well as the anatomical distributions of these features. Our analyses yielded several intriguing findings. First, metastases differed considerably in their response to treatment, with the tumor shrinkage fraction positively correlating with regression rate and negatively correlating with progression rate. Second, metastatic lesions within the same organ exhibited congruent response and progression dynamics, converging upon four organ-level phenotypes. Metastatic lesions in the liver exhibited high response and high progression probabilities (high-high phenotype), while lesions in the bone and brain/CNS had low response and high progression probabilities (low-high phenotype). These phenotypes appear to be determined by tumor local microenvironments and go beyond tumor biology, as we found a similar pattern in a biologically distinct tumor type, mHNSCC. Third, we found that organ-level progression sequence was closely associated with patient survival, and patients with the first progression in the liver had worse survival outcomes compared to patients with first progression in other organs.\n\nThis study quantified the degree of inter-lesion heterogeneity by modeling tumor regression and progression dynamics. By assuming first-order regression of drug-sensitive cancer cells (log-kill hypothesis), the empirical model adequately recapitulated the longitudinal size measurements on the lesion level. The first-order regression implies that drug-sensitive cancer cells may have only one rate-limiting step on the path to cell death37. Large tumors are often expected to have tumor regression potentially deviating from strict first-order kinetics because of their non-uniform drug distributions inside the tumor or only the surface tumor cells being actively proliferating and sensitive to treatments38,39,40. Our analyses did not find evidence to support these speculations. Baseline tumor burden did not correlate with tumor regression rates, restating the first-order regression. In contrast, tissue microenvironment matters more than the lesion size to tumor response to treatments. Despite large sizes, metastatic lesions in the liver had relatively higher regression rates than lesions at other organ sites.\n\nTumor progression rates showed much higher variation during treatment than their associated regression rates and accounted the majority of intrapatient heterogeneity. Lesion progression time, i.e., lead time to progression, was more closely determined by the progression rates rather than the response rate. This finding was in line with Stein et al., who reported that tumor progression rate was a stronger predictor of patient survival41. If validated prospectively, the progression rates would offer more appropriate efficacy endpoints in clinical trials than the current ones that focus on the early tumor response and shrinkage, such as overall response rate and best of response.\n\nAntibody therapies significantly increased tumor shrinkage fraction and showed lesion progression but did not considerably affect tumor-shrinking rates. These observations indicate that the therapeutic benefit of combined antibody therapies is primarily from tumor regrowth suppression rather than direct tumor killing effect. In renal cell carcinomas, bevacizumab significantly reduced tumor regrowth rate, which could become more apparent after progression, in line with our observations in mCRC42.\n\nMetastatic lesions with a lower fraction of non-responding tumor cells (F) also had a slower lesion progression rate. The small fraction of non-responding (resistant) tumor cells prior to treatment implies their low fitness compared to responding (sensitive) populations, which led to low tumor regrowth rates after progression. Interestingly, metastatic lesions in the liver appear to behave differently; they had higher probability to respond, but also faster lesion progression rates than lesions in the LN and lungs, suggesting unique ecological properties of liver lesions. Our analyses highlight the importance of tissue microenvironments to metastatic phenotypes. Metastatic lesions with higher responses were typically found in the liver, spleen, LN, and lungs. These organs are known to have discontinuous or fenestrated endothelial membranes, which may lead to higher drug exposure, potentially conferring high treatment responses43,44. In contrast, the organs bearing poorly-responding lesions are usually those with continuous endothelial membranes and thus more limited drug distribution, such as the muscle and brain/CNS45,46,47,48. Some organs that bear poorly responding metastatic lesions, such as kidney and muscle, have relatively dense tissue matrices. Dense tissue matrix could restrict the growth rate of metastatic lesions49,50 and also render them less responsive to cytotoxic chemotherapy probably due to limited drug distribution51,52.\n\nOn the other hand, organ-specific progression probabilities seem to closely relate to organ microenvironments. Metastatic lesions with higher progression potentials were often found in the liver, bone, and brain/CNS, which either are immune-privileged or tolerogenic organs13,14,28,29,30,31. Interestingly, high lesion progression in these organs also occurred during cytotoxic chemotherapies even though such therapies are usually not considered to involve the immune function for therapeutic effect (Fig.\u00a03f). Higher containment effect of tumor progression in immunocompetent organs implies the critical role anticancer immunity plays in long-term tumor control. Patients with highly progressive lesions, such as lesions in the liver and bones, had much worse survival outcomes and these patients likely required more effective and targeted therapeutics.\n\nTumor progression is a serious impediment to cancer treatment, but organ-level progression patterns remain poorly characterized. We found that early tumor progression in the liver, compared to early progression in other sites, predicts worse patient survival and more rapid subsequent progression. The liver\u2019s anatomical location, which may serve as a trafficking hub for CRC cells to spread to other organs, possibly underlies this finding53. By modeling large autopsy data sets in mCRC, Newton et al. highlighted that liver metastases could serve as tumor spreaders54, and that there are multidirectional paths of tumor spread during progression54,55. Although we did not estimate transit probabilities from site to site, we speculate it is likely that early progression in liver metastases could expand systemic peripheral tolerance and promote more frequent and rapid progression in other organs. Our population-level analysis supports this speculation and shows that liver metastases were often associated with a more pronounced tumor spread in the body.\n\nThe primary therapeutic benefit of antibody-targeted therapies was to delay tumor progression and systemic relapses, without clear preferential effect on any organ-specific metastases. As such, antibody therapies did not seem to affect lesion progression sequences, and the fraction of patients with the first progression in the liver were largely comparable to chemotherapy alone. Unfortunately, in patients with multiple progressive metastases (>\u20094 progressive organs), the therapeutic benefit of antibody therapies is minimal, and more effective treatments remain sorely needed for advanced patients with extensive metastases.\n\nOur study has limitations. First, the size of metastases was measured by radiological CT scans, but lesions in the bone are generally hard to assess, which could result in quantification bias and variability. Second, we also should note that patients in our dataset were from randomized control trials, who sometimes have different demographic characteristics with real-world patients. The average age of the patients in our dataset was 60.2, younger than the average age of diagnosis in real world patients15. Third, we pooled patients from multiple trials and these patients had very discrete surgical and treatment histories. Even though we controlled our analysis by considering these factors as covariates, direct trial-to-trial comparisons should be prevented. Machine learning has therefore identified distinct patterns for tumor progression, but future validation of these findings will be through acquisition and study of further independent datasets.\n\nIn conclusion, we quantified lesion response and progression heterogeneity by modeling the longitudinal size measurement of metastatic lesions. This study provided a broad characterization of phenotypic heterogeneity across metastatic lesions in mCRC, which could complement conventional molecular and cellular analyses to and promote a more comprehensive view of lesion-specific response heterogeneity and yield substantial implications for metastasis-targeting therapies.",
+            "section_image": []
+        },
+        {
+            "section_name": "Methods",
+            "section_text": "Multiple mCRC studies with longitudinal measurements of individual metastatic tumor information were included for the analyses. All datasets are accessible in Project Data Sphere, an open-access platform that aggregates cancer clinical trial data from biopharmaceutical companies, academic medical centers, and government organizations(https://www.projectdatasphere.org/). Patients under one of the following conditions were excluded: (1) no target lesion longitudinal measurements; (2) baseline tumor size measured more than 12 weeks before the treatment. Patients\u2019 demographics and survival information were collected if applicable. The size and anatomical site about target/non-target lesion and occurring time and anatomical sites of new lesions were all recorded and analyzed if any. Data from a Phase III study on panitumumab plus chemotherapy for mHNSCC were also collected from Project Data Sphere and were analyzed with same workflow for validation purpose. Data were processed in in R-4.1.0 and RStudio 2022.07.1 dplyr package. The clinical trials information was provided in Supplementary Data\u00a01.\n\nAll study protocols were approved by institutional review boards at each participating center, including the clinical trial review boards from Amgen Inc., Pfizer Inc., Sanofi Inc., and AstraZeneca Inc. All patients have been provided written informed consent before study-related procedures were performed. All data sharing plans have been approved by the data sponsors.\n\nThe longest diameter was converted to volume assuming the ellipsoidal shape of tumor (1) and the ratio of the tumor long versus short axis as 1.3156. An empirical tumor growth model (2) was used to recapitulate lesion-specific tumor growth dynamics.\n\nV is the tumor volume, V0 is the tumor baseline volume, t is the time. The model has three parameters for estimation: F is the fraction of non-responding tumor cells, with 1-F as the response depth; Kg is the progression rate and Kd is the regression rate. We fitted the model for all target lesions simultaneously using the Non-Linear Mixed Effect (NLME) method in Monolix2020R1 Lixoft. Stochastic approximation expectation-maximization (SAEM) algorithm57 was applied to search global optimum in the estimation. M3 method58 was applied for quantifying size below the quantification of limit (<\u2009200\u2009mm3)59. In the NLME method, the model parameters are described in (3)-(5).\n\nwhere \\(\\theta\\) is the population typical value, and \u03b7 is the random effect with a log-normal and logit-normal\u00a0distribution describing the difference between individuals and population average for each lesion j. Proportional error model was assumed. The initial values of Kg, Kd and F were 0.01 day\u22121, 0.01 day\u22121, and 0.1 (unitless).\n\nOf note, the volumetric conversion would make the thresholds for response and progression different from dimensional metrics. However, our model system assumes first-order dynamics of tumor regression and progression, and the response and progression sequence at the organ level remains unchanged regardless of the use of volumetric or dimension metrics.\n\nTumor growth dynamic parameters were further taken to predict the longitudinal profiles of response and progression for each target lesion. The longitudinal response and progression status for each target or non-target lesion were determined per RECIST V1.126. Target lesion time to response (when the lesion volume decreases \u2265\u200920% from baseline) and time to progression (when the lesion volume increases \u2265\u200930% from tumor nadir or at least 200\u2009mm3 increase from nadir) were derived using tumor growth model with NLME-estimated parameters on the individual lesion level. Non-target lesions responded when partial response or complete response was firstly observed during the treatment and progressed when progressive disease appeared in tumor evaluation. The time to progression for new lesions were defined as the detection time.\n\nCox proportional models were built to estimate lesion response and progression probabilities across organs and treatments in R-4.1.0 and RStudio 2022.07.1 coxme package. Inter-patient variability was adjusted in the Cox models as random effect. The potential covariate effects of treatment, line of therapy, age, gender, race, BMI, and surgical history were tested in mCRC patients and significant covariates (p-value <\u20090.05) were selected in the final model. For mHNSCC, we tested treatment, line of therapy, age, gender, race, BSA, and surgical/radiation history as potential covariates. Lesions without progression or response during the treatment were labeled as censored by the last day of that patient in the trial. New lesions were considered only in the progression hazard estimation.\n\nWe used the k-means clustering algorithm32 to categorize all the patients based on their organ progression sequence in Spyder (Python 3.8) in Anaconda using the SCIKIT-LEARN 1.0.2 software package. Elbow method33, Silhouette score34, AIC and BIC35, were applied to find optimal k. The Elbow method selects optimal k based on the inflexion point of the performance curve. The Silhouette method is based on the similarity of a data-point to its own identified cluster and dissimilarity to other clusters. The AIC and BIC provide complementary measures that balance model complexity and predictive performance. The optimal k was also selected to yield adequate separation of patient survival and interpretable features of lesion progression.\n\nGradient Boosting algorithm36 was applied to build a progression pattern predictive model in Spyder (Python 3.8) in Anaconda using the SCIKIT-LEARN 1.0.2 software package. The research samples were randomly divided into training and testing sets at a ratio of 4:1, with rotation of the test dataset to implement 5-fold cross validation. The initial value of the hyperparameters used in this model was determined by parameter grid search, using 5-fold cross-validation and F1-score as a metric (Supplementary Table\u00a04). The model outcome is the patient progression sequence classified in k-means clustering algorithm. Model inputs included patient clinical and demographic characteristics, as well as the baseline metastatic profiles, including the metastatic organs, metastatic numbers, metastatic target lesion baseline volume. Continuous predictors were normalized and categorical predictors were transformed to dummy variables using OneHotEncoder package softmax function. Performance index accuracy, precision, recall rate and area ROC curves were used to evaluate model performance.\n\nComparisons of continuous variables were performed using the two-tailed Mann\u2013Whitney test or Kruskal\u2013Wallis test. Multiple comparisons were adjusted by Dunn\u2019s test. PFS (defined as the start of therapies until RECIST-defined progression or death) and OS (defined as the start of therapies until patient death) among the groups were depicted using Kaplan\u2013Meier curves and compared using log-rank tests. All the statistical tests were performed in GraphPad Prism 9. No statistical method was used to predetermine sample size. Data were included or excluded based on pre-established criteria. No randomization was involved in our analysis and the investigators were not blinded to allocation during data analyses and outcome assessments.\n\nFurther information on research design is available in the\u00a0Nature Portfolio Reporting Summary linked to this article.",
+            "section_image": []
+        },
+        {
+            "section_name": "Data availability",
+            "section_text": "The raw clinical data that support the findings of this study are available in the Project Data Sphere, https://data.projectdatasphere.org/projectdatasphere/html/access. Access can be acquired on the Project Data Sphere website. The processed data generated in the study are provided in the Source Data file with this paper.\u00a0Source data are provided with this paper.",
+            "section_image": []
+        },
+        {
+            "section_name": "Code availability",
+            "section_text": "The modeling and machine learning algorithms codes were deposited at https://github.com/zhoujw14/Mapping-Metastasis.git.",
+            "section_image": []
+        },
+        {
+            "section_name": "References",
+            "section_text": "Welch, D. R. & Hurst, D. R. Defining the hallmarks of metastasis. Cancer Res. 79, 3011\u20133027 (2019).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nEccles, S. A. & Welch, D. R. Metastasis: Recent discoveries and novel treatment strategies. Lancet. 369, 1742\u20131757 (2007).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nAnderson, R. L. et al. A framework for the development of effective anti-metastatic agents. Nat. Rev. Clin. Oncol. 16, 185\u2013204 (2019).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSchmid, S. et al. Organ-specific response to nivolumab in patients with non-small cell lung cancer (NSCLC). Cancer Immunol. Immunother. 67, 1825\u20131832 (2018).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nOsorio, J. C. et al. Lesion-level response dynamics to programmed cell death protein (PD-1) blockade. J. Clin. Oncol. 37, 3546\u20133555 (2019).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nCrusz, S. M. et al. Heterogeneous response and progression patterns reveal phenotypic heterogeneity of tyrosine kinase inhibitor response in metastatic renal cell carcinoma. BMC Med. 14, 1\u20139 (2016).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nPires da Silva, I. et al. Site-specific response patterns, pseudoprogression, and acquired resistance in patients with melanoma treated with ipilimumab combined with anti\u2013PD-1 therapy. Cancer 126, 86\u201397 (2020).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMerz, M. et al. Deciphering spatial genomic heterogeneity at a single cell resolution in multiple myeloma. Nat. Commun. 13, 1\u201315 (2022).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWu, F. et al. Single-cell profiling of tumor heterogeneity and the microenvironment in advanced non-small cell lung cancer. Nat. Commun. 12, 1\u201311 (2021).\n\nADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nRusso, M. et al. Tumor heterogeneity and lesion-specific response to targeted therapy in colorectal cancer. Cancer Discov. 6, 147\u2013153 (2016).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKashyap, A. et al. Quantification of tumor heterogeneity: From data acquisition to metric generation. Trends Biotechnol. 40, 647\u2013676 (2021).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHuang, S. Reconciling non-genetic plasticity with somatic evolution in cancer. Trends Cancer 7, 1\u201314 (2021).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBinnewies, M. et al. Understanding the tumor immune microenvironment (TIME) for effective therapy. Nat. Med. 24, 541\u2013550 (2018).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nPao, W. et al. Tissue-specific immunoregulation: A call for better understanding of the \u201cimmunostat\u201d in the context of cancer. Cancer Discov. 8, 395\u2013402 (2018).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSiegel, R. L. et al. Colorectal cancer statistics, 2020. Ca. Cancer J. Clin. 70, 7\u201330 (2020).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nViale, P. H. The American Cancer Society\u2019s facts & figures: 2020 edition. J. Adv. Pract. Oncol. 11, 135 (2020).\n\n\n                    Google Scholar\u00a0\n                \n\nBiller, L. H. & Schrag, D. Diagnosis and treatment of metastatic colorectal cancer: a review. Jama 325, 669\u2013685 (2021).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nZhou, J., Li, Q. & Cao, Y. Spatiotemporal heterogeneity across metastases and organ-specific response informs drug efficacy and patient survival in colorectal cancer. Cancer Res. 81, 2522\u20132533 (2021).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nAsleh, K. et al. Proteomic analysis of archival breast cancer clinical specimens identifies biological subtypes with distinct survival outcomes. Nat. Commun. 13, 1\u201319 (2022).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMcDonald, K.-A. A. et al. Tumor heterogeneity correlates with less immune response and worse survival in breast cancer patients. Ann. Surg. Oncol. 26, 2191\u20132199 (2019).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSveen, A. et al. Intra-patient inter-metastatic genetic heterogeneity in colorectal cancer as a key determinant of survival after curative liver resection. PLoS Genet. 12, e1006225 (2016).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWilkerson, J. et al. Estimation of tumour regression and growth rates during treatment in patients with advanced prostate cancer: A retrospective analysis. Lancet Oncol. 18, 143\u2013154 (2017).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nDai, W. et al. Does tumor size have its prognostic role in colorectal cancer? Re-evaluating its value in colorectal adenocarcinoma with different macroscopic growth pattern. Int. J. Surg. 45, 105\u2013112 (2017).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKornprat, P. et al. Value of tumor size as a prognostic variable in colorectal cancer: A critical reappraisal. Am. J. Clin. Oncol. 34, 43\u201349 (2011).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSantullo, F. et al. Tumor size as a prognostic factor in patients with stage IIa colon cancer. Am. J. Surg. 215, 71\u201377 (2018).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nEisenhauer, E. A. et al. New response evaluation criteria in solid tumours: Revised RECIST guideline (version 1.1). Eur. J. Cancer 45, 228\u2013247 (2009).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMurphy, S. A. Consistency in a proportional hazards model incorporating a random effect. Ann. Stat. 22, 712\u2013731 (1994).\n\nArticle\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nIlkovitch, D. & Lopez, D. M. The liver is a site for tumor-induced myeloid-derived suppressor cell accumulation and immunosuppression. Cancer Res. 69, 5514\u20135521 (2009).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMundy, G. R. Metastasis to bone: Causes, consequences and therapeutic opportunities. Nat. Rev. Cancer 2, 584\u2013593 (2002).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nZhao, E. et al. Bone marrow and the control of immunity. Cell. Mol. Immunol. 9, 11\u201319 (2012).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nFabry, Z., Schreiber, H. A., Harris, M. G. & Sandor, M. Sensing the microenvironment of the central nervous system: immune cells in the central nervous system and their pharmacological manipulation. Curr. Opin. Pharmacol. 8, 496\u2013507 (2008).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSinaga, K. P. & Yang, M.-S. Unsupervised K-means clustering algorithm. IEEE Access 8, 80716\u201380727 (2020).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSyakur, M. A., Khotimah, B. K., Rochman, E. M. S. & Satoto, B. D. Integration k-means clustering method and elbow method for identification of the best customer profile cluster. In IOP conference series: materials science and engineering vol. 336 12017 (IOP Publishing, 2018).\n\nShahapure, K. R. & Nicholas, C. Cluster quality analysis using silhouette score. In 2020 IEEE 7th International Conference on Data Science and Advanced Analytics (DSAA) 747\u2013748 (IEEE, 2020).\n\nBurnham, K. P. & Anderson, D. R. Multimodel inference: understanding AIC and BIC in model selection. Sociol. Methods Res. 33, 261\u2013304 (2004).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nNatekin, A. & Knoll, A. Gradient boosting machines, a tutorial. Front. Neurorobot. 7, 21 (2013).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nYin, A. et al. A review of mathematical models for tumor dynamics and treatment resistance evolution of solid tumors. CPT Pharmacomet. Syst. Pharmacol. 8, 720\u2013737 (2019).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nNorton, L. The norton-simon hypothesis revisited. Cancer Treat. Rep. 70, 163\u2013169 (1986).\n\nCAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nFu, F., Nowak, M. A. & Bonhoeffer, S. Spatial heterogeneity in drug concentrations can facilitate the emergence of resistance to cancer therapy. PLoS Comput. Biol. 11, e1004142 (2015).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nByrne, H. M. & Chaplain, M. A. J. Growth of necrotic tumors in the presence and absence of inhibitors. Math. Biosci. 135, 187\u2013216 (1996).\n\nArticle\u00a0\n    CAS\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nStein, W. D. et al. Tumor regression and growth rates determined in five intramural NCI prostate cancer trials: The growth rate constant as an indicator of therapeutic efficacy. Clin. Cancer Res. 17, 907\u2013917 (2011).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nStein, W. D., Yang, J., Bates, S. E. & Fojo, T. Bevacizumab reduces the growth rate constants of renal carcinomas: a novel algorithm suggests early discontinuation of bevacizumab resulted in a lack of survival advantage. Oncologist 13, 1055\u20131062 (2008).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nClaesson-Welsh, L., Dejana, E. & McDonald, D. M. Permeability of the endothelial barrier: identifying and reconciling controversies. Trends Mol. Med. 27, 314\u2013331 (2021).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGifre-Renom, L., Daems, M., Luttun, A. & Jones, E. A. V. Organ-Specific Endothelial Cell Differentiation and Impact of Microenvironmental Cues on Endothelial Heterogeneity. Int. J. Mol. Sci. 23, 1477 (2022).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSarin, H. Physiologic upper limits of pore size of different blood capillary types and another perspective on the dual pore theory of microvascular permeability. J. Angiogenes. Res. 2, 1\u201319 (2010).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nParton, R. G., Schrotz, P., Bucci, C. & Gruenberg, J. Plasticity of early endosomes. J. Cell Sci. 103, 335\u2013348 (1992).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nAugustin, H. G., Kozian, D. H. & Johnson, R. C. Differentiation of endothelial cells: Analysis of the constitutive and activated endothelial cell phenotypes. Bioessays 16, 901\u2013906 (1994).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nCao, Y., Balthasar, J. P. & Jusko, W. J. Second-generation minimal physiologically-based pharmacokinetic model for monoclonal antibodies. J. Pharmacokinet. Pharmacodyn. 40, 597\u2013607 (2013).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKay, K., Dolcy, K., Bies, R. & Shah, D. K. Estimation of solid tumor doubling times from progression-free survival plots using a novel statistical approach. AAPS J. 21, 1\u201312 (2019).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nZharinov, G. M. et al. Prognostic value of tumor growth kinetic parameters in prostate cancer patients. Oncotarget 10, 5020 (2019).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLin, Z., Fan, Z., Zhang, X., Wan, J. & Liu, T. Cellular plasticity and drug resistance in sarcoma. Life Sci. 263, 118589 (2020).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nAmato, R. J. Chemotherapy for renal cell carcinoma. in. Semin. Oncol. 27, 177\u2013186 (2000).\n\nCAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nClark, A. M., Ma, B., Taylor, D. L., Griffith, L. & Wells, A. Liver metastases: Microenvironments and ex-vivo models. Exp. Biol. Med. 241, 1639\u20131652 (2016).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nNewton, P. K. et al. Spreaders and sponges define metastasis in lung cancer: A Markov chain Monte Carlo mathematical model. Cancer Res. 73, 2760\u20132769 (2013).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nQuinn, J. J. et al. Single-cell lineages reveal the rates, routes, and drivers of metastasis in cancer xenografts. Science 371, eabc1944 (2021).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nZhou, J., Liu, Y., Zhang, Y., Li, Q. & Cao, Y. Modeling tumor evolutionary dynamics to predict clinical outcomes for patients with metastatic colorectal cancer: a retrospective analysis. Cancer Res. 80, canres.1940.2019 (2019).\n\nSavic, R. & Lavielle, M. Performance in population models for count data, part II: A new SAEM algorithm. J. Pharmacokinet. Pharmacodyn. 36, 367\u2013379 (2009).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nAhn, J. E., Karlsson, M. O., Dunne, A. & Ludden, T. M. Likelihood based approaches to handling data below the quantification limit using NONMEM VI. J. Pharmacokinet. Pharmacodyn. 35, 401\u2013421 (2008).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nErdi, Y. E. Limits of tumor detectability in nuclear medicine and PET. Mol. Imaging Radionucl. Ther. 21, 23 (2012).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nDownload references",
+            "section_image": []
+        },
+        {
+            "section_name": "Acknowledgements",
+            "section_text": "We thank Mr. Timothy Qi and Dr. Tyler Dunlap from University of North Carolina at Chapel Hill, Eshelman School of Pharmacy for providing valuable suggestions and edits for the manuscript. Funding Source: National Institute of Health, R35GM119661.",
+            "section_image": []
+        },
+        {
+            "section_name": "Author information",
+            "section_text": "Division of Pharmacotherapy and Experimental Therapeutics, School of Pharmacy, University of North Carolina at Chapel Hill, Chapel Hill, NC, 27599, USA\n\nJiawei Zhou,\u00a0Amber Cipriani\u00a0&\u00a0Yanguang Cao\n\nUNC Health Medical Center, Department of Pharmacy, Chapel Hill, NC, 27514, USA\n\nAmber Cipriani\n\nSchool of Public Health, University of North Carolina at Chapel Hill, Chapel Hill, NC, 27599, USA\n\nYutong Liu\u00a0&\u00a0Quefeng Li\n\nDivision of Pharmaceutical Outcomes and Policy, School of Pharmacy, University of North Carolina at Chapel Hill, Chapel Hill, NC, 27599, USA\n\nGang Fang\n\nLineberger Comprehensive Cancer Center, School of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, NC, 27599, USA\n\nYanguang Cao\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nConceptualizations: J.Z., and Y.C.; methodology: J.Z., A.C., G.F., Q.L., and Y.C.; formal analysis: J.Z.; investigation: J.Z., Y.L., Q.L., and Y.C.; writing-original draft: J.Z., and Y.C.; writing-reviewing and editing: J.Z., A.C., G.F., Y.L., Q.L., and Y.C.; supervision: Y.C.\n\nCorrespondence to\n                Yanguang Cao.",
+            "section_image": []
+        },
+        {
+            "section_name": "Ethics declarations",
+            "section_text": "The authors declare no competing interests.",
+            "section_image": []
+        },
+        {
+            "section_name": "Peer review",
+            "section_text": "Nature Communications thanks Jennie Engstrand, Anita Sveen and the other anonymous reviewer(s) for their contribution to the peer review of this work. Peer review reports are available.",
+            "section_image": []
+        },
+        {
+            "section_name": "Additional information",
+            "section_text": "Publisher\u2019s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.",
+            "section_image": []
+        },
+        {
+            "section_name": "Source data",
+            "section_text": "",
+            "section_image": []
+        },
+        {
+            "section_name": "Rights and permissions",
+            "section_text": "Open Access  This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article\u2019s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article\u2019s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.\n\nReprints and permissions",
+            "section_image": []
+        },
+        {
+            "section_name": "About this article",
+            "section_text": "Zhou, J., Cipriani, A., Liu, Y. et al. Mapping lesion-specific response and progression dynamics and inter-organ variability in metastatic colorectal cancer.\n                    Nat Commun 14, 417 (2023). https://doi.org/10.1038/s41467-023-36121-y\n\nDownload citation\n\nReceived: 13 March 2022\n\nAccepted: 16 January 2023\n\nPublished: 26 January 2023\n\nVersion of record: 26 January 2023\n\nDOI: https://doi.org/10.1038/s41467-023-36121-y\n\nAnyone you share the following link with will be able to read this content:\n\nSorry, a shareable link is not currently available for this article.\n\n\n\n\n                            Provided by the Springer Nature SharedIt content-sharing initiative\n                        ",
+            "section_image": [
+                "https://data:image/svg+xml;base64,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"
+            ]
+        },
+        {
+            "section_name": "Associated content",
+            "section_text": "Collection",
+            "section_image": []
+        }
+    ]
+}

04b173197c545f10945642520e847eff78301d76d085be0937577c7b3426ca9d/metadata.json

@@ -0,0 +1,151 @@
+{
+    "title": "Hydrogen evolution with hot electrons on a plasmonic-molecular catalyst hybrid system",
+    "pre_title": "Plasmon-ligand-mediated hydrogen evolution with visible light",
+    "journal": "Nature Communications",
+    "published": "10 January 2024",
+    "supplementary_0": [
+        {
+            "label": "Supplementary Information",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-024-44752-y/MediaObjects/41467_2024_44752_MOESM1_ESM.pdf"
+        },
+        {
+            "label": "Peer Review File",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-024-44752-y/MediaObjects/41467_2024_44752_MOESM2_ESM.pdf"
+        }
+    ],
+    "supplementary_1": [
+        {
+            "label": "Source data",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-024-44752-y/MediaObjects/41467_2024_44752_MOESM3_ESM.zip"
+        }
+    ],
+    "supplementary_2": NaN,
+    "source_data": [
+        "/articles/s41467-024-44752-y#Sec22"
+    ],
+    "code": [],
+    "subject": [
+        "Chemical physics",
+        "Hydrogen energy",
+        "Nanoparticles"
+    ],
+    "license": "http://creativecommons.org/licenses/by/4.0/",
+    "preprint_pdf": "https://www.researchsquare.com/article/rs-2751820/v1.pdf?c=1704978912000",
+    "research_square_link": "https://www.researchsquare.com//article/rs-2751820/v1",
+    "nature_pdf": "https://www.nature.com/articles/s41467-024-44752-y.pdf",
+    "preprint_posted": "02 May, 2023",
+    "research_square_content": [
+        {
+            "section_name": "Abstract",
+            "section_text": "Plasmonic systems convert light into electrical charges and heat that mediate catalytic transformations. However, the debate about the involvement of hot carriers in the catalytic process remains shredded in controversy. Here, we demonstrate the direct use of plasmon hot electrons in the hydrogen evolution with visible light. A plasmonic nanohybrid system consisting of NiO/Au/[CoII(phen-NH2)2(H2O)2] (phen-NH2 = 1,10-Phenanthrolin-5-amine) that is unstable at water thermolysis temperatures was consciously assembled, ensuring that the plasmon contribution to the catalytic process is solely from hot carriers. With the combination of photoelectrocatalysis and advanced in situ spectroscopies, one could establish the reaction mechanism, which consisted of electron injection into the phenanthroline-ligands followed by two quick, concerted proton-coupled electron transfer steps resulting in the evolution of hydrogen. Light-driven hydrogen evolution with plasmons provides a sustainable route for producing green hydrogen, which modern society strives to achieve.Physical sciences/Chemistry/Photochemistry/PhotocatalysisPhysical sciences/Nanoscience and technology/Nanoscale devices/Nanophotonics and plasmonics",
+            "section_image": []
+        },
+        {
+            "section_name": "Additional Declarations",
+            "section_text": "There is NO Competing Interest.",
+            "section_image": []
+        },
+        {
+            "section_name": "Supplementary Files",
+            "section_text": "H2evolutionSI.docx",
+            "section_image": []
+        }
+    ],
+    "nature_content": [
+        {
+            "section_name": "Abstract",
+            "section_text": "Plasmonic systems convert light into electrical charges and heat, mediating catalytic transformations. However, there is ongoing controversy regarding the involvement of hot carriers in the catalytic process. In this study, we demonstrate the direct utilisation of plasmon hot electrons in the hydrogen evolution reaction with visible light. We intentionally assemble a plasmonic nanohybrid system comprising NiO/Au/[Co(1,10-Phenanthrolin-5-amine)2(H2O)2], which is unstable at water thermolysis temperatures. This assembly limits the plasmon thermal contribution while ensuring that hot carriers are the primary contributors to the catalytic process. By combining photoelectrocatalysis with advanced in situ spectroscopies, we can substantiate a reaction mechanism in which plasmon-induced hot electrons play a crucial role. These plasmonic hot electrons are directed into phenanthroline ligands, facilitating the rapid, concerted proton-electron transfer steps essential for hydrogen generation. The catalytic response to light modulation aligns with the distinctive profile of a hot carrier-mediated process, featuring a positive, though non-essential, heat contribution.",
+            "section_image": []
+        },
+        {
+            "section_name": "Introduction",
+            "section_text": "Plasmonic photocatalysis uses electrical charges formed during the plasmon resonance decay triggered by light absorption. A plasmon is a quantised oscillation of the electron density, and its decay can generate high-energy carriers (electrons and holes) in metals due to the interaction between plasmons and incident light, causing them to become \u2018hot\u2019 or have high kinetic energy1. These hot carriers can then be used to generate electrical current or to drive chemical reactions. However, hot electrons\u2019 involvement in catalysis remains disputed, despite reports of their participation in processes such as solar to chemical energy reactions2,3,4,5, epoxidations6,7, dehydrogenations8, ammonia electrosynthesis9, etc10,11,12. The scepticism surrounding their involvement relates to the hot carriers\u2019 ultrafast relaxation (ca. 1\u2009ps)13, and that several examples rationalise their participation as enhancers of the photothermal process. Therefore, the catalytic output is prone to errors since the surface temperature of plasmonic materials is notoriously tricky to measure accurately, thus underestimating the thermal contribution to the catalysis14,15. Despite the significant progress, it remains challenging to disentangle charge carrier catalysis from photothermal effects16,17,18,19.\n\nThe hot carriers\u2019 energy distribution is broad, with a significant fraction of the carriers having energies above the Fermi level of the metal caused by the non-Fermi-Dirac distribution20. More research is needed to fully understand plasmonics\u2019 hot carrier energy distribution dynamic behaviour21,22. Still, the ultrafast relaxation can be partially mitigated via ultrafast charge transfer to suitable acceptors consecutively23,24, or simultaneously25, forming this contribution scientific basis to demonstrate the direct involvement of hot carriers in the catalytic process.\n\nHerein, a plasmonic nanohybrid system consisting of NiO/Au/[CoII(phen-NH2)2(H2O)2] (phen-NH2\u2009=\u20091,10-Phenanthrolin-5-amine) was assembled and demonstrated to perform hydrogen evolution reaction (HER). NiO acted as a hole acceptor26,27,28,29, and the cobalt complex, a mimic of the HER catalyst reported by Luo et al.30 and the hydrolytic DNA cleavage agent by Sharma et al.31, as an electron acceptor. Water can be converted into hydrogen through thermolysis that requires temperatures of 500\u20131000\u2009\u00b0C32, a temperature at which our molecular system would decompose, suggesting the participation of hot electrons in the catalytic process. To further this finding, the reaction mechanism was monitored by combining photoelectrocatalysis, unbiased ultrafast spectroscopies, and in situ electrochemistry, followed by near ambient pressure X-ray photoelectron spectroscopy (NAP-XPS) studies. The results suggest a reaction mediated by the phenanthroline-ligands that accept the electrons from the plasmon and transfer them to the cobalt centre in two concerted proton-electron transfers (CPET) that significantly lowers the energy threshold of the steps as it avoids the formation of higher energy intermediates33. Recently, a study was published with a similar concept, namely a cobalt porphyrin supported on plasmonic nanoparticle that, on illumination, produced H234. Still, there is a clear distinction. In the present contribution, only the Au nanoparticles (Au NPs) are photoactive, contrasting with the published study where the catalyst and Au NPs are photoactive. Thus, their observation might be related to photonic enhancement instead of a plasmonic hot carrier. Moreover, the authors suggested a cooperative result between plasmon hot carriers and localised thermal effects, for which this study does not have evidence. This contribution also offers more extensive experimental support for the mechanism reaction involving the plasmon hot carrier and Co complex catalyst ligands, which are markedly different from what has been published on cobalt systems for HER. The proposed combined spectroscopically approach offers a robust methodology to measure the reaction mechanism on mesoporous electrodes that represent real electrodes more truthfully.",
+            "section_image": []
+        },
+        {
+            "section_name": "Results and discussion",
+            "section_text": "The molecular catalyst mimics the system proposed by Sharma et al.31 as a DNA hydrolytic cleavage agent. The prime difference between Sharma\u2019s system and the one presented herein is the presence of the amine group on the ligand, which is necessary to anchor the catalyst to the Au NPs. Consequently, the as-prepared catalyst has a cobalt centre coordinated with two 1,10-Phenanthrolin-5-amine ligands and a bidentate nitrate group. A second out-sphere nitrate ensures complex neutrality (Fig.\u00a01B), which is consistent with previously reported crystal structures31. Details on the catalyst preparation and characterisation can be found in supporting information (SI). The optical spectrum of the as-prepared catalyst in dimethylformamide is shown in Fig.\u00a0S4. It displays a strong absorption peak centred at 290\u2009nm with a shoulder at 360\u2009nm, characteristic of phenanthroline complexes35. The molecular catalyst thermally decomposes at 265\u2009\u00b0C, significantly below the temperature needed for water thermolysis.\n\nA Cobalt complex in dimethylformamide titration with water followed by in situ UV-Vis, with inset showing the effect of acid in the spectrum, it represents with and without acid in water; B proposed structure of the catalyst in water, which was used to anchor the catalyst to Au NPs; C UV-Vis spectra of Au NPs region before and after addition of the cobalt complex on thin film. On inset infrared spectra of the amino stretching region normalised by C-N stretching band intensity: (i) catalyst before anchoring; (ii) catalyst after anchoring it to the Au NPs; D photosystem structure used for the photo-electrocatalytic H2 evolution.\n\nCobalt nitrate complexes are known to exchange their nitrate ligands with water36 which is the solvent used to attach the complex to the Au NPs and perform the catalysis. Therefore, the complex was titrated with water to evaluate whether ligand exchange occurred. Figure\u00a01A shows the increase of the UV-Vis shoulder located at 360\u2009nm, with an increase in water content, saturating at around 20% water content. The exchange was also confirmed by the X-ray photoelectron spectroscopy (XPS) analysis. The N 1\u2009s region in Fig.\u00a0S13, acquired in the low vacuum after introducing the electrode in the Near Ambient Pressure (NAP) analysis chamber, displays a sharp peak centred at 398.6\u2009eV. Such a binding energy value can be assigned to the pyridinic nitrogen of the phenanthroline ligand37. Nitrate ligands are typically found at a binding energy of 408\u2009eV38, but there were no features in this region in the collected spectrum.\n\nFurther elaboration on ultra-high vacuum (UHV)-XPS analysis will be provided when corroborating evidence regarding anchoring the catalyst to the Au surface is presented. Notably, adding acid to the aqua complex did not change its optical absorption (Fig.\u00a01A inset), suggesting that the di-aqua complex resultant from the exchange of the nitrate ligand by water molecules is very stable. Figure\u00a01B shows the proposed catalyst structure after ligand exchange.\n\nThe attachment of the complex was followed by UV-Vis and infrared spectroscopies (Fig.\u00a01C). The Au NPs were prepared using the Turkevich method, as reported elsewhere39 and briefly described in SI. The UV-Vis of the Au NPs on glass shows the characteristic localised surface plasmon resonant (LSPR) peak at 535\u2009nm, corresponding to an average particle size of 8\u2009\u00b1\u20092\u2009nm (determined by dynamic light scattering (Fig.\u00a0S5) and atomic force microscopy (AFM) (Fig.\u00a0S6)), consistent with what we have published previously40. Note that NiO morphology (Fig.\u00a0S7) and electronic structure (Fig.\u00a0S8) did not change with Au NPs deposition and subsequent annealing.\n\nThe Au LSPR peak shifts to lower energy when the cobalt catalyst is added (Fig.\u00a01C), confirming the anchoring. Note that the LSPR peak absorption is sensitive to the surrounding dielectric medium41. Consequently, the surface modification by the catalyst should induce a shift in the LSPR maximum as absorbed. Additionally, it is possible to see the complex absorption shoulder located at 370\u2009nm, corroborating the attachment between the catalyst and Au NPs. Unfortunately, the glass support (FTO or cover glass) covers the rest of the complex UV-Vis band, precluding their measurement.\n\nThe 10-Phenanthrolin-5-amine ligand was purposely chosen to ensure selective coordination to the gold surface via the amino groups. This first supporting evidence came from XPS measured at low vacuum conditions, with a Co 2p signal related to the catalyst only when Au NPs were present. The Co 2p3/2 on Au/Co-cat and NiO/Au/Co-cat measured had a single contribution centred at around 780.5\u2009eV, consistent with Co is oxidation +242. The observation that the Co signal was only present when Au is present is a solid endorsement for the selective anchoring of the catalyst to the gold surface. The anchoring is believed to occur via the -NH2 groups43. This was corroborated by the disappearance of the amino bands in the infrared (Fig.\u00a01C inset). Before anchoring, the complex has three small peaks between 3450\u20133300\u2009cm\u22121 and 3250\u20133200\u2009cm\u22121 associated with N-H stretching modes of primary amino groups44. The correspondent bending modes between 1650\u20131580 cm\u22121 are also visible but somewhat overlapped by the water O-H bending mode. After attaching the catalyst to the Au NPs, the N-H bands disappear. The complete disappearance suggests that the catalyst coordinates to the Au NPs via both 1,10-Phenanthrolin-5-amine ligands, as shown schematically in Fig.\u00a01D. The infrared bands were normalised to the C-N stretch at 1280\u2009cm\u22121 intensity to enable direct comparison. The C-N band is unaffected by the attachment, making it suitable for normalisation. Unfortunately, the formed Au-N bonds are not infrared active, and the low loading prevented their detection with Raman spectroscopy. Therefore, UHV-XPS experiments were performed to substantiate the catalyst anchoring to the Au surface via the -NH2 group.\n\nThe UHV-XPS comparing the N 1\u2009s and Au 4\u2009f signals before and after anchoring the catalyst to the Au surface are presented in Fig.\u00a02A, B, respectively. Before attaching, the catalyst has two N 1\u2009s peaks: the N from the phenanthroline bonded to the cobalt centre at 398.8\u2009eV45,46 and N at 401.4\u2009eV ascribed to the -NH2 groups47. Note that the UHV-XPS also did not show a peak ascribed to the nitrate groups, corroborating its exchange by water molecules. Upon attaching, the N 1\u2009s signal related to the -NH2 group disappeared, and the N from phenanthroline shifted to 399.1\u2009eV and got broader (FWHM before 1.641 and after attaching 1.812). Note that one expects the N 1\u2009s from the amino group to shift to lower binding energies as it loses the protons at a rate of about 1\u2009eV per hydrogen atom lost48. These observations are consistent with attachment via the NH2 groups and formation of Au-N species, which cannot be distinguished from the N of the phenanthroline ligand because of signal-to-noise. The Au 4f7/2 had binding energy at 83.6\u2009eV for the sample with and without a catalyst, consistent with metallic gold49. There was only one species in all the samples, which is coherent with the idea that any electronic change due to NiO and Co-catalyst is delocalised over all the gold atoms, suggesting good electronic coupling50. Figure\u00a01D shows the schematic representation of the complete photosystem.\n\nUHV-XPS shows the changes due to the catalyst after anchoring it to the Au surface via the -NH2 functional groups. A XPS N 1\u2009s and B Au 4\u2009f regions.\n\nBulk electrolysis of the cobalt catalyst in water using 0.1\u2009M LiCl as a supporting electrolyte without acid (Fig.\u00a03A) shows no unique reduction peaks before the onset of the catalytic wave at \u22121.18 vs Ag/Ag/Cl. Initiating with the absence of reduction peaks, this discovery diverges from the observations made by Luo et al.30 (whose system served as inspiration for this study) and the system studied by Wang et al.51,52 exhibiting comparable rigidity and coordination. The onset potential for electrolytic hydrogen production without using acid in our catalyst was lower, differing by only ca. 15\u2009mV from Wang et al.52 system but by a noteworthy 190\u2009mV from the Luo et al.30 system.\n\nA Bulk electrolysis of cobalt catalysts in water and the presence of 3\u2009mM acetic acid, using glassy carbon as the working electrode, Pt wire as the counter electrode, Ag/AgCl as reference electrode, 0.1\u2009M LiCl as supporting electrolyte (pH\u2009=\u20095.2) and scan rate 50\u2009mV/s; B Cyclic voltammetry of the NiO/Au and NiO/Au/Co-catalyst in water using Pt wire as a counter electrode, Ag/AgCl as reference electrode and 0.1\u2009M LiCl as supporting electrolyte (pH\u2009=\u20095.2) (scan rate 50\u2009mV/s); C Effect of light in the chronoamperometry of the complete photosystem applying \u22120.65\u2009V potential with the inset showing the catalytic wave when the experiments were performed with 3\u2009mM acetic acid (pH\u2009=\u20093.5). A 532\u2009nm CW laser with 43.8\u2009mW/cm2 power was used as illumination. In inset NiO/Au/Co-catalyst in water (blue trace) and NiO/Au/Co-catalyst\u2009+\u2009acetic acid (red trace); D Effect of light in the chronoamperometry of the NiO/Au and NiO/Au/ligand applying \u22120.65\u2009V potential in 3\u2009mM acetic acid and 0.1\u2009M LiCl (pH\u2009=\u20093.5). A 532\u2009nm CW laser with 43.8\u2009mW/cm2 power was used as illumination.\n\nPredictably, the catalytic wave\u2019s amplitude is heightened with the presence of protons, underscoring proton availability as a pivotal factor in accessing catalytic performance within this system. Consequently, all catalytic data was obtained in a 3\u2009mM acetic acid environment. Although the introduction of acid did not result in new reduction peaks, it did cause a notable shift in the onset potential, reducing it by as much as 120\u2009mV - a deviation from the characteristics of the previously documented system. These observations imply the engagement of two concerted proton-electron transfer mechanisms as opposed to the conventional sequential reduction followed by protonation51. Subsequent sections will delve into additional evidence supporting this proposed mechanism.\n\nFigure\u00a03B shows the cyclic voltammetry of the NiO/Au and NiO/Au/Co-catalyst thin films in water. The NiO/Au shows a reduction peak centred at around \u22120.49\u2009V vs Ag/AgCl related to the reduction of gold surface oxygen and weak catalytic wave staring at \u22120.90\u2009V vs Ag/Ag/Cl. Adding the cobalt catalyst drastically reduced the peak at \u22120.49\u2009V vs Ag/Ag/Cl, suggesting that surface functionalisation by the catalyst decreases the amount of adsorbed oxygen on Au NPs. The cobalt loading in the complete system was determined to be 11.4\u2009\u03bcg/cm2 by inductively coupled plasma-optical emission spectrometry (ICP-OES), equating to 0.29 wt.% of Co. The low loading creates issues regarding spectroscopy signal-to-noise but ensures that the activity is primarily due to the catalyst and enables the detection of catalyst degradation evidence.\n\nThe complete system is susceptible to illumination at 532\u2009nm, as shown by the CV in Fig.\u00a0S17. Illumination of the entire system at plasmon resonance led to a slight decrease in offset potential (ca. 0.05\u2009V) but, more importantly, an increase in the catalytic wave photocurrent.\n\nChronoamperometry data at \u22120.65\u2009V vs Ag/Ag/Cl in the absence and presence of light (CW laser 532\u2009nm, selectively exciting only the Au plasmon) is presented in Fig.\u00a03C. The photosystem is responsive to the light, increasing the photocurrent by ca. \u221215\u2009\u03bcA (\u221219\u2009\u03bcA/cm2). The H2 production rate was estimated from three consecutive on/off light switches to be 3.1 nmol/(min.cm2). The increase in photocurrent was shown to be due to the evolution of H2, as confirmed by online quadrupole mass spectrometry (QMS) analysis (Fig.\u00a0S10). The response was found to be constant during the cycling of light on and off for the duration of the experiment (ca. 360\u2009min). The findings indicate that the photosystem is relatively stable, and the process is catalytic. The significant increase in evolved H2 relates to the presence of cobalt catalyst since neither the NiO/Au nor NiO/Au/ligand systems produce significant photocurrent (Fig.\u00a03D) and had no detectable hydrogen evolution by online QMS analysis. The drift in the baseline is related to heating effects from the illumination and plasmonic decay because it lacked synchronisation with the light on/off cycles.\n\nIt is clear from the data that the systems with plasmonic materials are responsive to the 532\u2009nm illumination. Light-mediated plasmon-catalysis is a very complex process due to many potential reaction enhancers. One possibility is the near-field enhancements caused by the local electric fields formed upon Au LSPR excitation. At the most basic level, near-fields can enhance charge separation and alignment of molecular dipoles53. However, such localised electric fields impact the hot carriers but cannot catalyse the reaction autonomously. The second option employs the plasmon-induced resonance energy transfer (PIRET) process, connecting the plasmon evanescent field to a semiconductor absorber through dipole-dipole interaction. However, these systems necessitate core-shell architectures (which are not applicable in this context)54,55 and the most substantial enhancements were observed with silver as the plasmonic material rather than gold56. The final option explores strong-correlated plasmon-molecule systems. Still, in this scenario, there must be an optical overlap between the plasmon and molecule50 which is again not present.\n\nCatalytic performance measurements were conducted using off-resonant excitation to assess local field enhancement contribution. Off-resonance excitation induces local effects such as elevated local near-fields57,58,59. However, in the context of hot carriers, off-resonance excitation generates low-energy carriers that are not conducive to driving photocatalytic processes60. Excitation at 650\u2009nm (off-resonant) caused no significant differences in the CV compared to experiments performed in the dark (Fig.\u00a0S17). This result was further supported by light switch chronoamperometry (Fig.\u00a0S18). The findings imply that local near-fields do not contribute significantly to enhancing catalysis. Consequently, the observed increase in catalytic output under resonant illumination is likely associated with hot electrons and heat rather than near-fields. This, however, does not rule out the potential for near-fields to assist catalysis by enhancing charge separation; they would likely influence hot carriers indirectly engaged in the catalytic process.\n\nHeat is an inherent factor in plasmonic catalysis due to the underutilisation of hot charges, leading to their recombination and local heat generation. Although it is acknowledged that the surface temperature of excited plasmonic materials exceeds that of the solution, determining the precise value poses a challenge due to the ultrafast dynamics of thermalisation. The molecular catalyst remains stable only up to 265\u2009\u00b0C, a temperature considerably lower than required for uncatalysed water thermolysis. Nevertheless, a noteworthy temperature range remains unexplored, primarily because experiments are conducted in an aqueous medium.\n\nTo disentangle the heat contribution, we performed light modulation chronoamperometry. A study by Maley et al.61 demonstrated that light absorption at the electrode surfaces within nanoparticle arrays led to significant localised temperature increases and altered solution flows. These thermal effects were anticipated to influence electrochemical currents through diverse mechanisms, encompassing enhancements in mass transfer, shifts in equilibrium redox potentials, and conventional temperature-dependent accelerations in kinetic rates for electrode processes. Their analysis suggests that mass transfer enhancements alone would result in substantial current increases applicable to electrochemical reactions involving dissolved reactants and products, both outer-sphere and inner-sphere reactants. Consequently, heat-induced effects exhibit a distinctive gradual rise and decay of the current during light modulation, as they operate on processes with time constants in the nanosecond range.\n\nThe light response of the entire system (NiO/Au-Co-catalyst) in comparison to the system without NiO (i.e., Au/Co-catalyst) offers insights into the role of heat in the process (see Fig.\u00a03C, D, compounded figure below). Removing NiO is expected to decrease the lifetime of the charge-separated state, generating more heat. However, contrary to the expectation that heat is the primary contributor to reactivity, we observed a fourfold increase in current induced by light when NiO was present. Additionally, examining the response of the Au/Co-catalyst to light modulation reveals a classic heat-mediated process with a relatively slow rise and decay, in contrast to the complete system. The entire system demonstrates a faster rise and decay to light modulation, indicating the involvement of hot carriers.\n\nLight absorption by planar electrodes randomly decorated with plasmonic structures acts as a uniform heat source delocalised across the electrode\u2212solution interface, resulting in heat dissipation in a linear geometry with significant temperature changes as a function of time. Consequently, the electrochemical response to light modulation provides a strategy to decouple heat from hot carriers\u2019 contributions to substantiate our mechanistic claim. Commonly, the experiments are performed by modulating the light intensity. However, the tested electrodes are quasi-transparent, making light-intensity modulation studies challenging. Thus, we opted to change the light ON/OFF cycle repetition rate to modulate electrode exposure to light.\n\nFigure\u00a04A shows the changes in measured photocurrent (\u0394i) as a function of light modulation repetition rate. Unsurprisingly, lower repetition rates (i.e. higher light exposure) resulted in more significant \u0394i. In a heat-mediated process, \u0394i is expected to scale with~t1/2 (t = time)61 inconsistent with the observed current transients, providing the first substantiation for a hot carrier-mediated process. Additionally, \u0394i in a heat-mediated electrochemical process follows a linear dependence with increased light exposure, independent of the process occurring via inner or outer-sphere reaction. Fig.\u00a04B shows that \u0394i does not offer a linear behaviour regarding light exposure, thus providing more explicit evidence for hot carriers\u2019 involvement.\n\nThe effect of light modulation in the chronoamperometry was performed at \u22120.65\u2009V vs Ag/AgCl with 3\u2009mM acetic acid (pH\u2009=\u20093.5) and 532\u2009nm CW laser with 43.8\u2009mW/cm2. The experiments were performed using a squared function at different repetition rates. A Changes in photocurrent (\u0394i) at different light modulation frequencies (8, 11, 16 and 33\u2009MHz) over different light ON/OFF cycles; and B \u0394i versus light modulation frequency.\n\nHaving established that the photosystem is responsive to light and able to evolve hydrogen, it is essential to determine whether plasmon hot electrons are involved in the process. Transient absorption spectroscopy (TAS) without an external electric field (unbiased) was performed to evaluate charge transfer in the photosystem. Excitation of the Au NPs LSPR results in a bleach signal and two small winglets on each side of the bleach to broaden the LSPR peak62 (see the representative spectrum in SI Fig.\u00a0S11). Figure\u00a05A shows the kinetic traces extracted at 490\u2009nm (edge of the positive winglet to the blue of the LSPR maximum) after excitation at 550\u2009nm of the Au NPs, NiO/Au and NiO/Au/Co-catalyst systems. Note that similar findings were obtained when performing the analysis on the winglet to the red of LSPR maximum. The kinetic traces were fitted with a rising edge and a double exponential decay. The rising edge is assigned to the electron-electron (e-e) scattering lifetime, the shorter exponential decay to electron-phonon (e-ph) scattering lifetime and the longer decay to photo-phonon (ph-ph) scattering lifetime13,14,63. Recently, we demonstrated that charge transfer can be established from changes in the e-ph lifetime. Both electron and hole transfer reduce the e-ph lifetime compared to the plasmon nanoparticles without charge transfer. Hot electrons reduce the e-ph by taking energy from the resonance. Hot holes decrease the e-ph by injecting cold electrons into the resonance, reducing the average electron temperature24,25.\n\nA Kinetic traces of Au, NiO/Au, NiO/Au/Co-catalyst, extracted at 490\u2009nm from the TAS measurements; B TIRAS map of the NiO/Au/Co-catalyst; C TIRAS kinetic trace extracted at 4705\u2009nm for NiO/Au and NiO/Au/Co-catalyst; D TIRAS kinetic trace extracted at 4705\u2009nm for Au/ligand and Au/Co-catalyst.\n\nThe Au NPs on glass have an e-e lifetime estimated to be 167\u2009\u00b1\u200949\u2009fs and an e-ph of 5.1\u2009\u00b1\u20090.4\u2009ps, which is within what has been published22,64,65. When attached to NiO (hole acceptor), the Au NPs e-e increased slightly to 198\u2009\u00b1\u200985\u2009fs with a noticeable decrease in e-ph lifetime to 3.4\u2009\u00b1\u20091.0\u2009ps, consistent with what is anticipated if holes are transferred from Au NPs to NiO. The system composed of Au/Co-catalyst has an e-e of 148\u2009\u00b1\u200930\u2009fs and a significantly shorter e-ph lifetime (4.2\u2009\u00b1\u20090.3\u2009ps) compared with Au NPs alone. Since the catalyst is expected to be the electron acceptor, the reduction in e-ph lifetime suggests that electrons are transferred from Au NPs to the catalyst. The complete photosystem had an e-e of 198\u2009\u00b1\u2009116\u2009fs and the most significant reduction in the e-ph from 5.1\u2009\u00b1\u20090.4\u2009ps (Au NPs) to 2.6\u2009\u00b1\u20091.0\u2009ps, suggesting the hot holes and electrons are transferred to the respective acceptors. In sum, the presence of electron and hole acceptors reduced the e-ph, consistent with charge transfer from Au NPs to the acceptors, with the most significant e-ph lifetime reduction observed when both electron and hole acceptors are present.\n\nSimply formulated, the decrease in e-ph lifetime relates to changes in electron average temperature, which are regulated by the total energy in the resonance. Therefore, the observed reduction in e-ph lifetime can be related to an energy transfer56. To ascertain that the changes observed with TAS measurements are related to hot carrier transfer, not energy transfer, complementary unbiased transient infrared absorption spectroscopy (TIRAS) spectroscopy studies were performed. Free carriers absorb strongly in the infrared domain due to forming a quasi-metallic state66. The signal is characterised by broad and featureless infrared absorption, often depicted as a background shift in the infrared spectrum67.\n\nA representative TIRAS data map after LSPR excitation at 550\u2009nm is shown in Fig.\u00a05B. Kinetic traces extracted between 4645\u20134700\u2009nm (2150\u20132130\u2009cm\u22121) are presented in Fig.\u00a05C. The kinetic traces were fitted with a rising edge and two exponential decays, ascribed to the lifetimes of the injection and recombination processes, respectively. NiO/Au shows a rising edge with a 196\u2009\u00b1\u2009104\u2009fs time component, suggesting fast injection of the hole. Most injected charge recombines within 417\u2009\u00b1\u2009117\u2009fs (~91%). The complete system displayed a similar injection time (100\u2009\u00b1\u200923\u2009fs) and increased recombination time (6.6\u2009\u00b1\u20094.2\u2009ps, 85% of the signal). However, an increase in signal amplitude is noticeable, suggesting that more charge is transferred when both acceptors are present, i.e., an enhancement of charge separation, resulting in more charge available for the catalysis. In both cases, 5\u201310% of the charge survives past 1\u2009ns, making it useful for catalytic transformations, including H2 evolution.\n\nThe free carrier signal detected by unbiased TIRAS confirms the transferrence of hot carriers to the acceptors, suggesting that the observed decrease in e-ph lifetime is less likely to be due to energy transfer. Since the Au plasmon resonance absorption is very far from the acceptors\u2019 absorptions, one can also discard the hypothesis of photonic enhancement as the corporate for catalytic performance. Noteworthy is that unbiased TIRAS measurements on a system without the NiO, namely with Au/Co-catalyst and Au/ligand, also show a broad featureless infrared absorption (Fig.\u00a0S12), characteristic of free carrier absorption, not localised charge. The kinetic traces in Fig.\u00a05D are noisy due to low signal and thus challenging to fit. However, qualitatively, the signal shows a short rising component (faster than the instrument response function (ca. 100\u2009fs)), suggesting a high-speed electron injection. In the case of the Au/ligand, the decay is swift, but when Co is present, the decay is comparatively slow, with the charge surviving past 1\u2009ns. The shape of the TIRAS signal indicates that it relates to free carries. This suggests that hot electrons are injected into the ligands due to the strong coupling between phenanthroline ligands and Au NPs; the charge is delocalised through the aromatic rings, acting as free carriers. The observation confirms ligand reduction under light irradiation occurs even without external bias, which is crucial for the hot electron-mediated mechanism.\n\nThe presence of Co improves hot electrons\u2019 lifetime due to some charge stabilisation. Still, the cobalt centre is the same since this would lead to the disappearance of the \u2018free carrier\u2019 infrared absorption behaviour, which only happens for at least 1\u2009ns. The TIRAS observation is also consistent with the electrochemistry\u2019s absence of cobalt reduction peaks. This is a peculiar observation because, from the speculated mechanisms of the analogous30 and relatable cobalt complexes68,69 gives a central role to the metal centre, often undergoing sequential reductions and protonations before evolving hydrogen. Conversely, the ligands are confined to a spectator role in the catalytic process.\n\nTo corroborate the peculiar finding, electrochemical experiments combined with in situ near ambient pressure (NAP)-XPS measurements were performed using the dip-and-pull approach70 in the absence (Fig.\u00a0S14) and presence (Fig.\u00a06) of acetic acid. The measurements were performed with the mesoporous films used for photocatalytic data. This significantly reduces the signal intensity and requires adaptation of the dip-and-pull method to reduce the amount of electrolyte trapped in the porous film. Briefly, only the lowest segment of the sample was dipped into the electrolyte solution, while the rest formed a liquid film via capillary forces. After some equilibration time (ca. 30\u2009min), the liquid film settled to a point where the photoemission signals of the electrode (O 1\u2009s, Co 2p and N 1\u2009s) were detectable together with the signal of liquid and gas phase water (O 1\u2009s). All spectra were corrected for charge and baseline, and the peaks were deconvoluted according to XPS fitting constraints71. Charging correction was performed using an advantageous C 1\u2009s signal.\n\nA O 1\u2009s signals (the inset magnifies the prominent peaks, highlighting the binding energy shift due to the potential applied); B Co 2p signals overlaid with a 400 counts offset for better visibility.\n\nFigures\u00a06A and S8A show the O 1\u2009s spectra acquired in situ at different applied potentials. Three prominent peaks, centred at approximately 531.0\u2009eV, 532.7\u2009eV and 534.9\u2009eV, are assigned to adsorbed hydroxyls, liquid water (thin electrolyte film on top of the electrode) and gas phase water, respectively72,73 (fitting parameters are reported in Table\u00a0S1). The slight shift of the O 1\u2009s central peak measured between the pure electrolyte and the same in the presence of acid (from 532.68 to 532.75\u2009eV under OCP conditions) may be due to the presence of acetic acid, whose contribution falls within the spectral range of liquid water peak74. A fourth peak, centred at 529.5\u2009eV, was detected in the absence of acid and assigned to lattice oxygen (electrode). Such a difference between the two conditions (no lattice oxygen detected in the presence of acid) suggests that the liquid electrolyte film in the absence of acid was thinner than in the presence of acid.\n\nPotential control and the availability of a continuous liquid film up to the position monitored by XPS were surveyed by shifting the O 1\u2009s signal according to the applied potential. Figures\u00a04A and S8A show a shift in the central peak of O 1\u2009s, centred at approximately 532.8\u2009eV and assigned to liquid water (thin electrolyte layer), as the potential was applied (detected as a positive binding energy shift, proportional to the potential applied to the WE), confirming experiment validity and thus access to Co oxidation state at different potentials. A difference of about 0.1\u20130.15\u2009V was detected between the applied potential chosen based on the catalysis and seen in the O1s NAP-XPS, which is assigned to calibration shifts of the reference electrode during long experimental times. Therefore, the values specified in the plots are applied for consistency, not measured voltages. It is essential to highlight that while the 532.7\u2009eV peak component displays a shift proportional to the applied voltage, components at 529.5 and 531.0\u2009eV do not.\n\nConfident that NAP-XPS experiments reflected the Co oxidation state at different potentials, one can proceed with the analysis of the Co 2p region (Fig.\u00a06B). Despite the low signal-to-noise ratio, due to Co 2p attenuation through the liquid electrolyte layer stabilised on the WE, the main features of Co 2p peaks can be detected on both investigated electrodes (see Figs.\u00a06B and S14B). The Co 2p3/2 prominent peak is centred at around 780\u2009eV, ascribed to Co2+ as expected75,76. The peak position does not shift with the applied potential, as the working electrode is set to ground potential during the experiment. Nevertheless, a peak shift due to the formation of a different cobalt species is not observed either in the presence or in the absence of acetic acid (Fig.\u00a0S8B). Indeed, the reduction of cobalt from 2+ to the metallic state leads to a shift of the binding energy of the 2p3/2 peak by about 2\u2009eV, from 780 to 778\u2009eV77,78. Figure\u00a04B also shows that at \u22120.65\u2009V vs Ag/AgCl, the cobalt signal decreases in intensity. Such a potential is sufficient for the evolution of H2 since bubbles were detected at the electrode and marked the onset of the catalytic wave (Fig.\u00a0S9). Thus, the decrease in signal-to-noise ratio relates to experimental conditions since the signal statistics are affected by the formation of H2 bubbles at the highest potential. Therefore, only Co2+ is present until H2 formation; thus, the reduction peak at lower voltages relates to the reduction of the phenanthroline ligands (Fig.\u00a0S9). Since cobalt reduction is required to evolve hydrogen, not seeing its drop suggests that the cobalt centre reduction and protonation steps are fast and just before the hydrogen evolution, i.e., changes at the cobalt not rate limiting and cannot be detected due to the NAP-XPS temporal resolution79.\n\nSimilar results occur in the absence of acid, but the onset potential of H2 evolution was shifted to 0.8\u2009V vs Ag/AgCl, i.e., about 0.15\u20130.2\u2009V, as observed in the catalysis (Fig.\u00a0S9). Furthermore, the Co 2p signal intensity detected was lower. The following hypotheses can be postulated to explain such a behaviour: (i) the liquid electrolyte film was thicker in this case, attenuating more the signal of the substrate; (ii) the surface of the electrode is slightly different in the absence of the acid, suggested by the more prominent electrode-related shoulder (centred at ca. 530.5\u2009eV) in the O 1\u2009s spectra (Fig.\u00a0S8).\n\nFigure\u00a07 shows a schematic representation of the two hypothetic catalytic cycles that can justify all the observations. Excitation of Au plasmon results in the formation of hot electrons and holes as part of its resonance decoherence via Landau damping. The hot holes are transferred to the NiO and react at the counter electrode, leading to O2 production. The electrons are transferred to the phenanthroline ligands. After both ligands acquire a charge (i.e., undergo reduction), two potential mechanisms are hypothesised: a simultaneous mechanism (blue arrow pathway) and a sequential mechanism (pink arrow pathway).\n\nThe figure shows two hypotheses consistent with the reported data: the simultaneous mechanism (blue arrow pathway) and the sequential mechanism (pink arrow pathway).\n\nIn the simultaneous mechanism, the catalyst experiences two rapid concerted proton-electron transfer (CPET) steps in the concurrent scenario. These steps lead to the reduction and protonation of the cobalt catalyst, ultimately resulting in H2 evolution. The swift nature of this mechanism poses a challenge as it prevents the detection of any changes in the Co 2p signal in NAP-XPS. On the other hand, the sequential mechanism involves sequential CPET processes. The initial CPET yields a CoIII-H hydride intermediate with an electronic structure similar to CoII. This interpretation aligns with the NAP-XPS observations and offers a more chemically sound reaction mechanism. Earlier investigations into Fe systems have identified analogous quasi-isoelectronic configurations80,81 associated with the single-electron reduction of the metal centre and the addition of a proton. This process effectively results in the formation of a hydride, accompanied by an increase in the metal centre\u2019s oxidation state by +1. Nevertheless, further investigations are required for confirmation, though this is beyond the scope of the current contribution.\n\nIn summary, a photosystem was proposed to confirm the direct involvement of hot electrons in a photocatalytic process, in this case, the H2 evolution process. The photosystem effectively mitigates the heat contribution by designing a catalyst that decomposes well below water thermolysis conditions, positioning heat as a mere enhancement rather than the primary cause for the observed H2 evolution. Off-resonance measurements conclusively eliminate near-field contributions as the direct catalyst of the reaction, although they may still play a role in extending the lifetime of hot electrons. The catalytic response to light modulation exhibits a shape consistent with the desired electron mechanism, contrasting with the detrimental impact of heat. Nevertheless, it was discovered that heat does have some positive influence on catalysis. Unbiased ultrafast spectroscopic measurements confirm charge transfer to respective acceptors.\n\nAdditionally, in conjunction with NAP-XPS under variable potential, a postulated reaction mechanism highlights the crucial role of cobalt catalyst ligands. These ligands accept plasmon hot electrons and, through CPET steps, reduce and protonate the metal centre, ultimately leading to hydrogen evolution. This study conclusively resolves the longstanding debate within the research community regarding the direct involvement of hot carriers in the photocatalytic process.",
+            "section_image": [
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-024-44752-y/MediaObjects/41467_2024_44752_Fig1_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-024-44752-y/MediaObjects/41467_2024_44752_Fig2_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-024-44752-y/MediaObjects/41467_2024_44752_Fig3_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-024-44752-y/MediaObjects/41467_2024_44752_Fig4_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-024-44752-y/MediaObjects/41467_2024_44752_Fig5_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-024-44752-y/MediaObjects/41467_2024_44752_Fig6_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-024-44752-y/MediaObjects/41467_2024_44752_Fig7_HTML.png"
+            ]
+        },
+        {
+            "section_name": "Methods",
+            "section_text": "The Au NPs were synthesised following the Turkevich method, as published by Piella et al.39. Briefly, sodium citrate tribasic dihydrate (Merck, ACS reagent\u2009\u2265\u200999%) 50\u2009mL (6.6\u2009mM) water solution was taken in a 100\u2009mL round bottom flask and stirred at 70\u2009\u00b0C in an oil bath. Then, 0.1\u2009mL (2.5\u2009mM) tannic acid (Merck, ACS reagent\u2009\u2265\u200999.5%) was added to the reaction mixture. Finally, 1\u2009mL of (25\u2009mM) HAuCl4 (Merck,\u2009\u2265\u200999.9%) was added instantly. After 5\u2009min, the reaction mixture changed from dark blue to wine. The colour change confirms the formation of the Au nanoparticles. The synthesised Au nanoparticles were stored in a fridge. The size of the Au nanoparticles was analysed using dynamic light scattering (DLS).\n\n200\u2009mg (1.02\u2009mmol) 1,10-Phenanthrolin-5-amine (Merck, 97%) and 148.42\u2009mg (0.51\u2009mmol) cobalt(II) nitrate hexahydrate (Merck, ACS reagent\u2009\u2265\u200998%) were refluxed in 20\u2009ml ethanol (Merck,\u2009\u2265\u200999.9% (GC)) for one hour. After that, the reaction mixture was filtered. The filtrate was placed in a clean beaker at room temperature for a few days without further disturbance, and we got a red precipitate. Finally, the obtained product was collected through filtration and dried in a vacuum desiccator. The yield 216\u2009mg (73.9%). ATR-IR 3342\u2009cm\u22121, 3402\u2009cm\u22121.\n\nThe NiO paste was purchased from Solaronix (Ni-Nanooxide N/SP,\u223c20 wt. %) and used as received. A small portion of the NiO paste was placed on a screen on top of the cleaned FTO glass and manually printed on the conducting side of the FTO glass. Then, the NiO-printed FTO glass plates were annealed at 500\u2009\u00b0C for 1\u2009h at a rate of 10\u2009\u00b0C/min.\n\nThe synthesised Au nanoparticles were sprayed manually on NiO films and annealed at 500\u2009\u00b0C for 1\u2009h at a rate of 10\u2009\u00b0C/min.\n\nThe annealed NiO-Au films were dipped into a 4\u2009mg/mL water solution of the [CoII(phen-NH2)2(H2O)2] catalyst for 3 days. Finally, we could get the self-assembled NiO/Au/[CoII(phen-NH2)2(H2O)2] composite system. The system was rinsed with water several times to remove unbounded catalyst molecules. For experimental purposes, we have sprayed Au nanoparticles only on FTO glass and on FTO/NiO and attached the molecular linker using the same procedure to deposit the catalyst.\n\nThe ICP-OES was used to estimate the cobalt amount in the sample. The film was digested in 4\u2009mL of nitric acid (FisherScientific, Nitric acid 65%) for several hours. Before measurement, the small probe was diluted 10 times with Milli-Q water containing 2% HNO3 and filtered with 0.2\u2009\u03bcm syringe filters (Whatman). Avio 200 Scott/Cross-Flow Configuration was used for ICP measurements. A calibration curve was formed for the measurements using a Cobalt Calibration Standard (CPAchem). Concentrations of 0, 0.1, 1 and 10\u2009ppm of the Co were used to create a 4-point linear regression. All measured values are within a relative standard deviation (RSD) of 2%.\n\nThe electrochemical data were measured using an EmStat potentiostat instrument. For the electrochemical experiments, a typical cylindrical closed cell was used. The FTO films were placed on the side of the cell so that it could face the light. Acetic acid (Merck, ACS reagent\u2009\u2265\u200999.7%) with a 3\u2009M concentration was used as a proton source.\n\nIn the presence of 0.1\u2009M LiCl (Merck, ACS reagent\u2009\u2265\u200999%) as a supporting electrolyte and glassy carbon as a working electrode, a Pt wire as a counter electrode, Ag/AgCl (3\u2009M KCl in water, Merck) as reference electrode was used for the bulk electrolysis experiment.\n\nThe photoelectrode exposed area to light is 0.79\u2009cm2. Plasmonic excitation was performed with a 532\u2009nm laser of 43.8\u2009mW/cm2 intensity. Pt wire counter electrode and Ag/AgCl (3\u2009M KCl in water) reference electrode were purchased from Redox.me and used as received. Lithium chloride purchased from Merck was used as a supporting electrolyte without further purification. Acetic acid (Merck, ACS reagent\u2009\u2265\u200999%) was used as a proton source.\n\nWe measured the gas produced during the photoelectrochemical reaction using a quadrupole mass spectrometer (QMS) (HPR 20) from Hiden Analytical. Continuous argon gas flow (15\u2009mL/min) through the cylindrical electrochemical cell during the measurement. Before applying the electrochemical potential, we saturated the electrochemical cell with argon to get a stable argon signal. Argon, O2, and H2 were measured continuously with the SEM detector throughout the measurement.\n\nA 40-fs pulsed laser with a 3\u2009kHz repetition rate was generated through the Libra Ultrafast Amplifier System designed by Coherent. An optical parametric oscillator (TOPAS- prime, Light Conversion) created the excitation beam. The signals were detected with a UV-NIR detector from Newport MS260i spectrograph with interchangeable gratings. The fundamental laser (probe, 795\u2009nm) passes through the delay stage (1\u20132\u2009fs step size) and is focused in a Sapphire optical window to generate visible light from 400 to 750\u2009nm. The instrument response function obtained for our system is ca. 95\u2009fs.\n\nA 40-fs pulsed laser with a 3\u2009kHz repetition rate was generated through the Libra Ultrafast Amplifier System designed by Coherent. Two optical parametric oscillators (TOPAS- prime, Light Conversion) created the excitation beam and the probe light in the mid-IR (3000\u201310000\u2009nm). The signals were detected with a Horiba iHR 320 spectrometer. The pump laser power was constantly monitored with less than a 2% standard deviation. The timing resolution, i.e., the instrument response function, is ca. 100\u2009fs.\n\nDetails about the experimental chamber and the three-electrode setup have been described elsewhere70. NAP-XPS experiments were carried out at the PHOENIX I beamline of the Swiss Light Source Synchrotron (SLS), making use of the solid-liquid interface endstation in a three-electrode setup using a gold counter electrode and an Ag/AgCl reference electrode controlled via a potentiostat (BioLogic Science Instruments SP-300)70. Linearly polarised light was used throughout the experiments. The as-introduced sample was first analysed under high vacuum conditions to acquire reference spectra. Then, the chamber was opened, and the beaker containing the pre-deaerated electrolyte was introduced (see Fig.\u00a0S1). The chamber was pumped down in a controlled way, using a needle valve, to avoid electrolyte spilling and favour the pressure equilibration (around 20\u2009mbar). Measurements were carried out using an excitation energy of 5000\u2009eV.\n\nDeconvolution of the O 1\u2009s spectra was performed after removing a Shirley background. Gaussian and Voigt-shaped components, whose positions were set according to past literature reports, were used to obtain the best correlation with experimental data (see Fig.\u00a0S15). Fitting parameters (peak positions, full width at half maximum \u2013FWHM- and % of Lorentian-Gaussian) are summarised in Table\u00a0S1.",
+            "section_image": []
+        },
+        {
+            "section_name": "Data availability",
+            "section_text": "The data related to the figures in the paper are provided as Excel files in Source data. Additional data supporting this study\u2019s findings are available from the corresponding author upon request.\u00a0Source data are provided with this paper.",
+            "section_image": []
+        },
+        {
+            "section_name": "References",
+            "section_text": "S\u00e1, J. et al. Direct observation of charge separation on Au localized surface plasmon. Energy Environ. Sci. 6, 3584\u20133588 (2013).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMubeen, S. et al. An autonomous photosynthetic device in which all charge carriers derive from surface plasmons. Nat. Nanotechnol. 8, 247\u2013252 (2013).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nAslam, U., Rao, V. G., Chavez, S. & Linic, S. Catalytic conversion of solar to chemical energy on plasmonic metal nanostructures. Nat. Catal. 1, 656\u2013665 (2018).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLi, R. et al. Unassisted highly selective gas-phase CO2 reduction with a plasmonic Au/p-GaN photocatalyst using H2O as an electron donor. ACS Energy Lett. 6, 1849\u20131856 (2021).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKumari, G., Zhang, X., Devasia, D., Heo, J. & Jain, P. K. Watching visible light-drive CO2 reduction on a plasmonic nanoparticle catalyst. ACS Nano 12, 8330\u20138340 (2018).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMarimuthu, A., Zhang, J. & Linic, S. Tuning selectivity in propylene epoxidation by plasmon mediated photo-switching of Cu oxidation state. Science 339, 1590\u20131593 (2013).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLinic, S., Aslam, U., Boerigter, C. & Morabito, M. Photochemical transformations on plasmonic metal nanoparticles. Nat. Mater. 14, 567\u2013576 (2015).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nVadai, M., Angell, D. K., Hayee, F., Sytwu, K. & Dionne, J. A. In-situ observation of plasmon-controlled photocatalytic dehydrogenation of individual palladium nanoparticles. Nat. Commun. 9, 4658 (2018).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nContreras, E. et al. Plasmon-assisted ammonia synthesis. J. Am. Chem. Soc. 144, 10743\u201310751 (2022).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nChen, K. & Wang, H. Plasmon-driven photocatalytic molecular transformations on metallic nanostructure surfaces: mechanistic insights gained from plasmon-enhanced Raman spectroscopy. Mol. Syst. Des. Eng. 6, 250\u2013280 (2021).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nFrontiera, R., Guenke, N. L. & van Duyne, R. Fano-line resonaces arising from long-lived molecule-plasmon interactions in colloidal nanoantennas. Nano Lett. 12, 5989\u20135994 (2012).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWilson, A. J., Mohan, V. & Jain, P. K. Mechanistic understading of plasmon-enhanced electrochemistry. J. Phys. Chem. C 123, 29360\u201329369 (2019).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBrongersma, M. L., Halas, N. J. & Nordlander, P. Plasmon-induced hot carrier science and technology. Nat. Nanotechnol. 10, 25\u201334 (2015).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSivan, Y., Un, I. W. & Dubi, Y. Assistance of metal nanoparticles in photocatalysis \u2013 nothing more than a classical heat source. Faraday Discuss. 214, 215\u2013233 (2019).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBaffou, G., Bordacchini, I., Baldi, A. & Quidant, R. Simple experimental procedures to distinguish photothermal from hot-carrier processes in plasmonics. Light Sci. Appl. 9, 108 (2020).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nZhan, C. et al. Disentangling charge carrier from photothermal effects in plasmonic metal nanostructures. Nat. Commun. 10, 2671 (2019).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nZhang, X. et al. Plasmon-enhanced catalysis: distinguishing thermal and nonthermal effects. Nano Lett. 18, 1714\u20131723 (2018).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nJain, P. K. Taking the heat off of plasmonics chemistry. J. Phys. Chem. C 123, 24347\u201324351 (2019).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKamarudheen, R., Aalbers, G. J. W., Hamans, R. F., Kamp, L. P. J. & Baldi, A. Distinguishing among all possible activation mechanisms of a plasmon-driven chemical reaction. ACS Energy Lett. 5, 2605\u20132613 (2020).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nRossi, T. P., Erhart, P. & Kuisma, M. Hot-carrier generation in plasmonic nanoparticles: the importance of atomic structure. ACS Nano 14, 9963\u20139971 (2020).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nFann, W. S., Strotz, R., Tom, H. W. K. & Bokor, J. Electron thermalization in gold. Phys. Rev. B 46, 13592 (1992).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLink, S. & El-Sayed, M. A. Spectral properties and relaxation dynamics of surface plasmon electronic oscillations in gold and silver nanodots and nanorods. J. Phys. Chem. B 103, 8410\u20138426 (1999).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nFurube, A., Du, L., Hara, K., Katoh, R. & Tachiya, M. Ultrafast plasmon-induced electron transfer from gold nanodots into TiO2 nanoparticles. J. Am. Chem. Soc. 129, 14852\u201314853 (2007).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nTagliabue, G. et al. Ultrafast hot-hole injection modifies hot-electron dynamics in Au/p-GaN heterostructures. Nat. Mater. 19, 1312\u20131318 (2020).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHattori, Y., Abdellah, M., Meng, J., Zheng, K. & S\u00e1, J. Simultaneous hot electron and hole injection upon excitation of gold surface plasmon. J. Phys. Chem. Lett. 10, 3140\u20133146 (2019).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nOdobel, F. & Pellegrin, Y. Recent advances in the sensitization of wide-band-gap nanostructured P-type semiconductors. Photovoltaic and photocatalytic applications. J. Phys. Chem. Lett. 4, 2551\u20132564 (2013).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGardner, J. M. et al. Light-driven electron transfer between a photosensitizer and a proton-reducing catalyst co-adsorbed to NiO. J. Am. Chem. Soc. 134, 19322\u201319325 (2012).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nTong, L. et al. Sustained solar hydrogen generation using a dye-sensitized NiO photocathode/BiVO4 tandem photo-electrochemical device. Energy Environ. Sci. 5, 9472\u20139475 (2012).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nNakamura, K. et al. Properties of plasmon-induced photoelectric conversion on a TiO2/NiO p\u2013n junction with Au nanoparticles. J. Phys. Chem. Lett. 7, 1004\u20131009 (2016).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLuo, S.-P., Tang, L.-Z. & Zhan, S.-Z. A cobalt(II) complex of 2,2-bipyridine, a catalyst for electro- and photo-catalytic hydrogen production in purely aqueous media. Inorg. Chem. Commun. 86, 276\u2013280 (2017).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSharma, S., Toupet, L. & Arjmand, F. De novo design of a hydrolytic DNA cleavage agent, mono nitratobis(phen)cobalt(II) aqua nitrate complex. N. J. Chem. 41, 2883\u20132886 (2017).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHydrogen Production: Thermochemical Water Splitting, accessed 20 February 2023; https://www.energy.gov/eere/fuelcells/hydrogen-production-thermochemical-water-splitting.\n\nSav\u00e9ant, J. M. Electrochemical approach to proton-coupled electron transfers: recent advances. Energy Environ. Sci. 5, 7718\u20137731 (2012).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSheng, H. et al. Strong synergy between gold nanoparticles and cobalt porphyrin induces highly efficient photocatalytic hydrogen evolution. Nat. Commun. 14, 1528 (2023).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nAl-Omair, M. A. Biochemical activities and electronic spectra of different cobalt phenanthroline complexes. Arab. J. Chem. 12, 1061\u20131069 (2019).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKatzin, L. I. & Gebert, E. Spectrophotometric investigation of cobaltous nitrate in organic solvents. J. Am. Chem. Soc. 72, 5455\u20135463 (1950).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLazar, P., Mach, R. & Otyepka, M. Spectrophotometric investigation of cobaltous nitrate in organic solvents. J. Am. Chem. Soc. 72, 5455\u20135463 (1950).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBeard, B. C. Cellulose nitrate as a binding energy reference in N(1s) XPS studies of nitrogen-containing organic molecules. Appl. Surf. Sci. 45, 221\u2013227 (1990).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nPiella, J., Bast\u00fas, N. G. & Puntes, V. Size-controlled synthesis of sub-10-nanometer citrate-stabilized gold nanoparticles and related optical properties. Chem. Mater. 28, 1066\u20135463 (2016).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSilveira, V. R., Bericat-Vadell, R. & S\u00e1, J. Photoelectrocatalytic conversion of nitrates to ammonia with plasmon hot electrons. J. Phys. Chem. C 127, 5425\u20135431 (2023).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGhosh, S. K., Nath, S., Kundu, S., Esumi, K. & Pal, T. Solvent and ligand effects on the localized surface plasmon resonance (LSPR) of gold colloids. J. Phys. Chem. B 108, 13963\u201313971 (2004).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBurness, J. H., Dillard, J. G. & taylor, L. T. An x-ray photoelectron spectroscopy study of cobalt (II) schiff base complexes and their oxygenation products. J. Am. Chem. Soc. 97, 6080\u20136088 (1975).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nPavliuk, M. V. et al. Nano-hybrid plasmonic photocatalyst for hydrogen production at 20% efficiency. Sci. Rep. 7, 8670 (2017).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nIR Spectroscopy Tutorial: Amines, accessed 9 August 2023 https://www.orgchemboulder.com/Spectroscopy/irtutor/aminesir.shtml.\n\nSoorholtz, M. et al. Local platinum environments in a solid analogue of the molecular periana catalyst. ACS Catal. 6, 2332\u20132340 (2016).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGunasekar, G. H. et al. A covalent triazine framework, functionalized with Ir/N- heterocyclic carbene sites, for the efficient hydrogenation of CO2 to formate. Chem. Mater. 29, 6740\u20136748 (2017).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBeamson, G., Briggs. D. The XPS of polymers database. https://www.worldcat.org/title/xps-of-polymers-database/oclc/52147743 (Accessed 26/10/2023)\n\nArtemenko, A. et al. Reference XPS spectra of amino acids. IOP Conf. Ser. Mater. Sci. Eng. 1050, 012001 (2021).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSeah, M. P., Smith, C. G. & Anthony, M. T. AES: energy calibration of electron spectrometers. I\u2014an absolute, traceable energy calibration and the provision of atomic reference line energies. Surf. Interface Anal. 17, 293\u2013308 (1990).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nRossi, T. P., Shegai, T., Erhart, P. & Antosiewicz, T. J. Strong plasmon-molecule coupling at the nanoscale revealed by first-principles modeling. Nat. Commun. 10, 3336 (2019).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWang, P. et al. Electronic and steric tuning of catalytic H2 evolution by cobalt complexes with pentadentate polypyridyl-amine ligands. J. Am. Chem. Soc. 140, 9219\u20139229 (2018).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWang, P. et al. Enhanced hydrogen evolution in neutral water catalyzed by a cobalt complex with a softer polypyridyl ligand. Angew. Chem. Int. Ed. 59, 12694\u201312697 (2020).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBoyd, S. J., Long, R. & English, N. J. Electric field effects on photoelectrochemical water splitting: perspectives and outlook. Energies 15, 1553 (2022).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKholmicheva, N. et al. Plasmonic nanocrystal solar cells utilizing strongly confined radiation. ACS Nano 8, 12549\u201312559 (2014).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKholmicheva, N., Romero, L. R., Cassidy, J. & Zamkov, M. Prospects and applications of plasmon-exciton interactions in the near-field regime. Nanophotonics 8, 613\u2013628 (2019).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nCushing, S. K. et al. Controlling plasmon-induced resonance energy transfer and hot electron injection processes in metal@TiO2 core-shell nanoparticles. J. Phys. Chem. C 119, 16239\u201316244 (2015).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSun, Q. et al. Direct imaging of the near field and dynamics of surface plasmon resonance on gold nanostructures using photoemission electron microscopy. Light 2, e118 (2013).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHe, J., Fan, C., Ding, P., Zhu, S. & Liang, E. Near-field engineering of Fano resonance in a plasmonic assembly for maximizing CARS enhancements. Sci. Rep. 6, 20777 (2016).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHermann, R. J. & Gordon, M. J. Quantitative comparison of plasmon resonances and field enhancements of near-field optical antennae using FDTD simulations. Opt. Exp. 26, 27668 (2018).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nTagliabue, G. et al. Quantifying the role of surface plasmon excitation and hot carrier transport in plasmonic devices. Nat. Commun. 9, 3394 (2018).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBard, A. J., Faulkner, L. R. Electrochemical Methods 2nd edn, (Wiley, 2001).\n\nvan Turnhout, L., Hattori, H., Meng, J., Zheng, K. & S\u00e1, J. Direct observation of a plasmon-induced hot electron flow in a multimetallic nanostructure. Nano Lett. 20, 8220\u20138228 (2020).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGroeneveld, R. H. M., Sprik, R. & Lagendijk, A. Femtosecond spectroscopy of electron-electron and electron-phonon relaxation in Ag and Au. Phys. Rev. B 51, 11433 (1995).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBigot, J.-Y., Merle, J.-C., Cregut, O. & Daunois, A. Electron dynamics in copper metallic nanoparticles probed with femtosecond optical pulses. Phys. Rev. Lett. 75, 4702\u20134705 (1995).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBrown, A. M. et al. Experimental and Ab initio ultrafast carrier dynamics in plasmonic nanoparticles. Phys. Rev. Lett. 118, 087401 (2017).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nAntila, L. J., Santomauro, F. G., Hammastr\u00f6m, L., Fernandes, D. L. A. & S\u00e1, J. Hunting for the elusive shallow traps in TiO2. Chem. Commun. 51, 10914\u201310916 (2015).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBerger, T. et al. Light-induced charge separation in anatase TiO2 particles. J. Phys. Chem. B 109, 6061\u20136068 (2005).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nQueyriaux, N. et al. Electrocatalytic hydrogen evolution with a cobalt complex bearing pendant proton relays: acid strength and applied potential govern mechanism and stability. J. Am. Chem. Soc. 142, 274\u2013282 (2020).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nEckenhoff, W. T., McNamara, W. R., Du, P. & Eisenberg, R. Cobalt complexes as artificial hydrogenases for reductive side of water splitting. Biochim. Biophys. Acta Bioenerg. 1827, 958\u2013973 (2013).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nNovotny, Z. et al. Probing the solid\u2013liquid interface with tender x rays: a new ambient-pressure x-ray photoelectron spectroscopy endstation at the Swiss Light Source. Rev. Sci. Instrum. 91, 023103 (2020).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMajor, G. H., Fernandez, V., Failey, N., Smith, E. F. & Linford, M. R. Guide to XPS data analysis: applying appropriate constraints to synthetic peaks in XPS peak fitting. J. Vac. Sci. Technol. A 40, 063201 (2022).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nAxnanda, S. et al. Using \u201ctender\u201d x-ray ambient pressure x-ray photoelectron spectroscopy as a direct probe of solid-liquid interface. Sci. Rep. 5, 9788 (2015).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nFavaro, M. et al. Unravelling the electrochemical double layer by direct probing of the solid/liquid interface. Nat. Commun. 7, 12695 (2016).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKrepelov\u00e1, A., Bartels-Rausch, T., Brown, M. A., Bluhm, H. & Ammann, M. Adsorption of acetic acid on ice studied by ambient-pressure XPS and partial-electron-yield NEXAFS spectroscopy at 230\u2013240 K. J. Phys. Chem. A 117, 401\u2013409 (2013).\n\nArticle\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nL\u00e1zaro-Mart\u00ednez, J. M., Lupano, L. V. L., Piehl, L. L., Rodr\u00edguez-castell\u00f3n, E. & Dall\u2019Orto, V. C. New insights about the selectivity in the activation of hydrogen peroxide by cobalt or copper hydrogel heterogeneous catalysts in the generation of reactive oxygen species. J. Phys. Chem. C 120, 29332\u201329347 (2016).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKhalil, T. E. et al. Synthesis, structure, X-ray photoelectron spectroscopy (XPS), and antimicrobial, anticancer, and antioxidant activities of Co (III) complexes based on the antihypertensive hydralazine. Appl. Org. Chem. 36, e6565 (2022).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBridge, M. E. & Lambert, R. M. Oxygen chemisorption, surface oxidation, and the oxidation of carbon monoxide on cobalt (0001). Surf. Sci. 82, 413\u2013424 (1979).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHyman, M. P. & Vohs, J. M. Reaction of ethanol on oxidized and metallic cobalt surfaces. Surf. Sci. 605, 383\u2013389 (2011).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMoonshiran, D. et al. Tracking the structural and electronic configurations of a cobalt proton reduction catalyst in water. J. Am. Chem. Soc. 138, 10586\u201310596 (2016).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMormohades, M. et al. Direct observation of key catalytic intermediates in a photoinduced proton reduction cycle with a diiron carbonyl complex. J. Am. Chem. Soc. 136, 17366\u201317369 (2014).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nAster, A. et al. Metal vs. ligand protonation and the alleged proton-shuttling role of the azadithiolate ligand in catalytic H2 formation with FeFe hydrogenase model complexes. Chem. Sci. 10, 5582\u20135588 (2019).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nDownload references",
+            "section_image": []
+        },
+        {
+            "section_name": "Acknowledgements",
+            "section_text": "the authors thank the Paul Scherrer Institute for providing access to the Phoenix beamline at the Swiss Light Source. The authors also thank Ms Johanna Andersson from Uppsala University for the lab access and her support with complex melting point measurements. The authors also thank Prof. Leif Hamarstr\u00f6m from Uppsala University for the productive discussions about molecular catalyst reaction mechanisms and CPET aspects. Finally, the authors thank the MyFab clean room facility at the Angstrom Laboratory-Uppsala University for access to the clean room instrumentation. This work was funded by the Olle Engkvists stiftelse [210-0007 (J.S.)]; Knut & Alice Wallenberg Foundation [2019-0071(J.S.)]; Swedish Research Council [2019-03597 (J.S.)]; and Polish Ministry and Higher Education [1/SOL/2021/2 (A.W.)].",
+            "section_image": []
+        },
+        {
+            "section_name": "Funding",
+            "section_text": "Open access funding provided by Uppsala University.",
+            "section_image": []
+        },
+        {
+            "section_name": "Author information",
+            "section_text": "Department of Chemistry-\u00c5ngstr\u00f6m, Physical Chemistry division, Uppsala University, Box 532, 751 20, Uppsala, Sweden\n\nAnanta Dey,\u00a0Amal Mendalz,\u00a0Robert Bericat Vadell,\u00a0Vitor R. Silveira\u00a0&\u00a0Jacinto S\u00e1\n\nPaul Scherrer Institut, CH-5232, Villigen PSI, Switzerland\n\nAnna Wach,\u00a0Paul Maurice Leidinger,\u00a0Thomas Huthwelker,\u00a0Zbynek Novotny\u00a0&\u00a0Luca Artiglia\n\nSOLARIS National Synchrotron Radiation Centre, Jagiellonian University, Krakow, Poland\n\nAnna Wach\n\nDepartment of Materials Science and Engineering, division of Applied Materials Science, Uppsala University, 75103, Uppsala, Sweden\n\nVitalii Shtender\n\nInstitute of Physical Chemistry, Polish Academy of Sciences, Marcina Kasprzaka 44/52, 01-224, Warsaw, Poland\n\nJacinto S\u00e1\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nA.D. and J.S. conceived the idea and designed experiments. A.D. and A.M. prepared the materials. A.D., A.M., R.B.V., V.R.S., A.W., P.M.L., Z.N., C.S., L.A. and J.S. performed the basic characterization, electrochemical studies, catalysis and advanced spectroscopies. A.D., R.B.V., V.R.S., A.W., P.M.L., Z.N., C.S., L.A. and J.S. analysed the experimental data and co-wrote the manuscript. All authors contributed to the discussions and manuscript preparation.\n\nCorrespondence to\n                Jacinto S\u00e1.",
+            "section_image": []
+        },
+        {
+            "section_name": "Ethics declarations",
+            "section_text": "The authors declare no competing interests.",
+            "section_image": []
+        },
+        {
+            "section_name": "Peer review",
+            "section_text": "Nature Communications thanks Emiliano Cort\u00e9s, Padmanabh Joshi and the other, anonymous, reviewer for their contribution to the peer review of this work. A peer review file is available.",
+            "section_image": []
+        },
+        {
+            "section_name": "Additional information",
+            "section_text": "Publisher\u2019s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.",
+            "section_image": []
+        },
+        {
+            "section_name": "Source data",
+            "section_text": "",
+            "section_image": []
+        },
+        {
+            "section_name": "Rights and permissions",
+            "section_text": "Open Access  This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article\u2019s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article\u2019s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.\n\nReprints and permissions",
+            "section_image": []
+        },
+        {
+            "section_name": "About this article",
+            "section_text": "Dey, A., Mendalz, A., Wach, A. et al. Hydrogen evolution with hot electrons on a plasmonic-molecular catalyst hybrid system.\n                    Nat Commun 15, 445 (2024). https://doi.org/10.1038/s41467-024-44752-y\n\nDownload citation\n\nReceived: 22 April 2023\n\nAccepted: 03 January 2024\n\nPublished: 10 January 2024\n\nVersion of record: 10 January 2024\n\nDOI: https://doi.org/10.1038/s41467-024-44752-y\n\nAnyone you share the following link with will be able to read this content:\n\nSorry, a shareable link is not currently available for this article.\n\n\n\n\n                            Provided by the Springer Nature SharedIt content-sharing initiative\n                        ",
+            "section_image": [
+                "https://data:image/svg+xml;base64,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"
+            ]
+        },
+        {
+            "section_name": "This article is cited by",
+            "section_text": "Nature Communications (2025)\n\nJournal of Sol-Gel Science and Technology (2025)\n\nScience China Chemistry (2025)\n\nJournal of Nanoparticle Research (2025)\n\nCommunications Chemistry (2024)",
+            "section_image": []
+        },
+        {
+            "section_name": "Associated content",
+            "section_text": "Collection",
+            "section_image": []
+        }
+    ]
+}

051f366d2a3bfb6b436ca3def5886dd6413dadc2fc06e01a744274e96766842f/peer_review/peer_review.md

@@ -0,0 +1,299 @@
+Peer Review File
+
+ZFP36-family RNA-binding proteins in regulatory T cells reinforce immune homeostasis.
+
+Corresponding Author: Dr Martin Turner
+
+This file contains all reviewer reports in order by version, followed by all author rebuttals in order by version.
+
+Version 0:
+
+Reviewer comments:
+
+Reviewer #1
+
+(Remarks to the Author)
+Review of MS: NCOMMS-24-61322-T entitled “ZFP36-family RNA-binding proteins in regulatory T cells reinforce immune homeostasis”. The study by Sáenz-Narciso and colleagues investigates combined deletion of Zfp36l1 and Zfp36l2 in Treg cells. Different from deletion with CD4-Cre, deletion with Foxp3-IRES-YFP-Cre causes the induction of effector memory CD4 and CD8 T cells and the development of fatal autoimmune or autoinflammatory disease. The disease manifests for example in a colitis phenotyp. The authors set out to find out how Treg cells are changed in their gene expression and function, and they make a number of remarkable observations including differential CTLA-4 localization, hypersensitivity to IL-2 and IL-7 and hypersensitivity to Ifng signaling. Finally, the authors rescue the fatal disease by genetic combination with one Ifng knockout allele. The manuscript shows a thorough investigation and clear presentation, although the direct relationship which targets cause which phenotypes remain unsolved, however, providing such causal proof may exceed the scope of this paper. I only have a few points to help improve the paper.
+Major points
+1. The authors make elegant approaches to compare what happens in Treg cells when Zip36l1 and Zip36l2 are knocked out– either in all Tregs of (male) mice, or in a fraction of Treg cells – in (female) mice. This approach should also be used to ask which Treg intrinsic phenotypes are rescued by heterozygous Ifng deletion. Is the competitive fitness restored, how are the phenotypes of Tfr, eTreg affected.?
+2. The authors intersect differentially regulated genes in Treg cells with the identified targets of Zip36l1/l2 proteins in CD4+ and CD8+ T cells, and they identify interesting categories. Although this is done in a very sophisticated manner, one does not get to know the drivers of the phenotypes. Within the deregulated targets, one would expect the stronger deregulated targets having stronger influences on the observed phenotype. Nevertheless, the authors seem to argue that numerous subtle changes together build the very pronounced phenotype. It would be informative if targets that can explain the phenotype, were also tested for deregulation in the right cell type/activation staget etc.. For example, are the SOCS family members involved in cytokine sensitivity? Which genes/targets can be involved in CTLA-4 localization?
+3. The authors prevent the development of disease with inactivation of one Ifng allele, a prototypic Zip36 target with functional AU-rich elements. However, this approach does not easily explain why inactivation of Zip36l1 and Zip36l2 with CD4-Cre does not cause a similar disease, if the excess of Ifng originates from conventional T cells. One possible explanation for the observed phenotype could be that ex-Treg cells with autoreactive TCRs develop, if Zip36l1/l2 RBPs are inactivated with Foxp3-Cre. Those may be superior, tissue localized Ifng producers, since Ifng expression is induced in the exTregs by TCR signals, and is no longer repressed by Zip36 proteins binding to AU-rich elements. Ex-Tregs would have downregulated YFP and would therefore not be captured in the approaches shown in the paper so far. This possibility should be addressed experimentally, or at least mentioned in the “Limitations of this study” section.
+
+Reviewer #2
+
+(Remarks to the Author)
+The study “ZFP36-family RNA-binding proteins in regulatory T cells reinforce immune homeostasis” by Sáenz-Narciso and Bell et al. investigates the roles of Zip36, Zip36l1 and Zip36l2 in mouse Treg cells. The authors use conditional Foxp3-specific knock-out mice for each factor (or double-KO for Zip36l1 and Zip36l2) to determine their individual roles. While deletion of Zip36 or Zip36l2 had no apparent effect, the absence of Zip36l1 in Tregs caused a systemic hyper immune activation, which was even more pronounced when Zfp36l1 and Zfp36l2 were both deleted at the same time.
+Mechanistically, the authors use bulk RNA-Seq and scRNA-Seq of Zfp36l1/l2-deficient Tregs to reveal potential underlying mechanisms. While the core Treg signature appears stable in KO Treg cells, decreased expression of CD25 and CD127 seems to dampen IL-2/IL-7 sensitivity of KO Tregs, likely dampening their fitness. In contrast, KO Tregs respond more to IFNg, while also producing more IFNg themselves. Finally, the authors show that most of the hyper immune phenotype of Treg-specific Zfp36l1/l2-deficient mice is due to excessive IFNg signalling, since genetic deletion of one allele of Ifng is sufficient to restore homeostasis.
+
+In summary, the study is likely to be of considerable interest for the readers in the field. However, the authors missed to work out the precise molecular mechanisms and targets of how Zfp36 family members control Treg cell function. Also it remains mechanistically unclear, why Zfp36l1/l2-deficient Tregs fail to control especially type 1 IFNg-producing T cells. Therefore, the paper mostly provides a comprehensive characterization of the respective KO mice.
+
+Specific points:
+
+Figure 1:
+
+The authors report that 9% of FYCI1 mice and 12% of FYCI1/2 develop symptoms. And as far as I understand, the authors focus their subsequent analysis entirely on these sick mice. What about the majority of mice that don’t develop symptoms? Are they completely healthy? Do they show any differences in effector CD4+/CD8+ cells? How do the Treg cells look like in these mice?
+It would be helpful to show a more comprehensive analysis of key clinical parameters (e.g. weight change over time, pathology scores) for all mice and (immune) phenotyping for symptomatic vs. non-symptomatic mice.
+Overall, the unanswered question remains: Why do some KO mice experience a loss of immune homeostasis, while others not?
+
+Figure 2:
+
+I am missing a clearly unbiased analysis of the bulk RNA-Seq data from WT vs. KO Tregs. Right now, the authors only use a custom pre-defined group of gene sets for their enrichment analysis, to basically pave the way for their findings regarding IL-2/IL-7 sensitivity. It feels like these findings were already known before the RNA-Seq.
+I think a more explorative approach using unbiased GSEA for example could reveal additional signaling pathways or metabolic alterations in KO Tregs, that might contribute to the diminished fitness and function of Zfp36l1/l2-deficient Tregs. Additionally, it remains unclear which molecular targets are specifically bound be Zfp36l1/l2 in Treg cells. The authors use published CLIP data from activated CD4+ T cells, however CLIP data directly from Treg cells would certainly be more helpful to elucidate the molecular mechanism. Such data would also be a valuable resource for the community.
+
+Figure 7:
+
+The analysis of cytokine expression by Tcon and Treg in Fig. 7i is really informative and reveals that there is a major dysregulation of type 1 IFN-g-producing effector T cells in FYCI 1/2 mice, while type 2 and type 3 immunity seems rather unaffected. It also shows that KO Tregs do not have a problem in producing IL-10, ruling out this as a potential mechanism underlying the breakdown of immune homeostasis.
+There is the concept that type 1 T cells (eg Th1) are primarily regulated by Th1-like Treg cells, that mimic Th1 cells in terms of e.g. T-bet expression etc. The question that comes to my mind here is: Do KO Tregs cannot acquire a proper and functional Th1-like Treg phenotype and is this the reason for their reduced capacity to suppress type 1 immune cells?
+I am asking this bc the paper is overall lacking the explanation why Zfp36l1/l2-deficient Tregs cannot properly suppress (Can they actually suppress normally in conventional in vitro suppression assays?) and why specifically not type 1 immune cells. More data addressing this question would certainly increase the scope of the paper immensely.
+
+Minor comment an Figure 7i and Sup. Fig. 8. The staining of RORgt+ Tregs doesn’t really show any pos. population to me. I suggest to stain for these Treg subsets (also the GATA3+ Treg) in intestinal Treg cells rather than Treg cells from Spleen/LN.
+
+Reviewer #3
+
+(Remarks to the Author)
+In this study, Saenz-Narciso and colleagues investigate the role of ZFP36 family RBPs in Tregs. They specifically ablate Zfp36l1, Zfp36l2, and Zfp36l1l2 in Tregs and find that the loss of either protein or the combined loss of both leads to autoimmunity, as evidenced by the expansion of activated CD4 and CD8 T cells expressing proinflammatory cytokines. Mechanistically, these RBPs regulated multiple aspects of Treg biology, including endocytosis and sensitivity to cytokines, both of which had significant implications for Treg stability and function. The authors further demonstrated that IFNg drives the inflammatory phenotype in mice lacking Zfp36l1l2 specifically in Tregs, noting that deletion of one allele of Ifng is sufficient to ameliorate this phenotype. This is an excellent study with high-quality data; however, a few issues require clarification.
+
+1. The authors state that only 9% of FYCI1 and 13% of FYCI1l2 mice develop autoimmunity. Why is the disease penetrance not 100%? It appears the authors focus solely on male mice. Is there a specific reason for not analysing homozygous knockout females?
+2. In female FYC112/+ mice, the Foxp3+YFP+ population was outcompeted by Foxp3+YFP-WT Tregs, suggesting that Tregs lacking I112 have reduced competitive fitness. Do knockout Tregs exhibit an activated phenotype in this context (with competition), or is there a skewed nTreg/eTreg ratio?
+
+3. At the RNA level, there is downregulation of Ctla4, which appears to correlate with low protein levels of total CTLA4, despite the claim of endocytosis. What is the underlying cause of the downregulation of CTLA4 protein?
+
+4. Is defective STAT5 signalling responsible for the lack of competitive fitness in FYC112 mice? Can the IL-2/Ab complex expand YFP+ Tregs in FYC112/+ mice?
+
+5. FYC112 Tregs produce more IFNg and IL-10 when stimulated. Is this phenotype a result of inflammation in these mice? Would YFP+ Tregs in FYC112/+ mice display the same phenotype? A major concern is distinguishing the actual phenotype of Tregs caused by I112 deletion from that induced by inflammation. For example, bulk RNA sequencing has been performed on Tregs from FYC112/+ mice (with no inflammation), while scRNA sequencing was conducted on Tregs from FYC112 mice (where inflammation is high). Can these two datasets be compared (using scRNA-seq pseudobulk analysis) to differentiate between inflammation-induced and genuine I112-dependent gene expression?
+
+6. What happens to Tregs in non-lymphoid tissues? Are there any defects in migration?
+
+Version 1:
+
+Reviewer comments:
+
+Reviewer #1
+
+(Remarks to the Author)
+The authors have adequately addressed all of my points and provided important new information.
+
+Reviewer #2
+
+(Remarks to the Author)
+The authors have addressed all my concerns. Thank you.
+
+Regarding Figure 1, I would suggest to include the histology data (now Sup. Fig. 1c) and the LN cellularity data (now Sup. Fig. 1d and 2a) into the main Figure, bc they are strong phenotypic data that don't need to be hidden in the supplement.
+
+The potential reasons why only a minor fraction of mice develops clinically symptoms, should be briefly discussed in the manuscript.
+
+Reviewer #3
+
+(Remarks to the Author)
+The authors have addressed all of my queries. I have no further concerns. This is an excellent study!
+Open Access This Peer Review File is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.
+In cases where reviewers are anonymous, credit should be given to 'Anonymous Referee' and the source.
+The images or other third party material in this Peer Review File are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
+To view a copy of this license, visit https://creativecommons.org/licenses/by/4.0/.
+Reviewer #1 (Remarks to the Author)
+
+Review of MS: NCOMMS-24-61322-T entitled “ZFP36-family RNA-binding proteins in regulatory T cells reinforce immune homeostasis”. The study by Sáenz-Narciso and colleagues investigates combined deletion of Zfp36l1 and Zfp36l2 in Treg cells. Different from deletion with CD4-Cre, deletion with Foxp3-IRES-YFP-Cre causes the induction of effector memory CD4 and CD8 T cells and the development of fatal autoimmune or autoinflammatory disease. The disease manifests for example in a colitis phenotype. The authors set out to find out how Treg cells are changed in their gene expression and function, and they make a number of remarkable observations including differential CTLA-4 localization, hypersensitivity to IL-2 and IL-7 and hypersensitivity to Ifn-g signaling. Finally, the authors rescue the fatal disease by genetic combination with one Ifng knockout allele. The manuscript shows a thorough investigation and clear presentation, although the direct relationship which targets cause which phenotypes remain unsolved, however, providing such causal proof may exceed the scope of this paper. I only have a few points to help improve the paper.
+
+We thank the reviewer for their positive comments and feedback on our manuscript which we have taken on board to improve the study and its presentation, we respond to each point below. Revisions/new text are highlighted in the revised manuscript.
+
+Major points
+1. The authors make elegant approaches to compare what happens in Treg cells when Zfp36l1 and Zfp36l2 are knocked out– either in all Tregs of (male) mice, or in a fraction of Treg cells – in (female) mice. This approach should also be used to ask which Treg intrinsic phenotypes are rescued by heterozygous Ifng deletion. Is the competitive fitness restored, how are the phenotypes of Tfr, eTreg affected?
+
+In the submitted version of the manuscript, we did not present data on the eTreg populations in FYC+ l1/2 female mice. In these mice, the proportions of eTregs amongst YFP+ Tregs were not different than those in FYC+ mice; thus, the expansion of eTreg in FYC l1/2 male mice is not a Treg intrinsic phenotype. We have now added this data as Supplementary fig. 3a.
+
+We have assessed whether the competitive disadvantage of YFP+ Tregs observed in FYC+ l1/2 mice is affected by heterozygous Ifng deletion. In Ifng+/− FYC+ l1/2 mice the proportions and numbers of YFP+ Tregs are not different to those in FYC+ l1/2 mice indicating that loss of one allele of Ifng does not restore the competitive fitness (Reviewer Figure 1).
+
+![Flow cytometry plots showing FOXP3 versus YFP expression in FYC+, FYC+ l1/2 and Ifng+/− FYC+ l1/2 mice, and proportions and numbers of Tregs gated as CD4+ FOXP3+ YFP+ or CD4+ FOXP3+ YFP+](page_153_1042_1142_312.png)
+
+Reviewer Figure 1. Loss of one allele of Ifng does not rescue the competitive disadvantage observed in YFP+ Tregs from FYC+/− l1/2 mice. Flow cytometry plots showing FOXP3 versus YFP expression in FYC+, FYC+ l1/2 and Ifng+/− FYC+ l1/2 mice (left panel); proportions and numbers of Tregs gated as CD4+ FOXP3+ YFP+ or CD4+ FOXP3+ YFP+ (right panel); key as shown.
+
+Additionally, we observed no difference in CD25 or CD127 expression between YFP+ Tregs from Ifng+/− FYC+ l1/2 and FYC+ l1/2 mice (Reviewer Figure 2). Thus, deletion of one allele of Ifng is insufficient to restore expression of these cytokine receptors.
+Reviewer Figure 2. Loss of one allele of Ifng does not rescue CD25 nor CD127 expression levels in YFP+ Tregs from FYC+ I1I2 mice. a) Representative histogram overlays of CD25 expression and CD25 gMFI on nTreg (left panel) and eTreg (right panel) from spleen; b) Representative histogram overlays of CD127 expression and CD127 gMFI on nTreg (left panel) and eTreg (right panel) from spleen; n=3; key as shown.
+
+We propose not to include these results in the revised manuscript.
+
+2. The authors intersect differentially regulated genes in Treg cells with the identified targets of Zfp36l1/l2 proteins in CD4+ and CD8+ T cells, and they identify interesting categories. Although this is done in a very sophisticated manner, one does not get to know the drivers of the phenotypes. Within the deregulated targets, one would expect the stronger deregulated targets having stronger influences on the observed phenotype. Nevertheless, the authors seem to argue that numerous subtle changes together build the very pronounced phenotype. It would be informative if targets that can explain the phenotype, were also tested for deregulation in the right cell type/activation stage etc.. For example, are the SOCS family members involved in cytokine sensitivity? Which genes/targets can be involved in CTLA-4 localization?
+
+ZFP36 proteins can regulate mRNA at multiple levels by direct binding to AU-rich elements in the 3'UTR of mRNAs. Thus, it is possible that multiple mechanisms contribute to the complex phenotype and only the effects on RNA decay can be inferred from our RNAseq data.
+
+Given the defect in CTLA-4 cycling we have performed a deeper analysis of the genes within the endocytosis gene set that are differentially expressed between naïve Treg from FYC+ and FYC+ I1I2 mice and that are present in the leading edge of our GSEA analysis (i.e., genes that show a trend towards increased expression). These data are now included in the new Supplementary Fig. S10 and S11 and we have added additional text to the results.
+
+In addition, we found that several genes that might play a role in endocytosis, and/or CTLA-4 cycling are not directly targeted by the ZFP36 family which prompted us to examine transcription factors and epigenetic regulators that might be important in their regulation. We identified transcriptional regulators that were frequently bound at the promotors of genes within the endocytosis GSEA leading edge (Supplementary Fig. 12). Many transcription factors were ZFP36-family targets by CLIP and several were known to have roles in Tregs, including Ets1, Stat1, Hopx, and Bcl11b.
+Overall, we identified numerous genes whose expression is altered by deletion of Zfp36l1/l2 that have the potential to impact on endocytosis, including both direct targets of the ZFP36 family, and genes regulated indirectly. How this complex network of alterations coalesces to result in the decreased cycling of CTLA-4 in Zfp36l1/Zfp36l2 KO Tregs will be an important area for future in-depth study.
+
+Regarding the changes in cytokine sensitivity, no SOCS family member showed a relevant increase at the mRNA level (*Reviewer Figure 3*).
+
+![Bar graphs showing normalized read counts for various SOCS family members in YFP+ nTreg from FYC+/+ and FYC+/I1I2 mice](page_256_670_1152_384.png)
+
+*Reviewer Figure 3. RNAseq from the SOCS family.*
+Normalised read counts of the SOCS family members in YFP+ nTreg from *FYC*+/+ and *FYC*+/I1I2 mice; n=6-7; key as shown. Read counts were normalised using size factors derived from the overall DESeq2 analysis of all genes; p values determined using t-test with FDR correction.
+
+We observed only a minor decease (1.3-fold change) in the expression of SOCS1 protein by flow cytometry in eTregs from *FYC I1I2* male mice (*Reviewer Figure 4a*). In our opinion, this is unlikely to account for the diminished signalling in Treg in the knockout mice in response to IL-2 and IL-7 but may contribute to the increased sensitivity to IFN\(_\gamma\) (https://pmc.ncbi.nlm.nih.gov/articles/PMC11700635/ ). To further investigate if the diminished SOCS1 levels were cell intrinsic we analysed LN from *FYC*+/I1I2 female mice. We observed no difference in SOCS1 levels between YFP+ control Tregs and YFP+ Tregs lacking *I1I2* (*Reviewer Figure 4b*). Thus, the diminished SOCS1 levels observed in *FYC I1I2* male mice are likely due to the inflammatory environment.
+Reviewer Figure 4. SOCS1 expression is not affected by the lack of Zfp36l1 and Zfp36l2.
+a) Representative histogram overlays of SOCS1 expression in nTreg (left panel) and eTreg (right panel) from LN from male FYC I1I2 and FYC mice; gMFI of SOCS1.
+b) Representative histogram overlays of SOCS1 expression in YFP+ nTreg (left panel) and eTreg (right panel) from LN from female FYC+ and FYC+ I1I2 mice; gMFI of SOCS1; key as shown.
+
+Further in-depth analysis of the SOCS family is limited by the lack of reagents for detecting these proteins with the required sensitivity and we suggest additional experiments to investigate this area are beyond the scope of this manuscript.
+
+3. The authors prevent the development of disease with inactivation of one Ifng allele, a prototypic Zfp36 target with functional AU-rich elements. However, this approach does not easily explain why inactivation of Zfp36l1 and Zfp36l2 with CD4-Cre does not cause a similar disease, if the excess of Ifng originates from conventional T cells. One possible explanation for the observed phenotype could be that ex-Treg cells with autoreactive TCRs develop, if Zfp36l1/l2 RBPs are inactivated with Foxp3-Cre. Those may be superior, tissue localized Ifng producers, since Ifng expression is induced in the exTregs by TCR signals, and is no longer repressed by Zfp36 proteins binding to AU-rich elements. Ex-Tregs would have downregulated YFP and would therefore not be captured in the approaches shown in the paper so far. This possibility should be addressed experimentally, or at least mentioned in the “Limitations of this study” section.
+
+One key difference between the CD4cre system and the Foxp3cre system we are using in our study is that in the CD4cre system Zfp36l1 and Zfp36l2 are absent in naïve and effector CD4 and CD8 T cells as well as Treg. In the manuscript by Cook et al. (https://pmc.ncbi.nlm.nih.gov/articles/PMC9832469/) their Figure 8 shows that in in Cd4-Cre+ Zfp36l1fl/fl Zfp36l2fl/fl mice there is defective priming of CD4 T cells. The mutant T cells are unable to generate an immune response and Cook et al. show Cd4-Cre+ Zfp36l1fl/fl Zfp36l2fl/fl mice are resistant to the induction of experimental autoimmune encephalitomyelitis (EAE).
+We have added new data (Figure 9) which addresses the reviewer’s question of whether Tregs that lack Zfp36l1 and Zfp36l2 can give rise to ex-Tregs.
+
+In Foxp3eGFP-creERT2 R26IISTOP-tdRFP I1/I2 mice we observed an increase in the number of ex-Tregs, indicating that these RBP are required for the maintenance of Treg stability (Figure 9a). The ratio of ex-Treg/Treg was higher in female Foxp3eGFP-creERT2 heterozygous R26IISTOP-tdRFP I1/I2 mice (Figure 9h), suggesting that the instability observed in I1/I2 deficient Tregs is cell intrinsic.
+
+Reviewer #2 (Remarks to the Author)
+The study „ZPF36-family RNA-binding proteins in regulatory T cells reinforce immune homeostasis” by Sáenz-Narciso and Bell et al. investigates the roles of Zfp36, Zfp36l1 and Zfp36l2 in mouse Treg cells. The authors use conditional Foxp3-specific knock-out mice for each factor (or double-KO for Zfp36l1 and Zfp36l2) to determine their individual roles. While deletion of Zfp36 or Zfp36l2 had no apparent effect, the absence of Zfp36l1 in Tregs caused a systemic hyper immune activation, which was even more pronounced when Zfp36l1 and Zfp36l2 were both deleted at the same time.
+Mechanistically, the authors use bulk RNA-Seq and scRNA-Seq of Zfp36l1/l2-deficient Tregs to reveal potential underlying mechanisms. While the core Treg signature appears stable in KO Treg cells, decreased expression of CD25 and CD127 seems to dampen IL-2/IL-7 sensitivity of KO Tregs, likely dampening their fitness. In contrast, KO Tregs respond more to IFNg, while also producing more IFNg themselves. Finally, the authors show that most of the hyper immune phenotype of Treg-specific Zfp36l1/l2-deficient mice is due to excessive IFNg signalling, since genetic deletion of one allele of Ifng is sufficient to restore homeostasis.
+
+In summary, the study is likely to be of considerable interest for the readers in the field. However, the authors missed to work out the precise molecular mechanisms and targets of how Zpf36 family members control Treg cell function. Also, it remains mechanistically unclear why Zfp36l1/l2-deficient Tregs fail to control especially type 1 IFNg-producing T cells. Therefore, the paper mostly provides a comprehensive characterization of the respective KO mice.
+
+We thank the reviewer for their positive comments and feedback on our manuscript which we have taken on board to improve the study and its presentation, we respond to each point below. Revisions/new text are highlighted in the revised manuscript.
+
+Specific points:
+
+Figure 1:
+
+The authors report that 9% of FYCI1 mice and 12% of FYCI1/2 develop symptoms. And as far as I understand, the authors focus their subsequent analysis entirely on these sick mice. What about the majority of mice that don’t develop symptoms? Are they completely healthy? Do they show any differences in effector CD4+/CD8+ cells? How do the Treg cells look like in these mice?
+It would be helpful to show a more comprehensive analysis of key clinical parameters (e.g. weight change over time, pathology scores) for all mice and (immune) phenotyping for symptomatic vs. non-symptomatic mice.
+Overall, the unanswered question remains: Why do some KO mice experience a loss of immune homeostasis, while others not?
+
+We have now clarified in the text the health status of the animals analysed. The majority of the data presented in the original version of the manuscript, except the histology shown in Figure 1e and the Kaplan Meier plot shown in Figure 8a, had been obtained from apparently healthy mice. For clarity, we have edited the text and moved the histology data from mice with clinical symptoms to Supplementary Fig. 1c. We now include data showing lymph node cellularity is increased in
+mice with clinical signs (Supplementary Fig. 1d). All of the apparently healthy male FYC I1/2 mice analysed also showed increased lymph node cellularity and increased proportions and numbers of CD4 and CD8 effector cells, GC B cells, and eTregs. Thus, all male mice have shown evidence for the loss of immune homeostasis but only some of them progressed to exceed the clinical severity limit by the time we analysed them.
+
+We can only speculate as to why 9-12% mice develop symptoms that exceed the severity limit, but consider potential effects from the local cage environment and the aggressive interactions between male mice that are cohoused, together with the overall specific-pathogen-free facility in which they are housed. The gut flora was altered Schaedler flora and since the opening of this barrier facility in 2009, no primary pathogens or additional agents listed in the FELASA recommendations (https://pubmed.ncbi.nlm.nih.gov/24496575/) have been confirmed during health monitoring surveys of the stock holding rooms. The transition from mild to moderate severity is governed by a number of assessment criteria as listed in the attached PDF (Limiting clinical signs in laboratory rodents), with marked piloerection and intermittent hunching being the most common limiting adverse effect resulting in humane killing of an animal. We have attached this as a Supplemental file to the resubmission.
+
+Figure 2: I am missing a clearly unbiased analysis of the bulk RNA-Seq data from WT vs. KO Tregs. Right now, the authors only use a custom pre-defined group of gene sets for their enrichment analysis, to basically pave the way for their findings regarding IL-2/IL-7 sensitivity. It feels like these findings were already known before the RNA-Seq. I think a more explorative approach using unbiased GSEA for example could reveal additional signaling pathways or metabolic alterations in KO Tregs, that might contribute to the diminished fitness and function of Zfp36l1/l2-deficient Tregs.
+
+We believe that the reviewer’s definition of a ‘clearly unbiased’ analysis would comprise, for example, a GSEA analysis using a curated group of gene sets such as KEGG, Hallmark, Reactome or Gene Ontology. During the exploration of our RNA-seq data, we performed numerous such analyses to aid us in identifying cellular functions and pathways that were altered in Zfp36l1/l2-deficient Tregs and warranted further investigation. We considered showing one of these in our manuscript, however, we decided to show a more focussed group of gene sets with direct relevance to T cells. We believe that this still represents a global analysis; whilst we may in part have been guided by the larger-scale analyses, we aimed to include gene sets for the major signalling and metabolic pathways, as well as numerous aspects of cell function, including DNA and RNA metabolism, the cell cycle and apoptosis, and gene sets more specific to T cell differentiation and function. The breadth of the gene sets chosen can be seen in Supplementary Table 4. Our decision to use this more focussed group of gene sets was based on a number of considerations:
+
+1. In our view, all gene set analyses are, by their nature, biased. For any curated group of gene sets, decisions will have been made on what those sets comprise, and the genes within them will be biased, for example, towards more widely studied cell types. We would, therefore, argue that the advantage of such curated gene sets in being unbiased is not as clear-cut as implied by the reviewer.
+
+2. No curated group of gene sets covers everything that might be of interest. Hallmark gene sets are the most limited; however, even for the larger groups of gene sets, this is an issue. For example, KEGG does not contain a general “Apoptosis” gene set, whilst Reactome does not include a general “cell adhesion” or “migration” gene set. None of these groups of gene sets include a “Treg signature”.
+
+3. The interpretation of large-scale gene set analyses is hampered by the inclusion of many irrelevant gene sets, as well as by the huge amount of redundancy inherent to gene ontology terms, and to a lesser extent to Reactome and KEGG pathways; together this means that the reader has to dig into the genes within each set to understand what they
+mean, and plough through a lot of repeated information to reach additional interesting pathways that are buried beneath. The redundancy can, in part, be mitigated by tools such as REViGO for gene ontology terms, but in our experience, substantial redundancy often still remains. We show, as an example (Reviewer Figure 5), the top 15 increased and decreased gene sets from an analysis of all KEGG pathways. The increased pathways include “platelet activation”, “cardiomyopathy”, “axon guidance” and “collecting acid duct secretion”, none of which are directly relevant to Tregs, “Human cytomegalovirus infection” and “Human papilloma virus infection” likely relate to altered immune cell function – but in our context are clearly not indicating that the cells are infected with these viruses. Regulation of the actin cytoskeleton is driven by increased expression of integrin genes, likely reflecting altered signalling and/or cell migration, both of which are covered more specifically by the pathways that we focus on. The decreased pathways show some redundancy, with two gene sets related to ribosome biogenesis, two to DNA repair, and several to amino acid metabolism. All of these examples require substantial additional effort to understand and interpret, and/or push other gene sets of interest out of the top few that are possible to show in a figure. Whilst such effort in interpretation was necessary for us in exploring and understanding our data, for a manuscript figure we preferred to show a more focussed analysis, enabling much clearer interpretation by the reader.
+
+![Bar chart showing top 15 increased and decreased pathways from all the KEGG pathways. Each pathway is represented by a bar, with color coding indicating NES (number of enriched genes), and whether the pathway is neither, ARE, CLIP target, or +ARE.](page_374_682_1002_324.png)
+
+Reviewer Figure 5. GSEA from KEGG pathways
+Top 15 increased and decreased pathways from all the KEGG pathways. Numbers in brackets indicate the total number of genes in the pathway; numbers in white indicate FDR-adjusted p values.
+
+4. Except for Hallmark gene sets, which have a relatively limited scope, none of the curated groups of gene sets are designed for GSEA, and thus, they include many gene sets that are either too large (in the case of gene ontology) or too small. We note that, in the KEGG example shown, many of the decreased pathways are close to the lower limit of 15 genes acceptable for GSEA, and that even above this limit, enrichment of such gene sets can be driven by very few genes and is more subject to artefacts. In some cases, our desire to include a particular pathway meant that a small gene set was unavoidable, with the smallest gene set we included comprising 23 genes (One-carbon metabolism). However, we aimed to minimise this and included only two gene sets with less than 30 genes, with the vast majority of our gene sets comprising between 50-250 genes.
+5. Performing the analysis in this way allowed us to add in gene sets that were of particular interest to us, such as the Treg signature, and an improved T cell signalling pathway gene set which we have curated based on our knowledge of T cell signalling. As the reviewer suspects, we were aware of the effects on cytokine signalling when we assembled our group of gene sets, and we included the gene sets derived from immunological genome project data for IL-2 and IL-7, and published data for IFNγ, in order to test the hypothesis that altered signalling impacts on downstream gene expression. This enabled us to use data derived from Tregs, and thus directly relevant to the signalling responses in our cells. The highly significant results we obtained using this data would not have been seen from any of the curated lists of gene sets, which either contain gene sets comprising the signalling machinery (KEGG and Reactome), or include very generic gene sets for altered or increased gene expression upon IL-2/STAT-5 or IFNγ signalling (Hallmark and GO biological process). We considered carefully whether we should include these gene sets together with the global GSEA, or in a separate analysis, alongside our other data on cytokine signalling. We chose to include them here because we wanted to be able to compare the GSEA statistics of these sets to the other gene sets included. This would not be possible if the analyses were performed separately, since, whilst the values may be broadly similar, the way in which the normalised enrichment scores and p-values are calculated means that they would not be directly comparable.
+
+We acknowledge that there will be additional altered pathways that do not appear in the analysis we show. However, we believe that the differences between the available curated sets, and the issues outlined above mean that no large-scale GSEA analysis is able to both comprehensively and concisely reveal all pathways with alterations. Our chosen group of gene sets provides good coverage of relevant pathways, whilst mitigating some of the problems inherent to such analyses.
+
+Additionally, it remains unclear which molecular targets are specifically bound be Zfp36l1/l2 in Treg cells. The authors use published CLIP data from activated CD4+ T cells, however CLIP data directly from Treg cells would certainly be more helpful to elucidate the molecular mechanism. Such data would also be a valuable resource for the community.
+
+We agree that iCLIP from Tregs would be perfect for identifying ZFP36L1/L2 targets. However, unfortunately the small numbers of Tregs that can be obtained ex vivo preclude the generation of good quality iCLIP data. Given this limitation, we believe that CLIP data from activated CD4+ T cells provides a good surrogate and will allow us to identify the majority of direct targets. Use of CLIP data from three different activation conditions also aimed to maximise our detection of target transcripts. Comparison of our target list generated from CD4+ T cells with a similar list generated using CD8+ cytotoxic T lymphocyte (CTL) iCLIP (https://doi.org/10.1038/s41467-022-29979-x) shows a large degree of overlap between these (Reviewer Figure 6a). Moreover, comparison of this dataset with the 4h-activated CD4+ CLIP data shows that there is a very high correlation between the sum of crosslinks over gene 3'UTRs between CD4+ and CD8+ T cells, with genes significant in only one of the datasets typically showing weaker detection (Reviewer Figure 6b). Whilst the correlation will, in part, be driven by the correlation between expression levels, this is itself a major determinant of whether a target is detected in a CLIP experiment. We would expect Treg CLIP data to show a similar, if not greater, correspondence with activated CD4+ T cell CLIP data.
+Reviewer Figure 6 CD4 and CD8 CLIP targets are correlated.
+a) Venn Diagram showing the overlap between CD4 and CD8 (CTL) CLIP targets
+b) Plot showing the correlation between the sum of crosslinks over gene 3'UTRs between CD4+ (4h-activated dataset) and CD8+ (CTL) T cells
+
+We do acknowledge that it is possible that some direct targets of ZFP36L1/L2 will be missed in the activated CD4+ T cell data, either due to Treg-specific gene expression, or differential targeting of a transcript in Treg compared to activated CD4+ T cells; this may also result in genes detected in the activated CD4+ T cells that are not targeted in Tregs. Therefore, in order to strengthen our analysis, we have identified transcripts that have an AU-rich element (defined as 2 x UAUU separated by up to 3 nucleotides) within their 3'UTR, providing a potential high-affinity binding site for the ZFP36 family. We have revised Fig. 2c,d,e, Supplementary Fig. 4, Supplementary Fig. 5a and Supplementary Tables 4 and 5, to incorporate this information. Whilst the presence of a potential AU-rich element does not necessarily mean that a transcript is a target, we believe that our approach is helpful for identifying genes that may be direct targets, as well as providing additional evidence for genes identified in the CLIP data.
+
+Figure 7: The analysis of cytokine expression by Tcon and Treg in Fig. 7I is really informative and reveals that there is a major dysregulation of type 1 IFN-g-producing effector T cells in FYC I1/2 mice, while type 2 and type 3 immunity seems rather unaffected. It also shows that KO Tregs do not have a problem in producing IL-10, ruling out this as a potential mechanism underlying the breakdown of immune homeostasis.
+There is the concept that type 1 T cells (eg Th1) are primarily regulated by Th1-like Treg cells, that mimic Th1 cells in terms of e.g. T-bet expression etc. The question that comes to my mind here is: Do KO Tregs cannot acquire a proper and functional Th1-like Treg phenotype and is this the reason for their reduced capacity to suppress type 1 immune cells?
+I am asking this bc the paper is overall lacking the explanation why Zfp36l1/l2-deficient Tregs cannot properly suppress (Can they actually suppress normally in conventional in vitro suppression assays?) and why specifically not type 1 immune cells. More data addressing this question would certainly increase the scope of the paper immensely.
+
+In addition to the dysregulation of type 1 IFNγ-producing effector T cells in FYC I1/I2 mice, we also observe greatly elevated levels of serum IgE, and increased proportions and numbers of both germinal centre B cells and T follicular helper cells in FYC I1/I2 mice, which shows that type 2 immunity is also affected. Furthermore, we also observe a selective increase in the number of “resident” cDC2 in the LN of FYC I1/I2 mice. As cDC2 can prime Th17 and Th2 responses (https://doi.org/10.1038/s41423-021-00741-5), this suggests the dysregulation is not limited to type 1 responses.
+To test if Tregs lacking Zfp36l1 and Zfp36l2 can suppress normally in a conventional *in vitro* suppression assay we performed this assay and observed no difference between control Tregs and Tregs lacking *Zfp36l1* and *Zfp36l2* in their ability to suppress the proliferation of CD4+ conventional T cells or CD8+ T cells *in vitro* (**Reviewer Figure 7**).
+
+![Flow cytometry plots showing suppression of CD4 T cell proliferation by Tregs from I112 and CD4cre I112 mice at different ratios](page_184_563_1082_312.png)
+
+**Reviewer Figure 7.** **Tregs lacking Zfp36l1 and Zfp36l2 suppress the proliferation of CD4 T cells in vitro.** Tregs from *Cd4cre I112* mice were sorted (CD4+CD25+). 50000 Naive CD4+ T cells were labelled with Cell Trace Violet and activated with anti-CD3 (1ug/ml) and anti-CD28 (0.5ug/ml) antibodies. CD4+ T cells were co-cultured with 50,000 Splenocytes from *Rag2*-deficient mice and different numbers of Tregs (50000;25000;12500). FACS analysis was performed at day three of culture.
+
+**Minor comment on Figure 7i and Sup. Fig. 8. The staining of RORgt+ Tregs doesn’t really show any pos. population to me. I suggest to stain for these Treg subsets (also the GATA3+ Treg) in intestinal Treg cells rather than Treg cells from Spleen/LN.**
+
+We have now revised the visualisation of the flow cytometry data in **Figure 7** and **Supplementary Fig.S18e** to display larger dots for each event and added the number of events for each data file. In addition, we include new data showing RORγt expression in CD4+ cells from mesenteric LN (**Supplementary Fig. S18e**). Together, we think this revised figure more clearly presents the findings.
+
+**Reviewer #3 (Remarks to the Author)**
+*In this study, Saenz-Narciso and colleagues investigate the role of ZFP36 family RBPs in Tregs. They specifically ablate Zfp36l1, Zfp36l2, and Zfp36l1l2 in Tregs and find that the loss of either protein or the combined loss of both leads to autoimmunity, as evidenced by the expansion of activated CD4 and CD8 T cells expressing proinflammatory cytokines. Mechanistically, these RBPs regulated multiple aspects of Treg biology, including endocytosis and sensitivity to cytokines, both of which had significant implications for Treg stability and function. The authors further demonstrated that IFNg drives the inflammatory phenotype in mice lacking Zfp36l1l2 specifically in Tregs, noting that deletion of one allele of IFNg is sufficient to ameliorate this phenotype. This is an excellent study with high-quality data; however, a few issues require clarification.*
+
+We thank the reviewer for their positive comments and feedback on our manuscript which we have taken on board to improve the study and its presentation, we respond to each point below.
+Revisions/new text are highlighted in the revised manuscript.
+
+1. The authors state that only 9% of FYC11 and 13% of FYC11/2 mice develop autoimmunity. Why is the disease penetrance not 100%?
+
+We would like to clarify that all the male mice analysed, even those that did not display overt clinical signs, showed increased lymph node cellularity, increased proportions and numbers of effector cells and an increased proportion of IFNγ-producing Treg. The penetrance of these phenotypes is 100%. We now include data showing lymph node cellularity is increased in mice with clinical signs (Supplementary Fig. 1d). The reason why only some mice go on to display clinical signs that exceed the severity limit is uncertain. It is likely that the variable tendency of male mice to fight when cohoused induces local infection and immune responses which may trigger clinical disease in some instances.
+
+It appears the authors focus solely on male mice. Is there a specific reason for not analysing homozygous knockout females?
+
+In order to generate homozygous female mice, we would need to cross hemizygous males (FYC 11/2 which have the potential to develop clinical symptoms) with heterozygous females. We tried breeding from hemizygous males crossed to heterozygous females but were unsuccessful in generating homozygous females (from two pairs and 30 pups which survived to weaning, only 1 homozygous female was identified from 10 females) so we did not continue for practical and welfare reasons.
+
+2. In female FYC11/2/+ mice, the Foxp3+YFP+ population was outcompeted by Foxp3+YFP-WT Tregs, suggesting that Tregs lacking I112 have reduced competitive fitness. Do knockout Tregs exhibit an activated phenotype in this context (with competition), or is there a skewed nTreg/eTreg ratio?
+
+We have now included data for the percentage of nTreg and eTreg in female mice; the ratio between FOXP3+ cre-negative and FOXP3+ cre-positive Treg is not different between FYC+/I112 and FYC+ mice (Supplementary Fig. 3a).
+
+3. At the RNA level, there is downregulation of Ctl4, which appears to correlate with low protein levels of total CTLA4, despite the claim of endocytosis. What is the underlying cause of the downregulation of CTLA4 protein?
+
+In Fig. 2b of the original manuscript the labelling of the MA plot showing the position of Ctl4 mRNA may have suggested that its amounts are decreased, but this is not the case. Our data show that Ctl4 is not decreased at the RNA level in FYC+/I112 mice. For clarity, we now show in Supplementary Fig. 8a of the manuscript no difference in the normalised read counts or splicing of the Ctl4 transcript in Treg from FYC+/I112 or FYC+ mice.
+
+In Fig. 3e of the original and revised manuscript we report a minor decrease (1.5-fold) in total CTLA-4 protein (geoMFI) detected by intracellular flow cytometry in the CD62Lhi nTreg subset in FYC+/I112 mice. We wish to emphasise that these cells were cultured *in vitro* in the presence of 2ng/ml IL-2.
+
+In the original manuscript Supplementary Fig. 4a showed intracellular CTLA-4 staining in samples that were fixed directly *ex vivo*. This showed no difference in the total CTLA-4 protein in YFP+ nTreg from FYC+/I112 mice compared to control FYC+ mice – the subset in which we observed the decreased cycling of CTLA-4. In the revised manuscript this data is now shown in Supplementary Fig. 6a. In the eTreg population (where there are no defects in cycling) we observed a very minor decrease (1.3-fold) in CTLA-4 geoMFI (*Supplementary Fig. 6a*). Many factors could be affecting the expression of CTLA-4, e.g. TCR signalling, FOXP3 and IL-2 upregulate CTLA-4 expression.
+Following internalisation, CTLA-4 can also be delivered to lysosomes where it is degraded.
+
+4. Is defective STAT5 signalling responsible for the lack of competitive fitness in FYC1I12 mice?
+We suggest that that the most clearcut way to test this hypothesis would be to introduce a constitutively active STAT5 transgene which is predicted to rescue the phenotype. Because the STAT5 transgene causes leukaemia this would have to be achieved using the conditional allele described here (https://pmc.ncbi.nlm.nih.gov/articles/PMC5071159/ ). We do not have these mice and, and time and resource consideration make this experiment unfeasible.
+
+Can the IL-2/Ab complex expand YFP+ Tregs in FYC1I12/+ mice?
+Based on the response of the cKO Treg from FYC+/ I1I2 mice to IL-2 in vitro, and the expansion of cKO Treg in male mice, it is reasonable expect the cKO Treg in FYC+/ I1I2 mice will respond if enough stimulus is provided. After considering this carefully, we have concluded that the substantial numbers of animals required to optimise dosing with the IL-2/Ab complex and subsequently to demonstrate a difference in the response to IL-2 in vivo is hard to justify against the limited mechanistic insight it will provide.
+
+5. FYC1I12 Tregs produce more IFNg and IL-10 when stimulated. Is this phenotype a result of inflammation in these mice? Would YFP+ Tregs in FYC1I12/+ mice display the same phenotype? A major concern is distinguishing the actual phenotype of Tregs caused by I1I2 deletion from that induced by inflammation. For example, bulk RNA sequencing has been performed on Tregs from FYC1I12/+ mice (with no inflammation), while scRNA sequencing was conducted on Tregs from FYC1I12 mice (where inflammation is high). Can these two datasets be compared (using scRNA-seq pseudobulk analysis) to differentiate between inflammation-induced and genuine I1I2-dependent gene expression?
+
+The analysis of female mice heterozygous for the FYC allele is used to identify cell intrinsic effects in Tregs caused by lack of the RBP in the absence of inflammation. We have measured IFNγ and IL-10 produced by Treg in these mice.
+
+Analysis of IFNγ produced by Tregs in FYC+/ I1I2 mice revealed a higher proportion of YFP+ Treg that were IFNγ+ compared to either YFP+ Treg in FYC+ mice (three-fold increase) or cre-negative cells in the same mouse (Supplementary Fig. 19a). In addition, we observed higher amounts (3.6-fold increase) of Ifng at the RNA level in the RNAseq data from YFP+ nTreg from FYC+ I1I2 mice compared to YFP+ nTreg from FYC+ mice (Supplementary Fig. 19b).
+
+This reflects a cell intrinsic effect meaning that the increase in proportion of IFNγ+ Treg observed in FYC male mice is not solely caused by inflammation. Nonetheless, the increase in proportion is smaller in the absence of inflammation compared with the increase observed in male mice (five-fold increase; Figure 7i), indicating that the phenotype is exacerbated by the inflammatory environment.
+
+We observed no difference in proportions of IL-10+ cells between YFP+ Treg from female FYC+ and FYC+/ I1I2 mice (Supplementary Fig. S19c). These data suggest that the increased proportions (two-fold increase) of IL-10+ FYC I1I2 Tregs are likely a result of the inflammation in male mice.
+
+As the reviewer states, the bulk RNA-seq is performed in the absence of inflammation, and thus the findings from this data show the cell-intrinsic effects of the deletion independently of inflammation. Pseudobulk analysis of the scRNA-seq data, which was performed on cells in inflammatory conditions, would, at best, serve only to indicate additional pathways that are downstream of inflammation. However, differences in the cell populations sorted for the two experiments, and in the sensitivity of pseudobulk compared with bulk RNA-seq analysis of cells from females will together add substantial noise to the data, making it very difficult to draw robust and meaningful conclusions from the analysis. This would not clarify the cell-intrinsic effects of Zfp36l1/I2 deletion.
+One of the major findings of the scRNA-seq was an increased proportion of cells displaying an IFNγ signature. Our analysis of the bulk RNA-seq data indicates that this phenotype is not driven solely by the inflammatory environment, since gene set enrichment analysis showed a trend (albeit not quite reaching significance) towards increased expression of genes that are increased upon IFNγ treatment (Fig. 2g). This is also supported by our data showing increased responsiveness to IFNγ in the absence of inflammation (Fig. 6). In addition to the increased expression of Ifng at the RNA and protein level in female heterozygous mice (see above), we also saw non-significant trends towards increased mRNA abundance of Gata3 and Cxcr3, and decreased Rorc mRNA (Reviewer Figure 8). This adds further evidence that the increased frequency of cells producing IFNγ is, in part, a cell intrinsic effect.
+
+![Bar graphs comparing normalized read counts for Ifng, Cxcr3, Gata3, and Rorc between FYC+ and FYC+ I1I2 mice.](page_184_579_1082_410.png)
+
+Reviewer Figure 8. Normalised read counts for *Ifng*, *Cxcr3*, *Gata3* and *Rorc* comparing nTreg from *FYC+* and *FYC+ I1I2* mice. Bars indicate the mean, and points indicate normalised counts for individual female mice analysed by bulk RNA-seq. Numbers above indicate FDR-adjusted p values; these and normalised counts were derived from DESeq2 analysis
+
+Our scRNA-seq data also allowed us to better define the skewing of the Tregs towards a more activated phenotype, characterising the transcriptome of the populations that are gained and lost by the *FYC I1I2* mice. In contrast to male *FYC I1I2* mice (Fig. 1c), the female heterozygous mice do not show an increase in the proportion of icre+ Tregs with an effector phenotype: we now show this data in Supplementary Fig. 3a. Thus, the expansion of effector relative to naive Tregs is driven by the inflammatory environment, and we would not expect to see the same level of skewing in the transcriptome of Tregs from female heterozygous mice. Our bulk RNA-seq was also performed only on naive Tregs, due to the difficulty in obtaining sufficient numbers of effector Tregs, limiting our capacity to draw conclusions about the effector population. Nonetheless, we are able to see some trends in our bulk RNA-seq that align with our findings in the scRNA-seq. Gene set enrichment analysis using the marker genes for each of the clusters identified in the scRNA-seq shows that, in the bulk RNA-seq, *FYC+ I1I2* Tregs are strongly enriched for genes that are highly expressed in cluster 5 (Reviewer Figure 9). This cluster represents naive Tregs (Supplementary Fig. 18b) but, in contrast to clusters 0 and 2, it is maintained in *FYC I1I2* mice. Whilst cluster 0 markers are also enriched in knockout Tregs in the bulk RNA-seq, these comprise only 5 genes, all of which are also markers for cluster 5. This suggests that the naive Tregs in female heterozygous mice already show some skewing towards this phenotype, although we are unable to draw conclusions as to the magnitude of this skewing. Of note, the cluster 5 markers include numerous genes that are known to be induced by IFNγ, such as Ly6a, Ly6c1, ligp1, Dap1l, lgtp and Samhd1, and over half of the *FYC I1I2* cells within this cluster were defined as having an
+IFNγ gene signature (Fig. 7e), suggesting that the skewing towards this phenotype may result, at least in part, from the increased responsiveness to IFNγ signalling.
+
+![Bar chart showing NES (normalised enrichment score) vs log10(FDR) for clusters 0-6](page_384_256_480_256.png)
+
+Reviewer Figure 9. GSEA analysis in bulk RNAseq from nTreg from FYC+ I1I2 females showed an enrichment in expression of cluster markers which defined cluster 5 in scRNAseq from FYC I1I2 male mice. GSEA was performed to assess enrichment in FYC+ I1I2 compared with FYC+ nTreg, using the marker genes defining each of the clusters identified in scRNAseq analysis of FYC I1I2 Treg. X axis represents –log10(FDR-adjusted p value), and colour represents normalised enrichment score from the GSEA analysis. Numbers to the right of each bar indicate the total number of marker genes for each cluster.
+
+In addition to the enrichment of cluster 5 marker genes, we also saw a significant enrichment of cluster 4 markers within the FYC+ I1I2 naive Tregs. This represents effector Tregs in the scRNA-seq and is the cluster containing the highest proportion of FYC I1I2 (relative to FYC) Tregs. This suggests that the naive Tregs in the heterozygous females already show some increased expression of genes characteristic of effector Tregs; however, our flow cytometry data show that this is insufficient to cause expansion of a bona fide effector Treg population.
+
+6. What happens to Tregs in non-lymphoid tissues? Are there any defects in migration?
+We have data from the liver and lamina propria from male mice showing the numbers of CD4+YFP+ Nrp1+ Treg are not different between FYC and FYC I1I2 mice (Reviewer Figure 10). We conclude that Treg are present in these non-lymphoid tissues. We have not measured migration directly but our data show that Tregs are present in spleen, lymph nodes, thymus and non-lymphoid tissue.
+
+![Flow cytometry plots showing expression of NRP1 versus CD62L on TCRβ+CD4+ YFP+ Treg in lamina propria and liver](page_384_1024_480_256.png)
+
+Reviewer Figure 10. FYC I1I2 Tregs are present in liver and lamina propria. Representative flow cytometry plots showing the expression of NRP1 versus CD62L on TCRβ+CD4+ YFP+ Treg in the lamina propria (upper panels) and liver (bottom panels) from FYC and FYC I1I2 mice; including proportions and numbers of CD62L+ NRP1+ Tregs in each tissue.
+Final response to reviewer’s comments
+
+Reviewer #1 (Remarks to the Author):
+
+The authors have adequately addressed all of my points and provided important new information.
+We thank the reviewer for their positive comments. No additional material was requested.
+
+Reviewer #2 (Remarks to the Author):
+
+The authors have addressed all my concerns. Thank you.
+
+Regarding Figure 1, I would suggest to include the histology data (now Sup. Fig. 1c) and the LN cellularity data (now Sup. Fig. 1d and 2a) into the main Figure, bc they are strong phenotypic data that don’t need to be hidden in the supplement.
+We thank the reviewer for their positive comments.
+
+These data are now included in the main Figure 1b,c.
+
+The potential reasons why only a minor fraction of mice develops clinically symptoms, should be briefly discussed in the manuscript.
+This is briefly discussed in the Discussion under “Limitations of the study”.
+
+Reviewer #3 (Remarks to the Author):
+
+The authors have addressed all of my queries. I have no further concerns. This is an excellent study!
+
+We thank the reviewer for their positive comments. No additional material was requested.

07ed9c6d9e6ca0cab5b9b0180c58f6b6054610c90c12780b9cc4a7e3a9a2f0e1/preprint/preprint.md

@@ -0,0 +1,294 @@
+Molecular Manipulation of Polyamide Nanostructures Reconciles the Permeance–Selectivity Threshold for Precise Ion Separation
+
+Gang Han
+hangang@nankai.edu.cn
+
+Nankai University    https://orcid.org/0000-0001-8943-569X
+Zhenxiang pan
+Nankai university    https://orcid.org/0000-0001-6818-553X
+Yalong Lei
+Nankai University
+Tiange Yan
+Nankai University
+Fuxin Zheng
+Nankai University
+Yu Liao
+Nankai University
+Jiang Zhan
+Nankai University
+Tong Zhang
+Nankai University
+LU SHAO
+Harbin Institute of Technology    https://orcid.org/0000-0002-4161-3861
+
+Article
+
+Keywords: Nanofiltration, Polyamide membrane, Interfacial polymerization, Ion separation, Permselectivity
+
+Posted Date: December 17th, 2024
+
+DOI: https://doi.org/10.21203/rs.3.rs-5431568/v1
+
+License: This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License
+Additional Declarations: There is NO Competing Interest.
+
+Version of Record: A version of this preprint was published at Nature Communications on August 4th, 2025. See the published version at https://doi.org/10.1038/s41467-025-62376-8.
+Molecular Manipulation of Polyamide Nanostructures Reconciles the Permeance–Selectivity Threshold for Precise Ion Separation
+
+Zhenxiang Pan a, Yalong Lei a, Tiange Yan a, Fuxin Zheng a, Yu Liao a, Jiang Zhan a, Tong Zhang a, Lu Shao b,*, Gang Han a, *
+
+a College of Environmental Science and Engineering, Tianjin Key Laboratory of Environmental Remediation and Pollution Control, Nankai University, 38 Tongyan Road, Tianjin, 300350, China
+b State Key Laboratory of Urban Water Resource and Environment, School of Chemistry and Chemical Engineering, Harbin Institute of Technology, Harbin, China
+
+* Corresponding author
+Tel: +86 022-23501117. Email: hangang@nankai.edu.cn
+Tel: +86 13100870576. Email: shaolu@hit.edu.cn
+
+Abstract
+Membrane nanofiltration (NF) has emerged as a prominent energy-efficient separation technology for widespread applications related to the water–energy nexus. However, state-of-the-art polyamide (PA) NF membranes are markedly constrained by a ubiquitous, pernicious tradeoff between water permeance and selectivity. Leveraging the prestigious structure-determining performance rationale, this work conceives a facile and robust molecular engineering approach
+that enables simultaneous improvements in water permeance and co-cation selectivity through synthetic molecular construction of a PA nanofilm with unique cationic triazolyl heterocyclic polyamide (CTHP) structures during scalable interfacial polymerization. Experimental data in conjunction with molecular simulations reveal that the CTHP structures instigate exquisite regulation of the PA subnanometer pore architecture and the specific binding affinity with water and ions, which not only affords precise ion sieving ability and advanced Donnan exclusion selectivity but also energetically facilitates the partitioning and transport of water molecules. The exemplified PA membrane exhibits unparalleled divalent cation rejections of over 99%, accompanied by a 9-fold increase in monovalent/divalent cation sieving selectivity, which is substantially greater than that of the pristine benchmark, a superior water permeation rate, and excellent chemical and operational stability, circumventing the permeance/selectivity threshold.
+
+We believe that the molecular engineering strategy implemented in this work holds broad prospects for the rational design and fabrication of semipermeable polymeric NF membranes for sustainable and precision separations.
+
+Keywords
+
+Nanofiltration; Polyamide membrane; Interfacial polymerization; Ion separation; Permselectivity
+
+Introduction
+
+Precision discrimination of target ions and molecules from complex aqueous mixtures of similar species remains a superior challenge in widespread applications such as water, clean energy, and
+resource reclamation \(^{1-3}\). Membrane nanofiltration (NF), featuring phase-free conversion separation, has evolved into a premier tool for sustainable water separation because of its high energy efficiency, low carbon footprint, compact design, and manufacturing scalability \(^{4,5}\). The rapid dissemination of NF technology relies on high-performance membranes that ideally have high values of both water permeance and selectivity to fully exploit the prominent process advantages, but such a combination is exceedingly difficult to achieve, particularly for polymeric membranes, as the material properties that affect solute transport would, in turn, affect water permeation \(^{6-8}\).
+
+Polyamide (PA) thin-film composite membranes are state-of-the-art NF membranes that are particularly attractive for water filtration in practical modules across all scales \(^{9-12}\). However, the deleterious tradeoff between water permeance and membrane selectivity consistently poses a stumbling block for further advancing their performance, where increasing water permeance is inevitably accompanied by a diminished ability to selectively reject solutes \(^{13,14}\). According to the prevailing membrane separation mechanisms, effective strategies for rationally regulating mass transfer across PA membranes hinge on well-defined pore sizes with a narrow size distribution, properly tuned interactions between PA and the permeants of interest, and a thin PA selective layer \(^{15-17}\). Innovative materials and fabrication methods that can precisely regulate PA chemistry and nanostructures have therefore become essential pursuits of academic research.
+A prevalent approach that has been widely adopted to increase the permselectivity of PA membranes toward charged species involves tuning the surface charges to strengthen electrostatic exclusion effects via in situ and/or post-synthetic modifications \(^{18, 19}\). However, most of the approaches reported thus far have focused primarily on promoting solute rejection and selectivity rather than overcoming the permeance/selectivity tradeoff threshold \(^{20, 21}\). Recently, superior size sieving ability and co-cation selectivity have been achieved by trailblazing studies that use interfacial modulators to narrow the pore size distribution of the PA layer \(^{22-25}\). Unfortunately, a significant decrease in water permeance is usually accompanied by a concomitant increase in water transport resistance \(^{23, 24}\). Many studies have focused on exploring advanced membrane materials ranging from biological ion channels and aquaporins to emerging microporous materials of zeolites, metal-organic frameworks, covalent organic frameworks, macrocycles, and porous organic cages, some of which achieve superior permselectivity with unusual combinations of high permeance and selectivity \(^{26-28}\). Although the practicality of these intriguing materials is markedly restricted by many daunting limitations that vary from inherent low structural stability to inferior material availability and the feasibility of membrane fabrication on a large scale, their unique structural features underscore the importance of well-defined pore sizes and exquisitely regulated mutual interactions in achieving exceptional molecular sieving capabilities and water transport rates \(^{29-31}\).
+
+Hereupon, we believe that multifunctional monomers with synthetically engineered chemistry that enable the formation of a novel PA structure that not only imparts small pores with a narrow size distribution but also provides low resistance for water transport are likely to achieve disruptive
+improvements in both water permeance and solute selectivity, successfully overcoming the formidable tradeoff in PA membranes. Unfortunately, there is currently a lack of rational material design and feasible membrane fabrication strategies to accomplish this arduous undertaking.
+
+Herein, we demonstrate a facile and robust molecular engineering approach for precise regulation of the mass transfer behavior of PA membranes to circumvent the deleterious permeance/selectivity tradeoff in ion differentiation. Our strategy is contingent on molecular-level control over the nanoporous structure of the PA nanofilm and its interactions with water and ions via in situ construction of cationic triazolyl heterocyclic polyamide (CTHP) structures during interfacial polymerization (IP) via synthetic triamino quaternary triazole ammonium (DAT-NH2) isomers. Experimental data and molecular simulations revealed that the CTHP structures endow the PA layer with well-defined subnanometer pores with a narrow size distribution and abundant positive charge but low intrinsic water transport resistance, which synergistically enhances steric hindrance sieving and Donnan exclusion and facilitates the permeation of water. The advantages of this molecularly engineered PA structure were demonstrated by its superior performance within precise ion separation (Fig. 1a), achieving a 9-fold increase in monovalent/divalent cation selectivity with a tripled water flux relative to the benchmark, reconciling the tradeoff threshold.
+
+In light of the diverse array of monomer chemistries, the implemented molecular design strategy provides a gateway to advance the rational design and fabrication of PA membranes with superior permselectivity for precise ion separations toward clean water and renewable energy.
+Results and Discussion
+
+Synthesis of the DAT-NH₂ monomer and fabrication of PA membranes
+
+Fig. 1b shows the conceptual diagram of the ideal PA structure we intend to construct to circumvent the permeance/selectivity tradeoff threshold in precision nanofiltration (NF). Specifically, the PA layer is synthetically designed with well-defined permeate–PA binding affinity and profound steric sieving selectivity imparted by pore size confinement to efficiently manipulate the enthalpy and entropy barriers for water and solute transport \(^{32, 33}\). To realize this design strategy, we molecularly constructed a multifunctional structure with primary amine dangling and a highly polarized triazolyl heterocyclic core bearing quaternary ammonium (i.e., DAT-NH₂, Fig. 1c). Our pursuit of this molecular structure was inspired by these trailblazing studies showing that the triazole derivative heterocyclic moieties may provide preferential water transport paths with a low energy barrier \(^{34}\), whereas the amine groups concurrently provide highly reactive sites to crosslink with trimesoyl chloride (TMC) to form a PA nanofilm with superior hydrophilicity and interconnected positively charged subnanometer pores.
+Fig. 1 | Synthetic engineering of the PA molecular structure with rationally designed monomers. a, Working principle of precision co-ion separation via nanofiltration. b, Schematic diagram of the interconnected subnanometer-sized pores in a desired PA nanofilm with high selectivity and low water transport resistance, and three-dimensional view of an amorphous cell of the PA (cell size: \(65 \times 65 \times 65\) Å\(^3\)). c, Synthetic reaction formula of DAT-NH\(_2\) isomers and the visualized conformation of their atomic electrostatic potential. d, \(^1\)H NMR spectra and liquid
+chromatography (upper left inset) of DAT-NH2 isomers. e, Schematic illustration of the interfacial polymerization between DAT-NH2/PEI and TMC at the water–hexane interface to form a PA nanofilm. f, Molecular structure of the DP-M PA nanofilm (left) and its corresponding chain structure derived from an amorphous cell generated by molecular dynamic (MD) simulations. g, FT-IR spectra of the DP-M and P-M PA nanofilms. h, N 1s XPS spectra of the DP-M PA nanofilm. The N1s core level spectrum was deconvoluted into three components located at 399.7, 400.4, and 401.7 eV corresponding to N–(C=O)–, N(H)–C–, and N+–C–, respectively.
+
+The molecularly designed DAT-NH2 was synthesized via a one-pot quaternization reaction between 3,5-diamino-1,2,4-triazole and 2-bromoethylamine in DMF and then purified by nonsolvent precipitation (Fig. 1c and Supplementary Fig. 1). Intriguingly, liquid chromatography–mass spectrometry (LC–MS) data reveals that isomers of 3,5-diamino-4/1-(2-aminoethyl)-1,2,4-triazole were formed during the reaction, where two distinct peaks with almost the same intensity and area ratio are observed in the LC chromatogram pattern (Fig. 1d), and two peaks at m/z = 143 show up in MS (Supplementary Fig. 2), which is in good agreement with the chemical structures of the DAT-NH2 isomers (C4N6H11+, Mw = 143). Proton nuclear magnetic resonance (1H NMR) spectra corroborate these results, where four 1H NMR peaks corresponding to the two types of protons in each isomer are spotted at 3.11 (labeled H I), 3.30 (labeled H I), 3.35 (labeled H II), and 3.99 (labeled H II) ppm. The area ratios of the two peaks (H I/H II) in the 1H NMR spectrum were measured to be 0.88 and 1.04 for the isomers, which is in close proximity to the theoretically
+expected values based on the DAT-NH2 chemical structure. The visualized atomic electrostatic potential image of DAT-NH2 intuitively proclaims the positive charge characteristics of the triazole ring, and quantitative analysis of the molecular van der Waals surface electrostatic potential (ESP) shows that the distribution area and intensity of the positive ESP of the DAT-NH2 isomers are greater than the negative values (Fig. 1c and Supplementary Fig. 3).
+
+A self-sustaining PA thin film with good stability immediately formed when DAT-NH2 was brought in contact with TMC at the water/hexane interface (Supplementary Fig. 4), indicating a high polymerization rate between DAT-NH2 and TMC. DFT calculations further confirmed the high nucleophilic substitution reactivity of the amino groups on DAT-NH2 toward acyl chloride (Supplementary Figs. 5 and 6). Comprehensive ESP and average local ionization energy (ALIE) analyses reveal that the backbone and chemical functional moieties of the PA structures formed by the two isomers with TMC are almost identical (Supplementary Fig. 7). Therefore, the DAT-NH2 isomers were directly used for PA membrane preparation without further purification. A continuous PA selective layer can also be synthesized via a similar scalable interfacial polymerization (IP) procedure on top of a polyethersulfone (PES) substrate to prepare robust PA membranes for NF tests (Supplementary Fig. 8). Unfortunately, we observed that the obtained DAT-NH2/TMC PA membrane experienced severe water swelling and thus relatively low permselectivity were obtained (Supplementary Fig. 9), likely owing to the superior hydrophilic nature of the cationic triazolyl heterocyclic structures. We thereby further modified the PA chemical structure by using
+PEI as a comonomer during IP to increase the membrane stability and separation performance (Fig. 1e,f). PEI is a benchmark monomer that is widely used for the synthesis of positively charged PA NF membranes (Supplementary Fig. 10). Synthesis condition optimization experiments confirmed that the DP-M membrane fabricated with a 0.06 wt% DAT-NH2 shows an optimum combination of water permeance and ion selectivity and excellent robustness (Supplementary Figs. 11 and 12), substantially exceeding that of the PA membrane formed solely by PEI (denoted as the P-M benchmark). The FT-IR peaks observed at 1621 cm^{-1}, 1806 cm^{-1}, and 1705 cm^{-1} are associated with the amide I band, which arises from the stretching vibration of C=O and the coupling with the bending of N-H in subtly different chemical environments (Fig. 1g and Supplementary Fig. 13), validating the formation of polyamide structure during IP. The quaternary ammonium peak at 401.8 eV in the XPS spectrum of DP-M (Fig. 1h and Supplementary Figs. 14 and 15) confirms the presence of DAT-NH2 moieties in the PA layer. It is noteworthy that a significant decline in the O/N ratio was observed by DP-M compared with that of the P-M benchmark, where the O/N ratio decreases from ~1.50 to 0.96 (Supplementary Table 1), suggesting a substantial increase in the PA crosslinking degree of DP-M. The elevated crosslinking degree constricts the space between the stacked polymer chains, thus diminishing the pore sizes and augmenting the mechanical strength of the PA film (Supplementary Fig. 12). According to the chemical characterization and molecular simulations, the DAT-NH2 modulated PA nanofilm of DP-M has a semirigid 3D polyamide network with a large amount of intrinsic positive charges and smaller chain space compared to the P-M benchmark (Fig. 1f and Supplementary Figs. 16 and 17).
+Morphological and structural analysis of the DP-M PA membrane
+
+Field emission scanning electron microscopy (FESEM) images corroborate the uniformity and integrity of the formed PA thin layer at the macroscopic scale (Fig. 2a,d and Supplementary Fig. 18). At a finer scale, the PA layer of DP-M appears a smooth and compact surface, whereas the counterpart of the P-M benchmark shows a crumpled surface with numerous ridged wrinkles unequivocally seen on top. This surface morphological discrepancy was further manifested by the atomic force microscopy (AFM) data, where the surface roughness of DP-M (Rq = 3.44 nm) is distinctly lower than that of P-M (Rq = 9.09 nm) (Fig. 2b,e and Supplementary Fig. 19). A smooth surface is conducive to alleviating the fouling tendency. Cross-sectional transmission electron microscopy (TEM) images showcase that DP-M has an extremely low PA thickness of 59 ± 2 nm (Fig. 2c,f and Supplementary Fig. 20), which is much thinner than that of the P-M benchmark (i.e., 88 ± 2 nm). The significantly reduced PA thickness might be attributed to the rapid formation of a relatively dense nascent PA film mediated by DAT-NH₂ at an initial stage of IP (Supplementary Fig. 21), which stymies the diffusion of aqueous monomers at the interface and thus suppresses the subsequent growth of the PA layer \(^{35,36}\). On the other hand, the positively charged structure of DAT-NH₂ may slow down the diffusion of PEI towards the interface via H-bonding interactions \(^{37-41}\). In the context of membrane filtration, a thinner PA selective layer spontaneously confers shorter transport pathways and lower water penetration resistance, which is favorable for achieving high water permeance.
+Fig. 2 | Morphological and structural features of PA membranes. a,d, Surface FESEM images. b,e, 2D and 3D AFM images. c,f, Cross-sectional TEM images. Top: P-M. Bottom: DP-M. g, Pore diameter distribution of PA membranes obtained by PEG rejection tests. h, Molecular dynamics (MD) simulations of the fractional free volume (FFV) of PA layers (left). The dark blue and gray colors represent the voids between the polymer chains and the space occupied by the polymer skeleton, respectively. Representative porous molecular structures of the DP-M and P-M PA
+networks (right). i, MD simulations of the pore diameter distribution of the PA nanofilms. j, Zeta potential as a function of pH. k, Summary of the MWCO, water contact angle (CA), root mean square roughness (Rq), FFV, polyamide layer thickness, and zeta potential (ZP) at pH = 6 for DP-M and P-M.
+
+The rejection tests with neutral solutes indicate that DP-M has a molecular weight cutoff (MWCO) of 245 Da, almost two times smaller than that of the P-M benchmark (MWCO = 479 Da, Supplementary Fig. 22). Correspondingly, a small effective mean pore diameter of 2.8 Å accompanied by a narrow size distribution is achieved by DP-M (Fig. 2g), whereas P-M shows a relatively larger mean pore size of 3.1 Å and broader pore size distribution, in accordance with our design strategy and the XPS results (Fig. 1b,h). MD simulations were performed to construct realistic structural models via simulated polymeric algorithms to glean molecular-level insights into the porous structure of the PA layer. As shown in Fig. 2h, the fractional free volumes (FFVs) of DP-M and P-M PA layers are approximately 10.3% and 21.6%, respectively. Moreover, the pore diameter analyses conducted via MD molecular simulations substantiate that most of the pores inside the DP-M PA layer are approximately 2.75 Å in length, which is significantly smaller than that of the P-M benchmark (i.e., 3.38 Å). Further analyses of the interior cavity diameters disclose a narrower range of pore sizes within DP-M (Fig. 2i and Supplementary Fig. 23), signifying the compact structure of the PA mediated by DAT-NH2 (Fig. 1c). Notably, the microscopic pore features derived from molecular simulations coincide well with those experimentally obtained
+from neutral solute rejection tests (Fig. 2g,i). The tight nanostructure of DP-M constricts membrane pores to dimensions more favorable for size sieving, with precise ionic and molecular sieving capabilities and a threshold of 2.8 Å. Furthermore, aligning with the chemical features of the CTHP structures in DP-M (Fig. 1g), the cationic DAT-NH₂ moieties adequately elevate the membrane hydrophilicity and positive charge density, as manifested by its smaller surface water contact angle (CA) and higher zeta potential (ZP) than that of the P-M benchmark (Supplementary Fig. 23 and Fig. 2j). In the realm of NF applications, the enhanced hydrophilicity facilitates surface partitioning and interior diffusion of water molecules, whereas the ameliorated positive charge density reinforces the electrostatic repulsion selectivity. Collectively, the advanced membrane characteristics gained by DP-M resonate with our intended design strategy illustrated in Fig. 1b, which underpins the significance of synthetic molecular engineering in precisely regulating the nanoporous structure and chemical features of the PA selective layer (Fig. 2k).
+
+Simultaneous improvement in water permeance and ion selectivity
+
+The well-defined subnanometer pores with a sharped size distribution and the inherent positive charges of the DP-M membrane would afford prominent molecular sieving and electrostatic repulsion selectivity in sustainable NF applications. We subsequently examined the mass transport behavior of a wide spectrum of inorganic salts through DP-M using a crossflow filtration system. In contrast to the acquiescent expectation that a decrease in the membrane pore size along with a downscaled FFV generally accompanied by a concomitant reduction in the water permeation rate,
+a substantial increase in the water permeance was achieved by DP-M, where the water permeation flux of DP-M is almost three times greater than that of the P-M benchmark at the same pressure (Fig. 3a), corresponding to an approximately 3-fold increase in the pure water permeance (PWP). The incongruence between the enhanced water permeance and the reduced pore sizes and FFVs likely stems from the molecularly constructed CTHP structures and the low thickness of the PA layer, which provide facilitated water transport pathways with low resistance.
+
+At the same time, DP-M shows a sharp size-exclusion cutoff of ~2.5 Å in the Stokes radius of cations (Fig. 3b), adhering to its densified catatonic PA molecular structure, which thereby facilitates the transport of smaller monovalent cations (i.e., Rb^+, K^+, Na^+, and Li^+) while sufficiently blocking larger divalent cations (i.e., Ni^{2+}, Ca^{2+}, Mg^{2+}, and Zn^{2+}), with rejection rates greater than 98.0% and high water flux of 50 L m^{-2} h^{-1} (LMH). Notably, judiciously high rejections of up to 99.1% towards MgCl_2 and MgSO_4 were specifically achieved by DP-M (Fig. 3c), exceeding most of the state-of-the-art NF membranes, and similar rejections and water permeance were maintained over a wide pressure range of 2–16 bar (Fig. 3d). In contrast, LiCl rejection monolithically increases from 45.7% to 81.2% when the pressure is raised from 2 to 16 bar (Supplementary Fig. 24). As a result, an ideal Li^+/Mg^{2+} selectivity of 35.8–42.9 was obtained on the basis of single salt rejections (Supplementary Fig. 25), demonstrating its promising capability for precise cation screening. In the same vein, the P-M benchmark is inferior in terms of both salt rejection and cation differentiation selectivity (Fig. 3b and Supplementary Fig. 26).
+Fig. 3 | Ultrafast and precision co-ion separation via a low-pressure NF. a, Pure water flux (PWF) of DP-M and P-M at different operation pressures. b, Water flux and ion rejections of DP-M for filtrating different cation solutions. (Feed salt concentration: 1000 ppm; test pressure: 6.0 bar). c, MgSO4 and MgCl2 rejections of P-M and DP-M (feed salt concentration: 1000 ppm, test pressure: 6.0 bar). d, Effect of operation pressure on the water permeance and MgCl2 rejection of DP-M (feed: 1000 ppm MgCl2). e, Effect of pH on the water flux and LiCl rejection of DP-M (feed: 1000 ppm LiCl, test pressure: 6.0 bar). f, Effect of operation time on the water flux and MgCl2 rejection of DP-M (feed: 1000 ppm MgCl2, test pressure: 6.0 bar). g, Effect of operation
+pressure on the water permeance and \( S_{\text{Li}^+/Mg^{2+}} \) of DP-M (feed: 2000 ppm binary mixture of MgCl$_2$ and LiCl with a MgCl$_2$/LiCl mass ratio of 20). **h**, Effects of the Mg$^{2+}$/Li$^+$ ratio on the water flux and \( S_{\text{Li}^+/Mg^{2+}} \) ratio of DP-M (feed: 2000 ppm binary mixture of MgCl$_2$ and LiCl with various MgCl$_2$/LiCl mass ratios; test pressure: 6.0 bar). **i**, Performance comparison of DP-M with other reported state-of-the-art PA NF membranes operated under cross-flow nanofiltration. The corresponding references for the data points in (i) are specified in Supplementary Table 2.
+
+The separation performances of conventional NF membranes are generally susceptible to the feed salt concentration and pH due to the electrostatic screening effects. Interestingly, DP-M consistently retains its water flux, salt rejection, and co-cation selectivity across a wide range of feed salt contents and pH values. As demonstrated by the cycling performance tests, DP-M maintains stable MgCl$_2$ rejections of over 95.8% and relatively low LiCl retentions of less than 59.8% when the feed salt content spanning from 1000 to 7000 mg/L, and the rejections recover to the initial values of the test (Supplementary Fig. 27), substantiating its strong electrostatic shielding resistance toward high ionic strength. There were also no obvious deteriorations in salt rejections and water flux as the feed pH escalated from 1 to 13 (Fig. 3e), underscoring the ability of DP-M to maintain excellent separation performance in both acidic and alkaline environments. The superior pH and salinity stabilities of DP-M are consistent with its highly ionizable cationic PA structure and outstanding size-sieving ability imparted by the well-defined pore sizes. Moreover, DP-M shows excellent structural durability and stability throughout long-term filtration
+for 120 h (Fig. 3f), where MgCl$_2$ rejection consistently surpasses 99%, with a stable water flux of ~51 LMH being retained.
+
+The co-cation sieving ability of DP-M was further illuminated by binary salt filtration tests using a mixture of LiCl and MgCl$_2$ as the probe feed. Similar to the single-salt tests (Supplementary Fig. 25), DP-M shows high Li$^+$/Mg$^{2+}$ co-cation selectivity of greater than 39.0 (\( S_{Li+/Mg2+} \)) for binary mixtures at different operation pressures (Fig. 3g), which signifies a 9-fold greater magnitude than that achieved by the P-M benchmark (\( S_{Li+/Mg2+} = 4.3 \)), accompanied by an approximately 2.5-fold increase in water permeance. The persistently high rejections toward divalent cations and co-cation selectivity are presumably ascribed to the advanced molecular sieving and electrostatic repulsion effects afforded by the exquisitely regulated subnanometer pores and inherent positive charges of DP-M. Furthermore, slight fluctuations in the co-cation selectivity are observed when the feed Mg$^{2+}$/Li$^+$ mass ratio alters from 1–120, where the \( S_{Li+/Mg2+} \) oscillates between 33.5 and 44.3 and the water flux is consistently higher than 51.4 LMH (Fig. 3h). Compared with other reported PA NF membranes with similar chemical and structural properties, DM-P exhibits upper-level water permeance and co-cation selectivity (i.e., Li$^+$/Mg$^{2+}$) (Fig. 3i). The successful breakthrough of the permeance/selectivity tradeoff underpins our membrane design strategy illustrated in Fig. 1b and exemplifies the great feasibility of synthetic molecular engineering in rational membrane design. The excelled water permeation rate and co-ion screening capability in tandem with the scalable fabrication bolster the remarkable potential of the implemented strategy for developing effective
+NF membranes for widespread applications pertaining to wastewater treatment, lithium extraction, recycling, and removal of heavy metal ions, in line with sustainable development.
+
+Regulatory mechanisms of facilitated water permeation and superior co-cation selectivity
+
+Given an average pore diameter of ~2.75 Å with narrow size distribution and the strong intrinsic positive charge of DP-M (Fig. 2k), we speculate that its superior rejection of divalent cations and the co-cation sieving ability lean upon the on-demand tuning of PA chemistry and nanoporous structure according to the size and valence differences between cations, which instigates unusual differences in energy barriers that cations need to overcome for dissolution and diffusion. To gain fundamental insights into the underlying mechanisms responsible for the intriguing mass transport behavior of DP-M, dynamic molecular simulations were performed to correlate the separation performance with the membrane chemical structure. The simulations initiated with DFT calculations to illuminate the mutual interactions between PA and cations by performing configuration optimization and cation–PA binding energy calculations (Supplementary Table 3).
+
+As displayed in Fig. 4a, the negative binding energies of hexahydrated Mg^{2+} with the DP-M PA fragments (i.e., −12.03 and −18.93 kcal/mol) are consistently lower than those with the P-M fragments (−22.17 and −26.01 kcal/mol), suggesting that the binding interactions between hexahydrated Mg^{2+} and P-M are relatively more stable. In other words, hexahydrated Mg^{2+} has a greater energy barrier towards DP-M than P-M, which signifies that it is more difficult for Mg^{2+} to pass through DP-M. On the contrary, the binding energy between hydrated Li^{+} and the DP-M
+fragments (−28.83 kcal/mol) is close to that between it and the P-M fragments (−29.65 kcal/mol) (Fig. 4b), suggesting that the transport of hydrated Li^+ through DP-M and P-M nearly remains energetically unchanged even though the pore sizes of DP-M are substantially diminished and narrowed. However, the binding energy gaps between Li^+ and Mg^{2+} in the DP-M fragments are 16.80 and 9.90 kcal/mol, respectively, which are markedly larger than those in the P-M fragments (i.e., 2.82 and 6.66 kcal/mol) (Fig. 4b), implying that DP-M has an overwhelming advantage over P-M to differentiate Li^+ and Mg^{2+} from the perspective of energy barrier.
+
+The interaction region indicator (IRI) was subsequently applied to perform an in-depth analysis of the specific type of interactions between the PA fragments and hexahydrated Mg^{2+}. As a slight modification of the reduced density gradient (RDG), IRI (IRI(l) = |\nabla\rho(l)|/[\rho(l)^a]) can effectively manifest the chemical bonding and weak interaction regions ^{42}. IRI visual representations of these two binding configurations and the PA molecular fragments of DP-M and P-M were constructed (Supplementary Fig. 28). The corresponding scatter plots reveal that the interaction forces involved are intricate and hard to distinguish (Supplementary Fig. 29). Therefore, the hydration layer of Mg^{2+} and the influence of TMC were shielded to better disclose the contributions of DAT-NH_2 moieties in the PA (Fig. 4c). Examining their respective IRI scatter plots, eminent peaks appear near sign(l_2)r values of −0.05 and 0.06 a.u. in the DP-M PA molecular fragments. From the electron density point of view, the peak at −0.05 a.u. corresponds to a weak interaction of higher strength, whereas the peak at 0.06 a.u. stems from a stronger spatial repulsion. Notably, the scatter
+plot of the weak interaction portion ascertains an anomalous peak at approximately 0.013 a.u. in DP-M (Supplementary Fig. 30), indicating that the CTHP structures derived from DAT-NH₂ may generate proprietary steric hindrance at the molecular level. The two distinct peaks in the scatter plot obtained from interaction decomposition confirm that CTHP instigates both attractive and repulsive interactions (Fig. 4d). Further analysis of the isosurface of the IRI visualization reveals that the peak near −0.05 a.u. is attributed to intramolecular H-bonding interactions. These H-bonds account for the in situ formation of the cyclic conformations in the CTHP structures (Fig. 4d), which dictate additional steric hindrance at approximately 0.013 a.u. (the peak near 0.013 is retained in Supplementary Fig. 31). Moreover, the peak at approximately 0.06 a.u. is associated with the strong repulsion induced by the overlap of the triazole rings in CTHP driven by van der Waals surfaces. Other than the narrowed subnanometer pore sizes, these anomalous intramolecular H-bonding structures provide additional steric hindrance at the molecular level, further amplifying the energy barrier of permeation acting on divalent cations.
+Fig. 4 | DFT atomistic simulations of the mutual interactions between PA and cations within NF. a, Binding energies of hydrated Mg^{2+} to the PA molecular fragments of DP-M and P-M (for each membrane, two binding energies were calculated by positioning the hydrated Mg^{2+} at two representative binding sites of the PA fragments). b, The binding energies between the hydrated Li^{+}/Mg^{2+} and the PA molecular fragments of P-M/DP-M. The numbers represent the calculated
+energy gaps between Li^+ and Mg^{2+} in P-M and DP-M. **c**, Interaction region indicator analysis of PA fragments interacting with Mg^{2+}. The effects of the hydration layer and TMC were shielded (the unit a.u. here represents energy, and 1 a.u. is approximately 27.21 eV). **d**, Interaction region indicator analysis of the PA fragments derived from DAT-NH_2 and the visualized structure diagram. **e**, The independent gradient model based on Hirshfeld partition (IGMH) was used to analyze the binding configurations between P-M/DP-M molecular fragments and hydrated Mg^{2+} and Li^+. **f**, Electrostatic potential analysis of the PA molecular fragments. **g**, van der Waals surface area ratio corresponding to various electrostatic potential intervals of the PA molecular fragments.
+
+The interactions of P-M and DP-M fragments with hydrated Li^+ and Mg^{2+} ions were also visualized using the independent gradient model based on Hirshfeld partitioning (IGMH) (Fig. 4e and Supplementary Fig. 32). Compared with the P-M fragments, IGMH analysis shows that weaker H-bonding and van der Waals interactions exist between the DP-M fragments and the water molecules in the hydration shell of hydrated Mg^{2+} ions, which are energetically unfavorable for stabilizing the configuration, thereby hindering the thermodynamic partitioning and kinetic diffusion of hydrated Mg^{2+} ions, resulting in high rejections. In addition, the lower intrinsic charge number of Li^+ relative to Mg^{2+} impairs its ability to effectively bind water molecules in the hydration shell (Supplementary Fig. 33), while the positively charged CTHP structures in DP-M further promote the escape of polar water molecules from the Li^+ hydration shell by forming extensive van der Waals interactions to counter the electrostatic confinement, as demonstrated in
+the DP-M/hydrated Li^+ configuration (Fig. 4e), aggrandizing the dehydration of the hydrated Li^+ ions during the permeation. According to DFT calculations, the system diameters of Mg^{2+} dodecahydrate and Li^+ hexahydrate are 11.49 and 8.42 Å (Supplementary Fig. 34a), and their hydration energies are approximately −1922.45 and −563.52 kJ/mol (Supplementary Fig. 34b), respectively, implying that hydrated Mg^{2+} faces greater challenges than hydrated Li^+ in overcoming the dehydration energy threshold at the same transmembrane pressure. This energetically promoted dehydration of Li^+ expedites its transport through DP-M, playing an important role in enhancing Li^+/Mg^{2+} selectivity particularly when the membrane pore sizes are diminished and the size distribution is constricted.
+
+In addition to the non-Coulombic interactions, long-range Coulombic electrostatic forces (Donnan exclusion) also play an imperative role in the cation–PA interactions, particularly considering the strongly positively charged structure of DP-M. Qualitative and quantitative analyses of the electrostatic potential (ESP) are conducted, and Fig. 4f and Supplementary Fig. 35 illustrate the distribution of van der Waals surface ESP of the PA molecular fragments of DP-M and P-M. It was found that DP-M exhibits a superior positive potential and this observation is reaffirmed by the quantitative calculation of the ESP region proportions, where DP-M shows a large positive potential region proportion of 97% and an average ESP value of 55.55 kcal/mol, far exceeding the respective value of P-M (Fig. 4g). The pronounced electrostatic repulsion between DP-M and the positively charged hexahydrated cations is unequivocally conferred by the CTHP structures of the
+PA layer. The ESP of hydrated Mg^{2+} is nearly twice as high as that of hydrated Li^{+} (193.29 vs. 98.53 kcal/mol, Supplementary Fig. 33), which inevitably invokes formidable Donnan exclusion selectivity towards the positively charged DP-M (55.55 kcal/mol). Overall, the intriguing co-cation screening ability of DP-M proceeds through a cooperative mechanism of steric hindrance and electrostatic repulsion, as corroborated by the comprehensive IRI, IGMH, and ESP analyses.
+
+Fig. 5 | Mechanistic insights into facilitated water permeation in DP-M. a, Schematic diagram of the advanced structure of an ultrapermeable PA membrane for precise differentiation of monovalent/divalent cations. b, DFT atomistic calculation of polarity differences between DP-M and P-M molecular fragments. c, ESP distributions of van der Waals surfaces of the water molecule
+obtained by DFT calculations. **d**, Three representative PA molecular fragments of DP-M and P-M (left) and schematic illustration of the distribution of water cluster in each fragment (right). **e**, Radial distribution functions between water and the three PA fragments. **f**, Binding energy (BE) and the number of water molecules around the three PA fragments. All the information and data described in (c–e) were obtained from MD simulations. **g**, Schematic illustration of the working principle of semipermeable DP-M for ultrafast co-cation separation (yellow and blue spheres represent monovalent and divalent cations, respectively).
+
+The permeation of water through the PA membrane is primarily governed by the chemical features and nanoporous structure of the PA selective layer, during which water molecules need to overcome certain energy barriers when dissolving and diffusing through. The mutual interactions of water molecules with the binding sites on PA networks thereby have substantial impacts on the water permeation rate (Fig. **5a**). To gain a fundamental understanding of the mechanisms governing the facilitated permeation of water through the DP-M membrane, DFT atomistic calculations and MD simulations were performed. DFT was employed to specifically evaluate the molecular interactions between water and the cationic DP-M PA molecular fragments by calculating the surface free energy (SFE, Supplementary Table 4) and the molecular polarity index (MPI), which reflect membrane hydrophilicity from the perspectives of interfacial thermodynamics and quantum chemistry, respectively. As displayed in Fig. **5b**, DP-M shows a higher MPI value than that of P-M (i.e., 56.0 vs. 19.7), which is consistent with the lower surface
+water contact angle of the former. Meanwhile, the SFE value of DP-M is greater than that of P-M (i.e., 50.5 vs. 41.4 kJ/m^2), indicating a clear preference of polar water molecules for wetting and partitioning into the PA fragments of DP-M (Fig. 5c and Supplementary Table 4). To further elucidate the diffusion behavior of water molecules within the PA layer, MD simulations were subsequently conducted to acquire the binding affinities of water molecules to the PA fragments at the molecular level (Fig. 5d). The radial distribution functions (RDFs) plots (Fig. 5e), drawn with the respect to the PA fragments labeled in green (I), blue (II), and orange (III), respectively, reveal that the peak of I-H_2O in the first coordination layer is significantly higher than those of II-H_2O and III-H_2O, closely aligning with the DFT data. Furthermore, the water bonding (WB) capacity calculated by RDFs follows an order of WB_I > WB_{II} > WB_{III}, while the computed binding energies (BE) of fragments I, II, and III with water are in the order of |BE_I| > |BE_{II}| > |BE_{III}| (Fig. 5f). Hereupon, the cationic triazolyl heterocyclic PA structures derived from DAT-NH_2 in DP-M have a relatively lower affinity to water clusters. Such energy metrics accentuate a thermodynamic inclination of DP-M to pronouncedly facilitate the transport of water by providing water binding sites with moderate resistance, aligned with our design strategy and the ideal PA nanostructure we intend to construct (Fig. 5g).
+
+Conclusion
+
+Nanofiltration membranes with superior water permeance and precise ionic and molecular sieving capabilities offer promising solutions to address numerous challenges associated with water
+scarcity and renewable energy. In this study, a facile and robust molecular engineering strategy was demonstrated to reconcile the longstanding permeance–selectivity tradeoff threshold in state-of-the-art PA nanofiltration membranes. Our approach is contingent on the exquisite regulation of the porous nanostructure and the mutual interactions of the PA selective layer with water molecules and ions by the in situ creation of cationic triazolyl heterocyclic polyamide (CTHP) structures during interfacial polycondensations via synthetic DTA-NH$_2$ isomers. The obtained PA membrane exhibited synchronously enhanced water permeance and subangular selectivity for cation separation, achieving unparalleled divalent cation rejections of over 99%, accompanied by a 9-fold increase in monovalent/divalent cation sieving selectivity and tripled water permeance in comparison with the pristine benchmark, as well as outstanding chemical and operational stability, circumventing the permeance/selectivity threshold. Experimental data in tandem with advanced molecular simulations confirm that the intriguing permselectivity springs from the intricacy of the porous and chemical structures of the CTHP-modulated PA layer, which not only investigates a substantial decrease in the effective mean pore size and narrows the size distribution but also affords a high positive charge density, significantly strengthening the size sieving and Donnan exclusion effects in nanofiltration. Coincidentally, the CTHP structures also provide preferential water binding sites with low energy barriers, energetically facilitating the accommodation and diffusion of water molecules in the PA layer, which eliminates the increased water transport resistance caused by pore size shrinkage. The developed synthetic engineering strategy sheds light on the rational design and fabrication of advanced polymer membranes for high-precision
+Methods
+
+Synthesis of DAT-NH2 isomers
+
+3,5-Diamino-4/1-(2-aminoethyl)-1,2,4-triazole (DAT-NH2) isomers were synthesized via a one-step quaternization reaction following the reaction path shown in Supplementary Fig. 1. In a typical synthesis, 4.24 g of 3,5-diamino-1,2,4-triazole (DAT, 42.8 mmol) and 8.77 g of 2-bromoethylamine (42.8 mmol) were dissolved in 90 mL of N,N-dimethylformamide (DMF) in a 150 mL round-bottom flask. The flask with the reaction mixture was then heated to 45 °C in a water bath and reacted at this temperature for 24 h under vigorous stirring. A pale green solution rapidly formed with increasing reaction time. When the reaction was complete, the resulting mixture was immediately transferred into a 500 mL beaker, and 180 mL of acetonitrile was then added to obtain a milky white suspension. The obtained flocculent precipitates were subsequently redissolved in 10 mL of DMF and then precipitated with 180 mL of acetonitrile. The white solids were collected via high-speed centrifugation. The above dissolution and precipitation treatment was repeated three times. Finally, the as-synthesized DAT-NH2 was vacuum dried at 40 °C overnight and then stored in a sealed container for subsequent characterization and membrane fabrication. Detailed information on chemicals can be found in the Supporting Information (Supplementary Texts 1).
+Preparation of the PA NF membrane
+
+For the fabrication of the PA thin-film composite NF membrane, the polyether sulfone (PES) substrate was first immersed in an amine monomer aqueous solution with 0.1 wt% sodium dodecyl sulfate (SDS) and 0.1 wt% Na2CO3 for 5 min. After the excess water on the top surface was removed via filter paper, the amine-monomer saturated PES substrate was sandwiched into a homemade frame with the top surface facing upward. Interfacial polymerization was initiated by carefully adding excessive 0.3 wt% TMC solution into the frame to cover the surface, which was allowed to react for 1 min. When the reaction was complete, the excess hexane solution was drained, and the resulting membrane was dried at 60 °C for 30 min. Specifically, DP-M represents a PA membrane that was prepared following the above synthesis procedure using a mixture of DAT-NH2 (0.06 wt%) and PEI (0.44 wt%) as the amine monomer. The P-M benchmark membrane was fabricated solely by using PEI as the amine monomer. All the as-synthesized PA membranes were stored in deionized water at 5 °C for further characterization and performance tests. The detailed preparation process of the PA nanofilm is included in the Supporting Information (Supplementary Texts 2).
+
+Characterization
+
+The successful synthesis of DAT-NH2 isomers was confirmed by mass spectrometry (MS, MSQ Plus, USA) and high-performance liquid chromatography (HPLC, Ultimate 3000 RS, USA). The chemical structure of DAT-NH2 was characterized by proton nuclear magnetic resonance (1H NMR)
+spectroscopy (Bruker AVANCE AV400, USA). The chemical features of the PA membranes were also analyzed via Fourier transform infrared spectroscopy (FT-IR, Nicolet IN10, Thermo Fisher, USA) and X-ray photoelectron spectroscopy (XPS, Escalab 250Xi, Thermo Fisher, USA). The membrane surface morphology and roughness were examined via field emission scanning electron microscopy (FESEM, Quanta 250 FEG, FEI, USA) and atomic force microscopy (AFM, Nano Wizard 4, Bruker, Germany). The membrane cross-sectional morphology was identified via high-resolution transmission electron microscopy (TEM, FEI Tecnai G2 F30, FEI, USA). Surface hydrophilicity was assessed via water contact angle measurements on a contact angle goniometer (HARKE-SPCA, HARKE, China). The surface zeta potential was measured via a SurPASS electrokinetic analyzer (Anton Paar, GmbH, Austria). The molecular weight cutoff (MWCO) and pore size distribution of the membrane were obtained via solute retention tests using polyethylene glycol (PEG) probes with different molecular weights. The detailed procedures for each measurement are included in the Supporting Information (Supplementary Texts 3–4).
+
+Nanofiltration performance tests
+
+The separation performance of the PA membranes was characterized in nanofiltration mode at 23 °C via a cross-flow filtration apparatus with an effective membrane filtration area of 6.0 cm². Before data collection, the membrane sample was conditioned at a pressure of 2 bar greater than the intended test pressure until the water flux stabilized. The pure water flux (\( J_w \), L m\(^{-2}\) h\(^{-1}\), abbreviated as LMH) was measured using deionized water as the feed, and the water permeance
+(A, LMH/bar) was calculated via Eq. (1).
+
+\[
+A = \frac{J_w}{\Delta P} = \frac{\Delta V}{\Delta t \times S \times \Delta P}
+\] (1)
+
+where \( \Delta P \) (bar) is the trans-membrane hydraulic pressure, \( \Delta V \) (L) is the volume of permeate water collected during a time interval of \( \Delta t \) (h), and \( S \) (\( m^2 \)) is the effective membrane filtration area.
+
+The membrane ion sieving ability was examined via rejection tests that were conducted under various conditions using a wide spectrum of inorganic salts as solutes. Specifically, Na$_2$SO$_4$, Li$_2$SO$_4$, MgSO$_4$, MgCl$_2$, LiCl, NaCl, RbCl, KCl, NiCl$_2$, CaCl$_2$, and ZnCl$_2$ solutions with different concentrations and compositions were used as the feed. The single salt rejection (R, %) was calculated via Eq. (2).
+
+\[
+R = \left( 1 - \frac{C_p}{C_f} \right) \times 100\%
+\] (2)
+
+where \( C_p \) and \( C_f \) are the salt contents of the permeate and feed, respectively. The salt concentration was determined via conductivity measurement via a SevenCompact™ S230 (Mettler Toledo) conductivity meter. Binary mixtures of MgCl$_2$ and LiCl with different mass ratios were used to evaluate the membrane selectivity for cocation fractionation. The Li$^+$/Mg$^{2+}$ separation factor (\( S_{Li^+/Mg^{2+}} \)) was calculated via Eq. (3).
+
+\[
+S_{Li^+/Mg^{2+}} = \left( \frac{C_{f\ Mg^{2+}} / C_{f\ Li^+}}{C_{p\ Mg^{2+}} / C_{p\ Li^+}} \right)
+\] (3)
+where \( C_{f\ Mg^{2+}} \) and \( C_{f\ Li^+} \) and where \( C_{p\ Mg^{2+}} \) and \( C_{p\ Li^+} \) represent the concentrations of Mg\( ^{2+} \) and Li\( ^+ \) in the feed and permeate, respectively. An inductively coupled plasma optical emission spectrometer (ICP–OES, iCAP 7000, Germany) was used to quantify the ion contents of the solution. Each data point was tested three times under the same conditions using randomly selected membrane samples, and the average value was reported.
+
+The long-term stability of the membrane was evaluated by monitoring the water flux and salt rejection for up to 120 h at 6 bar using 1000 ppm MgCl\(_2\) solution as the feed. The pH stability of the membrane was assessed by measuring the water flux and salt rejection in a feed pH range of 1–13 using 1000 ppm LiCl solution as the probe feed, during which the solution pH was adjusted via HCl and NaOH.
+
+Density functional theory (DFT) calculations and molecular dynamics (MD) simulations
+
+DFT atomistic calculations were conducted via ORCA quantum chemistry software (version 5.0.4) 43-45. The binding energy and ion hydration energy were obtained via single-point energy calculations. Electrostatic potential (ESP) 46, 47, average local ionization energy (ALIE) 48, interaction region indicator (IRI) 42, independent gradient model based on Hirshfeld partition (IGMH) 49, and molecular polarity index (MPI) 50 analyses were performed with the Multiwfn software package to gain insights into molecular electronic properties and interaction patterns 51. LAMMPS and GROMACS were employed for MD simulations 52, 53, where LAMMPS was used
+to calculate the free volume fraction of the PA nanofilm, whereas GROMACS was applied to analyze the distribution and binding energy of water molecules around specific PA molecular segments. These simulations enabled a detailed exploration of water–PA interactions, which is crucial for understanding the hydration behavior and separation performance of the membrane. Comprehensive details of the computational methods and protocols are provided in Supplementary Text 5 and Text 6.
+
+References
+
+1. Zhang, F., Fan, J.-b. & Wang, S. Interfacial Polymerization: From Chemistry to Functional Materials. Angewandte Chemie International Edition **59**, 21840-21856 (2020).
+2. Lu, X. & Elimelech, M. Fabrication of desalination membranes by interfacial polymerization: history, current efforts, and future directions. Chem. Soc. Rev. **50**, 6290-6307 (2021).
+3. Epsztein, R., DuChanois, R.M., Ritt, C.L., Noy, A. & Elimelech, M. Towards single-species selectivity of membranes with subnanometre pores. Nat. Nanotechnol. **15**, 426-436 (2020).
+4. Wang, K. et al. Tailored design of nanofiltration membranes for water treatment based on synthesis–property–performance relationships. Chem. Soc. Rev. **51**, 672-719 (2022).
+5. Sengupta, B. et al. Carbon-doped metal oxide interfacial nanofilms for ultrafast and precise separation of molecules. Science **381**, 1098-1104 (2023).
+6. Lu, C. et al. Dehydration-enhanced ion-pore interactions dominate anion transport and selectivity in nanochannels. Science Advances **9**, eadf8412 (2023).
+
+7. Ji, Y. et al. Electric Field-Assisted Nanofiltration for PFOA Removal with Exceptional Flux, Selectivity, and Destruction. Environ. Sci. Technol. (2023).
+
+8. Zhang, Y. et al. Ice-confined synthesis of highly ionized 3D-quasilayered polyamide nanofiltration membranes. Science **382**, 202-206 (2023).
+
+9. Freger, V. & Ramon, G.Z. Polyamide desalination membranes: Formation, structure, and properties. Prog. Polym. Sci. **122**, 101451 (2021).
+
+10. Tan, Z., Chen, S., Peng, X., Zhang, L. & Gao, C. Polyamide membranes with nanoscale Turing structures for water purification. Science **360**, 518-521 (2018).
+
+11. Zhao, G.-J. et al. Polyamide Nanofilms through a Non-Isothermal-Controlled Interfacial Polymerization. Adv. Funct. Mater. **n/a**, 2313026 (2024).
+
+12. Geise, G.M. Why polyamide reverse-osmosis membranes work so well. Science **371**, 31-32 (2021).
+
+13. Lopez, K.P., Wang, R., Hjelvik, E.A., Lin, S. & Straub, A.P. Toward a universal framework for evaluating transport resistances and driving forces in membrane-based desalination processes. Sci. Adv. **9**, eade0413 (2023).
+
+14. Culp, T.E. et al. Nanoscale control of internal inhomogeneity enhances water transport in desalination membranes. Science **371**, 72-75 (2021).
+
+15. Wang, Z. et al. Manipulating interfacial polymerization for polymeric nanofilms of
+composite separation membranes. Prog. Polym. Sci. **122**, 101450 (2021).
+
+16. Zhou, X. et al. Intrapore energy barriers govern ion transport and selectivity of desalination membranes. *Science Advances* **6**, eabd9045 (2020).
+
+17. Liu, C. et al. Interfacial polymerization at the alkane/ionic liquid interface. *Angew. Chem. Int. Ed.* **133**, 14757-14764 (2021).
+
+18. Raaijmakers, M.J.T. & Benes, N.E. Current trends in interfacial polymerization chemistry. *Prog. Polym. Sci.* **63**, 86-142 (2016).
+
+19. Shen, L. et al. Polyamide-based membranes with structural homogeneity for ultrafast molecular sieving. *Nat. Commun.* **13**, 500 (2022).
+
+20. Li, X. et al. Polycage membranes for precise molecular separation and catalysis. *Nat. Commun.* **14**, 3112 (2023).
+
+21. Yuan, B. et al. Self-assembled dendrimer polyamide nanofilms with enhanced effective pore area for ion separation. *Nat. Commun.* **15**, 471 (2024).
+
+22. Zhao, G., Gao, H., Qu, Z., Fan, H. & Meng, H. Anhydrous interfacial polymerization of sub-1 Å sieving polyamide membrane. *Nat. Commun.* **14**, 7624 (2023).
+
+23. Chen, W., Liu, M., Ding, M., Zhang, L. & Dai, S. Advanced thin-film composite polyamide membrane for precise trace short-chain PFAS sieving: Solution, environment and fouling effects. *Process Saf. Environ. Prot.* **169**, 493-503 (2023).
+
+24. Dai, R. et al. Nanovehicle-assisted monomer shuttling enables highly permeable and selective nanofiltration membranes for water purification. *Nat. Water* **1**, 281-290 (2023).
+25. Peng, H. & Zhao, Q. A Nano-Heterogeneous Membrane for Efficient Separation of Lithium from High Magnesium/Lithium Ratio Brine. Adv. Funct. Mater. **31**, 2009430 (2021).
+
+26. Hong, S. et al. Precision ion separation via self-assembled channels. Nat. Commun. **15**, 3160 (2024).
+
+27. Zhang, S. et al. Solar-driven membrane separation for direct lithium extraction from artificial salt-lake brine. Nat. Commun. **15**, 238 (2024).
+
+28. Zhou, Z. et al. Electrochemical-repaired porous graphene membranes for precise ion-ion separation. Nat. Commun. **15**, 4006 (2024).
+
+29. Peng, H., Liu, X., Su, Y., Li, J. & Zhao, Q. Advanced Lithium Extraction Membranes Derived from Tagged-Modification of Polyamide Networks. Angew. Chem. Int. Ed. **62**, e202312795 (2023).
+
+30. Peng, H., Su, Y., Liu, X., Li, J. & Zhao, Q. Designing Gemini-Electrolytes for Scalable Mg2+/Li+ Separation Membranes and Modules. Adv. Funct. Mater. **33**, 2305815 (2023).
+
+31. Guo, B.-B. et al. Double charge flips of polyamide membrane by ionic liquid-decoupled bulk and interfacial diffusion for on-demand nanofiltration. Nat. Commun. **15**, 2282 (2024).
+
+32. Shefer, I., Peer-Haim, O. & Epsztein, R. Limited ion-ion selectivity of salt-rejecting membranes due to enthalpy-entropy compensation. Desalination **541**, 116041 (2022).
+
+33. Wang, W. et al. Simultaneous Manipulation of Membrane Enthalpy and Entropy Barriers towards Superior Ion Separations. Angew. Chem. Int. Ed. **63**, e202408963 (2024).
+
+34. Istirokhatun, T. et al. Unveiling the impact of imidazole derivative with mechanistic
+insights into neutralize interfacial polymerized membranes for improved solute-solute selectivity. *Water Res.* **230**, 119567 (2023).
+
+35. Freger, V. Nanoscale Heterogeneity of Polyamide Membranes Formed by Interfacial Polymerization. *Langmuir* **19**, 4791-4797 (2003).
+
+36. Muscatello, J., Müller, E.A., Mostofi, A.A. & Sutton, A.P. Multiscale molecular simulations of the formation and structure of polyamide membranes created by interfacial polymerization. *J. Membr. Sci.* **527**, 180-190 (2017).
+
+37. Zhu, Q.-Y. et al. Can the mix-charged NF membrane directly obtained by the interfacial polymerization of PIP and TMC? *Desalination* **558**, 116623 (2023).
+
+38. Zhang, M. et al. Modulating interfacial polymerization with phytate as aqueous-phase additive for highly-permselective nanofiltration membranes. *J. Membr. Sci.* **657**, 120673 (2022).
+
+39. Karan, S., Jiang, Z. & Livingston, A.G. Sub–10 nm polyamide nanofilms with ultrafast solvent transport for molecular separation. *Science* **348**, 1347-1351 (2015).
+
+40. Košutić, K., Dolar, D., Ašperger, D. & Kunst, B. Removal of antibiotics from a model wastewater by RO/NF membranes. *Sep. Purif. Technol.* **53**, 244-249 (2007).
+
+41. Sarkar, P., Ray, S., Sutariya, B., Chaudhari, J.C. & Karan, S. Precise separation of small neutral solutes with mixed-diamine-based nanofiltration membranes and the impact of solvent activation. *Sep. Purif. Technol.* **279**, 119692 (2021).
+
+42. Lu, T. & Chen, Q. Interaction Region Indicator: A Simple Real Space Function Clearly
+Revealing Both Chemical Bonds and Weak Interactions**. Chemistry–Methods **1**, 231-239 (2021).
+
+43. Neese, F., Wennmohs, F., Becker, U. & Riplinger, C. The ORCA quantum chemistry program package. *The Journal of Chemical Physics* **152**, 224108 (2020).
+
+44. Stoychev, G.L., Auer, A.A. & Neese, F. Automatic Generation of Auxiliary Basis Sets. *J. Chem. Theory Comput.* **13**, 554-562 (2017).
+
+45. Neese, F. Software update: The ORCA program system—Version 5.0. *WIREs Computational Molecular Science* **12**, e1606 (2022).
+
+46. Manzetti, S. & Lu, T. The geometry and electronic structure of Aristolochic acid: possible implications for a frozen resonance. *J. Phys. Org. Chem.* **26**, 473-483 (2013).
+
+47. Lu, T. & Manzetti, S. Wavefunction and reactivity study of benzo[a]pyrene diol epoxide and its enantiomeric forms. *Struct. Chem.* **25**, 1521-1533 (2014).
+
+48. Zhang, J. & Lu, T. Efficient evaluation of electrostatic potential with computerized optimized code. *Phys. Chem. Chem. Phys.* **23**, 20323-20328 (2021).
+
+49. Lu, T. & Chen, Q. Independent gradient model based on Hirshfeld partition: A new method for visual study of interactions in chemical systems. *J. Comput. Chem.* **43**, 539-555 (2022).
+
+50. Liu, Z., Lu, T. & Chen, Q. Intermolecular interaction characteristics of the all-carboatomic ring, cyclo[18]carbon: Focusing on molecular adsorption and stacking. *Carbon* **171**, 514-523 (2021).
+
+51. Lu, T. & Chen, F. Multiwfn: A multifunctional wavefunction analyzer. *J. Comput. Chem.*
+33, 580-592 (2012).
+
+52. Hess, B., Kutzner, C., van der Spoel, D. & Lindahl, E. GROMACS 4: Algorithms for Highly Efficient, Load-Balanced, and Scalable Molecular Simulation. J. Chem. Theory Comput. **4**, 435-447 (2008).
+
+53. Thompson, A.P. et al. LAMMPS - a flexible simulation tool for particle-based materials modeling at the atomic, meso, and continuum scales. Comput. Phys. Commun. **271**, 108171 (2022).
+
+**Acknowledgments**
+
+The authors gratefully acknowledge the financial support from the National Natural Science Foundation of China (22125603), Fundamental Research Funds for the Central Universities (nos. 040-63243125 and 040-63233061), and the National Key Research and Development Project (nos. 2023YFC3708003 and 2023YFC3708000). Special thanks are also made to the Han Gang Research Lab members for their helpful suggestions related to the characterization of materials.
+Supplementary Files
+
+This is a list of supplementary files associated with this preprint. Click to download.
+
+• SupportinginformationNov112024.pdf

0845528f7ce8151c643f7cfb0afbbb91c8fd501364586b8b087110447c3361d9/preprint/preprint.md

@@ -0,0 +1,408 @@
+Continuous Visual Navigation with Ant-Inspired Memories
+
+Gabriel Gattaux
+gabriel.gattaux@univ-amu.fr
+
+Aix-Marseille University - Institute of Movement Science   https://orcid.org/0000-0002-9424-7543
+
+Antoine Wystrach
+CNRS - Université Paul Sabatier   https://orcid.org/0000-0002-3273-7483
+
+Julien Serres
+Aix Marseille University   https://orcid.org/0000-0002-2840-7932
+
+Franck Ruffier
+Aix-Marseille Univ, CNRS   https://orcid.org/0000-0002-7854-1275
+
+Article
+
+Keywords:
+
+Posted Date: December 5th, 2024
+
+DOI: https://doi.org/10.21203/rs.3.rs-5505975/v1
+
+License: This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License
+
+Additional Declarations: There is NO Competing Interest.
+
+Version of Record: A version of this preprint was published at Nature Communications on September 24th, 2025. See the published version at https://doi.org/10.1038/s41467-025-62327-3.
+Continuous Visual Navigation with Ant-Inspired Memories
+
+Gabriel Gattaux1*, Antoine Wystrach2, Julien R. Serres1,3, Franck Ruffier1
+1Aix Marseille Univ, CNRS, ISM, Marseille, France.
+2Univ Toulouse, CRCA, CBI, UMR CNRS-UPS 5169, Toulouse, France.
+3Institut Universitaire de France, IUF, Paris, France.
+
+*Corresponding author(s). E-mail(s): gabriel.gattaux@univ-amu.fr;
+
+Abstract
+Solitary foraging ants excel in following long visual routes in complex environments with limited sensory and neural resources—an ability that remains challenging for robots with minimal computational power. Here, we introduce a self-supervised, insect-inspired neural network that enables robust route-following on the compact, low-cost Antcar robot. The robot leverages key aspects of ant brain and behavior: (i) continuous, one-shot visual route learning using panoramic encoding in a mushroom body-inspired network, (ii) categorization of low-resolution egocentric panoramas via oscillatory movements, (iii) opponent-process control of angular and forward velocities based on visual familiarity, (iv) recognition of places of interest along routes, and (v) motivation-based memory modulation. Antcar autonomously followed routes between indoor or outdoor destinations, forward or backward, while remaining stable in both theoretical analysis and real-world testing despite occlusions and visual changes. Across 1.3 km of autonomous travel, Antcar achieved challenging route-following with sub-20 cm lateral error at speeds up to 150 cm/s, requiring only 148 kilobits of memory and processing panoramas every 62 ms. This efficient, brain-inspired architecture stands out from more sensor-intensive and computationally demanding methods, presenting a neuromorphic approach with valuable insights into insect navigation and practical robotic applications.
+
+Introduction
+Insect navigation has long intrigued researchers across various fields, from biology to robotics, driving the development of cutting-edge technologies for autonomous mobile robots [1–3]. Autonomous navigation remains a demanding and interdisciplinary challenge with applications ranging from space exploration to last miles delivery [4, 5], especially in scenarios where robots cannot rely on satellite systems [6]. Simultaneously, robots serve as valuable tools for studying insects navigation and brain structure, advancing neuromorphic engineering [7–11].
+
+In Robotics, visual teach-and-repeat methods combined with dead-reckoning techniques have gained in popularity [12–15]. However, experienced solitary foraging ants navigate along familiar routes using only visual memories, without relying on dead reckoning (so-called path integration in the insect literature) [16–18]. This behavior has inspired various robotic models, although current implementations are generally limited to short-range experiments of about ten meters, with modest computational efficiency, precision, and accuracy [19–23]. While ant-inspired models achieve results comparable to conventional computer vision approaches [13, 24], they struggle in dynamic environments where computational efficiency must be balanced with resource use.
+
+Fig. 1 Biological inspiration for robotic navigation. An ant in the foreground symbolizes nature’s efficient navigational strategies, while the Antcar robot in the background integrates these principles into a neuromorphic system. The blurred image captures only the large masses of the environment, similar to the low-pass spatial filter in the ant’s visual system, which retains these large features even when objects obstruct the view between the robot and the building. ©Tifenn Ripoll - VOST Collectif / Institut Carnot STAR.
+
+These challenges are partly due to early navigation models that emphasized hymenopteran behavior rather than underlying brain processes. Early models, referred to as perfect memory models, stored periodic snapshots at specific waypoints [25, 26]. Then, during autonomous route following (or exploitation), forced scanning movements compared acquired views to an image bank, using
+rotational image differences to establish the most familiar image and desired heading—a process known as the visual compass [27–32]. However, these approaches has revealed two main limitations when applied in robotics.
+
+The first limitation involves the cumulative storage of snapshots, which significantly increases memory and computational demands as the route lengthens, making it unsuitable for long-distance navigation. This issue was partially addressed by a neural network using the Infomax algorithm [33], which enables efficient encoding of increasing numbers of images without a corresponding rise in memory load [20, 31, 34]. However, Infomax requires substantial adjustments to synaptic weights for each input through a non-local learning mechanism, limiting its biological plausibility.
+
+In parallel, research on the Mushroom Body (MB), a key part of the insect brain, has highlighted its essential role in olfactory and visual learning [35, 36]. In the MB, learning occurs through synaptic depression between thousands of Kenyon Cells (KCs) – intrinsic neurons that sparsely encode sensory input – and a few Mushroom Body Output Neurons (MBONs), which modulate behavioral responses based on learned associations. These processed signals are then transmitted to downstream neural circuits, influencing decision-making [37]. The first MB model simulating visual route following used a Spiking Neural Network with 20,000 KCs and one MBON to compute familiarity [38]. Despite this advancement, a second limitation remains: a forced systematic scanning during navigation slows robotic movement [21]. Also, this limitation does not reflect natural ant behavior, where scanning occurs only occasionally [39–41].
+
+To address the second limitation, an early robotic implementation combined a klinokinesis model with perfect memory, enhancing short-distance route-following by replacing cumbersome scanning with alternating, ballistic left and right turns where familiarity adjusted turn amplitude [19] (later also observed in ants [42]).
+
+To move beyond the random, undirected movement of kinesis, a taxis model was proposed, simulating directed movement toward a stimulus. In this model, KC firing activity was categorized into two distinct MBONs based on left or right orientation relative to the goal [43, 44]. This approach mirrors how insects, through continuous lateral body oscillations, sample multiple directions based on their nest position [42, 45]. Subsequent robotic models for route following attempted to integrate this lateralized approach by splitting the visual field into separate left and right memories, but these implementations showed limited efficiency in real-world tasks [22, 46]. In ants, however, the entire field of view is sent to the MB, and memories are fundamentally binocular [47].
+
+Here, we propose the lateralized route memories model, an MB-inspired design with four MBONs: two dedicated to route following and two for recognizing route extremities (Fig. 2). During a one-shot outbound learning route, ant-like body oscillations are simulated through continuous in-silico rotation of the panoramic image, mimicking head movement. This simulated head orientation, relative to the dynamic local orientation of the route, categorizes views into left or right memory based on the polarity of the angular value, leading to a self-supervised model for route learning. This design also mimics dopaminergic feedback from motor centers, modulating MBON synapses based on the currently active KCs and the integration of left and right stimuli [44].
+
+In addition, our model incorporates key aspects of ant navigation not previously applied in MB models, such as adjusting forward speed by accelerating on familiar routes and slowing down in unfamiliar areas [39]. Our model also enables bi-directional route learning, allowing to retrace a route while moving backward or forward, recognizing visual memories from the outbound journey [48–51]. Embedded in the compact Antcar robot (Figs. 1 and 2a), the model was tested across 99 autonomous trajectories, covering 1.3 km indoors and outdoors, achieving median lateral and angular errors of 20 cm and 3°, respectively, with refresh rates of 16 Hz during exploitation and 38 Hz during learning. Our MB model showed strong robustness to visual changes, including light fluctuations and pedestrian interference. This performance demonstrates the potential of our MB model for efficient, adaptable visual navigation in complex environments with accessible hardware and minimal computing requirements.
+
+Results
+
+Our proposed MB model emulates ant visual processing by encoding panoramic images as ultra-low resolution neural representations, enabling efficient learning and route recognition with minimal computational demands (see Methods for details, Fig. 2b). The model operates in two main phases: learning (Fig. 2c) and exploitation (Fig. 2d). During the learning phase, our self-supervised model encodes the route using two MBONs and stores place-specific memories for the Nest and Feeder as route extremities (see Methods, Fig. 2c). In the exploitation phase, the robot processes each view through both memory pathways, yielding two familiarity values (left and right MBON activities). The lateralized difference of familiarities (\( \lambda_{diff} \)) directs steering, while the maximum familiarity value modulates forward speed. Additionally, a motivational control modulates motor gain, allowing the robot to stop or reverse based on a familiarity thresholds set by place-specific MBONs (see Methods, Fig. 2d).
+
+This study begins with an offline analysis of the proposed self-supervised MB model using two route MBONs to assess stability, followed by experimental route-following tasks in challenging indoor and outdoor environments. Next, a homing task is described, in which the robot follows a long outdoor route in reverse toward the starting area, designated as the Nest (N), and stops nearby, utilizing three MBONs. Finally, a shuttling task is introduced, where the robot, after a single learning trial with two route MBONs and two extremities MBONs for the Nest and Feeder, autonomously shuttles to and fro between these two locations, driving both forward and backward.
+Fig. 2 Overview of the Lateralized Route Memories model implemented in the Antcar robot. This figure illustrates the process from image encoding to navigation control in both learning and exploitation phases. a The Antcar robot: a compact car-like platform equipped with an omnidirectional camera and a (Global Positioning System - Real-Time Kinematic) GPS-RTK system for ground truth data. b The image encoding process mimics ant’s visual processing. Panoramic images (I) are captured, blurred, subsampled, and edge-filtered to create a low-resolution 32 × 32 pixels panorama (IS). The IS is then transformed into Projection Neurons (PN), which are expanded into Excitatory Post-Synaptic Projections (EP) and reduced into Action Potentials (AP) via a κ-WTA function, forming the Kenyon Cells (KC). c During learning, the robot follows a path (C) from a start point (N) with an oscillatory movement to simulate angular deviations (\( \dot{\theta}_e \)). Synaptic updates occur in the Mushroom Body Output Neurons (MBONs) through the modulation by Dopaminergic-like Neurons (DAN), associating visual inputs with route memories in a self-supervised manner, dependent on the sign of \( \dot{\theta}_e \). An internal oscillator adjusts the image to simulate different angular errors, while joystick inputs control learning dynamics. d During exploitation, the robot aims to minimize the lateral (d) and angular (\( \dot{\theta}_e \)) errors relative to the route. The encoded image activates the MBONs according to the learned synaptic weights, allowing the robot to determine the position of the route and adjust its steering angle and speed. Familiarity indexes (\( \lambda \)) of MBONs work in an opponent valence process to guide navigation; steering adjustments are based on differentiated familiarities, while the maximum familiarity modulates the speed. Specific MBONs related to start and end points alter motivational states to adjust route polarity or stop movement.
+
+Self-supervised lateralized route memories model
+
+We first evaluated the self-supervised model for route learning (using only two MBONs) with a dataset of indoor and outdoor parallel routes (Figs. 3c,f). Results demonstrated that, with a controlled oscillation amplitude during learning, the model accurately estimated its heading error based on the differential familiarity \( \lambda_{diff} \), handling angular deviations up to 135° indoors and 90° outdoors (Fig. 3a,d,g). Furthermore, the maximum familiarity index \( \lambda_{max} \), used as feedback for speed control, increased proportionally with heading error, enabling the robot to slow down when misaligned with the route. This behavior was consistent even when the robot was moved laterally off-route (Fig. 3a,b,d and e), indicating a higher visual contrast with larger landmarks.
+
+The model’s ability to identify heading error accurately across training oscillation amplitudes up to 135° (Fig. 3i; see also Supplementary note 1 and Fig. S1) suggests that this parameter may not require further tuning below this threshold. However, larger oscillation amplitudes increased computation time, especially on the Raspberry Pi platform (0.4s for ±45°, Fig. 3i). Notably, the familiarity difference index (Fig. 3g) closely matched the spatial derivative of the maximum familiarity index, corresponding to the catchment area and turn rate amplitude observed in ants (Fig. 3h, Supplementary note 1, 2, Fig. S1 and S2 [43]).
+
+This analysis helped establish the operational limits of our MB model, maintaining stable behavior within a lateral error (d) of 2 meters and an angular error (\( \dot{\theta}_e \)) within the learning oscillation amplitude, set here at 45°. For asymptotic stability (i.e., the system’s ability to return to equilibrium), we assumed a proportional relationship between \( \lambda_{diff} \) and \( \dot{\theta}_e \), supported by the Pearson correlation coefficient being close to 1 (Fig. 3i) and expressed as \( K_{diff} \cdot \lambda_{diff} = -\dot{\theta}_e \), where \( K_{diff} \) is a tuned negative gain. Integrating this relationship into the robot’s motion equations, we applied a Lyapunov function for stability analysis. Results confirmed that the system converged to equilibrium points at \( d^e = 0 \) and \( \dot{\theta}_e^e = 0 \), effectively correcting small deviations and enabling the robot to remain aligned with the learned route. The full derivation of these equations and Lyapunov stability proof are provided in the Methods (section 6) and Supplementary note 3,4 and Fig. S3.
+
+Route-following: robustness to visual changes
+
+The proposed self-supervised approach for route learning was validated through a series of indoor and outdoor route-following tasks in fully autonomous mode, with
+Fig. 3 Offline familiarity mapping for learning of indoor and outdoor routes. This figure illustrates the differentiation and maximum familiarity of route Mushroom Body Output Neurons (MBONs) during offline analysis of panoramic images and positional data from indoor (Mediterranean Flight Arena) and outdoor (Luminy Campus, Marseille, France) environments. The mapping was performed using an oscillation amplitude \( A \) of 45°. **a,d** Familiarity difference index (\( \lambda_{diff} \)) and **b,e** familiarity maximum index (\( \lambda_{max} \)) are mapped in the route’s frame of reference, showing variations with both lateral and angular errors (\( d \) and \( \theta_e \)). The defined operating area is highlighted in pink. **c** Overview of the indoor (top) and **f** outdoor (bottom) environments with the learned route highlighted in red. **g** Cross-sectional view of the familiarity difference index (\( \lambda_{diff} \)) and **h** familiarity maximum index (\( \lambda_{max} \)) against the angular error (\( \theta_e \)) when the lateral error (\( d \)) is null. Plotted for indoor (solid line) and outdoor (dotted line) conditions. **i** Pearson correlation coefficient illustrating the linear relationship between familiarity difference index (\( \lambda_{diff} \)) and angular error (\( \theta_e \)) as a function of oscillation amplitude \( A \). This evolution of the correlation coefficient also illustrates the learning time required for a single oscillation cycle for each image captured on board the robot.
+
+only two MBONs. After a first outbound route with online learning, where images were captured continuously to update synaptic weights in real-time, the robot demonstrated robust route-following in various configurations (Figs. 4 and 5). First, the Antcar robot successfully navigated convex and concave routes in a cluttered indoor environment of approximately 8 meters (median lateral error ±median absolute deviation (MAD) = 0.21 ± 0.09 m, angular error ±MAD = 3.4 ± 6.2°, Fig. 4a,g and Fig. 7a). Moreover, the robot showed resilience in a kidnapped robot scenario, realigning with the learned route after being displaced (lateral error ±MAD = 0.26 ± 0.14 m, angular error ±MAD = 6.45 ± 4.19°, Fig. 4b and Fig. 7a). Only one crash occurred when the robot exceeded theoretical angular limits (see Supplementary Fig. S5 ).
+
+Further tests assessed the robot’s adaptability to high and low light conditions (Figs. 4c,h and Figs. 4d,i). Despite a single learning trial under standard lighting (815 Lux), the robot accurately followed its route in high (1,340 Lux) and low (81 Lux) lighting, with similar lateral and angular errors across tests (Fig. 7). This indicates that the MB-based control system is robust to significant changes in illumination.
+
+In dynamic conditions with pedestrians and camera occlusions (Figs. 4e,f), the robot maintained reliable route-following when encountering pedestrians (lateral error ±MAD = 0.27 ± 0.15 m, angular error ±MAD = 4 ± 2.8°, Fig. 4e and Fig. 7a) and with dynamic occlusions (lateral error ±MAD = 0.22 ± 0.13 m, angular error ±MAD = 4.7 ± 3.3°, Fig. 4f and Fig. 7a). The presence of pedestrians and occlusions was reflected by the
+Fig. 4 Real world experiments of indoor route following in different conditions. The learned route in red is approximately 8m long. These experiments used two route MBONs. Environmental configurations and specific familiarity data are provided in the Supplementary Fig. S4 and video. From a to f, Route following results using the proposed self-supervised one-shot learning approach in different environmental conditions. From g to k, The visual environments in which the robot evolved during the experiments.
+
+loss of maximum familiarity and led to speed reductions and increased emerging oscillatory motion (15% slower than in the previous experiments, Figs. 4e,f and supplementary video), which was also observed near obstacles. These results underscore the system’s resilience under challenging conditions.
+
+Outdoor experiments demonstrated the model’s ability to maintain stable performance even over a long, 53-meter route and under altered environmental conditions. A route was learned and accurately recapitulated on a sunny day (lateral error ±MAD = 0.39 ± 0.13 m, angular error ±MAD = 5.8 ± 2.8°, Fig. 5a & 7a) and then retested the following day with parked cars removed (lateral error ±MAD = 1.3 ± 0.5 m, angular error ±MAD = 6.2 ± 3.2°; Fig. 5b & 7a). While the robot’s error margins were slightly broader on the second day, it remained well within acceptable limits over the entire route. To test Antcar’s maximum speed, a higher speed gain was applied during the second test (Fig. 5b), resulting in a cruising speed of 1.5 m/s compared to 1 m/s on the first day (see Supplementary Information note 5, Fig. S4 and Table S7).
+
+Homing: homeward route and stop
+
+Building on the validated route-following strategy, further tests refined the robot’s behavior, focusing on ant-like homing. Homing, by definition, is the ability to return to a specific location after displacement. To test this, we evaluated the robot’s ability to follow a 50 m outdoor route in reverse, stopping at a designated Nest area (point N in Fig. 6a). During learning, a 180° shift in the visual oscillation pattern simulated the “turn back and look” behavior observed in ants and led to homeward route following.
+
+The robot successfully followed the 50 m route in reverse under cloudy outdoor conditions (lateral error ±MAD = 0.9 ±0.5 m, angular error ±MAD = 6.3 ±4.2°, Fig. 6a and Fig. 7a). Although maximum familiarity was higher than in previous outdoor experiments (see Supplementary note 5, Fig. S4 and Table S7), overall accuracy remained stable and emerging oscillatory movements was demonstrated (see Supplementary Video).
+
+To enable autonomous stopping at the Nest, a place-specific MBON was used to learn ‘nest-views’ at the starting point of the route. Subsequent ‘recognition’ in this MBON, based on a familiarity threshold, acted as a motivational cue to halt route-following behavior and reducing the robot’s linear velocity. This mechanisms was sufficient for the robot to successfully reach and stop at the Nest area in 4 out of 5 trials, with a median stopping distance of 1.4 m (Fig. 6c, see also Supplementary Fig. S6b for detailed familiarities values over distance).
+
+Shuttling: foodward and homeward routes
+
+Reverse route-following is also commonly observed in ants and was successfully replicated on board Antcar. Homing ants can pull food items backward when it is too large to carry forward, maintaining body alignment with the outbound route learned forward, and
+Fig. 5 Real word experiments of outdoor route-following with shared memories. a First day experiments, learning and autonomous route with several cars along the road. b Second day experiments, autonomous routes using the memories from day one in an altered environment (without cars).
+
+using outbound memories with an opposite valance [50]. Shuttling tests show the robot’s ability to switch movement direction and drive backward while maintaining alignment with the outbound route (Fig. 6b).
+
+This foraging behavior was made possible by incorporating two additional place MBONs, which learned a series of panoramic views defining each endpoint of the route (Feeder and Nest). During shuttling, the model triggered a switch in motor gain polarity upon recognizing these panoramic views corresponding to the Feeder or Nest areas. In a cluttered indoor environment along a 6-meter learned route, the robot autonomously shuttled to and fro between the Feeder and the Nest, covering a total distance of 160 meters without interruption. Using a similar familiarity threshold on the two route-extremity MBONs, the robot detected the endpoints 22 times, achieving a median stopping distance of 0.31 m (Fig. 6d) (See Supplementary Fig. S6a for detailed familiarities values over distance).
+
+This continuous shuttling revealed distinct differences in error profiles between forward and backward movement (Fig. 6b). During forward motion, the robot maintained stable control with minimal deviations (lateral error ±MAD = 0.1 ± 0.03 m, angular error ±MAD = 1.26 ± 0.83°, Fig. 6b). However, during backward motion, the traction-driven setup amplified steering effects, resulting in slightly larger deviations from both accuracy and precision, though overall performance remained acceptable (lateral error ±MAD = 0.19 ± 0.08 m, angular error ±MAD = 2.7 ± 2.1°, Fig. 6b & 7a). The increased ‘motor’ variability led to lower visual recognition signal and thus usefully affected speed, which decreased by 14% compared to forward motion (see Supplementary note 5, Fig. S4 and Table S7). Nonetheless, the robot consistently realigned with the correct path after such minor deviations. These results highlight the model’s versatility across different driving dynamics, capability to implement inverted steering, and adaptability to variations in motor kinematics and propulsion.
+
+Performance summary
+
+Across all experiments, including both indoor and outdoor route-following, homing and shuttling tasks, the model demonstrated robust and stable navigation performance, completing 99 autonomous trajectories with a total of 1.3 km traveled. The theoretical limits of the system were validated, with convergence toward equilibrium points consistently achieved under various environmental conditions, even in the presence of noise (lateral error ±MAD = 0.22 ±0.10 m, angular error ±MAD = 3.8 ±2.4°, Fig. 7b). Lateral errors were within acceptable margins for both indoor and outdoor contexts, aligning within the standard widths of roads in France (5m) and typical indoor corridor (1.5m).
+
+Additionally, statistical analysis showed no significant differences in the lateral or angular errors across the eleven test scenarios (Kruskal-Wallis test, \( H = 1.20 \) for lateral error, p value \( \approx 1 \); \( H = 0.97 \) for angular error, p value \( \approx 1 \)), underscoring the system’s reliability across diverse conditions (see Statistical Information). These results highlight the robustness and adaptability of the MB model in both structured and dynamic environments, confirming its potential applicability in a variety of navigation contexts.
+
+Discussion
+
+Our study presents a robust, embedded, and biologically inspired Mushroom Body (MB) model capable of long-distance navigation in the real world with minimal sensor acuity and computational resources. Using fewer than a thousand pixels, the Antcar robot successfully followed routes at speeds up to 1.5 m/s—approximately eight times its body length—achieving continuous online learning in just 20 ms per image, with exploitation times of 75 ms and an extrapolated memory footprint of only 0.3 Mo per kilometer. By integrating ant-inspired lateralized memory with self-supervised panoramic learning through oscillations, our model sustained high navigational accuracy across dynamic lighting, cluttered, and altered environments, with a positional accuracy of approximately 20 cm. Offline analysis confirmed the model’s stability and alignment with defined limits, predicting robust real-time performance by reliably maintaining route alignment within learning oscillation bounds.
+
+The angular error between the agent’s head direction and the dynamic local route orientation (defined
+Fig. 6 Real word experiments of outdoor homing and indoor shuttling. a Homing experiments using two route MBONs and one motivation MBON for a 53m L-shaped route, in an outdoor cloudy environment. Autonomous route headed in the opposite direction. b Familiarity nest index (\( \lambda_N \)) over traveled distance with the fixed stopping condition (\( p = 0.2 \)). c Shuttling experiments using two route MBONs and two place MBONs in an indoor environment with artificial visual cues. Autonomous routes swing back (blue) and forth (black). d Familiarity nest (\( \lambda_N \)) and feeder (\( \lambda_F \)) index over the traveled distance, zoomed in to illustrate backward and forward movement.
+
+in the Methods as the Frenet frame [52]) emerged as both a challenge during exploitation—where the system minimizes this error—and a cue during learning, where the categorization process depends on its polarity. Our model demonstrated homing behavior using either a 180° shift in visual oscillation or by inverting motor gains, thus enabling forward and backward movements with only a single forward learning route. Additionally, visual place memories stored in supplementary MBONs, paired with a motivational control system, allowed the robot to recognize route endpoints and modulate motor gain, halting movement or reversing foraging motivation. With a single learning pass in one direction, the agent could follow the route forward, backward, and in reverse, controlled by oscillation parameters and motivational cues. Only motivational rules required adjustment to switch between route following, homing, and shuttling, underscoring the model’s flexibility.
+
+Our results surpass earlier ant-inspired familiarity-only models robots, which were generally limited to short indoor routes, slower linear speed (stop and scan), and lower efficiency [19–23]. Our model also markedly outperforms state-of-the-art visual teach-and-repeat methods, which report memory footprints of 3 Mo per kilometer and processing times around 400 ms [13]. Our model also achieves competitive results against teach-and-repeat systems incorporating odometry [14, 15].
+
+This lateralized MB model distinguishes itself through reduced time and space complexity for route direction processing compared to perfect memory, snapshot, and visual compass approaches [43]. Whereas time and space complexity increase with the number of images in perfect memory or snapshot models, our MB model maintains constant space complexity, relying only on the synaptic matrix size KCtoMBON. Additionally, in contrast to visual compass approaches, where computational complexity scales with in-silico scan range and resolution during exploitation (\( \mathcal{O}(n) \)), our MB model maintains a constant factor (\( \mathcal{O}(1) \)) since in-silico scanning is only required during learning. For instance, while a visual compass scanning a ±45° range at 1° resolution requires 90 comparisons per image, our model requires only two comparisons, eliminating the need for angular scanning in exploitation. Notably, our model produced commands five times faster than the visual compass approach on the same robot platform [21].
+
+Our contribution also aligns well with current biological observations, particularly highlighting the effectiveness of latent learning [53], where continuous learning bypass the need to control “when to learn” [31, 44]. The opposed event-triggered and snapshot-based learning models producing place learning [15, 54] where used here only to recognise place of interests such as the nest and the feeder to switch motivation, but were not engaged for route guidance. Also, our MB model prioritized body orientation within the local frame rather than divided the visual field [22, 46], aligning with biological observations in ants with unilateral visual impairment, showing that these insects store and recognise fundamentally binocular views [47]. Interestingly, the linear relationship observed between familiarity measures (and thus motor output) and angular error during exploitation closely mirrors experimental findings in ants [43]. This relationship enabled us to demonstrate the asymptotical stability of the system within a defined domain, ensuring the consistent and predictable behavior essential for a robotic navigation model [55].
+Fig. 7 Performance during route following overview a Detailed errors for each experiment. b Weighted bi-variate distribution for lateral (d) and angular errors (\( \theta_e \)) across 11 different experimental configurations.
+
+Furthermore, oscillatory learning behavior mirrors ant behavior, where initial routes involve slow, rotational movements, transitioning to direct paths on subsequent journeys [39]. These oscillations typically fall within \( \pm 10^\circ \), with peaks around \( \pm 45^\circ \) in unfamiliar terrain [40, 42]. The robot’s ability to slow down and produce emerging mechanical scanning upon entering unfamiliar areas (see Supplementary Video) are consistent with such naturalistic behaviors. Finally, Antcar’s homing capability was maintained even when navigating backward, closely mirroring ant behavior while dragging food [48–50, 56]. Overall, our attempt to integrate multiple MBONs, oscillations, “turn back and look” behavior, and motivational control mechanisms echoes insect mechanisms [2, 57], and the resulting expression when implemented in the robot echoes insect behaviours.
+
+This study addresses several core needs identified in research on embodied neuromorphic intelligence [6, 8], such as robustness to visual changes, adaptability to real-world environments, and support for extended route learning. Our algorithm’s efficiency allows computational power for additional tasks, making it valuable in GPS-compromised or SLAM-disrupted scenarios (SLAM stands for Simultaneous Localization And Mapping). The robot’s low-resolution, wide-angle vision proves resilient against moving objects that often disrupt SLAM. Our model is well-suited for dynamic environments or situations where odometry (e.g., visual, inertial, step-counting, or wheel-rotation) is unreliable.
+
+Interestingly, the semi-random encoding process, specifically the PNToKC synaptic projections, introduces a “fail-secure” memory-sharing mechanism. If synaptic weights for encoding differ, memory sharing becomes inaccessible, an advantageous feature for swarm robotics or cross-robot memory sharing.
+
+Future research could enhance this approach. Transitioning this model to a spiking neural network on neuromorphic hardware could further enhance computational efficiency and biological fidelity [11]. Additionally, incorporating obstacle avoidance [58], would improve performance in dynamic environments.
+
+In addition, a reduction of the visual field could correspond to more general cases, rendering in silico scanning impossible. In such scenarios, it would be necessary to estimate the angular error between the road frame and the agent. This could be achieved using a local angular path integration system (or odometry) during learning. As demonstrated by Collett et al. [59], showing that ants could utilize route segment odometry for navigation.
+
+Our approach does not cover beeline homing post-foraging or search behaviors near points of interest, although these could be added by adding path integration mechanisms [60] or using the current visual mechanism but adding “learning walk” behaviors around place of interest [44]. Additionally, fixed neural parameters across all experiments suggest an opportunity for further exploration by adjusting Kenyon Cell numbers or connectivity, or testing different MB learning mechanisms [61]. Expanding the number of MBONs, akin to the 34 in Drosophila [37], could enable more complex motivational states, multi-branch memory storage [53], and broader navigational abilities [62].
+
+Overall, inspired by the neuroethology of ants, our MB model provides an effective bridge between theoretical insights and practical applications in insect-inspired autonomous robotic navigation. This egocentric model confirms the neuromorphic architecture’s promise for autonomous systems, suggesting a scalable solution for both robotics and biological research applications.
+
+Methods
+
+This section describes the methodology used in the present study, focusing on the Encoding, Learning, and Exploitation processes of the proposed MB model (Figs. 2b-d). We also provide details on the hardware setup, control architecture, and stability analysis (See Supplementary Fig. S7 for the detailed route following neural network).
+
+Image Encoding
+
+Inspired by the visual system of ants [63], the model encoded real-world images into sparse, binary neural representations to efficiently handle visual input.
+The encoding function (Fig. 2b) processed panoramic images from a camera with a 220° vertical and 360° horizontal field of view. This wide field of view enabled the camera to capture from slightly below the horizon to nearly directly below itself. To enhance natural contrast, the green channel of each image was selected [63], followed by Gaussian smoothing (\( \sigma = 3 \) pixels) to reduce noise. The image was then downsampled to an ultra-low-resolution \( 32 \times 32 \) pixel thumbnail (0.145 pixel per degree), approximating the visual resolution of ants at 7.1° between adjacent photoreceptors.
+
+Next, a Sobel filter extracted edges, mimicking lateral inhibition as seen in insect optical lobes [64]. These processed images were flattened into 800 Visual Projection Neurons (PNs), comparable to the number of ommatidia in ants. The PNs were further expanded into Kenyon Cells (KCs) using a fixed, sparse pseudo-random synaptic matrix (PNtoKC). Each KC received input from four PNs, enhancing the visual encoding’s discriminative power within the Mushroom Body (MB) [65], forming an Excitatory Post Synaptic Projection (EP) vector of size \( u \).
+
+The EP vector size was set to \( u = 15,000 \) for the route MBONs (\( MBON_R \) and \( MBON_L \)), while for place-specific MBONs (\( MBON_N \) and \( MBON_F \)), which required fewer images, \( u \) was set to 5,000. A \( \kappa \)-Winner-Take-All (WTA) mechanism was applied to capture the highest contrasts, creating a high-dimensional, sparsified binary vector. This vector, referred to as the Action Potential (AP), consequently activated only 1% of KCs (\( \kappa = 0.01 \)), giving \( \overline{u} = u * \kappa \) active neurons. This final binary representation served as the encoded visual input.
+
+All parameters were predefined by literature and experimental tests, but not further optimized.
+
+Routes and places learning
+
+The learning process is governed by synaptic depression through anti-Hebbian learning.
+
+\[
+KCtoMBON_i = \begin{cases}
+0, & \text{if } AP_i = 1 \\
+KCtoMBON_i, & \text{otherwise}
+\end{cases}
+\] (1)
+
+For each MBONs, their synaptic weight matrix (\( KCtoMBON \)) dynamically adjusted their weight based on input from the \( AP \) layer described in equation 1 and from the mimicked dopaminergic feedback. Here, \( i \) represents the \( i^{th} \) neuron in the specified vector, with \( KCtoMBON_i \) and \( AP_i \) in \( \{0, 1\} \).
+
+The simulated oscillatory movements during learning were obtained by rotating each captured image in steps, creating a sweep of rotations (\( \theta_c \)) described by the following function:
+
+\[
+\theta_c(n) = A \cdot \sin \left( n \cdot \Delta \theta + \phi \right) \quad \text{for } n = 0, 1, 2, \ldots, \frac{2A}{\Delta \theta}
+\] (2)
+
+where \( A \) represents the oscillation amplitude, \( \Delta \theta \) the step size, and \( \phi \) the phase shift. The step size was fixed at \( \Delta \theta = 5^\circ \), with \( A = 45^\circ \) for route MBONs and \( A = 30^\circ \) for place MBONs. The phase shift was \( \phi = 180^\circ \) only for the homing task (Fig. 6).
+
+For route learning, the model assumed the robot perfectly aligned to the route being learned. The body rotation was estimated as \( \theta_e = \theta_c + \theta_s \), where therefore \( \theta_e = 0 \) during learning. The encoded binary image was categorized based on the polarity of \( \hat{\theta}_e \), such that:
+
+\[
+\begin{cases}
+Learn(AP, KCtoMBON_R), & \text{if } \hat{\theta}_e \leq 0 \\
+Learn(AP, KCtoMBON_L), & \text{if } \hat{\theta}_e \geq 0
+\end{cases}
+\] (3)
+
+Here, the function \( Learn() \) follows equation 1. Synaptic weights (KCtoMBON) were stored in CSR format, achieving significant data compression to 148 kilobits independently of the route length, reducing memory requirements by 99.97% from cumulative image storage. This self-supervised model continuously learned visual input at high throughput without memory overload, as only novel views (i.e., newly recruited KCs) modulated synapses. Several panoramic views were learned to define the start and finish areas in their respective MBONs, serving as motivational cues.
+
+Exploitation process and control architecture
+
+During exploitation, the model calculated familiarity scores (\( \lambda \)) by comparing the current input (\( AP \)) with each MBON’s synaptic weight matrix (\( KCtoMBON \)):
+
+\[
+\lambda = \frac{1}{\overline{u}} \sum_{i=1}^u AP_i \cdot KCtoMBON_i
+\] (4)
+
+This familiarity score, ranging from 0 (unfamiliar) to 1 (familiar), was used to assess route alignment. The lateralized difference in familiarities between the left and right MBONs (\( \lambda_{diff} = \lambda_L - \lambda_R \)), which indicates whether the current view is more oriented to the left or right of the route, guided the robot’s steering angle (\( \varphi \)). Meanwhile, the maximum familiarity (\( \lambda_{max} = \max(\lambda_L, \lambda_R) \)), representing how familiar the current view is, modulated its speed (\( v \)).
+
+Thus, the control input \( U \) was defined as:
+
+\[
+U = \begin{bmatrix} v \\ \varphi \end{bmatrix} = \begin{bmatrix} M \cdot K_v \cdot \operatorname{sat}(1 - \lambda_{max}) \\ M \cdot K_\varphi \cdot \lambda_{diff} \end{bmatrix}
+\] (1)
+
+Here, \( K_v \) and \( K_\varphi \) are proportional gains that control linear and angular velocities, while the saturation function (\( \operatorname{sat()} \)) establishes a minimum throttle level, ensuring minimum speed even at low familiarity levels. The motivational state (\( M \)) regulated transitions between behaviors based on a familiarity thresholds within place-specific MBONs. During route following, \( M \) was consistently set to 1. In homing experiments, where the objective was to stop at the nest, \( M \) initially started at 1 and switched to 0 once the familiarity of the nest-specific MBON (\( \lambda_N \)) fell below a fixed threshold (\( p = 0.2 \)), signaling arrival at the nest. For shuttling tasks, \( M \) alternated between values of 1 and \(-1\) as the
+robot reached each route extremity, driven by a familiarity thresholds of the two place-specific MBONs (\( \lambda_N \) and \( \lambda_F \)).
+
+Theoretical analysis of the robot stability
+
+Stability in mobile agents, biological or robotic, is essential for reliable, predictable behavior. In control theory, an agent’s motion is generally modeled as \( \dot{x} = f(x, U) \), where \( x \) is the state vector (e.g., position or velocity), \( U \) is the control input, and \( f \) describes system dynamics. A desired equilibrium point \( x_e \) is achieved by defining a control input \( U_e \) such that \( f(x_e, U_e) = 0 \), allowing the system to maintain stability and return to equilibrium after disturbances. Stability is typically assessed using a Lyapunov function [55], which ensures the system converges to a stable state over time.
+
+In contrast to conventional control approach, we applied a neuroethologically inspired control input derived from ant behavior, assessing stability via an a posteriori Lyapunov analysis. The robot’s motion was modeled in a Frenet frame, a moving reference frame coincident with the nearest point on the route, to minimize lateral and angular errors, defined by \( x = [d, \theta_e] \). Empirical data for stability assessment was collected in indoor and outdoor environments (paths of approximately 6 meters with 855 learned images each), providing distinct visual contexts (Figs. 2, 3). The robot’s equations of motion from a global to the Frenet frame are [66]:
+
+\[
+\begin{bmatrix}
+\dot{s} \\
+\dot{d} \\
+\dot{\theta}_e
+\end{bmatrix}
+=
+\begin{bmatrix}
+v \left( \cos \theta_e - \tan \varphi \sin \theta_e \right) \\
+v \left( \sin \theta_e + \tan \varphi \cos \theta_e \right) \\
+v \frac{\tan \varphi}{L}
+\end{bmatrix},
+\]
+
+where \( s \) is the arc length along the route, \( d \) is the lateral error, and \( \theta_e \) is the angular error.
+
+This kinematic model, along with by empirical observations (Fig. 3), enabled us to establish an asymptotically stable domain for lateral and angular errors (\( d \) and \( \theta_e \)), ensuring reliable route-following performance even with minor disturbances. The full theoretical stability proof and derivations of the model in the frenet frame are provided in the Supplementary note 3 and 4.
+
+Antcar robot and ground truth system
+
+The experiments were conducted using Antcar (Fig. 1 and Fig. 2a), a PiRacer AI-branded car-like robot. Antcar features four wheels, with two rear drive wheels powered by 37-520 DC motors (12V, 1:10 reduction rate) and a front steering mechanism controlled by an MG996R servomotor (9kg/cm torque, 4.8V). The robot’s chassis measures 13×24×19.6 cm and is powered by three rechargeable 18650 batteries (2600mAh, 12.6V output). Antcar’s primary sensor is a 220° Entaniya fisheye camera, mounted upward to capture panoramic images at 160 × 160px × 3 resolution and 30 Hz, processed using OpenCV on a Raspberry Pi 4 Model B (Quad-core Cortex-A72, 1.8GHz, 4GB RAM), running Ubuntu 20.04. Note that there was no closed-loop control on the wheel rotation speed. Raspberry Pi manages real-time performance and controls the motors through a custom ROS architecture.
+
+Real-time communication is facilitated by ROS Noetic, either via Wi-Fi (indoor) or a 4G dongle (outdoor). The robot can be controlled manually using a keyboard, joystick or with GPS waypoint, but in autonomous visual-only mode, it follows its own internal control law. Control inputs—steering angle (\( \varphi \)) and throttle (\( v \)) are processed using the PyGame library. Real-time data visualization and post-experiment monitoring are achieved via Foxglove.
+
+Antcar has a maximum velocity of 1.5 m/s and a maximum steering angle of 1 rad, with a wheelbase of 0.15 m. The robot’s configuration states \( q = (x, y, \theta) \) were tracked using different systems. Indoor experiments utilized eighteen Vicon™ motion capture cameras, with infrared markers on Antcar providing precise tracking at 50 Hz with 1 mm accuracy. Outdoor experiments employed a GPS-RTK system with a SparkFun GPS-RTK Surveyor, providing 14 mm accuracy at 2 Hz (GPS-RTK stands for Global Positioning System - Real-Time Kinematic). Ground speed and angular speed were calculated through position differentiation. The base station used for GPS corrections was a Centipede LLENX station located at 24 km (Aeroport Marseille Provence) from the experiment site in Marseille. Note that the ground truth acquisition system was run on the Rapserry Pi along with the mushroom body model.
+
+Lateral error was calculated by finding the nearest point on the learning route using the Euclidean distance, with the shortest distance representing the absolute lateral error. Angular error was defined as the absolute difference in heading between the nearest learning route point and the current position. The euclidean distance between the agent and the Nest or Feeder areas was calculated to estimate the distance when the robot switched behavior (i.e familiarity dropped below the threshold).
+
+Statistical informations
+
+The errors used for statistics were recorded at each command decision timing. Due to non-normality in error values (with outliers retained), Box-Cox transformations were applied to stabilize variance across experiments, reducing the impact of outliers caused by indoor obstacles that hid the robot from the motion capture system or by GPS-RTK inaccuracies outdoors. The groups was compared using the Kruskal-Wallis test [67], and median values are reported with median absolute deviation (MAD), as median ± MAD. The package python SciPy [68] was used for the statistics. The overall medians and bivariate distribution plots were weighted by the number of measurements per experiment for the Fig. 7.
+
+Acknowledgments
+
+The authors thank David Wood for revising the English in this study, Guillaume Caron for providing the camera reference, and Thomas Gaillard, Clément Serrasse, and Hamidou Diallo for their assistance during the robotic tests.
+Declarations
+
+• Funding: G.G. was supported by a doctoral fellowship grant from Aix Marseille University and the French Ministry of Defense (AID - Agence Innovation Défense, agreement #A01D22020549 ARM/DGA /AID). G.G., J.R.S. and F.R. were also supported by Aix Marseille University and the CNRS (Life Science, Information Science, and Engineering and Science & technology Institutes). The facilities for the experimental tests has been mainly provided by ROBOTEX 2.0 (Grants ROBOTEX ANR-10-EQPX-44-01 and TIRREX ANR-21-ESRE-0015).
+• Conflict of interest: the authors declare no competing interests.
+• Data availability: Upon publication
+• Code availability: Upon publication
+• Supplementary Video : https://youtu.be/OsI5Jyy6dF4
+• Author contribution: G.G., A.W., J.R.S., and F.R. designed this research work; G.G, A.W., J.R.S., and F.R. got funding for this study; G.G. performed experiments, collected and visualized the data; G.G., A.W., J.R.S., and F.R. analyzed data; G.G. wrote the first full draft. All authors reviewed the results and approved the final version of the manuscript.
+
+References
+
+[1] Franceschini, N. Small Brains, Smart Machines: From Fly Vision to Robot Vision and Back Again. Proceedings of the IEEE 102, 751–781 (2014).
+
+[2] Webb, B. & Wystrach, A. Neural mechanisms of insect navigation. Current Opinion in Insect Science 15, 27–39 (2016).
+
+[3] Denuelle, A. & Srinivasan, M. V. A sparse snapshot-based navigation strategy for UAS guidance in natural environments, 3455–3462 (IEEE, 2016).
+
+[4] Glick, P. E., Balaram, J. B., Davidson, M. R., Lyons, E. & Tolley, M. T. The role of low-cost robots in the future of spaceflight. Science Robotics 9, eadl1995 (2024).
+
+[5] Yang, G.-Z. et al. The grand challenges of Science Robotics. Science Robotics 3, eaar7650 (2018).
+
+[6] de Croon, G. C., Dupeyroux, J., Fuller, S. B. & Marshall, J. A. Insect-inspired ai for autonomous robots. Science Robotics 7, eabl6334 (2022).
+
+[7] Mangan, M. et al. A virtuous cycle between invertebrate and robotics research: Perspective on a decade of Living Machines research. Bioinspiration & Biomimetics 18, 035005 (2023).
+
+[8] Bartolozzi, C., Indiveri, G. & Donati, E. Embodied neuromorphic intelligence. Nature Communications 13, 1024 (2022).
+
+[9] Webb, B. Robots in invertebrate neuroscience. Nature 417, 359–363 (2002).
+
+[10] Franz, M. O. & Mallot, H. A. Biomimetic robot navigation. Robotics and Autonomous Systems (2000).
+
+[11] Sandamirskaya, Y., Kaboli, M., Conradt, J. & Celikel, T. Neuromorphic computing hardware and neural architectures for robotics. Science Robotics 7, eabl8419 (2022).
+
+[12] Simon, M., Broughton, G., Rouček, T., Rozsypálek, Z. & Krajník, T. in Performance Comparison of Visual Teach and Repeat Systems for Mobile Robots (eds Mazal, J. et al.) Modelling and Simulation for Autonomous Systems, Vol. 13866 3–24 (Springer International Publishing, Cham, 2023).
+
+[13] Stelzer, A., Vayugundla, M., Mair, E., Suppa, M. & Burgard, W. Towards efficient and scalable visual homing. The International Journal of Robotics Research 37, 225–248 (2018).
+
+[14] Nourizadeh, P., Milford, M. & Fischer, T. Teach and Repeat Navigation: A Robust Control Approach. 2024 IEEE International Conference on Robotics and Automation (ICRA) 2909–2916 (2024).
+
+[15] Van Dijk, T., De Wagter, C. & De Croon, G. C. H. E. Visual route following for tiny autonomous robots. Science Robotics 9, eadk0310 (2024).
+
+[16] Mangan, M. & Webb, B. Spontaneous formation of multiple routes in individual desert ants (Cataglyphis velox). Behavioral Ecology 23, 944–954 (2012).
+
+[17] Kohler, M. & Wehner, R. Idiosyncratic route-based memories in desert ants, Melophorus bagoti: How do they interact with path-integration vectors? Neurobiology of Learning and Memory 83, 1–12 (2005).
+
+[18] Wystrach, A., Schwarz, S., Schultheiss, P., Beugnon, G. & Cheng, K. Views, landmarks, and routes: how do desert ants negotiate an obstacle course? Journal of Comparative Physiology A 197, 167–179 (2011).
+
+[19] Kodzhabashev, A. & Mangan, M. Wilson, S. P., Verschure, P. F., Mura, A. & Prescott, T. J. (eds) Route Following Without Scanning. (eds Wilson, S. P., Verschure, P. F., Mura, A. & Prescott, T. J.) Biomimetic and Biohybrid Systems, Vol. 9222, 199–210 (Springer International Publishing, Cham, 2015).
+
+[20] Insect-Inspired Visual Navigation On-Board an Autonomous Robot: Real-World Routes Encoded in a Single Layer Network, Vol. ALIFE 2019: The 2019 Conference on Artificial Life of Artificial Life
+[21] Gattaux, G., Vimbert, R., Wystrach, A., Serres, J. R. & Ruffier, F. Antcar: Simple route following task with ants-inspired vision and neural model. Preprint at https://hal.science/hal-04060451v1 (2023).
+
+[22] Lu, Y., Cen, J., Maroun, R. A. & Webb, B. Embodied visual route following by an insect-inspired robot. Preprint at https://www.researchsquare.com/article/rs-4222706/latest (2024).
+
+[23] Jesusanmi, O. O. et al. Investigating visual navigation using spiking neural network models of the insect mushroom bodies. Frontiers in Physiology 15, 1379977 (2024).
+
+[24] Caron, G., Marchand, E. & Mouaddib, E. M. Photometric visual servoing for omnidirectional cameras. Autonomous Robots 35, 177–193 (2013).
+
+[25] Cartwright, B. A. & Collett, T. S. Landmark learning in bees: Experiments and models. Journal of Comparative Physiology ? A 151, 521–543 (1983).
+
+[26] Möller, R. & Vardy, A. Local visual homing by matched-filter descent in image distances. Biological cybernetics 95, 413–430 (2006).
+
+[27] Zeil, J., Hofmann, M. I. & Chahl, J. S. Catchment areas of panoramic snapshots in outdoor scenes. Journal of the Optical Society of America A 20, 450 (2003).
+
+[28] Wystrach, A., Cheng, K., Sosa, S. & Beugnon, G. Geometry, features, and panoramic views: ants in rectangular arenas. Journal of Experimental Psychology: Animal Behavior Processes 37, 420 (2011).
+
+[29] Gaffin, D. D. & Brayfield, B. P. Autonomous Visual Navigation of an Indoor Environment Using a Parsimonious, Insect Inspired Familiarity Algorithm. PLOS ONE 11, e0153706 (2016).
+
+[30] Philippides, A., Baddeley, B., Cheng, K. & Graham, P. How might ants use panoramic views for route navigation? Journal of Experimental Biology 214, 445–451 (2011).
+
+[31] Baddeley, B., Graham, P., Husbands, P. & Philippides, A. A Model of Ant Route Navigation Driven by Scene Familiarity. PLoS Computational Biology 8, e1002336 (2012).
+
+[32] Wystrach, A., Beugnon, G. & Cheng, K. Ants might use different view-matching strategies on and off the route. Journal of Experimental Biology 215, 44–55 (2012).
+
+[33] Linsker, R. Self-organization in a perceptual network. Computer 21, 105–117 (1988).
+
+[34] Wystrach, A., Mangan, M., Philippides, A. & Graham, P. Snapshots in ants? new interpretations of paradigmatic experiments. Journal of Experimental Biology 216, 1766–1770 (2013).
+
+[35] Heisenberg, M. Mushroom body memoir: From maps to models. Nature Reviews Neuroscience 4, 266–275 (2003).
+
+[36] Eichler, K. et al. The complete connectome of a learning and memory centre in an insect brain. Nature 548, 175–182 (2017).
+
+[37] Aso, Y. et al. Mushroom body output neurons encode valence and guide memory-based action selection in Drosophila. eLife 3, e04580 (2014).
+
+[38] Ardin, P., Peng, F., Mangan, M., Lagogiannis, K. & Webb, B. Using an Insect Mushroom Body Circuit to Encode Route Memory in Complex Natural Environments. PLOS Computational Biology 12, e1004683 (2016).
+
+[39] Haalck, L. et al. CATER: Combined Animal Tracking & Environment Reconstruction. SCIENCE ADVANCES (2023).
+
+[40] Deeti, S., Cheng, K., Graham, P. & Wystrach, A. Scanning behaviour in ants: An interplay between random-rate processes and oscillators. Journal of Comparative Physiology A (2023).
+
+[41] Wystrach, A., Philippides, A., Aurejac, A., Cheng, K. & Graham, P. Visual scanning behaviours and their role in the navigation of the australian desert ant melophorus bagoti. Journal of Comparative Physiology A 200, 615–626 (2014).
+
+[42] Clement, L., Schwarz, S. & Wystrach, A. An intrinsic oscillator underlies visual navigation in ants. Current Biology 33, 411–422 (2023).
+
+[43] Wystrach, A., Le Moël, F., Clement, L. & Schwarz, S. A lateralised design for the interaction of visual memories and heading representations in navigating ants. Preprint at https://www.biorxiv.org/content/10.1101/2020.08.13.249193v1 (2020).
+
+[44] Wystrach, A. Neurons from pre-motor areas to the Mushroom bodies can orchestrate latent visual learning in navigating insects. Preprint at https://www.biorxiv.org/content/10.1101/2023.03.09.531867v1 (2023).
+
+[45] Stürzl, W., Zeil, J., Boeddeker, N. & Hemmi, J. M. How Wasps Acquire and Use Views for Homing. Current Biology 26, 470–482 (2016).
+
+[46] Steinbeck, F. et al. Familiarity-taxis: A bilateral approach to view-based snapshot navigation. Adaptive Behavior 10597123231221312 (2024).
+
+[47] Schwarz, S., Clement, L., Haalck, L., Risse, B. & Wystrach, A. Compensation to visual impairments
+and behavioral plasticity in navigating ants. Proceedings of the National Academy of Sciences 121, e2410908121 (2024).
+
+[48] Ardin, P. B., Mangan, M. & Webb, B. Ant Homing Ability Is Not Diminished When Traveling Backwards. Frontiers in Behavioral Neuroscience 10 (2016).
+
+[49] Pfeffer, S. E. & Wittlinger, M. How to find home backwards? Navigation during rearward homing of Cataglyphis fortis desert ants. Journal of Experimental Biology 219, 2119–2126 (2016).
+
+[50] Schwarz, S., Clement, L., Gkanias, E. & Wystrach, A. How do backward-walking ants (Cataglyphis velox) cope with navigational uncertainty? Animal Behaviour 164, 133–142 (2020).
+
+[51] Freas, C. A. & Spetch, M. L. Terrestrial cue learning and retention during the outbound and inbound foraging trip in the desert ant, cataglyphis velox. Journal of Comparative Physiology A 205, 177–189 (2019).
+
+[52] Frenet, F. Sur les courbes à double courbure. Journal de mathématiques pures et appliquées 17, 437–447 (1852).
+
+[53] Clement, L., Schwarz, S. & Wystrach, A. Latent learning without map-like representation of space in navigating ants. Preprint at https://www.biorxiv.org/content/10.1101/2024.08.29.610243v1 (2024).
+
+[54] Möller, R. & Vardy, A. Local visual homing by matched-filter descent in image distances. Biological Cybernetics 95, 413–430 (2006).
+
+[55] Lyapunov, A. M. The general problem of the stability of motion. International journal of control 55, 531–534 (1992).
+
+[56] Webb, B. The internal maps of insects. Journal of Experimental Biology 222, jeb188094 (2019).
+
+[57] Aso, Y. et al. Mushroom body output neurons encode valence and guide memory-based action selection in drosophila. elife 3, e04580 (2014).
+
+[58] Schoepe, T. et al. Finding the gap: Neuromorphic motion-vision in dense environments. Nature Communications 15, 817 (2024).
+
+[59] Collett, T. S. & Collett, M. Route-segment odometry and its interactions with global path-integration. Journal of Comparative Physiology A 201, 617–630 (2015).
+
+[60] Stone, T. et al. An anatomically constrained model for path integration in the bee brain. Current Biology 27, 3069–3085 (2017).
+
+[61] Webb, B. Beyond prediction error: 25 years of modeling the associations formed in the insect mushroom body. Learning & Memory 31, a053824 (2024).
+
+[62] Sommer, S., Von Beeren, C. & Wehner, R. Multiroute memories in desert ants. Proceedings of the National Academy of Sciences 105, 317–322 (2008).
+
+[63] Aksoy, V. & Camlitepe, Y. Spectral sensitivities of ants—a review. Animal Biology 68, 55–73 (2018).
+
+[64] Wystrach, A., Dewar, A., Philippides, A. & Graham, P. How do field of view and resolution affect the information content of panoramic scenes for visual navigation? A computational investigation. Journal of Comparative Physiology A 202, 87–95 (2016).
+
+[65] Le Moël, F. & Wystrach, A. Vision is not olfaction: impact on the insect mushroom bodies connectivity. Preprint at https://www.biorxiv.org/content/10.1101/2024.08.31.610627v1 (2024).
+
+[66] Applonie, R. & Jin, Y.-F. A novel steering control for car-like robots based on lyapunov stability. 2019 American Control Conference (ACC) 2396–2401 (2019).
+
+[67] Kruskal, W. H. & Wallis, W. A. Use of ranks in one-criterion variance analysis. Journal of the American Statistical Association 47, 583–621 (1952).
+
+[68] Virtanen, P. et al. Scipy 1.0: fundamental algorithms for scientific computing in python. Nature methods 17, 261–272 (2020).
+Supplementary Files
+
+This is a list of supplementary files associated with this preprint. Click to download.
+
+• ContinuousvisualnavigationSupplementaryInformation.pdf
+• ContinuousvisualroutefollowingVF.mp4

089483f12365d3a631f34ad6f3c9ff251e698c756a3d693ffcf84f0664f35fd7/peer_review/peer_review.md

@@ -0,0 +1,327 @@
+Peer Review File
+
+Heteroaromatic Swapping in Aromatic Ketones
+
+Corresponding Author: Professor Junichiro Yamaguchi
+
+This manuscript has been previously reviewed at another journal. This document only contains information relating to versions considered at Nature Communications.
+
+This file contains all reviewer reports in order by version, followed by all author rebuttals in order by version.
+
+Version 0:
+
+Reviewer comments:
+
+Reviewer #3
+
+(Remarks to the Author)
+Aromatic ketones are fundamental structures widely existing in natural products and pharmaceuticals. Therefore, a concise and one-step method to directly convert other functional groups into hetero-aromatic ketones would undoubtedly improve efficiencies to prepare novel compounds in organic synthesis and medicinal chemistry. In this manuscript, Yamaguchi and his/her co-workers developed a hetero-aromatic swapping methodology from aromatic ketones by employing the classical Claisen/retro-Claisen reaction, which provided a potential late-stage functionalization protocol of bioactive molecules. However, its characteristic of thermodynamic control dramatically limited the further applications of this reaction. Hence, this reviewer strongly agrees reviewer 1’s comments on this manuscript about utility, novelty, and generalizable insight. But the detailed mechanism studies were interesting and informative for further reaction designs. Overall, present manuscript at this stage is not sufficient to be published on highly impactful Nat. Commun., unless the following additional concerns and suggestions from this reviewer were taken into considerations:
+
+1. For ketone synthesis, N-methoxy-N-methylamide (Weinreb amide) is common and useful. There is no significant improvement of author’s method considering that sodium hydride should be used here while lithium reagents are utilized in Weinreb ketone synthesis.
+
+2. For the mechanism of this swapping, the diketone intermediate 5 is vital to connect starting materials and products. So it’s quite important to prove its involvement. From the kinetic data and DFT calculations, it’s possible to trap it on NMR or MS at a lower temperature.
+
+3. To gain more insights on the mechanism, isotope labeling of oxygen or carbon in one ketone (such as 1a) might be conducted to confirm the proposed swapping.
+
+4. In the DFT calculations, please drawing all related chemical structures of intermediates and transition states in the energy profiles. It’s hard for readers to understand the mechanism without most structure information.
+
+5. The solvent THF is the potential ligand to sodium cation. This possibility should be also calculated. In addition, enolate could be tetramer, as can be seen from this study and related papers: Chem. 2023, 9, 1477-1494
+6. As the reaction is controlled by thermodynamics. Authors could propose some factors on aromatic groups to predict the reaction results, which might be more straightforward for chemists to apply this swapping into their synthesis. Electrophilicity and nucleophilicity employed in the Supporting information are one possible choice.
+
+7. For the discussion of electronegativity/nucleophilicity, now it is more convenient to use Mayr’s values and Yu’s FMO understanding of nucleophilicity=HMO, electrocity=LUMO (Asian J. Org. Chem. 2012, 1, 336 – 345.). So use LUMO energies to discuss this issue in the paper
+
+Version 1:
+
+Reviewer comments:
+
+Reviewer #3
+
+(Remarks to the Author)
+For the DFT calculation part, I advise to show the structures of all intermediates and transition states as structural formulas rather than ball-and-stick models, since ball-and-stick models are difficult for readers to understand.
+Response to Reviewer #3
+
+1. For ketone synthesis, N-methoxy-N-methylamide (Weinreb amide) is common and useful. There is no significant improvement of author’s method considering that sodium hydride should be used here while lithium reagents are utilized in Weinreb ketone synthesis.
+
+Response: We thank the reviewer for the thoughtful comparison with the Weinreb amide methodology. While Weinreb amides represent a well-established and reliable approach for synthesizing ketones, particularly via nucleophilic addition of organolithium or Grignard reagents, our transformation serves a fundamentally different purpose. Specifically, our method enables a direct *heteroaromatic swapping* of the aryl group in an aromatic ketone—rather than constructing a ketone from simpler components.
+
+To convert an aromatic ketone to a heteroaromatic ketone using the Weinreb strategy, a typical sequence would involve:
+
+(1) Baeyer–Villiger oxidation of the aryl ketone to give an ester,
+(2) hydrolysis to the corresponding acid,
+(3) coupling with N,O-dimethylhydroxylamine to form the Weinreb amide, and
+(4) nucleophilic addition of a heteroaryl organolithium or Grignard reagent.
+
+This multi-step sequence requires cryogenic conditions and strictly moisture-free setups for the organometallic addition.
+
+In contrast, our method achieves the skeletal transformation in a *single operation*, using bench-stable materials under mild conditions, without the need for low-temperature setup or air-sensitive reagents. These features underscore the synthetic practicality and operational simplicity of our approach, especially for late-stage diversification of complex molecules.
+
+To clarify this distinction, we have revised the Introduction to better highlight the conceptual and practical differences between the two methods, and added a schematic comparison in Figure 1 to assist the reader in appreciating the transformation pathway.
+Conventionally, converting an aromatic ketone into a heteroaromatic analogue via the Weinreb amide strategy involves a sequence of transformations: (1) Baeyer–Villiger oxidation of the aryl ketone to generate an ester; (2) hydrolysis to the corresponding carboxylic acid; (3) coupling with N,O-dimethylhydroxylamine to afford the Weinreb amide; and (4) nucleophilic addition of a heteroaryl organolithium or Grignard reagent. This approach requires low-temperature conditions and rigorously anhydrous setups for handling sensitive organometallic reagents. The development of a general and operationally simple one-step method for the direct conversion of aromatic ketones into heteroaromatic ketones at a late stage would markedly streamline the derivatization process and accelerate the discovery of novel molecules in medicinal chemistry.
+
+2. For the mechanism of this swapping, the diketone intermediate 5 is vital to connect starting materials and products. So it’s quite important to prove its involvement. From the kinetic data and DFT calculations, it’s possible to trap it on NMR or MS at a lower temperature.
+
+Response: We agree that the diketone intermediate 5 plays a central role in the proposed mechanism, and its detection would provide strong evidence for the reaction pathway. In the case of methyl picolinate (2A) as the heteroaromatic partner, we were unable to observe the corresponding diketone either by NMR or as an isolated product. However, when methyl isoquinolinate (2H) was used, the diketone intermediate was obtained in 16% yield after work-up. To further investigate, we monitored the reaction using in situ NMR spectroscopy. A set of peaks consistent with the proposed diketone structure was observed, To further investigate, we monitored the reaction using in situ NMR spectroscopy. A set of peaks consistent with the proposed diketone structure was observed, although full spectral agreement with the isolated compound could not be confirmed, likely due to peak shifts caused by the presence of the sodium enolate and overlap with other components in the mixture. Moreover, the intermediate was not clearly detected by mass spectrometry, and thus we consider this evidence tentative at this stage.
+1a: 0.20 mmol    2H (2.0 equiv)    NaH (2.0 equiv)    THF (0.20 M)    60 °C, 6 h    NMR yield
+3aH: 64%    4a: 33%    5aH: 16%
+
+1a: 0.20 mmol    2A-a (1.0 equiv)    NaH (2.0 equiv)    THF (0.20 M)    60 °C, 6 h    NMR yield
+3aA: 6%    4a-a: 4%    5aA: 25%
+
+1a: 0.20 mmol    2A-b (1.0 equiv)    NaH (2.0 equiv)    THF (0.20 M)    60 °C, 6 h    NMR yield
+3aA: 11%    4a-b: 6%    5aA: 39%
+
+1a: 0.10 mmol    2H (1.5 equiv)    NaH (2.0 equiv)    THF-d_6 (0.20 M)    RT    NMR yield
+3aH    4a
+
+Fig. in situ NMR analysis
+
+(with the red arrow indicating the sodium enolate form of the diketone)
+
+Nevertheless, given that the diketone-derived product is formed under these conditions, and in light of the \(^{18}\)O-labeling results (which confirm the origin of the carbonyl oxygen
+from the heteroaromatic ester), we believe the involvement of the diketone intermediate is well supported. These observations are discussed in the manuscript, and we hope they address the reviewer’s request for mechanistic clarification.
+
+3. To gain more insights on the mechanism, isotope labeling of oxygen or carbon in one ketone (such as 1a) might be conducted to confirm the proposed swapping.
+
+Response: We thank the reviewer for this excellent suggestion. In response, we conducted isotope labeling experiments using \(^{18}\mathrm{O}\)-labeled methyl picolinate (\(2\mathbf{A}\cdot^{18}\mathrm{O}\)) as the heteroaromatic component. After reaction with \(\mathbf{1a}\), the mixture was initially quenched with NH\(_4\)Cl solution and the products were isolated and analyzed by mass spectrometry to assess the incorporation of the \(^{18}\mathrm{O}\) label in the resulting ketones (\(3\mathbf{aA}\) and \(4\mathbf{a}\)). While the presence of \(^{18}\mathrm{O}\)-labeled \(4\mathbf{a}\) was clearly confirmed, \(3\mathbf{aA}\) did not show significant \(^{18}\mathrm{O}\) incorporation under these conditions. We attributed this to possible exchange of the ketone oxygen in \(3\mathbf{aA}\) with water due to its electron-deficient nature, rendering the carbonyl susceptible to hydration and subsequent exchange.
+
+To address this issue, we modified the workup procedure by avoiding acidic quenching and instead directly acetylated the enolate intermediate using acetic anhydride after completion of the reaction, thereby isolating the acetylated derivative (\(3\mathbf{aA}\)-Ac). Under these conditions, both \(3\mathbf{aA}\)-Ac and \(4\mathbf{a}\) were found to retain the \(^{18}\mathrm{O}\) label, as confirmed by mass spectrometric analysis. These results strongly support the incorporation of the oxygen atom from \(2\mathbf{A}\) into the carbonyl group of \(3\mathbf{aA}\), providing compelling evidence in favor of the proposed Claisen/retro-Claisen-type mechanism.
+
+This experiment and its outcome are now described in the manuscript (Fig. 3d), and detailed experimental procedures are included in the Supplementary Information.
+
+![Reaction schemes showing isotope labeling experiments with methyl picolinate and ketones](page_158_1047_1136_180.png)
+
+To confirm the origin of the carbonyl oxygen in the swapped product, we performed isotope-labeling experiments using \(^{18}\mathrm{O}\)-labeled methyl picolinate (\(2\mathbf{A}\cdot^{18}\mathrm{O}\), Fig. 3d). After reacting with \(\mathbf{1a}\), the mixture was initially quenched with NH\(_4\)Cl, and the resulting products were analyzed by mass spectrometry. While \(^{18}\mathrm{O}\) incorporation was clearly observed in \(4\mathbf{a}\), little to no labeling was detected in \(3\mathbf{aA}\),
+likely due to exchange with water under acidic conditions, facilitated by the electron-deficient nature of the ketone.
+
+To prevent this exchange, the reaction was instead quenched by direct acetylation of the enolate intermediate with acetic anhydride. Under these modified conditions, both 3aA-Ac and 4a retained the \(^{18}\)O label, as confirmed by mass spectrometry. These results clearly indicate that the carbonyl oxygen in 3aA originates from 2A, providing strong support for the proposed Claisen/retro-Claisen-type mechanism.
+
+4. In the DFT calculations, please drawing all related chemical structures of intermediates and transition states in the energy profiles. It’s hard for readers to understand the mechanism without most structure information.
+
+Response: We appreciate this suggestion. We have revised the energy diagrams in the Supplementary Information (Figures S8–S12) to include chemical structures for all relevant intermediates and transition states. This greatly enhances the clarity of the mechanistic rationale.
+
+Enolization of 5wA with (MeONa),
+
+![Energy diagram showing the enolization of 5wA with (MeONa), including chemical structures of intermediates and transition states, and Gibbs free energies.](page_370_693_1042_377.png)
+
+Figure S8. Enolization of 5wA with (MeONa), at the DLPNO-CCSD(T)/def2-TZVPP(D)//B3LYP-D3(BJ)/6-311+G(d,p) level of theory.
+
+Formation of 1_{enol} + 2 and 3_{enol} + 4 from 5wA with (MeONa)_i
+Black: Formation of \( \mathbf{1}_{\text{enol}} + 2 \)
+Red: Formation of \( \mathbf{3}_{\text{enol}} + 4 \)
+
+![Reaction energy diagram showing Gibbs free energy vs. reaction progress for the formation of 1_enol + 2 and 3_enol + 4 from 5wA with (MeONa), at the DLPNO-CCSD(T)/def2-TZVPP(D)//B3LYP-D3(BJ)/6-311+G(d,p) level of theory.](page_180_180_1087_799.png)
+
+Figure S9. Formation of \( \mathbf{1}_{\text{enol}} + 2 \) and \( \mathbf{3}_{\text{enol}} + 4 \) from \( \mathbf{5wA} \) with (MeONa); at the DLPNO-CCSD(T)/def2-TZVPP(D)//B3LYP-D3(BJ)/6-311+G(d,p) level of theory.
+Enolization of 5wA with (MeONa)₄
+
+![Enolization energy profile diagram showing Gibbs free energy vs. reaction coordinate for 5wA with (MeONa)₄, including intermediates Int1_enol, Int2_enol, TS_enol, and products 5wA_enol + MeOH.](page_186_184_1077_693.png)
+
+Figure S10. Enolization of 5wA with (MeONa)₄ at the DLPNO-CCSD(T)/def2-TZVPP(D)//B3LYP-D3(BJ)/6-311+G(d,p) level of theory. The suffix that specifies the MeONa cluster (e.g., “-(MeONa)₄”) is omitted for brevity.
+
+Formation of 1_enol + 2 and 3_enol + 4 from 5wA with (MeONa)₄
+Black: Formation of \(1_{\text{enol}} + 2\)
+Red: Formation of \(3_{\text{enol}} + 4\)
+
+![Reaction energy profile diagram showing Gibbs free energy (kcal mol⁻¹) versus reaction coordinate for the formation of 1_{enol} + 2 and 3_{enol} + 4 from 5wA with (MeONa)_4, including transition states TS1_{1-2}, TS1_{1-4}, TS2_{2-4}, TS2_{3-4}, Int1_{1-2}, Int1_{1-4}, Int2_{1-2}, Int2_{1-4}, Int3_{3-4}, and Int3_{3-2}.](page_186_186_1077_799.png)
+
+Figure S11. Formation of \(1_{\text{enol}} + 2\) and \(3_{\text{enol}} + 4\) from \(5wA\) with \((\text{MeONa})_4\) at the DLPNO-CCSD(T)/def2-TZVPP(D)/B3LYP-D3(BJ)/6-311+G(d,p) level of theory. The suffix that specifies the MeONa cluster (e.g., “-(MeONa)_4”) is omitted for brevity.
+
+Formation of \(1 + 2_{\text{ONa}}\) and \(3 + 4_{\text{ONa}}\) from \(5wA\) and NaOH
+Figure S12. Formation of \( \mathbf{1} + 2_{\mathrm{ONa}} \) and \( \mathbf{3} + 4_{\mathrm{ONa}} \) from \( 5wA \) with NaOH at the DLPNO-CCSD(T)/def2-TZVPP(D)//B3LYP-D3(BJ)/6-311+G(d,p) level of theory.
+
+5. The solvent THF is the potential ligand to sodium cation. This possibility should be also calculated. In addition, enolate could be tetramer, as can be seen from this study and related papers: Chem. 2023, 9, 1477–1494
+
+Response: We appreciate the helpful suggestion. While adding explicit THF molecules stabilizes the system by 4.0 kcal mol\(^{-1}\), the overall energy profile is qualitatively the same as the original energy profile without explicit solvent molecules. In particular, the most important conclusion from the computational study remains the same: formation of enolate of **1** and **2** is kinetically favorable, whereas that of enolate of **3** and **4** is thermodynamically favorable. The energy profile is summarized in Section 8-5 (Supplementary Information).
+
+**8-5. Comparison between the reaction profile with and without explicit THF molecules**
+
+To discuss the solvent coordination to Na\(^+\), two explicit THF molecules have been added. Our preliminary calculations showed that adding two THF molecules adequately captures the stabilization energy at a minimal computational cost. Adding explicit THF molecules stabilized the
+system by 4.0 kcal mol\(^{-1}\); the overall profile, however, remained similar to that without explicit THF molecules: formation of enolate of **1** and **2** is kinetically favorable, whereas that of enolate of **3** and **4** is thermodynamically favorable.
+
+<table>
+  <tr>
+    <th rowspan="2">Enolization of</th>
+    <th colspan="2">Without THF</th>
+    <th colspan="2">With THF</th>
+  </tr>
+  <tr>
+    <th>5wA + MeONa</th>
+    <th>Int1<sub>enol</sub></th>
+    <th>TS<sub>enol</sub></th>
+    <th>Int2<sub>enol</sub></th>
+    <th>5wA + MeOH</th>
+  </tr>
+  <tr>
+    <td rowspan="6">Formation of enolate of **1** and **2**</td>
+    <td>5wA + MeONa</td>
+    <td>0.0</td>
+    <td>0.9</td>
+    <td>3.3</td>
+    <td>-19.7</td>
+    <td>-21.3</td>
+    <td>-4.0</td>
+    <td>1.6</td>
+    <td>1.4</td>
+    <td>-20.4</td>
+    <td>-24.3</td>
+  </tr>
+  <tr>
+    <td>Int1<sub>1-2</sub></td>
+    <td>-2.9</td>
+    <td>-2.9</td>
+    <td>-3.3</td>
+    <td>-9.0</td>
+    <td>1.2</td>
+    <td>-11.4</td>
+    <td>-10.5</td>
+    <td>-3.6</td>
+    <td>-14.7</td>
+    <td>-14.9</td>
+  </tr>
+  <tr>
+    <td>TS1<sub>1-2</sub></td>
+    <td>-3.2</td>
+    <td>-3.2</td>
+    <td>-3.2</td>
+    <td>-11.2</td>
+    <td>-3.6</td>
+    <td>-14.7</td>
+    <td>-14.9</td>
+  </tr>
+  <tr>
+    <td>Int2<sub>1-2</sub></td>
+    <td>-11.2</td>
+    <td>-11.2</td>
+    <td>-11.2</td>
+    <td>-14.7</td>
+    <td>-14.9</td>
+  </tr>
+  <tr>
+    <td>TS2<sub>1-2</sub></td>
+    <td>-3.6</td>
+    <td>-3.6</td>
+    <td>-3.6</td>
+    <td>-14.7</td>
+    <td>-14.9</td>
+  </tr>
+  <tr>
+    <td>Int3<sub>1-2</sub></td>
+    <td>-14.7</td>
+    <td>-14.7</td>
+    <td>-14.7</td>
+    <td>-14.9</td>
+    <td>-14.9</td>
+  </tr>
+  <tr>
+    <td rowspan="6">Formation of enolate of **3** and **4**</td>
+    <td>5wA + MeONa</td>
+    <td>0.0</td>
+    <td>-3.0</td>
+    <td>-2.5</td>
+    <td>-8.9</td>
+    <td>3.0</td>
+    <td>-14.9</td>
+    <td>-12.4</td>
+    <td>-4.0</td>
+    <td>-2.5</td>
+    <td>-4.5</td>
+    <td>-12.2</td>
+    <td>-0.1</td>
+    <td>-19.1</td>
+    <td>-18.4</td>
+  </tr>
+  <tr>
+    <td>Int1<sub>3-4</sub></td>
+    <td>-3.0</td>
+    <td>-3.0</td>
+    <td>-2.5</td>
+    <td>-8.9</td>
+    <td>3.0</td>
+    <td>-14.9</td>
+    <td>-12.4</td>
+    <td>-2.5</td>
+    <td>-4.5</td>
+    <td>-12.2</td>
+    <td>-0.1</td>
+    <td>-19.1</td>
+    <td>-18.4</td>
+  </tr>
+  <tr>
+    <td>TS1<sub>3-4</sub></td>
+    <td>-4.5</td>
+    <td>-4.5</td>
+    <td>-4.5</td>
+    <td>-12.2</td>
+    <td>-0.1</td>
+    <td>-19.1</td>
+    <td>-18.4</td>
+  </tr>
+  <tr>
+    <td>Int2<sub>3-4</sub></td>
+    <td>-12.2</td>
+    <td>-12.2</td>
+    <td>-12.2</td>
+    <td>-18.4</td>
+  </tr>
+  <tr>
+    <td>TS2<sub>3-4</sub></td>
+    <td>-0.1</td>
+    <td>-0.1</td>
+    <td>-0.1</td>
+    <td>-18.4</td>
+  </tr>
+  <tr>
+    <td>Int3<sub>3-4</sub></td>
+    <td>-19.1</td>
+    <td>-19.1</td>
+    <td>-19.1</td>
+    <td>-18.4</td>
+  </tr>
+  <tr>
+    <td>3<sub>enol</sub> + 4</td>
+    <td>-18.4</td>
+    <td>-18.4</td>
+    <td>-18.4</td>
+  </tr>
+</table>
+
+Table S8. Relative Gibbs energies (in kcal mol\(^{-1}\)) without and with explicit THF molecules at the DLPNO-CCSD(T)/def2-TZVPP//B3LYP-D3(BJ)/6-311+G(d,p) level of theory
+
+We believe that a significant part of the solvent effect have been captured by the implicit solvent model, and therefore, the discussion using the profiles without explicit solvent molecules is, at least, qualitatively correct. We will briefly discuss the consideration of explicit solvent molecules in the main text:
+
+These findings strongly suggest that the reaction proceeds under thermodynamic control. These analyses have been conducted without explicit solvent (THF) molecules. Although it is possible that THF coordinates to the sodium cation, the overall profile is similar to that without explicit molecules (see Supplementary Table S8).
+It appears that the suggested reference refers to the study already cited as Ref. 43 (45) in the original (revised) manuscript: Luo, F. et al. Direct insertion into the C–C bond of unactivated ketones with
+
+NaH-mediated aryne chemistry. Chem 9, 2620–2636 (2023). We haven't made the changes in line with the suggestion.
+
+6. As the reaction is controlled by thermodynamics. Authors could propose some factors on aromatic groups to predict the reaction results, which might be more straightforward for chemists to apply this swapping into their synthesis. Electrophilicity and nucleophilicity employed in the Supporting information are one possible choice.
+
+Response: Regarding the prediction of reaction outcomes based on the electronic nature of the aromatic substituents, we have already conducted a Hammett analysis, which revealed a positive \( \varphi \) value (\( \varphi > 0 \)), indicating that electron-deficient aromatic rings promote the transformation. Although the reaction proceeds to some extent with electron-donating groups, near-complete conversion to the product side is generally observed only in cases where strongly electron-withdrawing substituents, such as a para-cyano group, are present. These results suggest that the reaction equilibrium is governed by the electronic properties of the aryl group, and we believe this provides a reliable basis for anticipating reaction outcomes in related substrates.
+
+7. For the discussion of electronegativity/nucleophilicity, now it is more convenient to use Mayr’s values and Yu’s FMO understanding of nucleophilicity = HOMO, electrophilicity = LUMO (Asian J. Org. Chem. 2012, 1, 336–345). So use LUMO energies to discuss this issue in the paper.
+
+Response: We appreciate the suggestion of the interesting and easily-accessible electrophilicity and nucleophilicity indices. We have analyzed the so-called orbital energy of frontier molecular orbitals and found that the suggested indices reproduce the trend of the indices calculated in the original manuscript as well as experiment (see Supplementary Table S5 and Section 8-2 for details). The analysis using the suggested index is discussed in the main text:
+
+…, rendering enolate 3 a weaker nucleophile (more details, see Supplementary Table S5).40,41 **We also found that Mayr’s nucleophilicity and electrophilicity indices**42 **evaluated through the fragment molecular orbital approach**43 **reproduce the same trend.** As a result,…
+with two additional references:
+
+(42) Streidl, N., Denegri, B., Kronia, O., Mayr, H. A Practical Guide for Estimating Rates of Heterolysis Reactions. Acc. Chem. Res. **43**, 1537–1549 (2010).
+
+(43) Zhuo, L.-G., Liao, W., Yu, Z.-X. A Frontier Molecular Orbital Theory Approach to Understanding the Mayr Equation and to Quantifying Nucleophilicity and Electrophilicity by Using HOMO and LUMO Energies. Asian J. Org. Chem. **1**, 336–345 (2012).
+
+We hope that the revised manuscript, additional experimental and computational studies, and point-by-point responses satisfactorily address the concerns raised. We believe that our findings provide a robust and mechanistically informed platform for aromatic-to-heteroaromatic skeletal editing, and we respectfully resubmit our manuscript for further consideration.
+
+Sincerely yours,
+
+[Contact Information]
+
+Prof. Dr. Junichiro Yamaguchi
+
+Department of Applied Chemistry, Waseda University, Wasedatsurumakicho 513, Shinjuku, Tokyo 162-0041, Japan.
+
+Phone : +81-3-5286-3225    E-mail: junyamaguchi@waseda.jp
+Response to Reviewer #3
+
+1. For the DFT calculation part, I advise to show the structures of all intermediates and transition states as structural formulas rather than ball-and-stick models, since ball-and-stick models are difficult for readers to understand.
+
+Response: We appreciate the reviewer’s suggestion regarding the presentation of the DFT-calculated intermediates and transition states. In the revised manuscript, we have replaced the ball-and-stick models with clear structural formulas for all intermediates and transition states in the Supplementary Information (Figures S9–S13).

08a17e35a9f30287f46ac8334be1320dddc37a6065b4ba95076e9f30f36e016a/peer_review/peer_review.md

@@ -0,0 +1,191 @@
+MS CETSA deep functional proteomics uncovers novel DNA repair programs leading to Gemcitabine resistance
+
+Corresponding Author: Dr Pär Nordlund
+
+This file contains all reviewer reports in order by version, followed by all author rebuttals in order by version.
+
+Version 0:
+
+Reviewer comments:
+
+Reviewer #1
+
+(Remarks to the Author)
+Innate and acquired drug resistance is a major challenge in cancer therapy, driven by genetic or epigenetic changes that rewire key cellular pathways like apoptosis or DNA repair. While genomics and transcriptomics have previously been used to identify resistance mechanisms, these methods are often insufficient to provide a systems-level understanding. In this manuscript, Liang et al. take an innovative approach by using MS-CETSA to investigate drug resistance mechanisms in both cancer cell lines and patient samples. The authors previously developed MS-CETSA, which measures drug-protein and protein-protein interactions over time, serving as a proxy for pathway modulation and cell state changes.
+
+Here, they focus on the cellular response to gemcitabine, a widely used cancer drug, in resistant and sensitive diffuse large B cell lymphoma cell lines. Using IMPRINTS-CETSA, they show that the initial response (3 hrs) to gemcitabine is similar in both resistant and sensitive cell lines, with RRM1 stabilisation, consistent with gemcitabine’s role as an RNR inhibitor, and RPA complex stabilisation at replication forks, leading to ATR-CHK1 activation. At later time points (5-8 hrs), the responses diverge: sensitive cells activate apoptosis, while resistant cells activate cell cycle progression (CDKs) and TLS bypass of gemcitabine lesions. They also identify a novel DNA repair protein ensemble, termed Auxiliary DNA Damage Repair (ADDR), which promotes drug resistance to gemcitabine and other DNA-damaging agents.
+
+Given the gemcitabine-induced ATR activation, the authors demonstrate that inhibiting ATR in resistant cells destabilises ADDR and TLS factors, resulting in drug hypersensitivity and overcoming resistance. This aligns with ongoing clinical trials of ATR inhibitors combined with gemcitabine, providing insight into the synthetic lethality observed in these treatments. Finally, the authors briefly show that their CETSA method can be used to analyse drug responses in patient samples, identifying apoptotic gene activation linked to drug sensitivity.
+
+This is a well-written and original manuscript. The use of IMPRINTS-CETSA to explore drug resistance mechanisms in cancer is promising and could inform decisions on personalised treatments for resistant cancers. However, several points need to be addressed before publication.
+
+Major Points
+1. Although the manuscript aims to use MS-CETSA to identify therapy resistance in patients, it only analyses samples from two gemcitabine-naive patients. The authors should expand their study to include drug-resistant cancer samples to better demonstrate the ability to identify resistance mechanisms.
+
+2. Given that resistance in cancer is often heterogeneous, the authors need to explain how MS-CETSA can be interpreted when analysing complex cell populations from patients with multiple resistance mechanisms.
+
+3. MS-CETSA may overlook resistance mechanisms caused by mutations or downregulation in genes involved in nucleoside analogue activation (e.g., nucleoside transporters or kinases), which could prevent downstream drug engagement. A combination of MS-CETSA with genomics and transcriptomics would provide a more comprehensive view of resistance mechanisms. The authors should comment on this limitation.
+
+4. Why is RRM1/2 engagement/stabilisation only observed in resistant cells at 8 hours (Fig 3B)?
+Minor Points
+1. The methodology is complex, particularly for non-experts. A figure illustrating the key concepts of CETSA and its integration with protein interaction networks would be helpful (e.g., modifying Fig 1A to better explain the CETSA concept such as protein stabilisation/destabilisation).
+
+2. The manuscript uses too many acronyms and initialisms, making it harder to follow. The authors should consider simplifying them (e.g., IMPRINTS-CETSA, MS-CETSA, ITDR-CETSA, IMPRINTS, PRINTS, CCAE, ADDR, RESP). For instance, the authors have described “Protein Interaction States” in previous papers without using the acronym “PRINTS.”
+
+Reviewer #2
+
+(Remarks to the Author)
+In this work, Liang and colleagues use MS-CETSA to profile the response of four diffuse large B-cell lymphoma (DLBCL) cell lines to gemcitabine in a time-resolved manner. This included two cell lines that are drug-sensitive and two that are drug-resistant. At early time points, the authors identify DNA damage response proteins with altered thermal stability in all cell lines in agreement with the appearance of stalled replication forks upon treatment with gemcitabine. At later time points, gemcitabine was shown to induce apoptosis in sensitive cells, while resistant cells progressed in their cell cycle, likely through activation of translesion synthesis polymerases. Inhibition of the translesion synthesis pathway was able to partially re-sensitize cells to gemcitabine, as was inhibition of ATR (a kinase involved in sensing DNA damage). In parallel, the authors identify a group of proteins that they term auxiliary DNA damage repair proteins (ADDR) that might be involved in a more general resistance mechanism to DNA damaging drugs. Overall, this manuscript describes a thorough and impressive characterization of the mechanism of action and resistance of gemcitabine and provides hints to revert resistance. I have no major comments to the methodology used, and have only some suggestions that could further improve the manuscript:
+1. It would be interesting to profile a bit further the cell lines used. Are there any mutations and/or baseline proteome changes between the cell lines that could hint at the effects seen? Particularly, why are these cell lines able to induce ADDR and TLS, while the sensitive cells are not?
+2. Since the manuscript describing the CETSA apoptosis ensemble (CCAE) of proteins is not available, the authors could provide that list of proteins as a supplementary table or pre-print the other manuscript. Can anything be learned from the specific proteins in this CCAE that change (n=24) vs the ones that do not change (n=23) in sensitive cells? Is there a particular apoptosis-inducing drug in the other manuscript that gives this particular signal?
+
+Best regards,
+André Mateus
+
+Reviewer #3
+
+(Remarks to the Author)
+In this study, Liang et al. employed Integrated Modulation of Protein Interaction States-Cellular Thermal Shift Assay (IMPRINT-CETSA), a system-wide functional proteomic approach, to investigate cellular resistance to gemcitabine in diffuse large B-cell lymphoma (DLBCL) cell lines. The authors found that gemcitabine activates similar proteomic responses across all tested DLBCL cell lines. Notably, by 5-8 hours, the responses in sensitive and resistant cell lines start to bifurcate, where sensitive cell lines activate apoptosis, indicated by 24 hits in the core CETSA apoptosis ensemble (CCAE). In contrast, resistant cell lines exhibit a translesion synthesis (TLS) response, represented by an increase in the thermal stability and abundance of TLS-related proteins. Furthermore, the authors identified an auxiliary DNA damage repair (ADDR) response protein ensemble that is specifically activated in resistant cells. The authors proposed that both TLS and ADDR responses promote ATR-dependent drug resistance. Finally, they showed the use of CETSA profiles to predict drug sensitivity in patients
+
+In summary, this study uses time-dependent CETSA to investigate differential drug responses, providing a system-level profiling of gemcitabine-induced proteomic responses with mechanistical insights. However, earlier studies on gemcitabine resistance mechanisms (PMID: 24376779, 29849128, 16603639) and ATR-induced TLS response upon DNA damage (PMID: 31189884, 39215012) limit the novelty of the current work. Moreover, some of the conclusions and claims require additional evidence and explanations. Below are some suggestions to strengthen the manuscript.
+
+Major points
+
+1. The interpretations of thermal stability shifts require more experimental evidence and literature supports. For example, in Figure 2, replication protein A subunits (RPA1, RPA2, and RPA3) exhibit thermal stabilization upon gemcitabine treatment. How can this phenomenon imply ssDNA binding? Furthermore, in Figure 5, the thermal destabilization of Polδ is interpreted as an outcome of its release from replication fork. How can these changes in protein thermal stability be linked to specific biochemical events?
+2. To categorize thermal stability shift, it would be helpful to have a quantitative metric or at least provide a solid workflow for how the classification is made. The illustration in Figure 1C seems to suggest that “temperature-dependency” is taking into account for defining CETSA destabilization or stabilization. If that is the case, why the authors interpret the CETSA profile of CHEK1 in Figure 2 as thermal destabilization while its CETSA profile seems to be complex and could be hard to define. For example, the thermal stability decreases first and then increases as temperature rises for 1 hour, 48 nM gemcitabine treatment. Similarly, how do the authors interpret CETSA profile of DNMT1 in sensitive cells in Figure 2?
+3. In Figure 2, the CETSA profile of DNMT1 protein seems to differ between sensitive and resistant cells. What’s the explanation for this difference?
+4. The results in Figure 2 indicate that in addition to the known gemcitabine target, RRM1, CETSA can detect the activation of ATR/CHEK1 and alteration in DNMT1 as general effects of gemcitabine. It is not clear how the authors reason that the resistance mechanisms occur downstream of these events (lines 181-183).
+
+Minor points:
+
+1. Is there any data for RPA3 treated with 280 nM gemcitabine?
+2. The plot is misaligned for PCLAF in Figure5 (SUDHL4, 24 uM gemcitabine, 8h).
+3. The interpretation of the CESTA profile in Figure 1C needs to be improved and more detailed.
+4. If the authors only want to address the early response (1hr) in Fig. 1D, it may not be necessary to include results from other timepoints.
+5. Figure 2 presentation needs to be improved. The current order of figure layout is different from the order of description in the main text.
+6. The results in Figure 5C are unclear. What are the molecular weights of Polk and its degraded form? The staining is quite weak. How the authors determine which bands correspond to the cleaved Polk?
+7. Throughout the whole article, there isn’t any explanation of the asterisk labels regarding the statistical significance. Also, in Figure 5B, what are the groups being compared in the statistic tests are not mentioned.
+
+Version 1:
+
+Reviewer comments:
+
+Reviewer #1
+
+(Remarks to the Author)
+I am satisfied with the manuscript improvements and replies to my comments and happy to support publication.
+
+Reviewer #2
+
+(Remarks to the Author)
+The authors have addressed my suggestions. I congratulate the authors on the nice work!
+
+Reviewer #3
+
+(Remarks to the Author)
+The authors have addressed most of my questions.
+credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.
+In cases where reviewers are anonymous, credit should be given to 'Anonymous Referee' and the source.
+The images or other third party material in this Peer Review File are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
+To view a copy of this license, visit https://creativecommons.org/licenses/by/4.0/
+Reviewer #1 (Remarks to the Author):
+
+Innate and acquired drug resistance is a major challenge in cancer therapy, driven by genetic or epigenetic changes that rewire key cellular pathways like apoptosis or DNA repair. While genomics and transcriptomics have previously been used to identify resistance mechanisms, these methods are often insufficient to provide a systems-level understanding. In this manuscript, Liang et al. take an innovative approach by using MS-CETSA to investigate drug resistance mechanisms in both cancer cell lines and patient samples. The authors previously developed MS-CETSA, which measures drug-protein and protein-protein interactions over time, serving as a proxy for pathway modulation and cell state changes.
+
+Here, they focus on the cellular response to gemcitabine, a widely used cancer drug, in resistant and sensitive diffuse large B cell lymphoma cell lines. Using IMPRINTS-CETSA, they show that the initial response (3 hrs) to gemcitabine is similar in both resistant and sensitive cell lines, with RRM1 stabilisation, consistent with gemcitabine’s role as an RNR inhibitor, and RPA complex stabilisation at replication forks, leading to ATR-CHK1 activation. At later time points (5-8 hrs), the responses diverge: sensitive cells activate apoptosis, while resistant cells activate cell cycle progression (CDKs) and TLS bypass of gemcitabine lesions. They also identify a novel DNA repair protein ensemble, termed Auxiliary DNA Damage Repair (ADDR), which promotes drug resistance to gemcitabine and other DNA-damaging agents.
+
+Given the gemcitabine-induced ATR activation, the authors demonstrate that inhibiting ATR in resistant cells destabilises ADDR and TLS factors, resulting in drug hypersensitivity and overcoming resistance. This aligns with ongoing clinical trials of ATR inhibitors combined with gemcitabine, providing insight into the synthetic lethality observed in these treatments. Finally, the authors briefly show that their CETSA method can be used to analyse drug responses in patient samples, identifying apoptotic gene activation linked to drug sensitivity.
+
+This is a well-written and original manuscript. The use of IMPRINTS-CETSA to explore drug resistance mechanisms in cancer is promising and could inform decisions on personalised treatments for resistant cancers. However, several points need to be addressed before publication.
+
+Major Points
+1. Although the manuscript aims to use MS-CETSA to identify therapy resistance in patients, it only analyses samples from two gemcitabine-naive patients. The authors should expand their study to include drug-resistant cancer samples to better demonstrate the ability to identify resistance mechanisms.
+
+We agree with the reviewer that adding more clinical data could be very interesting. However, due to the focus of the work on defining the mechanism for gemcitabine resistance in DLBCL, and the large bulk of data already in the manuscript, we feel a more extensive characterization of effects in clinical samples is outside the scope of this first study. Nevertheless, we think that the identified resistance mechanism is very likely relevant for the clinical situation, as the used cell lines are directly patient derived cells (in contrast to many other mechanistic studies with generated cell lines, where the resistance has been evolved in the lab). Also, the fact that we identify the same resistance mechanism in cell lines derived from two different patients, support that this is indeed a program that is in effect in a clinical situation. As a note, if we wanted to include additional gemcitabine DLBCL patient samples that are gemcitabine treated and resistant, 1) these would be rare cases since Gemcitabine is not the first line of treatment and 2) a biopsy is rarely collected in such a case.
+
+2. Given that resistance in cancer is often heterogeneous, the authors need to explain how MS-CETSA can be interpreted when analysing complex cell populations from patients with multiple resistance mechanisms.
+
+The unique strength of the MS-CETSA technology lies in the fact that changes in protein interaction states are measured on a proteome-wide scale, therefore multiple resistance mechanisms can potentially be assessed in the same experiment. In this study we focused on an overlapping program in the two resistant cell lines (which are derived from two different patients), although there might be additional induced and well-defined programs present in each cell line. Therefore, we envisage that studies of further patient derived cell lines by us and others will allow for more induced resistance programs to be identified (e.g. by overlapping responses) to establish a repertoire of available resistance programs for gemcitabine in DLBCL. It remains to be seen if this repertoire of biochemical resistance programs is small (and can be saturated) or large, but our impression from this study and other unpublished studies in our lab is that there are some dominant programs that are seen in multiple cell lines. Based on the repertoire of programs, we envisage that in real heterogenous patient samples, although each program is just present in low to medium stoichiometry (20-50%), it will be possible to identify which programs are dominating in the sample, to potentially guide therapeutic decisions. After such a treatment, new resistance programs might be enriched and should be accessible to detect with CETSA. We have now added a comment on this issue in the manuscript (refer to pg no. in MS).
+
+3. MS-CETSA may overlook resistance mechanisms caused by mutations or downregulation in genes involved in nucleoside analogue activation (e.g., nucleoside transporters or kinases), which could prevent downstream drug
+engagement. A combination of MS-CETSA with genomics and transcriptomics would provide a more comprehensive view of resistance mechanisms. The authors should comment on this limitation.
+
+We have now added an analysis of mutations as well as quantitative proteomics of the cell line (SFig 8C-F) and conclude that the information content on resistance mechanisms in such static omics data is relatively sparse as compared to time dependent CETSA data.
+
+Considering drug target engagement, the comparable isothermal dose response curves in sensitive and resistant cells treated with gemcitabine (Fig 1E) suggest similar drug target engagement between these cells. Additionally, this study includes one dataset that accurately recapitulates the reviewer's described scenario: a key player of the nucleotide metabolism, SAMHD1, is knocked out, thus possibly preventing downstream drug target engagement. In these cells the gemcitabine-induced thermal stabilization of RRM1 was abolished (SFig 5F), highlighting that MS-CETSA can detect such loss of drug target engagement.
+
+4. Why is RRM1/2 engagement/stabilisation only observed in resistant cells at 8 hours (Fig 3B)?
+
+We have now clarified that at 8h only the catalytic subunit of RNR, RRM1 is shifting in both sensitive and resistant cells (grey node) (SFig 8B), which was already observed as early as 1h and in all cell lines (Fig 1D). The radical subunit, RRM2, appeared as hit in response to 8h gemcitabine treatment in resistant cells (SFig 8B, green node) through its level changes, which we described later as part of ADDR ensemble (Fig 5A).
+
+Minor Points
+1. The methodology is complex, particularly for non-experts. A figure illustrating the key concepts of CETSA and its integration with protein interaction networks would be helpful (e.g., modifying Fig 1A to better explain the CETSA concept such as protein stabilisation/destabilisation).
+
+We have now modified Figure 1 to include an illustration of the concept of CETSA thermal stabilization/destabilization and how this translates to the IMPRINTS bar graphs shown and discussed in most of the following figures.
+
+2. The manuscript uses too many acronyms and initialisms, making it harder to follow. The authors should consider simplifying them (e.g., IMPRINTS-CETSA, MS-CETSA, ITDR-CETSA, IMPRINTS, PRINTS, CCAE, ADDR, RESP). For instance, the authors have described "Protein Interaction States" in previous papers without using the acronym "PRINTS."
+
+We have now simplified the text.
+Reviewer #2 (Remarks to the Author):
+
+In this work, Liang and colleagues use MS-CETSA to profile the response of four diffuse large B-cell lymphoma (DLBCL) cell lines to gemcitabine in a time-resolved manner. This included two cell lines that are drug-sensitive and two that are drug-resistant. At early time points, the authors identify DNA damage response proteins with altered thermal stability in all cell lines in agreement with the appearance of stalled replication forks upon treatment with gemcitabine. At later time points, gemcitabine was shown to induce apoptosis in sensitive cells, while resistant cells progressed in their cell cycle, likely through activation of translesion synthesis polymerases. Inhibition of the translesion synthesis pathway was able to partially re-sensitize cells to gemcitabine, as was inhibition of ATR (a kinase involved in sensing DNA damage). In parallel, the authors identify a group of proteins that they term auxiliary DNA damage repair proteins (ADDR) that might be involved in a more general resistance mechanism to DNA damaging drugs. Overall, this manuscript describes a thorough and impressive characterization of the mechanism of action and resistance of gemcitabine and provides hints to revert resistance. I have no major comments to the methodology used, and have only some suggestions that could further improve the manuscript:
+1. It would be interesting to profile a bit further the cell lines used. Are there any mutations and/or baseline proteome changes between the cell lines that could hint at the effects seen? Particularly, why are these cell lines able to induce ADDR and TLS, while the sensitive cells are not?
+
+As mentioned in response to reviewer 1 we have now added and analyzed mutational data of these cell lines as well as added static quantitative proteomics data of the 4 cell lines (baseline proteomics data)(SFig 8C-F). However, the GO analysis of these data do not give any clues to a resistance mechanism.
+
+2. Since the manuscript describing the CETSA apoptosis ensemble (CCAE) of proteins is not available, the authors could provide that list of proteins as a supplementary table or pre-print the other manuscript. Can anything be learned from the specific proteins in this CCAE that change (n=24) vs the ones that do not change (n=23) in sensitive cells? Is there a particular apoptosis-inducing drug in the other manuscript that gives this particular signal?
+
+The manuscript defining the CCAE motif is now out, Ramos et al Cell Reports, 2024;43(10):114784. Some of the proteins that are not part of the overlap is due to them not being well measured or that shifts are slightly below the cutoff. To allow the reader to access this for each protein we have now included the CCAE IMPRINTS profiles comparing the responses at 8h gemcitabine treatment in sensitive and resistant versus the Ramos et al CCAE proteins (SFig 3). There might be smaller effects in the gemcitabine data for chromatin proteins and mitochondrial proteins, but due to the low numbers it is hard to associate this to a distinct mechanistic difference.
+
+Best regards,
+André Mateus
+Reviewer #3 (Remarks to the Author):
+
+In this study, Liang et al. employed Integrated Modulation of Protein Interaction States-Cellular Thermal Shift Assay (IMPRINT-CETSA), a system-wide functional proteomic approach, to investigate cellular resistance to gemcitabine in diffuse large B-cell lymphoma (DLBCL) cell lines. The authors found that gemcitabine activates similar proteomic responses across all tested DLBCL cell lines. Notably, by 5-8 hours, the responses in sensitive and resistant cell lines start to bifurcate, where sensitive cell lines activate apoptosis, indicated by 24 hits in the core CETSA apoptosis ensemble (CCAE). In contrast, resistant cell lines exhibit a translesion synthesis (TLS) response, represented by an increase in the thermal stability and abundance of TLS-related proteins.
+Furthermore, the authors identified an auxiliary DNA damage repair (ADDR) response protein ensemble that is specifically activated in resistant cells. The authors proposed that both TLS and ADDR responses promote ATR-dependent drug resistance. Finally, they showed the use of CETSA profiles to predict drug sensitivity in patients
+
+In summary, this study uses time-dependent CETSA to investigate differential drug responses, providing a system-level profiling of gemcitabine-induced proteomic responses with mechanistical insights. However, earlier studies on gemcitabine resistance mechanisms (PMID: 24376779, 29849128, 16603639) and ATR-induced TLS response upon DNA damage (PMID: 31189884, 39215012) limit the novelty of the current work. Moreover, some of the conclusions and claims require additional evidence and explanations. Below are some suggestions to strengthen the manuscript.
+
+Major points
+
+1. The interpretations of thermal stability shifts require more experimental evidence and literature supports. For example, in Figure 2, replication protein A subunits (RPA1, RPA2, and RPA3) exhibit thermal stabilization upon gemcitabine treatment. How can this phenomenon imply ssDNA binding? Furthermore, in Figure 5, the thermal destabilization of Polδ is interpreted as an outcome of its release from replication fork. How can these changes in protein thermal stability be linked to specific biochemical events?
+
+CETSA shifts strongly support changes of protein interaction states through interactions with e.g. a protein, a nucleic acid or a metabolite, or a PTM change, in the cell at a certain time point. However, the exact molecular origin of the stability effect is challenging to access experimentally and will in many cases remain tentative and be defined by the known interaction made by a protein. As an example, even though we have shown that the destabilisation of CHEK1 is concomitant with phosphorylation, i.e. activation, we don’t know which specific interactions are changed to induce this destabilisation. Could be primarily interactions around the phosphorylated residue but could also be interactions with other proteins, etc. So in essence we relate the shifts to a functional state of a protein (activated Chek1, single strand bound RPA, activated CDK1, Dai Cell. 2018 173:1481, etc), rather than prove the exact molecular background for the change. We feel, for understanding cellular process, the possibility to follow transitions between different functional states of proteins at the proteome-level is of great value, as shown in the present study. There are a few instances where statements around states might not have been clear so we have now reformulated a few sentences in the text in line with this.
+
+2. To categorize thermal stability shift, it would be helpful to have a quantitative metric or at least provide a solid workflow for how the classification is made. The illustration in Figure 1C seems to suggest that “temperature-dependency” is taking into account for defining CETSA destabilization or stabilization. If that is the case, why the authors interpret the CETSA profile of CHEK1 in Figure 2 as thermal destabilization while its CETSA profile seems to be complex and could be hard to define. For example, the thermal stability decreases first and then increases as temperature rises for 1 hour, 48 nM gemcitabine treatment. Similarly, how do the authors interpret CETSA profile of DNMT1 in sensitive cells in Figure 2?
+
+As discussed in response to reviewer 1, we have now modified Figure 1 to include an illustration of the concept of CETSA thermal stabilisation/destabilisation and how this translates to the IMPRINTS bar graphs shown and discussed in most of the following figures.
+
+The proteins in Figure 2 are based on being common hits in resistant and sensitive cells with similar CETSA profiles across the various time points, albeit with different intensities where some (e.g. CHEK1 at 1h) don't pass hit selection criteria).
+
+The CETSA profile for DNMT1 in sensitive cells in Figure 2 with only effect at the higher temperature is a bit different to the resistant cells, but its presence in resistant cells give support for, at least at the late timepoint, there is a very significant shift in DNMT1 in sensitive cells at the highest temperature. DMNT1 shifts are further discussed in the next question below.
+
+3. In Figure 2, the CETSA profile of DNMT1 protein seems to differ between sensitive and resistant cells. What’s the explanation for this difference?
+
+We agree with the reviewer that there appears to be a very significant difference in the profiles in sensitive and resistant cells for DNMT1. This could be interesting and inform on structural/interaction differences of the protein
+in the two types of cells. However, when, as mentioned above, the molecular causes for the shifts are hard to follow up, we selected to not address this difference specifically in the text. Conceptually, there is a possibility that the different profiles reflect a difference in the protein population/states that shifts, or potentially (but less likely, when the profile follows the phenotype), that the intrinsic stability of the DNMT1 protein is different in the different strains, effecting the temperature range in which the protein shifts due to interaction changes.
+
+4. The results in Figure 2 indicate that in addition to the known gemcitabine target, RRM1, CETSA can detect the activation of ATR/CHK1 and alteration in DNMT1 as general effects of gemcitabine. It is not clear how the authors reason that the resistance mechanisms occur downstream of these events (lines 181-183).
+
+We have now added a brief schematic to clarify this (SFig 9) where, in resistant cells the activation ATR subsequently activate a signaling pathway (as stated in Discussion section, details remain to be elucidated) that lead to the hallmarks of the resistance response, the ADDR, TLS and the opening of cell cycle checkpoints. This signaling pathway does not seem to be operative in sensitive cells, or alternatively, apoptosis is happening in parallel and attenuates this pathway.
+
+Minor points:
+1. Is there any data for RPA3 treated with 280 nM gemcitabine?
+The RPA3 data for 280nM gemcitabine condition was incomplete with missing quantitative measurements for some conditions and was therefore removed during our data analysis cleaning step.
+
+2. The plot is misaligned for PCLAF in Figure5 (SUDHL4, 24 uM gemcitabine, 8h).
+The Figure has been adjusted
+
+3. The interpretation of the CESTA profile in Figure 1C needs to be improved and more detailed.
+We have now modified Figure 1 to include an illustration of the concept of CETSA thermal stabilisation/destabilisation and how this translates to the IMPRINTS bar graphs shown and discussed in most of the following figures.
+
+4. If the authors only want to address the early response (1hr) in Fig. 1D, it may not be necessary to include results from other timepoints.
+The additional timepoints are intentionally depicted to demonstrate the consistent CETSA profiles over time.
+
+5. Figure 2 presentation needs to be improved. The current order of figure layout is different from the order of description in the main text.
+We have matched the figure layout with the description in the main text.
+
+6. The results in Figure 5C are unclear. What are the molecular weights of PolK and its degraded form? The staining is quite weak. How the authors determine which bands correspond to the cleaved PolK?
+We have now replaced image with a clearer one and moved this section to supplementary data (SFig 6). The observed molecular weights of the degraded form of PolK were ~65-70kDa. Since their appearance and disappearance correlated with gemcitabine and zVAD-FMK treatment, respectively, we judged these bands as specific for PolK. However, as we did not further confirm and validate these bands as cleaved PolK, we have now removed this statement from the manuscript and instead only focus on the main PolK band where we now included a quantification showing the level decrease and rescue by zVAD-FMK.
+
+7. Throughout the whole article, there isn’t any explanation of the asterisk labels regarding the statistical significance. Also, in Figure 5B, what are the groups being compared in the statistic tests are not mentioned. Figure legends have been improved to include type of statistical tests and compared groups.

09b108cd9ad0034506ae629f2fb23d3bf7f4cd8c58a23feee09b7b89729c5397/metadata.json

@@ -0,0 +1,191 @@
+{
+    "title": "Hydrodynamic manipulation of nano-objects by optically induced thermo-osmotic flows",
+    "pre_title": "Hydrodynamic manipulation of nano-objects by optically induced thermo-osmotic flows",
+    "journal": "Nature Communications",
+    "published": "03 February 2022",
+    "supplementary_0": [
+        {
+            "label": "Supplementary Information",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-022-28212-z/MediaObjects/41467_2022_28212_MOESM1_ESM.pdf"
+        },
+        {
+            "label": "Peer Review File",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-022-28212-z/MediaObjects/41467_2022_28212_MOESM2_ESM.pdf"
+        },
+        {
+            "label": "Description of additional Supplementary File",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-022-28212-z/MediaObjects/41467_2022_28212_MOESM3_ESM.pdf"
+        },
+        {
+            "label": "Supplementary Movie 1",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-022-28212-z/MediaObjects/41467_2022_28212_MOESM4_ESM.mp4"
+        },
+        {
+            "label": "Supplementary Movie 2",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-022-28212-z/MediaObjects/41467_2022_28212_MOESM5_ESM.mp4"
+        },
+        {
+            "label": "Supplementary Movie 3",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-022-28212-z/MediaObjects/41467_2022_28212_MOESM6_ESM.mp4"
+        },
+        {
+            "label": "Supplementary Movie 4",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-022-28212-z/MediaObjects/41467_2022_28212_MOESM7_ESM.mp4"
+        },
+        {
+            "label": "Supplementary Movie 5",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-022-28212-z/MediaObjects/41467_2022_28212_MOESM8_ESM.mp4"
+        },
+        {
+            "label": "Supplementary Movie 6",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-022-28212-z/MediaObjects/41467_2022_28212_MOESM9_ESM.mp4"
+        },
+        {
+            "label": "Supplementary Movie 7",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-022-28212-z/MediaObjects/41467_2022_28212_MOESM10_ESM.mp4"
+        },
+        {
+            "label": "Supplementary Movie 8",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-022-28212-z/MediaObjects/41467_2022_28212_MOESM11_ESM.mp4"
+        },
+        {
+            "label": "Supplementary Movie 9",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-022-28212-z/MediaObjects/41467_2022_28212_MOESM12_ESM.mp4"
+        },
+        {
+            "label": "Supplementary Movie 10",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-022-28212-z/MediaObjects/41467_2022_28212_MOESM13_ESM.mp4"
+        },
+        {
+            "label": "Supplementary Movie 11",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-022-28212-z/MediaObjects/41467_2022_28212_MOESM14_ESM.mp4"
+        },
+        {
+            "label": "Supplementary Movie 12",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-022-28212-z/MediaObjects/41467_2022_28212_MOESM15_ESM.mp4"
+        }
+    ],
+    "supplementary_1": NaN,
+    "supplementary_2": NaN,
+    "source_data": [],
+    "code": [
+        "https://github.com/Molecular-Nanophotonics/TrackerLab",
+        "https://github.com/Molecular-Nanophotonics/Hydrodynamic-Manipulation-of-Nano-Objects-by-Thermo-Osmotic-Flows"
+    ],
+    "subject": [
+        "Fluid dynamics",
+        "Nanofluidics"
+    ],
+    "license": "http://creativecommons.org/licenses/by/4.0/",
+    "preprint_pdf": "https://www.researchsquare.com/article/rs-879955/v1.pdf?c=1644690399000",
+    "research_square_link": "https://www.researchsquare.com//article/rs-879955/v1",
+    "nature_pdf": "https://www.nature.com/articles/s41467-022-28212-z.pdf",
+    "preprint_posted": "24 Sep, 2021",
+    "research_square_content": [
+        {
+            "section_name": "Abstract",
+            "section_text": "The manipulation of nano-objects at the microscale is of great technological significance to construct new functional materials, to manipulate tiny amounts of liquids, to reconfigure sensorial systems or to detect minute concentrations of analytes in medical screening. It is commonly approached by the generation of potential energy landscapes, for example, with optical fields or by using pressure driven microfluidics. Here we show that strong hydrodynamic boundary flows enable the trapping and manipulation of nano-objects near surfaces. These thermo-osmotic flows are induced by modulating the van der Waals interaction at a solid-liquid interface with optically generated temperature fields. We use a thin gold film on a glass substrate to provide localized but reconfigurable point-like optical heating. Convergent boundary flows with velocities of tens of micrometres per second are observed and substantiated by a quantitative physical model. The hydrodynamic forces acting on suspended nanoparticles and attractive van der Waals or depletion induced forces enable precise positioning and guiding of the nanoparticles. Fast multiplexing of flow fields further provides the means for parallel manipulation of many nano-objects and the generation of complex flow fields. Our findings have direct consequences for the field of plasmonic nano-tweezers as well as other thermo-plasmonic trapping schemes and pave the way for a general scheme of nanoscopic manipulation with boundary flows.NanosciencePlasma and Fluidsnano-objectsmicrofluidicshydrodynamicsthermo-ostic flow",
+            "section_image": []
+        },
+        {
+            "section_name": "Additional Declarations",
+            "section_text": "There is NO Competing Interest.",
+            "section_image": []
+        },
+        {
+            "section_name": "Supplementary Files",
+            "section_text": "Video6.mp4Supplementary Video 6SI.pdfVideo3.mp4Supplementary Video 3Video2.mp4Supplementary Video 2Video8.mp4Supplementary Video 8Video7.mp4Supplementary Video 7Video1.mp4Supplementary Video 1Video5.mp4Supplementary Video 5Video4.mp4Supplementary Video 4",
+            "section_image": []
+        }
+    ],
+    "nature_content": [
+        {
+            "section_name": "Abstract",
+            "section_text": "Manipulation of nano-objects at the microscale is of great technological importance for constructing new functional materials, manipulating tiny amounts of fluids, reconfiguring sensor systems, or detecting tiny concentrations of analytes in medical screening. Here, we show that hydrodynamic boundary flows enable the trapping and manipulation of nano-objects near surfaces. We trigger thermo-osmotic flows by modulating the van der Waals and double layer interactions at a gold-liquid interface with optically generated local temperature fields. The hydrodynamic flows, attractive van der Waals and repulsive double layer forces acting on the suspended nanoparticles enable precise nanoparticle positioning and guidance. A rapid multiplexing of flow fields permits the parallel manipulation of many nano-objects and the generation of complex flow fields. Our findings have direct implications for the field of plasmonic nanotweezers and other thermo-plasmonic trapping systems, paving the way for nanoscopic manipulation with boundary flows.",
+            "section_image": []
+        },
+        {
+            "section_name": "Introduction",
+            "section_text": "The control and manipulation of nano-objects is a key element for future nanophotonics1,2,3,4,5, material science4,6,7, biotechnology2,8,9, or even quantum sensing10. Analytes dissolved in liquids, for example, need to be delivered, concentrated, separated, or locally confined for further studies to become eventually processed and removed. Photonic elements including plasmonic nanostructures require precise positioning or controlled rearrangements to serve as adaptive functional structures. Key elements of the control at the micro- and nanoscale are often either pressure-driven fluidics transporting liquid volume and solutes or the generation of potential energy landscapes or force fields. The latter is achieved with optical3 and plasmonic tweezers11,12, magnetic fields13, or using electrokinetic14 or opto-electronic15 effects. Especially in the field of plasmonic tweezers and nanoantennas, where light is used to excite collective electron motion in noble-metals, the Joule losses lead to the unavoidable generation of heat at boundaries as an unwanted side effect16,17. Yet, such optically generated temperature fields seem also suitable for the manipulation of nano-objects in liquids, for example, for the trapping of nanoparticles18 and single molecules19 or protein aggregates20 as well as for manufacturing active particles21,22,23,24. Those techniques rely on a drift of molecules and particles in optically generated temperature gradients termed thermophoresis or suggest thermo-electric effects25 relying on a thermally induced charge separation. In addition, thermo-electrohydrodynamic effects using time-varying electric fields have been proposed for rapid particle transport26,27 and convective effects that arise from temperature-induced density changes in the large liquid cells have been reported28,29,30,31.\n\nHere we report on a fundamental physical process that is able to provide versatile trapping and manipulation of nano-objects and fluids near surfaces in the simplest geometries. Contrary to most other techniques, our scheme is based on hydrodynamic flows generated by optically induced thermo-osmosis. Thermo-osmosis relies on a perturbation of the interfacial interactions at a solid\u2013liquid boundary and is present in all experiments involving temperature gradients in plasmonic structures including plasmonic tweezers. We show that local temperature gradients on a thin gold film induce strong interfacial flows of several 10\u2013100\u2009\u03bcm\u2009s\u22121 in its direct vicinity (10\u2009nm) that results in a flow pattern reminiscent of convection. Based on a fully quantitative analysis of our experimental results we reveal that these thermo-osmotic flows on gold\u2013water interfaces are induced by a temperature-induced perturbation of the van der Waals (vdW) interactions. Nano-objects suspended in the liquid are, therefore, dragged by the hydrodynamic forces originating from these flows. Utilizing attractive vdW interactions of the nano-object with the gold surface or temperature-induced depletion, we trap and manipulate different types of nano-objects near the surface. The fast heating at small scales allows us to multiplex flow fields and to manipulate multiple objects with great precision. Our detailed analysis of the flow fields, the localization accuracy of nano-objects, and a comparison with numerical and theoretical predictions provide a quantitative understanding of these effects and paves the way for controlling boundary layer dynamics to manipulate objects at the smallest length scales in solutions.",
+            "section_image": []
+        },
+        {
+            "section_name": "Results and discussion",
+            "section_text": "Our experiments rely on a simple sample geometry with a gold film (50\u2009nm) that is deposited on a microscopy glass coverslip (Fig.\u00a01a). The sample chamber contains a suspension of gold nanoparticles (AuNPs) or other nano-objects (polystyrene NPs and ellipsoids) with a controlled amount of salt (NaCl), surfactants (SDS, ...), or polymers (PEG). The gold film is heated locally in an inverted microscopy setup by a highly focused laser (532\u2009nm) using beam steering optics (Acusto-Optic Deflector, AOD). The nano-objects are observed using darkfield illumination with an oil-immersion darkfield condenser (NA 1.2) and a \u00d7100 oil-immersion objective set to NA 0.6 (Fig.\u00a01b). Additional details of the experimental setup and sample preparation are provided in the Methods section, Supplementary Note\u00a01 and Fig.\u00a0S1.\n\na The sample consists of two glass slides that confine a 3\u2009\u03bcm thin liquid film of gold nanoparticles (AuNPs) with radius R dispersed in aqueous NaCl solution. The lower glass slide carries a 50\u2009nm Au film that is locally heated by optical absorption of a focused laser of \u03bb\u2009=\u2009532\u2009nm wavelength. The inset depicts the DLVO (Derjaguin, Landau, Verwey, Overbeek) potential as described in the main text and Supplementary Notes 4 for a R\u2009=\u2009125\u2009nm AuNP in 10\u2009mM NaCl. b The experimental setup comprises an inverted optical microscope equipped with a steerable focused laser of \u03bb\u2009=\u2009532\u2009nm wavelength controlled by an acusto-optic deflector. The AuNPs are observed using darkfield illumination with an oil-immersion darkfield condenser (NA 1.2) and a \u00d7100 oil-immersion objective set to NA 0.6. Images are recorded with an EMCCD camera, typically, with an inverse framerate of \u03c4\u2009=\u200920\u2009ms.\n\nThe trapping of nano-objects as detailed in the following is comprising two effects. (i) The vertical confinement of the suspended objects as achieved by an attractive interaction of the suspended nano-objects with the gold surface, which is found to be the vdW interaction for gold nanoparticles and can be replaced by depletion forces for other materials. (ii) The generation of thermo-osmotic boundary flows that are induced by the local heating and the corresponding perturbation of the liquid\u2013solid interactions. This boundary flow is directed radially inwards to the heated spot and provides a confining hydrodynamic force on suspended objects at the heating spot.\n\nConsider a single AuNP with a radius of R\u2009=\u2009125\u2009nm that is suspended in an aqueous solution of NaCl at a concentration of c0\u2009=\u200910\u2009mM and diffusing in a thin liquid film of about 3\u2009\u03bcm thickness over a 50\u2009nm Au film (Fig.\u00a01a). Exploring the diffusion of the particle we observe a restriction of the z-positions to a thin layer close to the gold film (see Supplementary Notes\u00a02, 3 with Figs.\u00a0S2, S3 for details). The gold particle never defocuses under these conditions while it does in deionized (DI) water (Supplementary Video\u00a01). This restricted out-of-plane motion is the result of interactions comprising an attractive vdW contribution, a repulsion of the electrostatic double layers of the particle and surface32 as described by the DLVO theory and the gravitational potential (see Supplementary Notes\u00a04, 5 for details):\n\nThe total potential for the experimental situation is depicted in Fig.\u00a02a for different salt concentrations (Figs. S4\u2013S8 and Supplementary Note\u00a04 for parameters) as a function of the surface-to-surface distance d\u2009=\u2009z\u2009\u2212\u2009R, where z denotes the distance of the particle center from the surface. The stronger screening of the surface charges at the gold film and the AuNP at higher salt concentration increase the importance of attractive vdW interactions to create this secondary minimum in the DLVO part of the potential. This potential influences the observed dynamics and leads also to a stronger hydrodynamic coupling of the particle to the nearby gold surface33. The in-plane D\u2225, Eq. (2), and out-of-plane D\u22a5 diffusion coefficient (see Supplementary Note\u00a06 and Figs.\u00a0S11, S12 for details) are modulated with the distance z of the particle from the wall:\n\nOver the course of a diffusion trajectory, the particle samples different regions with different diffusion coefficients according to its probability density \\(p(d)\\propto \\exp (-V(d,{c}_{0})/({k}_{{{{{{{{\\rm{B}}}}}}}}}T))\\) to be at a distance d from the surface (filled regions in Fig.\u00a02a). The observed in-plane diffusion coefficient is, thus, a weighted average of the diffusion coefficient over the different vertical positions d. Using p(d) we can calculate the corresponding salt concentration dependence of the in-plane diffusion coefficient and compare that to the experimental results. Figure\u00a02b shows that the experimentally observed D\u2225 is decreasing with increasing salt concentration due to the hydrodynamic coupling in fair agreement with the theoretical predictions (for three different Hamaker constants for the AuNP gold surface interaction). If one additionally calculates the mean distance \u3008d\u3009 of the particle from the surface using p(d), it can be seen that the minimum of the DLVO potential significantly affects the mean distance only for c0\u2009>\u20093\u2009mM. For c0\u2009=\u200910\u2009mM, the potential minimum corresponds approximately to the mean distance \u3008d\u3009, indicating that the particle is completely trapped near the DLVO potential minimum and is not delocalized over the entire film thickness as at lower concentrations. Using the calculated mean distance as a function of concentration c0, it is also possible to estimate the mean distance \u3008d\u3009 of the particles from the surface in the experiments, which is about 1.5\u2009\u03bcm and 0.9\u2009\u03bcm at the lowest NaCl concentrations (Fig.\u00a02c). At a concentration of c0\u2009=\u200910\u2009mM the particle is hovering at a distance of \u3008d\u3009\u2009=\u200920\u2009nm surface. Note that this corresponds to values of z/(2R)\u2009\u2248\u20090.58, which is far below the commonly explored region of the hydrodynamic coupling of colloids to walls33 allowing to experimentally explore new terrains also in the field of hydrodynamic wall coupling of colloids.\n\na Plot of the DLVO potential, Eq. (1), between a 250\u2009nm AuNP and a 50\u2009nm Au film on a glass surface as function of surface\u2013surface distance d for different NaCl concentrations c0. The shaded curves display the calculated probability density for finding the particle at this distance at the different salt concentrations (see Supplementary Notes 4 for details). The vertical dashed lines correspond to the mean distance of the particle as calculated from the probability density for a 3\u2009\u03bcm liquid film height. b The measured diffusion coefficient \u3008D\u2225\u3009/D0 parallel to the Au film with respect to the bulk diffusion coefficient D0 as function of the NaCl concentration c0. Symbols correspond to the experimental values measured without additional laser heating of the gold film. The lines reflect the theoretical prediction including a distance dependent diffusion coefficient for three different Hamaker constants (dotted: AH\u2009=\u20094\u2009\u00d7\u200910\u221220\u2009J, dash-dotted: 5\u2009\u00d7\u200910\u221220\u2009J, and dashed: 6\u2009\u00d7\u200910\u221220\u2009J) of gold according to a Boltzmann weighting (see main text). The error bars have been estimated from pipetting errors and errors from the MSD fitting. c Relation between the mean distance \u3008d\u3009 and the NaCl concentration c0. The symbols and the horizontal lines denote the calculated distances for measured concentrations. The arrows indicate the secondary minimum of the corresponding DLVO potentials.\n\nWhen tightly focusing the light of 532\u2009nm wavelength to the gold film, a part of the incident energy (about 30%) is absorbed and converted into heat that perturbs the liquid\u2013solid interactions. The temperature rise at the gold surface can be determined using a thin nematic liquid crystal (5CB) film and substantiated by finite element simulations with the complete three-dimensional temperature profile in the solution (see Fig.\u00a03a, b and Supplementary Notes\u00a07, 8, Figs.\u00a0S13\u201318 for details).\n\na Relative temperature increment \\({{\\Delta }}\\,T/{{\\Delta }}\\,{T}_{\\max }\\) in the xz-plane of the sample as obtained from numerical simulations. The simulation is detailed in Supplementary Note 7. The white lines correspond to the gold and glass interface, respectively. b Experimentally obtained temperature increment \\({{\\Delta }}{T}_{\\max }\\) as a function of the incident laser power P0 (green data points) compared to the simulated values (green curve). The experimental data have been obtained with the help of a liquid crystal as explained in Supplementary Note 8. The error bars have been obtained from uncertainties of the power measurement and the estimation of phase transition radius. c Measured thermo-osmotic flow field in the xy-plane in close proximity to the gold film (z\u2009<\u2009500\u2009nm) for a laser power of 2 mW (\\({{\\Delta }}\\,{T}_{\\max }=40\\,{{{{{{{\\rm{K}}}}}}}}\\)) for a NaCl concentration c0\u2009=\u20090 mM (DI water). The grayscale image indicates the magnitude of the velocity. The arrows display the direction of the flow. d Measured thermo-osmotic flow field in the xz-plane for the same parameters as in c. The grayscale image indicates the magnitude of the velocity. The arrows display the direction of the flow. e Illustration of the measured flow field planes in c and d. f, g The x- and z-component of the measured flow velocities for the same parameters as in c and d compared to the simulation results in h and i.\n\nThese local temperature perturbations of the solid\u2013liquid interactions at the interface induce a thermo-osmotic flow34,35 (Supplementary Note\u00a09, Figs.\u00a0S19\u2013S22). Taking a liquid volume element close to the solid from the cold side and exchanging that with one on the hot side would not only transport heat since the liquid volumes have different temperatures, but also additional free energy as the liquid has a different interaction with the solid in these regions. The flow is induced in an ultrathin boundary layer corresponding in thickness to the length scale of liquid\u2013solid interactions. Since the characteristic interaction length of liquid\u2013solid interactions is only a few nanometers, the boundary flow on the substrate can be collapsed into a quasi-slip hydrodynamic boundary condition:\n\nwhere \u03b7 is the viscosity of the liquid, h(z) is the excess enthalpy, T the temperature and \u2207\u2225T is the temperature gradient parallel to the surface. The integral can be summarized to a thermo-osmotic coefficient \u03c7. The thermo-osmotic coefficient \u03c7, therefore, contains all information about the interfacial interaction between the liquid and the solid. If \u03c7\u2009<\u20090 the liquid is driven to the cold, whereas for \u03c7\u2009>\u20090, the liquid is driven to the hot. These boundary flows are present at all liquid\u2013solid interfaces with tangential temperature gradients, though, they are commonly overlooked. They become particularly important for plasmonic and thermo-plasmonic trapping16, as those techniques rely on the dynamics of molecules and particles in the direct vicinity of plasmonic nanostructures. The boundary flow drives the flow field inside the fluid film. The resulting volumetric flow field can be tracked experimentally by single AuNPs in DI water, where the particles are not confined to a surface layer as reported in the previous section. We analyze the in-plane (xy) position of the particle and its z-position, where the latter is estimated from the radius r0 of the defocussed particle images (see Supplementary Video\u00a02 and Supplementary Note\u00a02 for details). The measured velocity distributions in the xy-plane near the gold layer and in the xz-plane are shown in Fig.\u00a03c and d, respectively. The x- and z-component of the measured flow velocities are depicted in Fig.\u00a03f, g. The flow profiles compare well to the simulation results shown in Fig.\u00a03h, i, although some discrepanices exist especially close to r\u2009=\u20090. Particle velocities are highest in this region and can easily reach several 10\u2009\u03bcm\u2009s\u22121, which increases the particle localization errors for finite exposure time and finally blurs the velocity profile additionally. From the velocity measurements, we extract a thermo-osmotic coefficient on the order of \u03c7\u2009~\u200910\u2009\u00d7\u200910\u221210\u2009m2\u2009s\u22121 (see Supplementary Note\u00a09 for details). We can break down the contributions to this value with Eqs. (4) and (5) to estimate the double layer and vdW contributions using the experimental parameters. Note that AuNPs do not show thermophoresis due to their high thermal conductivity and thus isothermal surface (see Supplementary Note\u00a010 for details)\n\nFor the electrostatic contribution we used a zeta potential of \u03b6\u2009=\u2009\u221230\u2009mV36 and a static permittivity of \u03b5\u2009=\u200980\u2009\u03b50. An estimate of the vdW contribution can be given by:\n\nin terms of the Hamaker constant AH between water and the Au film with \u03b2\u2009=\u20090.2\u2009\u00d7\u200910\u22123\u2009K\u22121 being the thermal expansion coefficient of water and d0\u2009=\u20090.2\u2009nm for the cut-off parameter34 (see Supplementary Note\u00a09 for details). The sum of both contributions \u03c7\u2009=\u2009\u03c7E\u2009+\u2009\u03c7vdW\u2009=\u200910.1\u2009\u00d7\u200910\u221210\u2009m2\u2009s\u22121 matches well the experimental result and suggests that thermo-osmosis at gold\u2013water interfaces is governed by vdW interactions. The obtained quasi-slip velocities are ranging up to 80\u2009\u03bcm\u2009s\u22121 and provide, due to their omnipresence, a unique tool for nanofluidics. These thermo-osmotic flows are induced without any external pressure difference. They can be controlled by the light intensity of the heating laser and are quickly switched due to the extremely fast heat conduction at these length scales. Moreover, the finding of the vdW dominated thermo-osmotic flows suggest that such contributions must be present in any plasmonic trapping experiment with extended gold structures12,16,17,27.\n\nUsing \\({F}_{x}^{{{{{{{{\\rm{TF}}}}}}}}}=6\\pi \\eta R\\,{\\gamma }_{\\parallel }\\,{v}_{x}\\) and \\({F}_{z}^{{{{{{{{\\rm{TF}}}}}}}}}=6\\pi \\eta R\\,{\\gamma }_{\\perp }{v}_{z}\\), where \u03b3\u2225 and \u03b3\u22a5 are the correction factors for the friction coefficient of a sphere close to a surface, we are able to extract the hydrodynamic forces that are exerted on the AuNP tracers (see Eq. (2), Fig.\u00a0S9 and Supplementary Notes\u00a04, 6 for details). The lateral forces allow to confine objects at the heating spot, yet the hydrodynamic force normal to the surface (z-direction) is repulsive without any additional interaction. Finally, such boundary flows with substantial vertical velocity gradients also exhibit vorticity (see Supplementary Note\u00a09, Fig.\u00a0S23 for details) that generates a torque on suspended objects causing them to rotate37.\n\nThe vertical hydrodynamic drag force \\({F}_{z}^{{{{{{{{\\rm{TF}}}}}}}}}\\) is superimposed with an attractive force due to the DLVO potential when increasing the NaCl concentration. The surface-to-surface distance between AuNP and gold film and the depth of the appearing secondary DLVO potential minimum can be controlled by the NaCl concentration. At a NaCl concentration of about c0\u2009=\u200910\u2009mM, the attractive potential has a depth of about 10\u2009kBT (see Fig.\u00a02a) and is strong enough to compete with the vertical drag force and additional optical forces on the AuNP to trap the particle above the heating spot.\n\nSupplementary Video\u00a03 demonstrates this trapping of a AuNP above the hot spot on the Au surface. This is purely the result of the hydrodynamic drag forces generated by the thermo-osmotic flow and the attractive vdW interaction between the AuNP and the Au film. This observation is substantiated by a quantitative evaluation of the lateral trap stiffness and vertical forces, as depicted in Fig.\u00a04a, b. The fluctuations of the particle in the hydrodynamic flow arise from a balance of the restoring hydrodynamic currents and the diffusive currents. Analysis of the lateral position histograms (inset in Fig.\u00a04c for 1.25\u2009mW) yields an effective stiffness of the trap (Fig.\u00a04) that well matches the predictions based on the thermo-osmotic flow (Fig.\u00a03h, i), when converting the lateral flow speeds into forces using the previously mentioned Stokes friction force for \\({F}_{x}^{{{{{{{{\\rm{TF}}}}}}}}}\\) (see Supplementary Note\u00a011 and Fig.\u00a0S27 for details). The hydrodynamic trapping stiffness increases linearly up to a heating power of about 1.8\u2009mW. At this power, the vertical forces become strong enough to let the particle escape the secondary DLVO minimum, which is visible from the z-position time traces displayed in Fig.\u00a04d. The AuNPs are then observed to move vertically out of the DLVO potential to follow the flow inside the sample and to eventually return to the boundary flow via sedimentation (Fig.\u00a04e, Supplementary Video\u00a04). The forces which eject the particle from the potential comprise the hydrodynamic and the optical forces due to the radiation pressure from the heating laser leaked through the film (see Supplementary Note\u00a012 and Fig.\u00a0S28 for details). We have evaluated the individual contributions in simulations. They are shown together with the hydrodynamic force and the total vertical force as compared to the attractive force of the DLVO potential (Fig.\u00a04b) and provide quantitative agreement (threshold heating power of 2.25\u2009mW) with our experimental results. Note that while the stationary distribution of particles in the vertical direction is not influenced by the diffusive dynamics, the escape rate from the potential well is heavily altered by the fact that the vertical diffusion coefficient D\u22a5 of the particle is decreasing to zero when approaching the gold film. This is enhancing the trapping times considerably (see Fig.\u00a0S10 and Supplementary Note\u00a05 for details) but also increases the time required for the particle to enter the DLVO minimum by diffusion.\n\na The lateral trap stiffness kx obtained from the variance \\({\\sigma }_{x}^{2}\\) of the experimental lateral position histograms as function of the laser power P0 (blue data points, \\({k}_{x}={k}_{{{{{{{{\\rm{B}}}}}}}}}T/{\\sigma }_{x}^{2}\\), see the inset in c for the lateral position histogram at P0\u2009=\u20091.25\u2009mW). The blue solid line represents the simulation result obtained from the lateral velocity field in Fig.\u00a03h (see Supplementary Note 9 for details). Within the dashed area the particle intermittently escapes the trapping potential in vertical direction and transitions into a regime of instant vertical repulsion for higher laser powers (shaded area). The error bars have been estimated from uncertainties of the power measurement and errors from the histogram fitting. b The z-component of the thermo-osmotic drag force \\({F}_{z}^{{{{{{{{\\rm{TF}}}}}}}}}\\) (blue line), the optical force \\({F}_{z}^{{{{{{{{\\rm{OF}}}}}}}}}\\) (green line) and the total force \\({F}_{z}^{{{{{{{{\\rm{OF}}}}}}}}}+{F}_{z}^{{{{{{{{\\rm{TF}}}}}}}}}\\) (black dashed line) as function the incident laser power P0 for a NP located at x\u2009=\u20090, d\u2009=\u200930\u00a0(z\u2009=\u2009d\u2009+\u2009R). The attractive DLVO force \\({F}_{z}^{{{{{{{{\\rm{DLVO}}}}}}}}}\\) is independent of the incident laser power and depicted as horizontal, red line. c Trajectory of a AuNP for a heating laser power of 1.25\u2009mW (see Supplementary Video\u00a02 for details). The inset shows the corresponding lateral distribution histogram. d Time traces of the z-position for three different laser powers. e Trajectory of a AuNP for a heating laser power of 2.5\u2009mW, which is above the threshold power of 2.25\u2009mW.\n\nThe observed trapping is, hence, a vdW assisted thermo-hydrodynamic process. Vertical confinement is achieved by vdW attraction and double layer repulsion, while lateral confinement is the result of thermo-osmotic flows induced in an ultrathin sheet of liquid at the interface. No additional contributions, for example, due to convective flows with similar flow patterns (see Supplementary Note\u00a013, Figs.\u00a0S29\u2013S31 for details) or thermo-electric effects are required for a quantitative description25,31,38,39. Precise tuning of the DLVO potential enables the trapping of even smaller AuNPs (Fig.\u00a05a, Supplementary Video\u00a05).\n\na AuNP with 50 nm radius trapped at a NaCl concentration of 30 mM (Supplementary Video\u00a05). b Manipulation of two AuNPs by a multiplexed laser beam (Supplementary Video\u00a06). c Control of three AuNPs (Supplementary Video\u00a07). d Actuation of a single AuNP on a circular trajectory by a steerable laser beam (Supplementary Video\u00a08). The green and white, dashed arrows denote the moving direction of the laser focus and the particle, respectively. e Generation of thermo-viscous flows by rotating the laser focus on a circle with a rotation frequency of f\u2009=\u2009500\u2009Hz at high laser powers (Supplementary Video\u00a09). Note that laser movement (green arrow) and the thermo-viscous flow (white, dashed arrow) and have opposite directions. f Attraction of a AuNP and PS NPs in 5\u2009mM SDS due to depletion (Supplementary Video\u00a010). g An AuNP (125\u2009nm radius) trapped in an ensemble of polystyrene (PS) NPs of the same size at 10\u2009mM NaCl (Supplementary Video\u00a011), where the PS particles are repelled due to thermophoresis. h Attraction of PS ellipsoids (2.39\u2009\u03bcm major-axis length, 0.34\u2009\u03bcm minor-axis length) in 5\u2009mM SDS (Supplementary Video\u00a012).\n\nThe speed of heat diffusion, which is about 4 orders of magnitude faster than the particle diffusion40 allows us to introduce a flow field multiplexing. We switch the heating location between different positions inducing thermo-osmotic flow fields for time periods of about 100\u2009\u03bcs. With the help of this multiplexing, we are able to hold multiple R\u2009=\u2009125\u2009nm AuNPs (Fig.\u00a05b, c) at distances of less than 1\u2009\u03bcm, which would not be possible with continuous heating of close-by locations (Supplementary Videos\u00a06 and 7). A trapped AuNP can also be guided along the predefined path over the Au film as fast as 10\u2009\u03bcm\u2009s\u22121 (Fig.\u00a05d, Supplementary Video\u00a08). At larger manipulation speeds (f\u2009\u00a0>\u2009100\u2009Hz) and higher heating power (P0\u2009>\u200910\u2009mW) the thermo-osmotic attraction to the heating spot is combined with thermo-viscous flows41,42. These flows originate from the temperature-dependent viscosity \u03b7(T) of the liquid and are directed opposite to the scanning direction of the laser42. The result of this combination of thermo-osmosis and thermo-viscous flows is a rotating ring-like particle structure (Fig.\u00a05e and Supplementary Video\u00a09).\n\nThese different effects that can be exploited in a simple planar geometry give rise to numerous applications including, for example, a freely configurable nanoparticle on mirror geometry for plasmonic sensing5,43. The multiplexing of the local flow fields may be helpful to construct more complex effective flow fields for efficient transport of analytes without external pressure.\n\nSo far, the presented manipulation is based on thermo-osmotic flows that drive the lateral motion of suspended colloids and vertical confinement due to the secondary minimum of the DLVO potential between AuNP and Au film. While the thermo-osmotic flows are characteristic for all systems containing a heated gold\u2013water interface including all previous studies on thermo-plasmonic trapping, the DLVO potential minimum is much weaker for other materials like polymer colloids or macromolecules due to their smaller vdW attraction. Often, those systems even show a repulsion from the heat source due to thermophoresis, which is not present for AuNPs (see Supplementary Note\u00a010, Figs.\u00a0S24\u2013S26 for details). A more generalized strategy, therefore, needs additional attractive contributions, which confine suspended colloids or molecules to regions close to the gold surface to take advantage of the thermo-osmotic flow.\n\nSuch attractive contributions can arise from depletion interactions21,44. Thereby a temperature gradient repels dissolved molecules from the heated regions generating a concentration gradient that drives suspended nano-objects to the heating spot. To demonstrate this effect we use the surfactant sodium dodecyl sulfate (SDS) at a concentration of 5 mM well below the critical micelle concentration (8.2\u2009mM) to avoid complications of micelle formation. We suspend additional polystyrene particles (PS) and AuNPs of the same size (R\u2009=\u2009125\u2009nm) in the solution and compare their dynamics to a solution with AuNPs and PS particles without SDS but 10\u2009mM NaCl. Remarkably, the heated spot is attractive for both AuNPs and for PS NPs (Fig.\u00a05f, Supplementary Video\u00a010) in the SDS solution showing even PS colloidal crystal growth, while only the AuNP is trapped in the NaCl solution and the PS particles are repelled by thermophoresis (Fig.\u00a05g, Supplementary Video\u00a011).\n\nThe observations in NaCl are readily explained by the fact that the AuNP is confined in the DLVO minimum as demonstrated above but the PS particle is not due to a 10 times lower Hamaker constant (see Supplementary Note\u00a04 for details). The PS particle samples the whole liquid film thickness equally and not preferentially the region close to the Au film and experiences an additional thermophoretic drift velocity given by:\n\nwhere DT is the thermophoretic mobility34 and\u2009\u2207\u2009T the temperature gradient (Fig.\u00a05g, Supplementary Video\u00a011). For \u03c7\u2009>\u20090 the particle is driven to the cold. From Eq. (4) we find \u03c7\u2009\u2248\u2009\u03c7E\u2009=\u20091.28\u2009\u00d7\u200910\u221210\u2009m2\u2009s\u22121 and DT\u2009\u2248\u20090.3\u2009\u03bcm2\u2009K\u22121\u2009s\u22121, where we have used a measured zeta potential of \u03b6\u2009\u2248\u2009\u221238\u2009mV (see Methods section for details). The vdW contribution, \u03c7vdW to either the thermophoretic drift or the attraction to the gold surface can be neglected due to the smaller Hamaker constant of PS. From the stationary probability distribution of the PS NP we find a Soret coefficient of ST\u2009\u2248\u20090.24\u2009K\u22121 in agreement with our theoretical prediction ST\u2009=\u2009DT/D\u2225\u2009\u2248\u20090.21\u2009K\u22121.\n\nIn the SDS solution, the additional surfactant molecules now undergo thermophoresis to yield a concentration gradient in which suspended colloidal particles drift. The lower concentration in the heated regions promotes an effective attractive interaction of suspended colloids with the gold surface due to depletion forces. The total drift velocity u is then described by21,34,44.\n\nwhere uD denotes the depletion-induced drift velocity (see Supplementary Note\u00a014 and Fig.\u00a0S32 for details). Here Rm is the size of a SDS molecule, c0 their concentration in units of mol\u2009l\u22121, NA the Avogadro constant and \\({S}_{{{{{{{{\\rm{T}}}}}}}}}^{{{{{{{{\\rm{m}}}}}}}}}\\) the Soret coefficient of SDS. For Rm\u2009=\u20092\u2009nm45, c0\u2009=\u20095\u2009mM and \\({S}_{{{{{{{{\\rm{T}}}}}}}}}^{{{{{{{{\\rm{m}}}}}}}}}=0.03\\,{{{{{{{{\\rm{K}}}}}}}}}^{-1}\\)\u200946 we find\u2009\u22120.43\u2009\u03bcm2\u2009K\u22121\u2009s\u22121 for the additional depletion contribution, which exceeds the thermophoretic mobility, DT\u2009\u2248\u20090.3\u2009\u03bcm2\u2009K\u22121\u2009s\u22121, rendering the overall mobility negative. The PS NPs and the AuNPs are thus driven to the heated Au film surface (Fig.\u00a05f, Supplementary Video\u00a010) which allows for further transport in the thermo-osmotic boundary flow. Additional contributions, for example, thermo-electric fields may even enhance the attractive components. Overall, this concept is readily transferred to other objects as shown in Fig.\u00a05h and Supplementary Video\u00a012, where we have trapped ellipsoidal PS particles in a 5\u2009mM solution of SDS. Note that as compared to other schemes, our approach always includes thermo-osmotic boundary flows.\n\nSummarizing, we have demonstrated that thermo-hydrodynamic boundary flows can manipulate nano-objects with unprecedented flexibility in a very simple sample geometry. These flows are the key for future thermo-optofluidic implementations with an extensive range of applications in the fields of (i) nanoparticle sorting and separation; (ii) assembly of nanophotonic circuits47 and plasmonic quantum sensors4,5; (iii) biotechnology on-chip laboratories48 and (iv) manufacturing of nanomaterials4,6 and functional nanosurfaces49,50. We have substantiated our experimental findings of thermo-osmotic flow-assisted trapping with a quantitative theoretical description. A flow field multiplexing scheme has been further developed to allow for the simultaneous manipulation of many individual nano-objects and the generation of complex effective flow patterns. Our concept can be combined with other thermally induced effects such as thermophoresis, depletion forces and thermo-viscous flows to form a fully-featured nanofluidic system-on-a-chip. Besides direct consequences for the field of plasmonic nanotweezers and other thermo-plasmonic trapping schemes, the use of thermo-hydrodynamic flows as a tool for nanofluidic applications will extend the limits at the forefront of nanotechnology and help to develop AI and feedback-controlled schemes for material synthesis.",
+            "section_image": [
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-022-28212-z/MediaObjects/41467_2022_28212_Fig1_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-022-28212-z/MediaObjects/41467_2022_28212_Fig2_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-022-28212-z/MediaObjects/41467_2022_28212_Fig3_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-022-28212-z/MediaObjects/41467_2022_28212_Fig4_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-022-28212-z/MediaObjects/41467_2022_28212_Fig5_HTML.png"
+            ]
+        },
+        {
+            "section_name": "Methods",
+            "section_text": "The experimental setup (Fig.\u00a01b, Fig.\u00a0S1) consists of an inverted microscope (Olympus, IX71) with a mounted piezo translation stage (Physik Instrumente, P-733.3). The microparticles are heated by a focused, continuous-wave laser at a wavelength of 532\u2009nm (CNI, MGL-III-532). The beam diameter is increased by a beam expander and sent to an acousto-optic deflector (AA Opto-Electronic, DTSXY-400-532) and a lens system to steer the laser focus in the sample plane. The deflected beam is focused by an oil-immersion objective (Olympus, UPlanApo \u00d7100/1.35, Oil, Iris, NA 0.5\u20131.35) to the sample plane (w0\u2009\u2248\u20090.8\u2009\u03bcm beam waist in the sample plane). The sample is illuminated with an oil-immersion darkfield condenser (Olympus, U-DCW, NA 1.2\u20131.4) and a white-light LED (Thorlabs, SOLIS-3C). The scattered light is imaged by the objective and a tube lens (250\u2009mm) to an EMCCD (electron-multiplying charge-coupled device) camera (Andor, iXon DV885LC). The variable numerical aperture of the objective was set to a value below the minimum aperture of the darkfield condenser. The dichroic beam splitter (D) was selected to reflect the laser wavelength (Omega Optical, 560DRLP) and a notch filter (F) is used to block any remaining back reflections from the laser (Thorlabs, NF533-17). The acousto-optic deflector (AOD), as well as the piezo stage, are driven by an AD/DA (analog-digital/digital-analog) converter (J\u00e4ger Messtechnik, ADwin-Gold II). A LabVIEW program running on a desktop PC (Intel Core i7 2600 4\u2009\u00d7\u20093.40\u2009GHz CPU) is used to record and process the images as well as to control the AOD feedback via the AD/DA converter. Typically, images have been recorded with an inverse framerate of \u03c4\u2009=\u200920\u2009ms.\n\nThe sample consists of two glass coverslips (22\u2009\u00d7\u200922\u2009mm) confining a thin liquid film. First, the coverslips were thoroughly cleaned in an ultrasonic bath with Hellmanex III (1%), acetone, isopropanol, and Milli-Q water followed by 3\u2009min plasma cleaning (PDC-32G, Harrick Plasma Inc.). Then, one of the coverslips was coated with a 50\u2009nm gold film using a thermal evaporator (UNIVEX 300, Leybold GmbH) and a\u00a05\u2009nm chrome adhesion layer. Subsequently, the edges of the uncoated coverslip were covered with a thin layer of PDMS for sealing. The particle solution used for the experiments was prepared by dispersing gold nanoparticles (Cytodiagnostics Inc.), PS particles (microParticles GmbH) in different solutions of NaCl and SDS (Sigma\u2013Aldrich). Finally, 0.3\u2009\u03bcl of the mixed particle suspension is pipetted in the middle of one of the coverslips and the other is placed on top. Depending on the area covered by the liquid, typically about 10\u2009\u00d7\u200910\u2009mm, the resulting liquid film height is about 3\u2009\u03bcm.\n\nThe zeta potential of the particles has been estimated using electrophoretic light scattering (Malvern Instruments Ldt., Zetasizer Nano ZS).\n\nThe numerical simulation presented in this study were computed using COMSOL Multiphysics 5.2 using the Heat Transfer in Fluids and the Non-Isothermal Laminar Flow interface. The corresponding geometry is depicted in Fig.\u00a0S13 in Supplementary Note\u00a07 together with the used parameters as well as Supplementary Note\u00a09, Fig.\u00a0S21. Numerical simulation results are displayed in Figs.\u00a0S14\u2013S17 and Fig.\u00a0S22, S23.",
+            "section_image": []
+        },
+        {
+            "section_name": "Data availability",
+            "section_text": "The datasets produced in this study are available from the corresponding author upon reasonable request.",
+            "section_image": []
+        },
+        {
+            "section_name": "Code availability",
+            "section_text": "The Python code for the single particle tracking software is available at the corresponding GitHub Repository. Python scripts used to analyze the datasets and the files for the COMSOL simulations are available at the GitHub repository for this paper.",
+            "section_image": []
+        },
+        {
+            "section_name": "References",
+            "section_text": "Marag\u00f2, O. M., Jones, P. H., Gucciardi, P. G., Volpe, G. & Ferrari, A. C. Optical trapping and manipulation of nanostructures. Nat. Nanotechnol. 8, 807\u2013819 (2013).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGao, D. et al. Optical manipulation from the microscale to the nanoscale: fundamentals, advances and prospects. Light Sci. Appl. 6, 17039 (2017).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBradac, C. Nanoscale optical trapping: a review. Adv. Opt. Mat. 6, 1800005 (2018).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nXavier, J., Vincent, S., Meder, F. & Vollmer, F. Advances in optoplasmonic sensors\u2014combining optical nano/microcavities and photonic crystals with plasmonic nanostructures and nanoparticles. Nanophotonics 7, 1\u201338 (2018).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLi, G.-C., Zhang, Q., Maier, S. A. & Lei, D. Plasmonic particle-on-film nanocavities: a versatile platform for plasmon-enhanced spectroscopy and photochemistry. Nanophotonics 7, 1865\u20131889 (2018).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLin, L., Peng, X. & Zheng, Y. Reconfigurable opto-thermoelectric printing of colloidal particles. Chem. Commun. 53, 7357\u20137360 (2017).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nXie, Y. et al. Microfluidic isolation and enrichment of nanoparticles. ACS Nano 14, 16220\u201316240 (2020).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nFavre-Bulle, I. A., Stilgoe, A. B., Scott, E. K. & Rubinsztein-Dunlop, H. Optical trapping in vivo: theory, practice, and applications. Nanophotonics 8, 1023\u20131040 (2019).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nChoudhary, D., Mossa, A., Jadhav, M. & Cecconi, C. Bio-molecular applications of recent developments in optical tweezers. Biomolecules 9, 23 (2019).\n\nArticle\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKayci, M., Chang, Huan-Cheng & Radenovic, A. Electron spin resonance of nitrogen-vacancy defects embedded in single nanodiamonds in an ABEL trap. Nano Lett. 14, 5335\u20135341 (2014).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nJuan, M. L., Righini, M. & Quidant, R. Plasmon nano-optical tweezers. Nat. Photonics 5, 349\u2013356 (2011).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nZhang, Y. et al. Plasmonic tweezers: for nanoscale optical trapping and beyond. Light Sci. Appl. 10, 59 (2021).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nDe Vlaminck, I. & Dekker, C. Recent advances in magnetic tweezers. Annu. Rev. Biophys. 41, 453\u2013472 (2012).\n\nArticle\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWang, Q. & Moerner, W. E. An adaptive anti-Brownian electrokinetic trap with real-time information on single-molecule diffusivity and mobility. ACS Nano 5, 5792\u20135799 (2011).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWu, M. C. Optoelectronic tweezers. Nat. Photonics 5, 322\u2013324 (2011).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKotnala, A. & Gordon, R. Quantification of high-efficiency trapping of nanoparticles in a double nanohole optical tweezer. Nano Lett. 14, 853\u2013856 (2014).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nJiang, Q., Rogez, Benoi^t, Claude, Jean-Benoi^t, Baffou, G. & Wenger, J. \u00e9r\u00f4me Quantifying the role of the surfactant and the thermophoretic force in plasmonic nano-optical trapping. Nano Lett. 20, 8811\u20138817 (2020).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBraun, M. & Cichos, F. Optically controlled thermophoretic trapping of single nano-objects. ACS Nano 7, 11200\u201311208 (2013).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBraun, M., Bregulla, A. P., G\u00fcnther, K., Mertig, M. & Cichos, F. Single molecules trapped by dynamic inhomogeneous temperature fields. Nano Lett. 15, 5499\u20135505 (2015).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nFr\u00e4nzl, M. et al. Thermophoretic trap for single amyloid fibril and protein aggregation studies. Nat. Methods 16, 611\u2013614 (2019).\n\nArticle\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nJiang, H.-R., Wada, H., Yoshinaga, N. & Sano, M. Manipulation of colloids by a nonequilibrium depletion force in a temperature gradient. Phys. Rev. Lett. 102, 208301 (2009).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBregulla, A. P., Yang, H. & Cichos, F. Stochastic localization of microswimmers by photon nudging. ACS Nano 8, 6542\u20136550 (2014).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKhadka, U., Holubec, V., Yang, H. & Cichos, F. Active particles bound by information flows. Nat. Commun. 9, 3864 (2018).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nFr\u00e4nzl, M., Mui\u00f1os-Landin, S., Holubec, V. & Cichos, F. Fully steerable symmetric thermoplasmonic microswimmers. ACS Nano 15, 3434\u20133440 (2021).\n\nArticle\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLin, L. et al. Opto-thermoelectric nanotweezers. Nat. Photonics 12, 195\u2013201 (2018).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nNdukaife, J. C. et al. Long-range and rapid transport of individual nano-objects by a hybrid electrothermoplasmonic nanotweezer. Nat. Nanotechnol. 11, 53\u201359 (2016).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHong, C., Yang, S. & Ndukaife, J. C. Stand-off trapping and manipulation of sub-10 nm objects and biomolecules using opto-thermo-electrohydrodynamic tweezers. Nat. Nanotechnol. 15, 908\u2013913 (2020).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nDonner, J. S., Baffou, G., McCloskey, D. & Quidant, R. Plasmon-assisted optofluidics. ACS Nano 5, 5457\u20135462 (2011).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nRoxworthy, B. J., Bhuiya, A. M., Vanka, S. P. & Toussaint, K. C. Understanding and controlling plasmon-induced convection. Nat. Commun. 5, 3173 (2014).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nChen, J. et al. Thermal optofluidics: principles and applications. Adv. Opt. Mater. 8, 1900829 (2020).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nCiraulo, B., Garcia-Guirado, J., de Miguel, I., Ortega Arroyo, J. & Quidant, R. Long-range optofluidic control with plasmon heating. Nat. Commun. 12, 2001 (2021).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nIsraelachvili, J. N. Intermolecular and Surface Forces, 3rd edn. (Elsevier, Academic Press, 2011).\n\nLin, B., Yu, J. & Rice, S. A. Direct measurements of constrained Brownian motion of an isolated sphere between two walls. Phys. Rev. E 62, 3909\u20133919 (2000).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nW\u00fcrger, A. Thermal non-equilibrium transport in colloids. Rep. Prog. Phys. 73, 126601 (2010).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBregulla, A. P., W\u00fcrger, A., G\u00fcnther, K., Mertig, M. & Cichos, F. Thermo-osmotic flow in thin films. Phys. Rev. Lett. 116, 188303 (2016).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGiesbers, M., Mieke Kleijn, J. & Cohen Stuart, M. A. The electrical double layer on gold probed by electrokinetic and surface force measurements. J. Colloid Interface Sci. 248, 88\u201395 (2002).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBluemink, J. J., Lohse, D., Prosperetti, A. & Van Wijngaarden, L. A sphere in a uniformly rotating or shearing flow. J. Fluid Mech. 600, 201\u2013233 (2008).\n\nArticle\u00a0\n    ADS\u00a0\n    MathSciNet\u00a0\n    MATH\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLin, L. et al. Light-directed reversible assembly of plasmonic nanoparticles using plasmon-enhanced thermophoresis. ACS Nano 10, 9659\u20139668 (2016).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLin, L. et al. Opto-thermophoretic assembly of colloidal matter. Sci. Adv. 3, 1700458 (2017).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBaffou, G., Cichos, F. & Quidant, R. Applications and challenges of thermoplasmonics. Nat. Mater. 19, 946\u2013958 (2020).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWeinert, F. M., Kraus, J. A., Franosch, T. & Braun, D. Microscale fluid flow induced by thermoviscous expansion along a traveling wave. Phys. Rev. Lett. 100, 164501 (2008).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWeinert, F. M., Mast, C. B. & Braun, D. Optical fluid and biomolecule transport with thermal fields. Phys. Chem. Chem. Phys. 13, 9918 (2011).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nChikkaraddy, R. et al. Single-molecule strong coupling at room temperature in plasmonic nanocavities. Nature 535, 127\u2013130 (2016).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMaeda, Y. T., Buguin, A. & Libchaber, A. Thermal separation: interplay between the Soret effect and entropic force gradient. Phys. Rev. Lett. 107, 038301 (2011).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSyshchyk, O. et al. Influence of temperature and charge effects on thermophoresis of polystyrene beads. Eur. Phys. J. E 39, 129 (2016).\n\nArticle\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nVigolo, D., Buzzaccaro, S. & Piazza, R. Thermophoresis and thermoelectricity in surfactant solutions. Langmuir 26, 7792\u20137801 (2010).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nZhang, Q., Yu, H., Barbiero, M., Wang, B. & Gu, M. Artificial neural networks enabled by nanophotonics. Light Sci. Appl. 8, 42 (2019).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nXavier, J., Yu, D., Jones, C., Zossimova, E. & Vollmer, F. Quantum nanophotonic and nanoplasmonic sensing: Towards quantum optical bioscience laboratories on chip. Nanophotonics 10, 1387\u20131435 (2021).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nPeng, J. et al. Scalable electrochromic nanopixels using plasmonics. Sci. Adv. 5, eaaw2205 (2019).\n\nMohammadi, E., Tittl, A., Tsakmakidis, K. L., Raziman, T. V. & Curto, A. G. Dual nanoresonators for ultrasensitive chiral detection. ACS Photonics 8, 1754\u20131762 (2021).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nDownload references",
+            "section_image": []
+        },
+        {
+            "section_name": "Acknowledgements",
+            "section_text": "We acknowledge financial support by the German Research Foundation (Deutsche Forschungsgemeinschaft, DFG) through the Collaborative Research Center TRR 102 \u201cPolymers under multiple constraints: restricted and controlled molecular order and mobility\u201d (DFG project number 189853844, SFB TRR 102, to F.C.) and project CI 33/14-1 (DFG project number 242631004 to F.C.). This work was funded by the Federal Ministry for Economic Affairs and Energy based on a resolution of the German Bundestag (BMWi, STARK program, project number 46SKD023X to F.C.) and is cofinanced from tax revenues on the basis of the budget passed by the Saxon state parliament (SMWK). We thank Andrea Kramer for carefully reading the manuscript.",
+            "section_image": []
+        },
+        {
+            "section_name": "Funding",
+            "section_text": "Open Access funding enabled and organized by Projekt DEAL.",
+            "section_image": []
+        },
+        {
+            "section_name": "Author information",
+            "section_text": "Peter Debye Institute for Soft Matter Physics, Molecular Nanophotonics Group, Universit\u00e4t Leipzig, Linn\u00e9str. 5, 04103, Leipzig, Germany\n\nMartin Fr\u00e4nzl\u00a0&\u00a0Frank Cichos\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nM.F. and F.C. designed the experiments. M.F. performed the experiments. M.F. and F.C. analyzed the experimental data. M.F. and F.C. implemented and evaluated the numerical calculations. M.F. and F.C. wrote the manuscript. All authors discussed the results and commented on the manuscript.\n\nCorrespondence to\n                Frank Cichos.",
+            "section_image": []
+        },
+        {
+            "section_name": "Ethics declarations",
+            "section_text": "The authors declare no competing interests.",
+            "section_image": []
+        },
+        {
+            "section_name": "Peer review",
+            "section_text": "Nature Communications thanks Justus Ndukaife, and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available",
+            "section_image": []
+        },
+        {
+            "section_name": "Additional information",
+            "section_text": "Publisher\u2019s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.",
+            "section_image": []
+        },
+        {
+            "section_name": "Rights and permissions",
+            "section_text": "Open Access  This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article\u2019s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article\u2019s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.\n\nReprints and permissions",
+            "section_image": []
+        },
+        {
+            "section_name": "About this article",
+            "section_text": "Fr\u00e4nzl, M., Cichos, F. Hydrodynamic manipulation of nano-objects by optically induced thermo-osmotic flows.\n                    Nat Commun 13, 656 (2022). https://doi.org/10.1038/s41467-022-28212-z\n\nDownload citation\n\nReceived: 06 September 2021\n\nAccepted: 10 January 2022\n\nPublished: 03 February 2022\n\nVersion of record: 03 February 2022\n\nDOI: https://doi.org/10.1038/s41467-022-28212-z\n\nAnyone you share the following link with will be able to read this content:\n\nSorry, a shareable link is not currently available for this article.\n\n\n\n\n                            Provided by the Springer Nature SharedIt content-sharing initiative\n                        ",
+            "section_image": [
+                "https://data:image/svg+xml;base64,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"
+            ]
+        },
+        {
+            "section_name": "This article is cited by",
+            "section_text": "Nature Photonics (2025)\n\nNature Communications (2025)\n\nNature Communications (2025)\n\nCommunications Physics (2025)\n\nNature Communications (2024)",
+            "section_image": []
+        }
+    ]
+}

0a12712672277ba47bb32c7ead3c46d22335c62c595bf9b51a8e924c159b8d24/metadata.json

@@ -0,0 +1,144 @@
+{
+    "title": "Volumetric extrusive rates of silicic supereruptions from the Afro-Arabian large igneous province",
+    "pre_title": "Magma flux of silicic supereruptions from the Afro-Arabian large igneous province",
+    "journal": "Nature Communications",
+    "published": "02 November 2021",
+    "supplementary_0": [
+        {
+            "label": "Supplementary Information",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-021-26468-5/MediaObjects/41467_2021_26468_MOESM1_ESM.pdf"
+        },
+        {
+            "label": "Peer Review File",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-021-26468-5/MediaObjects/41467_2021_26468_MOESM2_ESM.pdf"
+        },
+        {
+            "label": "Description of Additional Supplementary Files",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-021-26468-5/MediaObjects/41467_2021_26468_MOESM3_ESM.pdf"
+        },
+        {
+            "label": "Supplementary Data 1",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-021-26468-5/MediaObjects/41467_2021_26468_MOESM4_ESM.xlsx"
+        }
+    ],
+    "supplementary_1": NaN,
+    "supplementary_2": NaN,
+    "source_data": [],
+    "code": [],
+    "subject": [
+        "Geochemistry",
+        "Volcanology"
+    ],
+    "license": "http://creativecommons.org/licenses/by/4.0/",
+    "preprint_pdf": "https://www.researchsquare.com/article/rs-469569/v1.pdf?c=1637612865000",
+    "research_square_link": "https://www.researchsquare.com//article/rs-469569/v1",
+    "nature_pdf": "https://www.nature.com/articles/s41467-021-26468-5.pdf",
+    "preprint_posted": "12 May, 2021",
+    "research_square_content": [
+        {
+            "section_name": "Abstract",
+            "section_text": "The Main Silicics phase of the Afro-Arabian large igneous province preserves some of the largest volcanic eruptions on Earth, with six units totaling >8,600 km3 dense rock equivalent (DRE). The large volumes of rapidly emplaced individual eruptions present a case study for examining the tempo of generation and emplacement of voluminous silicic magmas. We use high-precision 206Pb/238U zircon dating to differentiate individual eruption ages and show that the largest sequentially dated eruptions occurred within a timeframe of 48 \u00b1 34 kyr (29.755 \u00b1 0.023 Ma to 29.707 \u00b1 0.025 Ma), yielding a maximum magma flux of 3.09 x 10-1 km3/yr for 4,339 km3 DRE and making this sequence the highest known flux of silicic volcanism on Earth. The Main Silicics phase of volcanism occurred within a timeframe of 130 \u00b1 150 kyr (29.80 \u00b1 0.80 Ma to 29.67 \u00b1 0.13 Ma), yielding a maximum magma flux of 3.05 x 10-2 km3/yr. We also provide a robust tie-point for calibration of the geomagnetic polarity timescale by integrating recalculated 40Ar/39Ar data with our high-precision 206Pb/238U ages to yield new constraints on the duration of the C11n.1r Subchron.VolcanologyGeochemistryvolcanosearth sciencesilicic supereruptionsmagma",
+            "section_image": []
+        },
+        {
+            "section_name": "Figures",
+            "section_text": "Figure 1Figure 2Figure 3Figure 4",
+            "section_image": [
+                "https://assets-eu.researchsquare.com/files/rs-469569/v1/dfc3fa4875f3af5408df32ea.jpg%3FmaxDims%3D150x150&w=256&q=75.png",
+                "https://assets-eu.researchsquare.com/files/rs-469569/v1/d55c6860ebb45094548b493b.jpg%3FmaxDims%3D150x150&w=256&q=75.png",
+                "https://assets-eu.researchsquare.com/files/rs-469569/v1/5f25dfb8ab073b6d812cca84.jpg%3FmaxDims%3D150x150&w=256&q=75.png",
+                "https://assets-eu.researchsquare.com/files/rs-469569/v1/b56589271fb0dd750492b649.jpg%3FmaxDims%3D150x150&w=256&q=75.png"
+            ]
+        },
+        {
+            "section_name": "Full Text",
+            "section_text": "Due to technical limitations, full-text HTML conversion of this manuscript could not be completed. However, the manuscript can be downloaded and accessed as a PDF.",
+            "section_image": []
+        },
+        {
+            "section_name": "Additional Declarations",
+            "section_text": "There is NO Competing Interest.",
+            "section_image": []
+        },
+        {
+            "section_name": "Supplementary Files",
+            "section_text": "SupplementaryInformation1.docxSupplementary Information 1SupplementaryInformation2.xlsxSupplementary Information 2SupplementaryInformation3.docxSupplementary Information 3",
+            "section_image": []
+        }
+    ],
+    "nature_content": [
+        {
+            "section_name": "Abstract",
+            "section_text": "The main phase of silicic volcanism from the Afro-Arabian large igneous province preserves some of the largest volcanic eruptions on Earth, with six units totaling >8,600 km3 dense rock equivalent (DRE). The large volumes of rapidly emplaced individual eruptions present a case study for examining the tempo of voluminous silicic magma generation and emplacement. Here were report high-precision 206Pb/238U zircon ages and show that the largest sequentially dated eruptions occurred within 48\u2009\u00b1\u200934 kyr (29.755\u2009\u00b1\u20090.023\u2009Ma to 29.707\u2009\u00b1\u20090.025\u2009Ma), yielding the highest known long-term volumetric extrusive rate of silicic volcanism on Earth. While these are the largest known sequential silicic supereruptions, they did not cause major global environmental change. We also provide a robust tie-point for calibration of the geomagnetic polarity timescale by integrating 40Ar/39Ar data with our 206Pb/238U ages to yield new constraints on the duration of the C11n.1r Subchron.",
+            "section_image": []
+        },
+        {
+            "section_name": "Introduction",
+            "section_text": "Many of the largest silicic eruptions on Earth occur in large igneous provinces (LIPs), with total eruptive volumes often exceeding 1000\u2009km3 dense rock equivalent (DRE) for individual events (e.g., ~132\u2009Ma Paran\u00e1-Etendeka, ~30\u2009Ma Afro-Arabia, ~1.6\u2009Ga Gawler Range), which are likely to be emplaced in rapid succession1,2,3. Although LIPs are generally considered to represent the most productive magmatic systems on Earth4, uncertainty about volume estimates and imprecise or inaccurate age data for individual events preclude robust estimates of magma flux and volcanic output1,5. The silicic component of LIPs is largely understudied relative to their mafic counterpart and long-term rates of silicic magma generation have important implications for the energy budget and thermal evolution of the Earth. The Northern Yemen section of the Afro-Arabian LIP is an ideal testbed for using high-precision 206Pb/238U zircon dating to quantify the long-term volumetric extrusive rate of a series of flood volcanic eruptions, with three silicic supereruptions (1015\u2009kg or ~450\u2009km3 of magma6,7) occurring within a 70\u2013310\u2009kyr timeframe at ca. 29.7\u2009Ma8,9,10.\n\nOligocene volcanism in Northern Yemen (Fig.\u00a01) has been divided into three phases based on field observations, whole-rock geochemical correlations, and 40Ar/39Ar dating8,9,10: Main Basalts (31\u201329.7\u2009Ma), Main Silicics (29.7\u201329.5\u2009Ma), and Upper Bimodal (29.6\u201327.7\u2009Ma). The Main Basalts phase is characterized by effusive basaltic volcanism and volumetrically represents 60\u201370% of the total erupted volume of Afro-Arabian lavas9,11. The Main Silicics phase saw the rapid emplacement of seven silicic pyroclastic units and the Upper Bimodal phase includes small-volume basaltic and rhyolitic eruptions9. The Northern Yemen section has excellent exposure and well-characterized stratigraphic relationships from field mapping, paleomagnetic studies, and correlations with distal deep sea co-ignimbrite ash deposits8,9,12, while the Ethiopian section has been extensively faulted from active rifting with significant erosion around the volcanic plateau margin13.\n\na Schematic volcanic stratigraphy and paleomagnetic sampling profiles of volcanic units emplaced during Oligocene bimodal volcanism in Northern Yemen (modified from Ukstins Peate et al.9). Unit thicknesses and lithologies are from Ukstins Peate et al.9 and paleomagnetic data are from Riisager et al.8. Section abbreviations, from west to east, are: Esc Escarpment, BM Bayt Mawjan, A Section A, BB Bayt Baws, JS Jabal Shahirah, SK Shibam Kawkabam, WD Wadi Dhar, JK Jabal Kura\u2019a. \u0399gnimbrite is abbreviated as Ig. Sites are annotated with magnetic polarity data8 where white and black are reverse and normal polarity, respectively. Sites outlined in boxes denote those dated by 40Ar/39Ar (refs. 8,10,13) or 206Pb/238U geochronology (data presented here) and ages are shown in detail in Fig.\u00a02. Ages and sites denoted with an asterisk (*) are from correlative units in Ethiopia10. Sampling locations are shown in b9 with the Sana\u2019a region, Yemen indicated with a star.\n\nWe focus on the Main Silicics phase, which contains some of the largest known silicic eruptions on Earth, with an estimated minimum total eruptive volume of ~8600\u2009km3 DRE emplaced in present-day Yemen and Ethiopia over a period from 29.7 to 29.5\u2009Ma1,9. Volcano\u2013stratigraphic correlations in Yemen9 suggest the emplacement of the Jabal Kura\u2019a Ignimbrite (1600\u2009km3 DRE; ~29.6\u2009Ma) and Escarpment Ignimbrite (360\u2009km3 DRE; ~29.6\u2009Ma) was followed by a brief period of subsidence and erosion and then the rapid emplacement of the Green Tuff (60\u2009km3 DRE; 40Ar/39Ar age\u2009=\u200929.59\u2009\u00b1\u20090.12\u2009Ma8; Fig.\u00a02), SAM Ignimbrite (2300\u2009km3 DRE; 40Ar/39Ar age\u2009=\u200929.47\u2009\u00b1\u20090.14\u2009Ma10, Sana\u2019a Ignimbrite (1600\u2009km3 DRE; ~29.5\u2009Ma; Fig.\u00a02), and Iftar Alkalb caldera collapse mega-breccia (2700\u2009km3 DRE; 40Ar/39Ar age\u2009=\u200929.48\u2009\u00b1\u20090.13\u2009Ma8; Fig.\u00a02). The Green Tuff has been interpreted as representing the initial airfall deposit preceding the emplacement of the SAM Ignimbrite based on the sharp upper contact between the units with no evidence of a time gap during emplacement9. These bracketed 40Ar/39Ar ages indicate that all four units, with a cumulative estimated minimum total eruptive volume of ~6700\u2009km3 DRE, were emplaced in rapid succession within a timeframe of 70\u2013310\u2009kyr8,9,10, but there are no robust estimates of magma generation rates or magma flux over this time interval.\n\nComposite stratigraphy is shown using the average thickness of each unit9.\u00a0Four units have been correlated to Indian Ocean tephra layers12 and are annotated by the colored symbols. Paleomagnetic data8 are indicated where white\u2009=\u2009reverse polarity and black\u2009=\u2009normal polarity. Dashed lines show the approximate locations of the paleomagnetic reversals in the stratigraphy. Minor Unit #4 and AMPH 2 are from different sample localities and both underlie the Bayt Mawjan Ignimbrite, but their stratigraphic order relative to each other is uncertain. Oligocene flood volcanic units overly sedimentary basement (Tawilah Group sandstone)9. Symbols for 40Ar/39Ar ages8,10,11 are colored based on polarity. The gray field highlights the 40Ar/39Ar and 206Pb/238U ages with associated uncertainties of two pulses of Afro-Arabian silicic volcanism. Error bars for 206Pb/238U and 40Ar/39Ar ages are 2\u03c3 (Supplementary Data and Supplementary Information, respectively). The Escarpment Ignimbrite, Green Tuff, SAM and Sana\u2019a Ignimbrites, and Iftar Alkalb are a set of normal to reversed polarity that encompass the duration of the C11n.1r Subchron and are compared to the GPTS of Cande and Kent14 as reported in the 2020 Geologic Time Scale42. Benthic foraminiferal \u03b418O and \u03b413C curves are from the 2020 Geologic Time Scale42.\n\nPrevious paleomagnetism and 40Ar/39Ar studies8,9 indicate that the Main Silicics phase eruptions are a set of normal to reversed polarity units that encompass the duration of the C11n.1r Subchron, although overlapping ages for individual eruptions, due to analytical uncertainties, are currently unable to distinguish between the geomagnetic polarity time scale (GPTS) of Cande and Kent14 and Huestis and Acton15. While there are several cooling events identified in the Oligocene \u03b418O and \u03b413C chemostratigraphy16,17, the uncertainties of these ages also hinder the correlation of the Afro-Arabian silicic eruptions to any isotopic perturbations. In contrast to existing 40Ar/39Ar ages, the 0.1% precision of state-of-the-art chemical abrasion thermal ionization mass spectroscopy (CA-TIMS) U-Pb ages of zircons18 can distinguish between the ages of these units outside analytical uncertainty. These new high-precision 206Pb/238U zircon ages are crucial to quantifying the rapid emplacement of voluminous Afro-Arabian silicic magmas in order to understand the transient nature of silicic supereruptions, demonstrating that these eruptions had little to no observed impact on long-term climate change, and constraining the duration of the C11n.1r Subchron.",
+            "section_image": [
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-021-26468-5/MediaObjects/41467_2021_26468_Fig1_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-021-26468-5/MediaObjects/41467_2021_26468_Fig2_HTML.png"
+            ]
+        },
+        {
+            "section_name": "Results",
+            "section_text": "Zircon crystals from the Escarpment, SAM and Sana\u2019a Ignimbrites, and Iftar Alkalb were analyzed by cathodoluminescence (CL) imaging and laser ablation inductively coupled mass spectrometry (LA-ICP-MS) in order to distinguish petrochemical populations prior to CA-TIMS dating. The Escarpment Ignimbrite contains elongate prismatic crystals (typically 50\u2013120\u2009\u03bcm in length and, rarely, up to 150\u2009\u03bcm) and smaller equant crystals (50\u201375\u2009\u03bcm in length). Some prismatic crystals have oscillatory zoning with U-rich non-luminescent cores (CL dark). The SAM Ignimbrite contains elongate prismatic crystals that are both smaller (30\u201375\u2009\u03bcm, rarely up to 125\u2009\u03bcm) and less numerous than those found in the Escarpment Ignimbrite. Few crystals have subtle oscillatory zoning and one larger crystal ~120\u2009\u03bcm in length has a non-luminescent, oscillatory zoned core with a lighter overgrowth rim. Crystals in the SAM Ignimbrite have a weakly paramagnetic behavior, likely due to abundant Fe-Ti oxide and apatite inclusions. The Sana\u2019a Ignimbrite contains small elongate prismatic crystals (30\u201375\u2009\u03bcm) with subtle to no oscillatory zoning. Zircon is abundant in Iftar Alkalb as anhedral to euhedral elongate prismatic and equant crystals that range in length from 30 to 120\u2009\u03bcm. Internal morphologies are variable in Iftar Alkalb with populations of non-luminescent and luminescent zircon crystals with no oscillatory zoning, crystals with non-luminescent cores and lighter rims, and a few crystals with strong oscillatory zoning (see\u00a0Supplementary Information for CL images).\n\nIn total, 273 laser ablation spot analyses were conducted on 79 crystals from the Escarpment Ignimbrite, 46 crystals from the SAM Ignimbrite, 31 crystals from the Sana\u2019a Ignimbrite, and 95 crystals from Iftar Alkalb to identify xenocrysts (crystals that are several million years older than the relevant magma pulse and considered unrelated to the magma system19) and antecrysts (crystals that grew earlier and were incorporated in a later pulse19,20). The median uncertainty of a single LA-ICP-MS 206Pb/238U spot analysis is 3\u2009Ma, too imprecise to distinguish antecryst populations for this magmatic system but adequate to determine older xenocrystic zircon crystals. Every unit except the Escarpment Ignimbrite contains >10% zircon crystals with LA-ICP-MS 206Pb/238U ages >33\u2009Ma. The Sana\u2019a Ignimbrite and Iftar Alkalb contain significant proportions of older zircons (30 and 29%, respectively), although in the Sana\u2019a Ignimbrite this may be due to the low sample number (n\u2009=\u200931). There is no correlation between age and trace element (U, Th, Y, HREE) concentrations. CL dark zircon crystals in the Escarpment Ignimbrite and Iftar Alkalb have among the highest HREE concentrations and europium anomalies (Eu/Eu*) in each respective unit and the ages of the cores and rims of the few zircon crystals with clear zonation were indistinguishable outside uncertainty (Supplementary Data). The evolution of Eu/Eu* in zircons from the Escarpment, SAM and Sana\u2019a Ignimbrites, and Iftar Alkalb requires 50\u201360% fractional crystallization of feldspar to produce the observed range of zircon compositions. These findings are consistent with previous modeling of whole-rock compositions of ash shards from correlated deep-sea tephras (Fig.\u00a02), which required a minimum of 60% fractional crystallization of plagioclase, anorthoclase, augite, magnetite, and ilmenite to generate the observed compositional variation12.\n\nThirty-two grains that showed no sign of inclusions and yielded consistent U-Pb laser ablation dates were plucked from their respective grain mounts for high-precision CA-ID-TIMS geochronology (Supplementary Data). Preference was given to zircon crystals that captured the full range of compositions found in each unit. Six zircon crystals from the Escarpment Ignimbrite yielded a weighted mean 206Pb/238U date of 29.755\u2009\u00b1\u20090.023\u2009Ma (mean squared weighted deviation (MSWD)\u2009=\u20090.62; Fig.\u00a03). Excluding the oldest zircon crystal from the SAM Ignimbrite (which was older than the 206Pb/238U age of the underlying unit and inferred to be an antecryst), the remaining eight zircon crystals yielded a weighted mean date of 29.728\u2009\u00b1\u20090.017\u2009Ma (MSWD\u2009=\u20090.34). Six zircon crystals from the Sana\u2019a Ignimbrite yielded a weighted mean date of 29.707\u2009\u00b1\u20090.025\u2009Ma (MSWD\u2009=\u20090.65; Fig.\u00a03), excluding three zircon crystals older than 29.745\u2009Ma, also inferred to be antecrysts. The weighted mean 206Pb/238U dates have been interpreted as the eruption age of each respective unit. Weighted mean dates for the SAM and Sana\u2019a Ignimbrites calculated with the older zircon crystals are 29.733\u2009\u00b1\u20090.030\u2009Ma (MSWD\u2009=\u20092.40) and 29.793\u2009\u00b1\u20090.042 (MSWD\u2009=\u20098.96), respectively.\n\na The gray field highlights the ages and associated uncertainties (2\u03c3) of the Escarpment Ignimbrite, Green Tuff, SAM and Sana\u2019a Ignimbrites, and Iftar Alkalb. Ranked single-zircon and 206Pb/238U dates are shown for the Escarpment, SAM, and Sana\u2019a Ignimbrites. Horizontal gray bars outlined in black indicate the weighted mean 206Pb/238U ages with 95% confidence interval. Error bars for 40Ar/39Ar ages are 2\u03c3 (Supplementary Information). b Minimum total eruptive volume DRE (km3) values are from on-land and correlated deep-sea tephra layers found in Ocean Drilling Program cores from the Indian Ocean, Leg 1151,9,12.\n\nAlthough Iftar Alkalb is the stratigraphically youngest unit dated, nine zircon crystals were consistently older (29.731\u2009\u00b1\u20090.089\u201330.320\u2009\u00b1\u20090.094\u2009Ma; Fig.\u00a03) than the weighted mean ages of the other units and so no date was assigned. We attribute this to the emplacement mechanism of the caldera collapse breccia with abundant mega-clasts of underlying stratigraphy contributing xenolithic material or antecrysts that are recording an earlier stage of zircon crystallization. Zircon morphologies (Supplementary Information) and compositions (Fig.\u00a04 and Supplementary Data) were highly variable for Iftar Alkalb and further work is necessary to evaluate these complexities.\n\na\u00a0Zircon crystals are\u00a0denoted by age, dating method, and inclusion in final age calculations.\u00a0Zircons >33\u2009Ma (from preliminary LA-ICP-MS dating, average 2\u03c3 uncertainty \u00b13\u2009Ma) are denoted by diamond symbols. Non-luminescent (CL-dark) zircon crystals from the Escarpment Ignimbrite and Iftar Alkalb are denoted by black symbols. a\u2013e show Th/Y versus Eu/Eu* in detail for the Escarpment Ignimbrite (b), SAM Ignimbrite (c), Sana\u2019a Ignimbrite (d), and Iftar Alkalb (e). Error bars in b\u2013e are 2\u03c3 (Supplementary Data).\n\nSanidine from the Green Tuff, SAM Ignimbrite, and Iftar Alkalb were previously dated via the 40Ar/39Ar method8,10. Those dates have been recalculated using a 28.201\u2009Ma monitor age for the Fish Canyon Tuff sanidine21. Recalculations (Supplementary Information) yield a 29.78\u2009\u00b1\u20090.12\u2009Ma age for the Green Tuff, 29.66\u2009\u00b1\u20090.14\u2009Ma age for the SAM Ignimbrite, and 29.67\u2009\u00b1\u20090.08\u2009Ma age for Iftar Alkalb (Fig.\u00a02). Previous 40Ar/39Ar ages8,10,11 from the Shibam Kawkabam Ignimbrite (30.35\u2009\u00b1\u20090.13\u2009Ma), Kura\u2019a Basalt (30.22\u2009\u00b1\u20090.26\u2009Ma), Akraban Andesite (29.80\u2009\u00b1\u20090.08\u2009Ma), an overlying small-volume rhyolitic tuff (28.58\u2009\u00b1\u20090.14\u2009Ma) and ignimbrite (28.18\u2009\u00b1\u20090.10\u2009Ma), and the Bayt Mawjan Ignimbrite (27.85\u2009\u00b1\u20090.12\u2009Ma) have also been recalculated. The 206Pb/238U zircon ages are in agreement with the recalculated 40Ar/39Ar ages and are compiled and presented here as a revised chronostratigraphy of the Northern Yemen flood volcanics (Fig.\u00a02).",
+            "section_image": [
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-021-26468-5/MediaObjects/41467_2021_26468_Fig3_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-021-26468-5/MediaObjects/41467_2021_26468_Fig4_HTML.png"
+            ]
+        },
+        {
+            "section_name": "Discussion",
+            "section_text": "Elements that are normally incompatible during magma differentiation (e.g., U, Nb, Th, Y, and Hf) and the europium anomaly (Eu/Eu*) in rare earth element patterns resulting from feldspar fractionation are useful indicators of magma differentiation. Assuming that both elements remain incompatible, more differentiated rhyolites will evolve towards higher Th/Y ratios while Eu/Eu* will decrease with continued feldspar crystallization22. With a few exceptions, zircons dated via CA-TIMS for these units show the same trend: the least evolved zircon with the highest Eu/Eu* and lowest Th/Y values are older than the most evolved zircon by 0.01\u2009\u00b1\u20090.16\u2009Ma in the Escarpment Ignimbrite, 0.02\u2009\u00b1\u20090.09\u2009Ma in the SAM Ignimbrite, and 0.07\u2009\u00b1\u20090.17\u2009Ma in the Sana\u2019a Ignimbrite (Fig.\u00a04). Thus, ages for zircon crystals spanning the full geochemical ranges are statistically indistinguishable, suggesting that these large volume magmas were rapidly differentiated within 103\u2013104 years once the magmas reached Zr saturation. These estimates are based on LA-ICP-MS single-spot analyses and whole-grain CA-TIMS zircon ages because the small crystal sizes and presence of mineral inclusions made multiple spot analyses difficult. Eu/Eu* and Th/Y are not correlated for zircons in Iftar Alkalb and there is no age relationship between the most and least evolved zircons (Fig.\u00a04), further supporting that the zircons in Iftar Alkalb are of a mixed xenolithic or antecrystic origin.\n\nMagma flux rates (km3/yr) were calculated for 100 and 400\u2009kyr of residence for the Escarpment, SAM, and Sana\u2019a Ignimbrites based on the age difference between the most and least evolved zircon crystals in each unit. For 100\u2009kyr residence, magma flux rates are 3.6\u2009\u00d7\u200910\u22123\u2009km3/yr, 2.4\u2009\u00d7\u200910\u22122\u2009km3/yr, and 1.6\u2009\u00d7\u200910\u22122\u2009km3/yr for the Escarpment, SAM, and Sana\u2019a Ignimbrites, respectively. For 400\u2009kyr residence, magma flux rates are 9.0\u2009\u00d7\u200910\u22124\u2009km3/yr, 6.0\u2009\u00d7\u200910\u22123\u2009km3/yr, and 4.0\u2009\u00d7\u200910\u22123\u2009km3/yr for the Escarpment, SAM, and Sana\u2019a Ignimbrites, respectively. Upper estimates of 3.6\u2009\u00d7\u200910\u22123\u20132.4\u2009\u00d7\u200910\u22122\u2009km3/yr for 100\u2009kyr residence are similar to those calculated for other rapidly assembled large-volume silicic systems (e.g., Yellowstone supereruptions23,24 and Fish Canyon Tuff25). The most conservative estimates using 400\u2009kyr residence (9.0\u2009\u00d7\u200910\u22124\u20136.0\u2009\u00d7\u200910\u22124\u2009km3/yr) are similar to but lower than the minimum calculated magma flux from Yellowstone (2.8\u2009\u00d7\u200910\u22123\u2009km3/yr for the 280\u2009km3 Mesa Falls Tuff23) and significantly lower than that of Taupo (>0.33\u2009km3/yr for the 530\u2009km3 Oruanui eruption26).\n\nU-Pb zircon dating shows that three sequential eruptions of Afro-Arabian silicic volcanics\u2014the Escarpment Ignimbrite, the Green Tuff and SAM Ignimbrite, and Sana\u2019a Ignimbrite\u2014were collectively emplaced within a timespan of 48\u2009\u00b1\u200934\u2009kyr (calculated using the square root of the sum of the uncertainties), yielding a long-term volumetric extrusive rate of 5.27\u2009\u00d7\u200910\u22122\u20133.08\u2009\u00d7\u200910\u22121\u2009km3/yr for 4320\u2009km3 DRE. The estimated minimum total eruptive volume for the entirety of the Main Silicics phase is 8620\u2009km3 DRE over a duration of 130\u2009\u00b1\u2009150\u2009kyr, constrained by the ages of the Akraban Andesite and Iftar Alkalb, which yield a lower extrusive rate of 3.05\u2009\u00d7\u200910\u22122\u20136.63\u2009\u00d7\u200910\u22122\u2009km3/yr. Northern Yemen unit volume estimates1,9 are minimum values accounting for the lateral distribution and measured thickness in the studied field areas (Fig.\u00a01) and correlations to Indian Ocean deep-sea tephra layers located >2700\u2009km away from Yemen12. While there was wide-scale silicic volcanism following the termination of the main pulse of flood basalt emplacement13,27, unit volume estimates outside of Northern Yemen remain sparse. Extrusive rates for other regions of the Afro-Arabian province, such as the Ethiopian stratigraphy, are difficult to constrain due to pervasive post-emplacement faulting. Notably, a series of silicic supereruptions in the Tana Basin, Ethiopia28 have recently been dated at 31.108\u2009\u00b1\u20090.020\u201330.844\u2009\u00b1\u20090.027\u2009Ma with an estimated minimum eruptive volume of 2000\u20133000\u2009km3, corresponding to a long-term volumetric extrusive rate of 0.8\u20131.1\u2009\u00d7\u200910\u22122\u2009km3/yr.\n\nLong-term volumetric extrusive rates of basaltic and andesitic systems are thought to be higher than those of silicic systems by up to two orders of magnitude4. Average extrusive rates in silicic systems are calculated to be highest for continental arcs (4.90\u2009\u00b1\u20090.15\u2009\u00d7\u200910\u22123\u2009km3/yr) followed by oceanic arcs (4.50\u2009\u00b1\u20090.79\u2009\u00d7\u200910\u22123\u2009km3/yr), continental rifts (4.48\u2009\u00b1\u20090.86\u2009\u00d7\u200910\u22123\u2009km3/yr), continental hotspots (1.29\u2009\u00b1\u20090.25\u2009\u00d7\u200910\u22123\u2009km3/yr), and continental volcanic fields (6.47\u2009\u00b1\u20091.96\u2009\u00d7\u200910\u22124\u2009km3/yr). The extrusive rate of the Main Silicics phase of the Northern Yemen section of the Afro-Arabian province is most similar to\u2014but notably higher than\u2014the extrusive rates of the central Taupo volcanic zone (1.28\u2009\u00d7\u200910\u22122\u2009km3/yr; ref. 29), the silicic portion of Kamchatka (1.05\u2009\u00d7\u200910\u22122\u2009km3/yr; ref. 30), and Quaternary phonolites from the Kenya rift valley (1.20\u2009\u00d7\u200910\u22122\u2009km3/yr; ref. 31). Our findings are consistent with observations at other large-volume silicic systems that record rapid periods of differentiation and magma reservoir assembly superimposed on lower background fluxes. While some silicic systems have produced more voluminous individual eruptions (e.g., Fish Canyon Tuff with 4500\u2009km3 DRE32) and larger cumulative eruptive volumes over longer time intervals (e.g., Paran\u00e1-Etendeka LIP with 20,000\u201335,000\u2009km3 over 6\u2009Myr33,34), the eruptions of the Main Silicics phase in Northern Yemen represent the largest long-term volumetric extrusive rate of silicic volcanism on Earth.\n\nSome volcanic provinces appear to coincide with major global environmental change and mass extinctions (e.g., Siberian Traps, Karoo-Ferrar, Emieshan, and Central Atlantic LIPs), yet others, even those with silicic supereruptions (e.g., Paran\u00e1-Etendeka LIP), do not35. Models for volcanism-driven environmental change predict years of cooling from SO2 injection into the stratosphere from a single eruption and/or tens of thousands of years of warming from CO2 emissions36. Several of the Afro-Arabian silicic supereruptions have been correlated to 10\u201315\u2009cm-thick tephra layers located >2700\u2009km away in the Indian Ocean12 (Fig.\u00a02), suggesting volcanic fallout on a near-global scale. However, the timing of these supereruptions in relation to several Rupelian-aged cooling events that have been identified in Chrons C12 (Oi1a, Oi1b, and Oi237,38) and C10 (Oi2* and Oi2a37,38) indicate that the perturbations in \u03b418O and \u03b413C pre-date the eruptions16,17 (Fig.\u00a02). Other silicic supereruptions, such as the ~31\u2009Ma caldera-forming eruptions in the Tana Basin28 and ~28\u2009Ma eruption of the Fish Canyon Tuff32, likewise do not coincide with global cooling events. The correlation between volcanic eruptions and isotopic perturbations rely on the precision of the eruption ages, resolution of the climate proxy data (\u00b10.2\u2030)16, and the sensitivity of the climate proxies to the effects of individual volcanic eruptions. While the Afro-Arabian Main Silicics phase eruptions represent the largest known long-term volumetric extrusive rate of silicic volcanism, they did not cause major global climate change at the current resolution of these data. Challenges remain in discerning the various roles of the tempo, volatile budget, eruption mechanism, and volume of magma extruded from LIPs and their effect on global environmental change. However, robust temporal constraints continue to provide critical insight into this relationship.\n\nPrevious efforts have been made to correlate Oligocene Afro-Arabian volcanic deposits with the GPTS8,39 but those were unable to unambiguously distinguish between the GPTS of Cande and Kent14 and Huestis and Acton15. Recent studies on the Oligocene magnetic polarity sequence have utilized astronomical age models38, radio-isotope age models37, recalculations of the Cande and Kent14 GPTS using updated 40Ar/39Ar flux monitor ages40, and a combination of all three37. One of the lingering issues with distinguishing between an appropriate method for determining the Rupelian age (33.9\u201328.1\u2009Ma) is the lack of tie points from radio-isotopic dates. The Rupelian/Chattian boundary Global Boundary Stratotype Section and Point records a nearly continuous record of astronomically tuned magnetostratigraphy for the Oligocene but only provides one tie point for the Rupelian for the uppermost Chron C12r with a gap between 31.8\u2009\u00b1\u20090.2 and 27.0\u2009\u00b1\u20090.1\u2009Ma37,41. The 2012 Geologic Time Scale for the Paleogene37 favored an integrated radio-isotope, GPTS, and cyclostratigraphy model with sixth-order polynomial fit to produce a complete C-sequence. The C11n.1r Subchron is estimated to have a duration of 0.050\u2009Ma with a \u22120.654\u2009Ma discrepancy between radio-isotopic and astronomic age models37. The only discrepancy between the combined age model of the 2012 Geologic Time Scale and new 2020 Geologic Time Scale for the time range of interest is a shift of the base of Chron C12n to 30.977 from 31.034\u2009Ma37,42.\n\nWe propose that the 29.728\u2009\u00b1\u20090.017\u2009Ma 206Pb/238U zircon age of the SAM Ignimbrite and 29.67\u2009\u00b1\u20090.13\u2009Ma 40Ar/39Ar sanidine age of Iftar Alkalb\u2014further constrained to 29.67\u2009\u00b1\u20090.13\u2009Ma by the 29.707\u2009\u00b1\u20090.025\u2009Ma 206Pb/238U age of the Sana\u2019a Ignimbrite\u2014can be used as tie points for the GPTS. Our chronostratigraphy and magnetostratigraphy are definitively in agreement with the Cande and Kent14 GPTS (Fig.\u00a02). Discrepancies between our results and the 2020 Geologic Time Scale arise from the sparsity of radio-isotope dates for the Rupelian coupled with the short duration of the C11n.1r Subchron. Our findings are within the 0.654\u2009Ma discrepancy between the radio-isotopic and astronomic age models and could thus serve as robust tie points for future time scale calibrations.",
+            "section_image": []
+        },
+        {
+            "section_name": "Methods",
+            "section_text": "Samples from the Sana\u2019a area of Northern Yemen were previously collected and described in Ukstins Peate et al.9 (Fig.\u00a01). Paleomagnetic data was measured on 587 oriented drill cores collected at 71 sites8 (Fig.\u00a01). Zircon U-Pb petrochronology was undertaken at the Boise State University Isotope Laboratory. Zircon crystals from the Escarpment, SAM and Sana\u2019a Ignimbrites, and Iftar Alkalb were separated using standard magnetic and heavy liquid techniques and annealed at 900\u2009\u00b0C for 60\u2009h. Zircons were imaged using a JEOL T-300 scanning electron microscope fitted with a Gatan Mini CL detector and JEOL back-scattered electron detector under 15\u2009kV probe current and 2\u2009mA accelerating voltage operating conditions (Supplementary Information). Trace element analyses and preliminary U-Pb dating for 31\u201395 crystals per unit (Supplementary Data) were performed using a ThermoElectron X-Series II quadrupole ICP-MS and New Wave Research UP-213 Nd:YAG UV (213\u2009\u03bcm) laser ablation system with a 10\u2009Hz at 5\u2009J/cm2 pulsed laser and 15\u2009\u03bcm spot size. NIST SRM-610 and SRM-612 glasses were used as standards for trace element concentrations and Ple\u0161ovice zircon standard43 was used for U-Pb calibration. Zircon standards were measured every 10 unknowns; glass standards were analyzed at the beginning of two 109-spot cycles.\n\nA total of 32 crystals from the 4 units were selected for CA-TIMS analysis on the basis of morphology, zoning, chemistry, and preliminary 206Pb/238U dates. Zircon crystals were chemically abraded18 in 120\u2009\u03bcL of 29\u2009M hydrofluoric acid (HF) at 180\u2013200\u2009\u00b0C for 12\u2009h and then rinsed in 3.5\u2009M HNO3 in an ultrasonic bath for 60\u2009min. The residual crystals were rinsed twice in ultrapure H2O and transferred to Teflon PFA microcapsules and spiked with ET535 mixed U-Pb isotope tracer solution44,45. The spiked residual crystals were dissolved in Parr vessels in 120\u2009\u03bcL of 29\u2009M HF at 220\u2009\u00b0C for 48\u2009h, dried, and redissolved in 6\u2009M HCl at 180\u2009\u00b0C overnight46. Pb and U were purified from the chloride matrix using HCl-based anion-exchange chromatography and dried with 2\u2009\u03bcL of 0.05\u2009N H3PO4. High-precision isotope dilution U and Pb isotope ratio measurements were made using a single Re filament silica gel technique on an Isotopx Isoprobe-T multi-collector TIMS equipped with an ion-counting Daly detector (Supplementary Data). Dates are calculated using the decay constants of Jaffey et al.47. Analytical uncertainties on dates are reported to be 2\u03c3 and propagated using the algorithms of Schmitz and Schoene48.",
+            "section_image": []
+        },
+        {
+            "section_name": "Data availability",
+            "section_text": "Supplementary Information contains cathodoluminescence (CL) images of zircon crystals analyzed by LA-ICP-MS and CA-TIMS and details on the recalculation of 40Ar/39Ar ages. Supplementary Data contains details on the LA-ICP-MS trace element concentrations and 206Pb/238U dates for zircon crystals dated by CA-TIMS. The full dataset of LA-ICP-MS trace element concentrations for all zircon crystals analyzed in this study are available in the PANGAEA database. Samples collected by I.A.U. are housed at the University of Auckland.",
+            "section_image": []
+        },
+        {
+            "section_name": "References",
+            "section_text": "Bryan, S. E. et al. The largest volcanic eruptions on Earth. Earth Sci. Rev. 102, 207\u2013229 (2010).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBryan, S. E. & Ernst, R. E. Revised definition of large igneous provinces (LIPs). Earth Sci. Rev. 86, 175\u2013202 (2008).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nCoffin, M. F. & Eldholm, O. Volcanism and continental break-up: a global compilation of large igneous provinces. Geol. Soc. Lond. Spec. Publ. 68, 17\u201330 (1992).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWhite, S. M., Crisp, J. A. & Spera, F. J. Long-term volumetric eruption rates and magma budgets. Geochem. Geophys. 7, 1\u201320 (2006).\n\n\n                    Google Scholar\u00a0\n                \n\nBryan, S. E., Riley, T. E., Jerram, D. A., Stephens, C. J. & Leat, P. T. Silicic volcanism: an undervalued component of large igneous provinces and volcanic rifted margins. Geol. Soc. Am. Spec. Pap. 362, 99\u2013120 (2002).\n\n\n                    Google Scholar\u00a0\n                \n\nMason, B. E., Pyle, D. M. & Oppenheimer, C. O. The size and frequency of the largest explosive eruptions on Earth. Bull. Volcanol. 66, 735\u2013748 (2004).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSelf, S. The effects and consequences of very large explosive volcanic eruptions. Philos. Trans. R. Soc. A 364, 2073\u20132097 (2006).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nRiisager, P. et al. Paleomagnetism and 40Ar/39Ar Geochronology of Yemeni Oligocene volcanics: implications for timing and duration of Afro-Arabian traps and geometry of the Oligocene paleomagnetic field. Earth Planet. Sci. Lett. 237, 647\u2013672 (2005).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nUkstins Peate, I. et al. Volcanic stratigraphy of large-volume silicic pyroclastic eruptions during Oligocene Afro-Arabian flood volcanism in Yemen. Bull. Volcanol. 68, 135\u2013156 (2005).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nUkstins, I. A. et al. Matching conjugate volcanic rifted margins: 40Ar/39Ar chrono-stratigraphy of pre- and syn-rift bimodal flood volcanism in Ethiopia and Yemen. Earth Planet. Sci. Lett. 198, 289\u2013306 (2002).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBaker, J., Snee, L. & Menzies, M. A brief Oligocene period of flood volcanism in Yemen: implications for the duration and rate of continental flood volcanism at the Afro-Arabian triple junction. Earth Planet. Sci. Lett. 138, 39\u201355 (1996).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nUkstins Peate, I., Kent, A. J. R., Baker, J. A. & Menzies, M. A. Extreme geochemical heterogeneity in Afro-Arabian Oligocene tephras: preserving fractional crystallization and mafic recharge processes in silicic magma chambers. Lithos 102, 260\u2013278 (2008).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nRooney, T. O. The Cenozoic magmatism of East-Africa: Part I \u2014 Flood basalts and pulsed magmatism. Lithos 286\u2013287, 264\u2013301 (2017).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nCande, S. C. & Kent, D. V. Revised calibration of the geomagnetic polarity timescale for the Late Cretaceous and Cenozoic. J. Geophys. Res. 100, 6093\u20136095 (1995).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHuestis, S. P. & Acton, G. D. On the construction of geomagnetic timescales from non-prejudicial treatment of magnetic anomaly data from multiple ridges. Geophys. J. Int. 129, 176\u2013182 (1997).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nCramer, B. S., Toggweiler, J. R., Wright, J. D., Katz, M. E. & Miller, K. G. Ocean overturning since the Late Cretaceous: Inferences from a new benthic foraminiferal isotope compilation. Paleoceanography 24, PA4216 (2009).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nZachos, J., Pagani, M., Sloan, L., Thomas, E. & Billups, K. Trends, rhythms, and aberrations in global climate 65 Ma to present. Science 292, 686\u2013693 (2001).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMattinson, J. M. Zircon U\u2013Pb chemical abrasion (\u201cCA-TIMS\u201d) method: combined annealing and multi-step partial dissolution analysis for improved precision and accuracy of zircon ages. Chem. Geol. 220, 47\u201366 (2005).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMiller, J. S., Matzel, J. E. P., Miller, C. F., Burgess, S. D. & Miller, R. B. Zircon growth and recycling during the assembly of large, composite arc plutons. J. Volcanol. Geotherm. Res. 167, 282\u2013299 (2007).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHildreth, W. & Wilson, C. J. N. Compositional zoning of the Bishop Tuff. J. Petrol. 48, 951\u2013999 (2007).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKuiper, K. F. et al. Synchronizing rock clocks of Earth history. Science 320, 500\u2013504 (2008).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nRivera, T. A., Storey, M., Schmitz, M. D. & Crowley, J. L. Age intercalibration of 40Ar/39Ar sanidine and chemically distinct U/Pb zircon populations from the Alder Creek Rhyolite Quaternary geochronology standard. Chem. Geol. 345, 87\u201398 (2013).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nRivera, T. A., Schmitz, M. D., Jicha, B. R. & Crowley, J. L. Zircon petrochronology and 40Ar/39Ar sanidine dates for the Mesa Falls Tuff: crystal-scale records of magmatic evolution and the short lifespan of a large Yellowstone magma chamber. J. Petrol. 57, 1677\u20131704 (2016).\n\nCAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWotzlaw, J.-F., Bindeman, I. N., Stern, R. A., D\u2019Abzac, F.-X. & Schaltegger, U. Rapid heterogeneous assembly of multiple magma reservoirs prior to Yellowstone supereruptions. Nat. Sci. Rep. 5, 14026 (2015).\n\nADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWotzlaw, J.-F. et al. Tracking the evolution of large-volume silicic magma reservoirs from assembly to supereruption. Geology 41, 867\u2013870 (2013).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nAllan, A. S. R., Morgan, D. J., Wilson, C. J. N. & Millet, M.-A. From mush to eruption in centuries: assembly of the super-sized Oruanui magma body. Contrib. Mineral. Petrol. 166, 143\u2013164 (2013).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKrans, S. R., Rooney, T. O., Kappelman, J., Yirgu, G. & Ayalew, D. From initiation to termination: a petrostratigraphic tour of the Ethiopian Low-Ti Flood Basalt Province. Contrib. Mineral. Petrol. 173, 37 (2018).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nPrave, A. R. et al. Geology and geochronology of the Tana Basin, Ethiopia: LIP volcanism, Super eruptions and Eocene-Oligocene environmental change. Earth Planet. Sci. Lett. 443, 1\u20138 (2016).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWilson, C. J. N., Blake, S., Charlier, B. L. A. & Sutton, A. N. The 26.5 ka Oruanui eruption, Taupo Volcano, New Zealand: development, characteristics and evacuation of a large rhyolitic magma body. J. Petrol. 47, 35\u201369 (2006).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nErlich, E. N. & Volynets, O. N. Chapter 3 acid volcanism in Kamchatka. Bull. Volcanol. 42, 175\u2013254 (1979).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nCrisp, J. A. Rates of magma emplacement and volcanic output. J. Volcanol. Geotherm. Res. 20, 177\u2013211 (1984).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLipman, P., Dungan, M. & Bachmann, O. Comagmatic granophyric granite in the Fish Canyon Tuff, Colorado: implications for magma-chamber processes during a large ash-flow eruption. Geology 25, 915\u2013918 (1997).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSim\u00f5es, M. S., Lima, E. F., Rossetti, L. M. M. & Sommer, C. A. The low-Ti high-temperature dacitic volcanism of the southern Paran\u00e1-Etendeka LIP: geochemistry, implications for trans-Atlantic correlations and comparison with other Phanerozoic LIPs. Lithos 342\u2013343, 187\u2013205 (2019).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGibson, S. A., Thompson, R. N. & Day, J. A. Timescales and mechanisms of plume\u2013lithosphere interactions: 40Ar/39Ar geochronology and geochemistry of alkaline igneous rocks from the Paran\u00e1\u2013Etendeka large igneous province. Earth Planet. Sci. Lett. 251, 1\u201317 (2006).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBlack, B. A., Weiss, B. P., Elkins-Tanton, L. T., Veselovskiy, R. V. & Latyshev, A. Siberian Traps volcaniclastic rocks and the role of magma-water interactions. Geol. Soc. Am. Bull. 127, 1437\u20131452 (2015).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSelf, S., Schmidt, A. & Mather, T. A. Emplacement characteristics, time scales, and volcanic gas release rates of continental flood basalt eruptions on Earth. Geol. Soc. Am. Spec. Pap. 505, 319\u2013337 (2014).\n\nVandenberghe, N. et al. in The Geologic Time Scale (eds Gradstein, F. M., Ogg, J. G., Schmitz, M. D. & Ogg, G. M.) 855\u2013921 (Elsevier, 2012).\n\nP\u00e4like, H. et al. The heartbeat of the Oligocene climate system. Science 314, 1894\u20131898 (2006).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nRochette, P. et al. Magnetostratigraphy and timing of the Oligocene Ethiopian traps. Earth Planet. Sci. Lett. 164, 497\u2013510 (1998).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGuidry, E. P. et al. Oligocene\u2013Miocene magnetostratigraphy of deep-sea sediments from the equatorial Pacific (IODP Site U1333). Geol. Soc. Lond. Spec. Pub. 373, 13\u201327 (2013).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nCoccioni, R. et al. Integrated stratigraphy of the Oligocene pelagic sequence in the Umbria-Marche basin (northeastern Apennines, Italy): a potential Global Stratotype Section and Point (GSSP) for the Rupelian/Chattian boundary. Geol. Soc. Am. Bull. 120, 487\u2013511 (2008).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSpeijer, R. P., P\u00e4like, H., Hollis, C. J., Hooker, J. J. & Ogg, J. G. in The Geologic Time Scale (eds Gradstein, F. M., Ogg, J. G., Schmitz, M. D. & Ogg, G. M.) 1087\u20131140 (Elsevier, 2020).\n\nSl\u00e1ma, J. et al. Ple\u0161ovice zircon \u2013 a new natural reference material for U-Pb and Hf isotopic microanalysis. Chem. Geol. 249, 1\u201335 (2008).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nCondon, D. J., Schoene, B., McLean, N. M., Bowring, S. A. & Parrish, R. R. Metrology and traceability of U-Pb isotope dilution geochronology (EARTHTIME Tracer Calibration Part I). Geochim. Cosmochim. Acta 164, 464\u2013480 (2015).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMcLean, N. M., Condon, D. J., Schoene, B. & Bowring, S. A. Evaluating uncertainties in the calibration of isotopic reference materials and multi-element isotopic tracers (EARTHTIME Tracer Calibration Part II). Geochim. Cosmochim. Acta 164, 481\u2013501 (2015).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMacdonald, F. A. et al. Cryogenian of Yukon. Precambrian Res. 319, 114\u2013143 (2017).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nJaffey, A. H., Flynn, K. F., Glendenin, L. E., Bentley, W. C. & Essling, A. M. Precision measurement of half-lives and specific activities of U 235 and U 238. Phys. Rev. C 4, 1889\u20131906 (1971).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSchmitz, M. D. & Schoene, B. Derivation of isotope ratios, errors, and error correlations for U-Pb geochronology using 205Pb-235U-(233U)-spiked isotope dilution thermal ionization mass spectrometric data. Geochem. Geophys. 8, 1\u201320 (2007).\n\nADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nDownload references",
+            "section_image": []
+        },
+        {
+            "section_name": "Acknowledgements",
+            "section_text": "This material is based on work supported by the National Science Foundation under Grant Nos. EAR-1759200 and EAR-1759353. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation. We thank the AGeS program, members of the Boise State University Isotope Geology Laboratory for support with sample preparation and B.D. Cramer for insightful discussions.",
+            "section_image": []
+        },
+        {
+            "section_name": "Author information",
+            "section_text": "Department of Earth and Environmental Sciences, University of Iowa, 115 Trowbridge Hall, Iowa City, IA, 52242, USA\n\nJennifer E. Thines\n\nSchool of Environment, The University of Auckland, Private Bag 92 019, Auckland, New Zealand\n\nIngrid A. Ukstins\n\nDepartment of Geosciences, Boise State University, 1295 University Drive, Boise, ID, 83706, USA\n\nCorey Wall\u00a0&\u00a0Mark Schmitz\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nJ.E.T., I.A.U. and M.S. designed the research project as part of the AGeS2 Geochronology Program. Sample material was provided by I.A.U. C.W. and J.E.T. prepared samples and analyzed the data with help from M.S. J.E.T. wrote the manuscript with support from I.A.U. and M.S. Figure\u00a01b is modified from Ukstins Peate et al.9 with additional data from Riisager et al.8 and used with permission by I.A.U. Progress was overseen by I.A.U, the PhD thesis advisor of J.E.T.\n\nCorrespondence to\n                Jennifer E. Thines.",
+            "section_image": []
+        },
+        {
+            "section_name": "Ethics declarations",
+            "section_text": "The authors declare no competing interests.",
+            "section_image": []
+        },
+        {
+            "section_name": "Additional information",
+            "section_text": "Peer review information Nature Communications thanks Simon Barker and Anthony Prave for their contribution to the peer review of this work. Peer reviewer reports are available.\n\nPublisher\u2019s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.",
+            "section_image": []
+        },
+        {
+            "section_name": "Rights and permissions",
+            "section_text": "Open Access  This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article\u2019s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article\u2019s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.\n\nReprints and permissions",
+            "section_image": []
+        },
+        {
+            "section_name": "About this article",
+            "section_text": "Thines, J.E., Ukstins, I.A., Wall, C. et al. Volumetric extrusive rates of silicic supereruptions from the Afro-Arabian large igneous province.\n                    Nat Commun 12, 6299 (2021). https://doi.org/10.1038/s41467-021-26468-5\n\nDownload citation\n\nReceived: 27 April 2021\n\nAccepted: 04 October 2021\n\nPublished: 02 November 2021\n\nVersion of record: 02 November 2021\n\nDOI: https://doi.org/10.1038/s41467-021-26468-5\n\nAnyone you share the following link with will be able to read this content:\n\nSorry, a shareable link is not currently available for this article.\n\n\n\n\n                            Provided by the Springer Nature SharedIt content-sharing initiative\n                        ",
+            "section_image": [
+                "https://data:image/svg+xml;base64,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"
+            ]
+        }
+    ]
+}

0a57595338b1390798459ee5d63f40d3108931b635b32aecd97babb2ece11f54/metadata.json

@@ -0,0 +1,205 @@
+{
+    "title": "An ancestral interaction module promotes oligomerization in divergent mitochondrial ATP synthases",
+    "pre_title": "An ancestral interaction module promotes oligomerization in divergent mitochondrial ATP synthases",
+    "journal": "Nature Communications",
+    "published": "11 October 2022",
+    "supplementary_0": [
+        {
+            "label": "Supplementary Information",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-022-33588-z/MediaObjects/41467_2022_33588_MOESM1_ESM.pdf"
+        },
+        {
+            "label": "Peer Review File",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-022-33588-z/MediaObjects/41467_2022_33588_MOESM2_ESM.pdf"
+        },
+        {
+            "label": "Description of Additional Supplementary Files",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-022-33588-z/MediaObjects/41467_2022_33588_MOESM3_ESM.pdf"
+        },
+        {
+            "label": "Supplementary Movie 1",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-022-33588-z/MediaObjects/41467_2022_33588_MOESM4_ESM.mp4"
+        },
+        {
+            "label": "Supplementary Movie 2",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-022-33588-z/MediaObjects/41467_2022_33588_MOESM5_ESM.mov"
+        },
+        {
+            "label": "Supplementary Movie 3",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-022-33588-z/MediaObjects/41467_2022_33588_MOESM6_ESM.mp4"
+        },
+        {
+            "label": "Reporting Summary",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-022-33588-z/MediaObjects/41467_2022_33588_MOESM7_ESM.pdf"
+        }
+    ],
+    "supplementary_1": [
+        {
+            "label": "Source Data",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-022-33588-z/MediaObjects/41467_2022_33588_MOESM8_ESM.zip"
+        }
+    ],
+    "supplementary_2": NaN,
+    "source_data": [
+        "https://www.rcsb.org/structure/8AP7",
+        "https://www.rcsb.org/structure/8AP8",
+        "https://www.rcsb.org/structure/8AP9",
+        "https://www.rcsb.org/structure/8AP6",
+        "https://www.rcsb.org/structure/8APA",
+        "https://www.rcsb.org/structure/8APB",
+        "https://www.rcsb.org/structure/8APC",
+        "https://www.rcsb.org/structure/8APD",
+        "https://www.rcsb.org/structure/8APE",
+        "https://www.rcsb.org/structure/8APF",
+        "https://www.rcsb.org/structure/8APG",
+        "https://www.rcsb.org/structure/8APH",
+        "https://www.rcsb.org/structure/8APJ",
+        "https://www.rcsb.org/structure/8APK",
+        "https://www.ebi.ac.uk/emdb/EMD-15560",
+        "https://www.ebi.ac.uk/emdb/EMD-15561",
+        "https://www.ebi.ac.uk/emdb/EMD-15562",
+        "https://www.ebi.ac.uk/emdb/EMD-15559",
+        "https://www.ebi.ac.uk/emdb/EMD-15563",
+        "https://www.ebi.ac.uk/emdb/EMD-15564",
+        "https://www.ebi.ac.uk/emdb/EMD-15565",
+        "https://www.ebi.ac.uk/emdb/EMD-15566",
+        "https://www.ebi.ac.uk/emdb/EMD-15567",
+        "https://www.ebi.ac.uk/emdb/EMD-15568",
+        "https://www.ebi.ac.uk/emdb/EMD-15570",
+        "https://www.ebi.ac.uk/emdb/EMD-15571",
+        "https://www.ebi.ac.uk/emdb/EMD-15572",
+        "https://www.ebi.ac.uk/emdb/EMD-15573",
+        "https://www.rcsb.org/structure/6TDU",
+        "https://www.rcsb.org/structure/6TDV",
+        "https://www.rcsb.org/structure/6B2Z",
+        "https://www.rcsb.org/structure/6F5D",
+        "https://www.rcsb.org/structure/6ZNA",
+        "/articles/s41467-022-33588-z#Sec22"
+    ],
+    "code": [],
+    "subject": [
+        "Cryoelectron microscopy",
+        "Molecular evolution"
+    ],
+    "license": "http://creativecommons.org/licenses/by/4.0/",
+    "preprint_pdf": "https://www.researchsquare.com/article/rs-1196040/v1.pdf?c=1665486645000",
+    "research_square_link": "https://www.researchsquare.com//article/rs-1196040/v1",
+    "nature_pdf": "https://www.nature.com/articles/s41467-022-33588-z.pdf",
+    "preprint_posted": "30 Dec, 2021",
+    "research_square_content": [
+        {
+            "section_name": "Abstract",
+            "section_text": "Mitochondrial ATP synthase forms stable dimers arranged into oligomeric assemblies that generate the inner-membrane curvature essential for efficient energy conversion. Here, we report cryo-EM structures of the intact ATP synthase dimer from Trypanosoma brucei in ten different rotational states. The model consists of 25 subunits, including nine lineage-specific, as well as 36 lipids. The rotary mechanism is influenced by the divergent peripheral stalk, conferring a greater conformational flexibility. Proton transfer in the lumenal half-channel occurs via a chain of five ordered water molecules. The dimerization interface is formed by subunit-g that is critical for interactions but not for the catalytic activity. Although overall dimer architecture varies among eukaryotes, we find that subunit-g together with subunit-e form an ancestral oligomerization motif, which is shared between the trypanosomal and mammalian lineages. Therefore, our data defines the subunit-g/e module as a structural component determining ATP synthase oligomeric assemblies.",
+            "section_image": []
+        },
+        {
+            "section_name": "Additional Declarations",
+            "section_text": "There is NO Competing Interest.",
+            "section_image": []
+        },
+        {
+            "section_name": "Supplementary Files",
+            "section_text": "Video1.mp4Video 1. Overall structure and subunit-e/g module of trypanosomal ATP synthase dimersVideo2.mp4Video 2. Rotary cycle of T. brucei ATP synthaseVideo3.mp4Video 3. Rotary cycle of T. brucei ATP synthase",
+            "section_image": []
+        }
+    ],
+    "nature_content": [
+        {
+            "section_name": "Abstract",
+            "section_text": "Mitochondrial ATP synthase forms stable dimers arranged into oligomeric assemblies that generate the inner-membrane curvature essential for efficient energy conversion. Here, we report cryo-EM structures of the intact ATP synthase dimer from Trypanosoma brucei in ten different rotational states. The model consists of 25 subunits, including nine lineage-specific, as well as 36 lipids. The rotary mechanism is influenced by the divergent peripheral stalk, conferring a greater conformational flexibility. Proton transfer in the lumenal half-channel occurs via a chain of five ordered water molecules. The dimerization interface is formed by subunit-g that is critical for interactions but not for the catalytic activity. Although overall dimer architecture varies among eukaryotes, we find that subunit-g together with subunit-e form an ancestral oligomerization motif, which is shared between the trypanosomal and mammalian lineages. Therefore, our data defines the subunit-g/e module as a structural component determining ATP synthase oligomeric assemblies.",
+            "section_image": []
+        },
+        {
+            "section_name": "Introduction",
+            "section_text": "Mitochondrial ATP synthase consists of the soluble F1 and membrane-bound Fo subcomplexes and occurs in dimers that assemble into oligomers to induce the formation of inner-membrane folds, called cristae. The cristae are the sites for oxidative phosphorylation and energy conversion in eukaryotic cells. Dissociation of ATP synthase dimers into monomers results in the loss of native cristae architecture and impairs mitochondrial function1,2. While cristae morphology varies substantially between organisms from different lineages, ranging from flat lamellar in opisthokonts to coiled tubular in ciliates and discoidal in euglenozoans3, the mitochondrial ATP synthase dimers represent a universal occurrence to maintain the membrane shape4.\n\nATP synthase dimers of variable size and architecture, classified into types I to IV have recently been resolved by high-resolution cryo-EM studies. In the structure of the type-I ATP synthase dimer from mammals, the monomers are only weakly associated5,6, and in yeast insertions in the membrane subunits form tighter contacts7. The structure of the type-II ATP synthase dimer from the alga Polytomella sp. showed that the dimer interface is formed by phylum-specific components8. The type-III ATP synthase dimer from the ciliate Tetrahymena thermophila is characterized by parallel rotary axes, and a substoichiometric subunit, as well as multiple lipids were identified at the dimer interface, while additional protein components that tie the monomers together are distributed between the matrix, transmembrane, and lumenal regions9. The structure of the type-IV ATP synthase with native lipids from Euglena gracilis also showed that specific protein-lipid interactions contribute to the dimerization, and that the central and peripheral stalks interact with each other directly10. Finally, a unique apicomplexan ATP synthase dimerizes via 11 parasite-specific components that contribute ~7000 \u00c52 buried surface area11, and unlike all other ATP synthases, that assemble into rows, it associates in higher oligomeric states of pentagonal pyramids in the curved apical membrane regions. Together, the available structural data suggest a diversity of oligomerization, and it remains unknown whether common elements mediating these interactions exist or whether dimerization of ATP synthase occurred independently and multiple times in evolution4.\n\nThe ATP synthase of Trypanosoma brucei, a representative of kinetoplastids and an established medically important model organism causing the sleeping sickness, is highly divergent, exemplified by the pyramid-shaped F1 head containing a phylum specific subunit12,13. The dimers are sensitive to the lack of cardiolipin14 and form short left-handed helical segments that extend across the membrane ridge of the discoidal cristae15. Uniquely among aerobic eukaryotes, the mammalian life cycle stage of T. brucei utilizes the reverse mode of ATP synthase, using the enzyme as a proton pump to maintain mitochondrial membrane potential at the expense of ATP16,17. In contrast, the insect stages of the parasite employ the ATP-producing forward mode of the enzyme18,19.\n\nGiven the conservation of the core subunits, the different nature of oligomerization and the ability to test structural hypotheses biochemically, we reasoned that investigation of the T. brucei ATP synthase structure and function would provide the missing evolutionary link to understand how the monomers interact to form physiological dimers.\n\nHere, we address this question by combining structural, functional, and evolutionary analysis of the T. brucei ATP synthase dimer.",
+            "section_image": []
+        },
+        {
+            "section_name": "Results",
+            "section_text": "We purified ATP synthase dimers from cultured T. brucei procyclic trypomastigotes by affinity chromatography with a recombinant natural protein inhibitor TbIF120 and subjected the sample to cryo-EM analysis (Supplementary Figs.\u00a01 and 2). Using masked refinements, maps were obtained for the membrane region, the rotor, and the peripheral stalk. To describe the conformational space of the T. brucei ATP synthase, we resolved ten distinct rotary substates, which were refined to 3.5\u20134.3\u2009\u00c5 resolution. Finally, particles with both monomers in rotational state 1 were selected, and the consensus structure of the dimer was refined to 3.2\u2009\u00c5 resolution (Supplementary Table\u00a01, Supplementary Figs.\u00a02 and 3).\n\nUnlike the wide-angle architecture of dimers found in animals and fungi, the T. brucei ATP synthase displays an angle of 60\u00b0 between the two F1/c-ring subcomplexes. The model of the T. brucei ATP synthase includes all 25 different subunits, nine of which are lineage-specific (Fig.\u00a01a, Supplementary Fig.\u00a04, and Supplementary Movie\u00a01). We named the subunits according to the previously proposed nomenclature21,22,23 (Supplementary Table\u00a02). In addition, we identified and modelled 36 bound phospholipids, including 24 cardiolipins (Supplementary Fig.\u00a05). Both detergents used during purification, n-dodecyl \u03b2-D-maltoside (\u03b2-DDM) and glyco-diosgenin (GDN) are also resolved in the periphery of the membrane region (Supplementary Fig.\u00a06).\n\na Front and side views of the composite model with both monomers in rotational state 1. The two F1/c10-ring complexes, each augmented by three copies of the phylum-specific p18 subunit, are tied together at a 60\u00b0-angle. The membrane-bound Fo region displays a unique architecture and is composed of both conserved and phylum-specific subunits. b Side view of the Fo region showing the lumenal interaction of the ten-stranded \u03b2-barrel of the c-ring (grey) with ATPTB12 (pale blue). The lipid-filled peripheral Fo cavity is indicated. c Close-up view of the bound lipids within the peripheral Fo cavity with cryo-EM density shown. d Top view into the decameric c-ring with a bound pyrimidine ribonucleoside triphosphate, assigned as UTP, although not experimentally detected. Map density shown in transparent blue, interacting residues shown.\n\nIn the catalytic region, F1 is augmented by three copies of subunit p18, each bound to subunit-\u03b112,13. Our structure shows that p18 is involved in the unusual attachment of F1 to the peripheral stalk. The membrane region includes eight conserved Fo subunits (b, d, f, 8, i/j, k, e, and g) arranged around the central proton translocator subunit-a. We identified those subunits based on the structural similarity and matching topology to their yeast counterparts (Fig.\u00a02). For subunit-b, a single transmembrane helix superimposes well with bH1 from yeast and anchors the subunits-e and -g to the Fo (Fig.\u00a02a, b). In yeast and bovine ATP synthases bH1 and transmembrane helices of subunits-e and -g are arranged in the same way as in our structure and contribute to a characteristic wedge in the membrane domain5. The long helix bH2, which constitutes the central part of the peripheral stalk in other organisms is absent in T. brucei (Fig.\u00a02c). No alternative subunit-b24 is found in our structure.\n\na Top view of the membrane region with T. brucei subunits (coloured) overlaid with S. cerevisiae structure (grey transparent). Close structural superposition and matching topology allowed the assignment of conserved subunits based on matching topology and location. b Superposition of subunits-b, -e and -g with their S. cerevisiae counterparts PDB 6B2Z (S. cerevisiae mitochondrial ATP synthase) confirms their identity. c Schematic representation of transmembrane helices of subunit-b and adjacent subunits in T. brucei, E. gracilis PDB 6TDV (E. gracilis mitochondrial ATP synthase, membrane region)10 and S. cerevisiae PDB 6B2Z (S. cerevisiae mitochondrial ATP synthase)7 ATP synthases. PC \u2013 phosphatidylcholine.\n\nThe membrane region contains a peripheral subcomplex, formed primarily by the phylum-specific ATPTB1,6,12 and ATPEG3 (Fig.\u00a01b). It is separated from the conserved core by a membrane-intrinsic cavity, in which nine bound cardiolipins are resolved (Fig.\u00a01c), and the C-terminus of ATPTB12 interacts with the lumenal \u03b2-barrel of the c10-ring. The \u03b2-barrel, which has previously been reported also in the ATP synthase from E. gracilis10, extends from the c10-ring approximately 15\u2009\u00c5 to the lumen (Fig.\u00a01a and Supplementary Fig.\u00a07). The cavity of the decameric c-ring contains density consistent with disordered lipids, as observed in other ATP synthases5,6,7, and in addition near the matrix side, 10 Arg66c residues coordinate a ligand density, which is consistent with a pyrimidine ribonucleoside triphosphate (Fig.\u00a01d). We assign this density as uridine-triphosphate (UTP), due to its large requirement in the mitochondrial RNA metabolism of African trypanosomes being a substrate for post-transcriptional RNA editing25, and addition of poly-uridine tails to gRNAs and rRNAs26,27, as well as due to low abundance of cytidine triphosphate (CTP)28. The nucleotide base is inserted between two Arg82c residues, whereas the triphosphate region is coordinated by another five Arg82c residues, with Tyr79\u03b4 and Asn76\u03b4 providing asymmetric coordination contacts. The presence of a nucleotide inside the c-ring is surprising, given the recent reports of phospholipids inside the c-rings in mammals5,6 and ciliates9, indicating that a range of different ligands can provide structural scaffolding.\n\nThe trypanosomal peripheral stalk displays a markedly different architecture compared to its yeast and mammalian counterparts. In the opisthokont complexes, the peripheral stalk is organized around the long bH2, which extends from the membrane ~15\u2009nm into the matrix and attaches to OSCP at the top of F15,7. By contrast, T. brucei lacks the canonical bH2 and instead, helices 5-7 of divergent subunit-d and the C-terminal helix of extended subunit-8 bind to a C-terminal extension of OSCP at the apical part of the peripheral stalk (Fig.\u00a03a). The interaction between OSCP and subunit-d and -8 is stabilized by soluble ATPTB3 and ATPTB4. The peripheral stalk is rooted to the membrane subcomplex by a transmembrane helix of subunit-8, wrapped on the matrix side by helices 8-11 of subunit-d. Apart from the canonical contacts at the top of F1, the peripheral stalk is attached to the F1 via an euglenozoa-specific C-terminal extension of OSCP, which contains a disordered linker and a terminal helix hairpin extending between the F1-bound p18 and subunits -d and -8 of the peripheral stalk (Fig.\u00a03a and Supplementary Movies\u00a02, 3). Another interaction of F1 with the peripheral stalk occurs between the stacked C-terminal helices of subunit-\u03b2 and -d (Fig.\u00a03b), the latter of which structurally belongs to F1 and is connected to the peripheral stalk via a flexible linker.\n\na N-terminal OSCP extension provides a permanent central stalk attachment, while the C-terminal extension provides a phylum-specific attachment to the divergent peripheral stalk. b The C-terminal helices of subunits -\u03b2 and -d provide a permanent F1 attachment. c Substeps of the c-ring during transition from rotational state 1 to 2. d F1 motion accommodating steps shown in (c). After advancing along with the rotor to state 1e, the F1 rotates in the opposite direction when transitioning to state 2a. e Tilting motion of F1 and accommodating bending of the peripheral stalk.\n\nTo assess whether the unusual peripheral stalk architecture influences the rotary mechanism, we analyzed 10 classes representing different rotational states. The three main states (1\u20133) result from three ~120\u00b0 rotation steps of the rotor relatively to the static Fo. In all classes F1 is in a similar conformation, corresponding to the catalytic dwell, observed previously also in the crystal structure of T. brucei F1-ATPase13. In accordance with the ~120\u00b0 rotation of the central stalk, the conformations and nucleotide occupancy of the catalytic interfaces of the individual \u03b1\u03b2 dimers differ between the main states, showing ADP and ATP in the \u201cloose\u201d and \u201ctight\u201d closed conformations, respectively, and empty nucleotide binding site in the \u201copen\u201d conformation. We identified five (1a\u20131e), four (2a\u20132d), and one (3) classes of the respective main states. The rotor positions of the rotational states 1a, 2a, and 3 are related by steps of 117\u00b0, 136\u00b0, and 107\u00b0, respectively. Throughout all the identified substeps of the rotational state 1 (classes 1a to 1e) the rotor turns by ~33\u00b0, which corresponds approximately to the advancement by one subunit-c of the c10-ring (Fig.\u00a03c). While rotating along with the rotor, the F1 headpiece lags behind, advancing by only ~13\u00b0. During the following transition from 1e to 2a, the rotor advances by ~84\u00b0, whereas the F1 headpiece rotates ~22\u00b0 in the opposite direction (Fig.\u00a03d). This generates a counter-directional torque between the two motors, which is consistent with a power-stroke mechanism. This counter-directional torque may occur in all three main rotational state transitions. However, it was observed only in the main state 1, because it was captured in more substeps than the remaining two states, presumably as a consequence of the symmetry mismatch between the decameric c-ring and the \u03b13\u03b23 hexamer29. Within the four classes of the state 2 the rotor advances by 23\u00b0 and F1 returns close to its position observed in class 1a, where it is found also in the only observed class of the state 3. Albeit with small differences in step size, this mechanism is consistent with a previous observation in the Polytomella ATP synthase8. However, due to its large, rigid peripheral stalk, the Polytomella ATP synthase mainly displays rotational substeps, whereas the Trypanosoma F1 also displays a tilting motion of ~8\u00b0 revealed by rotary states 1a and 1b (Fig.\u00a03e and Supplementary Movie\u00a02). The previously reported hinge motion between the N- and C-terminal domains of OSCP8 is not found in our structures, instead, the conformational changes of the F1/c10-ring subcomplex are accommodated by a 5\u00b0 bending of the apical part of the peripheral stalk. (Fig.\u00a03e and Supplementary Movies\u00a02, 3). Together, the structural data indicate that the divergent peripheral stalk attachment confers greater conformational flexibility to the T. brucei ATP synthase.\n\nThe mechanism of proton translocation involves sequential protonation of E102 of subunits-c, rotation of the c10-ring with neutralized E102c exposed to the phospholipid bilayer, and release of protons on the other side of the membrane. The sites of proton binding and release are separated by the conserved R146 contributed by the horizontal helix H5 of subunit-a and are accessible from the cristae lumen and mitochondrial matrix by aqueous half-channels (Fig.\u00a04a). Together, R146 and the adjacent N209 coordinate a pair of water molecules in between helices H5 and H6 (Fig.\u00a04b). A similar coordination has been observed in the Polytomella ATP synthase8. The coordination of water likely restricts the R146 to rotamers that extend towards the c-ring, with which it is thought to interact.\n\na Subunit-a (green) with the matrix (orange) and lumenal (light blue) channels, and an ordered phosphatidylcholine (PC1; blue). E102 of the c10-ring shown in grey. b Close-up view of the highly conserved R146a and N209a, which coordinate two water molecules between helices H5-6a. c Sideview of the lumenal channel with proton pathway (light blue) and confining phosphatidylcholine (blue). d Chain of ordered water molecules in the lumenal channel. Distances between the W1\u2013W5 (red) are 5.2, 3.9, 7.3, and 4.8\u2009\u00c5, respectively. e The ordered waters extend to H155a, which likely mediates the transfer of protons to D202a.\n\nIn our structure, the lumenal half-channel, which displays a local resolution of 2.55\u2009\u00c5 (Supplementary Fig.\u00a03), is filled with a network of resolved water densities, ending in a chain of five ordered water molecules (W1\u2013W5; Fig.\u00a04c\u2013e). The presence of ordered water molecules in the aqueous channel is consistent with a Grotthuss-type mechanism for proton transfer, which would not require long-distance diffusion of water molecules5. However, because some distances between the observed water molecules are too large for direct hydrogen bonding, proton transfer may involve both coordinated and disordered water molecules. The distance of 7\u2009\u00c5 between the last resolved water (W1) and D202a, the conserved residue that is thought to transfer protons to the c-ring, is too long for direct proton transfer. Instead, it may occur via the adjacent H155a. Therefore, our structure resolves individual elements participating in proton transport (Fig.\u00a04d, e).\n\nThe lumenal proton half-channel in the mammalian5,6 and apicomplexan11 ATP synthase is lined by the transmembrane part of bH2, which is absent in T. brucei. Instead, the position of bH2 is occupied by a fully ordered phosphatidylcholine in our structure (PC1; Fig.\u00a04a, c). Therefore, a bound lipid replaces a proteinaceous element in the proton path.\n\nDespite sharing a set of conserved Fo subunits, the T. brucei ATP synthase dimer displays a markedly different dimer architecture compared to previously determined structures. First, its dimerization interface of 3600 \u00c52 is smaller than that of the E. gracilis type-IV (10,000 \u00c52) and the T. thermophila type-III ATP synthases (16,000 \u00c52). Second, unlike mammalian and fungal ATP synthase, in which the peripheral stalks extend in the plane defined by the two rotary axes, in our structure the monomers are rotated such that the peripheral stalks are offset laterally on the opposite sides of the plane. Due to the rotated monomers, this architecture is associated with a specific dimerization interface, where two subunit-g copies interact homotypically on the C2 symmetry axis (Fig.\u00a05a and Supplementary Movie\u00a01). Both copies of H1-2g extend horizontally along the matrix side of the membrane, clamping against each other (Fig.\u00a05c, e). This facilitates formation of contacts between an associated transmembrane helix of subunit-e with the neighbouring monomer via subunit-a\u2019 in the membrane, and -f\u2019in the lumen, thereby further contributing to the interface (Fig.\u00a05b). Thus, the ATP synthase dimer is assembled via the subunit-e/g module. The C-terminal part of the subunit-e helix extends into the lumen, towards the ten-stranded \u03b2-barrel of the c-ring (Supplementary Fig.\u00a07a). The terminal 23 residues are disordered with poorly resolved density connecting to the detergent plug of the c-ring \u03b2-barrel (Supplementary Fig.\u00a07b). This resembles the lumenal C-terminus of subunit-e in the bovine structure5, indicating a conserved interaction with the c-ring. In mammals, a mechanism, in which retraction of subunit-e upon calcium exposure pulls out the lipid plug and induces disassembly of the c-ring, which triggers permeability transition pore (PTP) opening, has been proposed6.\n\na Side view with dimerizing subunits coloured. The dimer interface is constituted by b subunit-e\u2019 contacting subunit-a in the membrane and subunit-f in the lumen, c subunits e and g from both monomers forming a subcomplex with bound lipids. d Subunit-g and -e form a dimerization motif in the trypanosomal (type-IV) ATP synthase dimer (this study), the same structural element forms the oligomerization motif in the porcine ATP synthase tetramer. The structural similarity of the pseudo-dimer (i.e., two diagonal monomers from adjacent dimers) in the porcine structure with the trypanosomal dimer suggests that type I and IV ATP synthase dimers have evolved through divergence from a common ancestor. e The dimeric subunit-e/g structures are conserved in Sus scrofa PDB 6ZNA (S. scrofa mitochondrial ATP synthase) and T. brucei (this work) and contain a conserved GXXXG motif (orange) mediating interaction of transmembrane helices. f Models of the ATP synthase dimers fitted into subtomogram averages of short oligomers15: matrix view, left; cut-through, middle, lumenal view, right; EMD-3560 (in situ structure of T. brucei mitochondrial ATP synthase).\n\nThe e/g module is held together by four bound cardiolipins in the matrix leaflet, anchoring it to the remaining Fo region (Fig.\u00a05c). The head groups of the lipids are coordinated by polar and charged residues with their acyl chains filling a central cavity in the membrane region at the dimer interface (Fig.\u00a05c and Supplementary Fig.\u00a05f). Cardiolipin binding has previously been reported to be obligatory for dimerization in secondary transporters30 and the depletion of cardiolipin synthase resulted in reduced levels of ATP synthase in bloodstream trypanosomes14.\n\nInterestingly, for yeasts, early blue native gel electrophoresis31 and subtomogram averaging studies2 suggested subunit-g as potentially dimer-mediating, however the e/g modules are located laterally opposed on either side of the dimer long axis, in the periphery of the complex, ~8.5\u2009nm apart from each other. Because the e/g modules do not interact directly within the yeast ATP synthase dimer, they have been proposed to serve as membrane-bending elements, whereas the major dimer contacts are formed by subunit-a and -i/j7. In mammals, the e/g module occupies the same position as in yeasts, forming the interaction between two diagonal monomers in a tetramer5,6,32, as well as between parallel dimers33. The comparison with our structure shows that the overall organization of the intra-dimeric trypanosomal and inter-dimeric mammalian e/g module is structurally similar (Fig.\u00a05d). Furthermore, kinetoplastid parasites and mammals share conserved GXXXG motifs in subunit-e34 and -g (Supplementary Fig.\u00a08), which allow close interaction of their transmembrane helices (Fig.\u00a05e), providing further evidence for subunit homology. However, while the mammalian ATP synthase dimers are arranged perpendicularly to the long axis of their rows along the edge of cristae35, the T. brucei dimers on the rims of discoidal cristae are inclined ~45\u00b0 to the row axis15. Therefore, the e/g module occupies equivalent positions in the rows of both evolutionary distant groups (Fig.\u00a05f and ref. 33).\n\nTo validate structural insights, we knocked down each individual Fo subunit by inducible RNA interference (RNAi). All target mRNAs dropped to 5\u201320% of their original levels after two and four days of induction (Fig.\u00a06a and Supplementary Fig.\u00a09a). Western blot analysis of whole-cell lysates resolved by denaturing electrophoresis revealed decreased levels of Fo subunits ATPB1 and -d suggesting that the integrity of the Fo moiety depends on the presence of other Fo subunits (Fig.\u00a06c, d). Immunoblotting of mitochondrial complexes resolved by blue native polyacrylamide gel electrophoresis (BN-PAGE) with antibodies against F1 and Fo subunits revealed a strong decrease or nearly complete loss of dimeric and monomeric forms of ATP synthases four days after induction of RNAi of most subunits (b, e, f, i/j, k, 8, ATPTB3, ATPTB4, ATPTB6, ATPTB11, ATPTB12, ATPEG3, and ATPEG4), documenting an increased instability of the enzyme or defects in its assembly. Simultaneous accumulation in F1-ATPase, as observed by BN-PAGE, demonstrated that the catalytic moiety remains intact after the disruption of the peripheral stalk or the membrane subcomplex (Fig.\u00a06b\u2013d and Supplementary Fig.\u00a09b).\n\na Growth curves of non-induced (solid lines) and tetracycline-induced (dashed lines) RNAi cell lines grown in the presence (black) or absence (brown) of glucose. The insets show relative levels of the respective target mRNA at indicated days post-induction (DPI) normalized to the levels of 18S rRNA (black bars) or \u03b2-tubulin (white bars). b Immunoblots of mitochondrial lysates from indicated RNAi cell lines resolved by BN-PAGE probed with antibodies against indicated ATP synthase subunits (n\u2009=\u20092). Positions of molecular weight (MW) marker are shown. c Immunoblots of whole cell lysates from indicated RNAi cell lines probed with indicated antibodies (n\u2009=\u20093). Positions of MW marker are shown. d Quantification of three independent replicates of immunoblots in (c). Values were normalized to the signal of the loading control Hsp70 and to non-induced cells. Plots show individual values, means, and standard deviations (SD; error bars).\n\nIn contrast to the other targeted Fo subunits, the downregulation of subunit-g with RNAi resulted in a specific loss of dimeric complexes with concomitant accumulation of monomers (Fig.\u00a06b), indicating that it is required for dimerization, but not for the assembly and stability of the monomeric F1Fo ATP synthase units. Transmission electron microscopy of thin cell sections revealed that the ATP synthase monomerization in the subunit-gRNAi cell line had the same effect on mitochondrial ultrastructure as nearly complete loss of monomers and dimers upon knockdown of subunit-8. Both cell lines exhibited decreased cristae counts and aberrant cristae morphology (Fig.\u00a07a, b), including the appearance of round shapes reminiscent of structures detected upon deletion of subunit-g or -e in Saccharomyces cerevisiae1. These results indicate that monomerization prevents the trypanosomal ATP synthase from assembling into short helical rows on the rims of the discoidal cristae15, as has been reported for impaired oligomerization in counterparts from other eukaryotes2,36.\n\na Transmission electron micrographs of sections of non-induced or 4 days induced RNAi cell lines. At least 70 micrographs were obtained in each category. Mitochondrial membranes and cristae are marked with blue and red arrowheads, respectively. Top panel shows examples of irregular, elongated and round cross-sections of mitochondria quantified in (b). Scale bars: 500\u2009nm. b Cristae numbers per vesicle from indicated induced (IND) or non-induced (NON) cell lines counted separately in irregular, elongated and round mitochondrial cross-section. Boxes and whiskers show 25th to 75th and 5th to 95th percentiles, respectively. The numbers of analyzed cross-sections are indicated for each data point. Unpaired two-sided t-test, p-values are shown in the graph. c Mitochondrial membrane polarization capacity of non-induced or RNAi-induced cell lines two and four DPI measured by Safranine O. Black and grey arrow indicate addition of ATP and oligomycin, respectively. d ATP production in permeabilized non-induced (0) or RNAi-induced cells 2 and 4 DPI in the presence of indicated substrates and inhibitors. The graphs show individual values of two technical replicates of n\u2009=\u20092 (subunit-8), n\u2009=\u20093 (ATPTB4), or n\u2009=\u20094 (subunit-g) independent experiments and means (bars) and SD (error bars) of the averaged values of the technical replicates. Gly3P DL-glycerol phosphate; KCN potassium cyanide; CATR carboxyatractyloside.\n\nDespite the altered mitochondrial ultrastructure, the subunit-gRNAi cells showed only a very mild growth phenotype, in contrast to all other RNAi cell lines that exhibited steadily slowed growth from day three to four after the RNAi induction (Fig.\u00a07a and Supplementary Fig.\u00a09a). This is consistent with the growth defects observed after the ablation of Fo subunit ATPTB119 and F1 subunits-\u03b1 and p1812. Thus, the monomerization of ATP synthase upon subunit-g ablation had only a negligible effect on the fitness of trypanosomes cultured in glucose-rich medium, in which ATP production by substrate level phosphorylation partially compensates for compromised oxidative phosphorylation37.\n\nMeasurement of oligomycin-sensitive ATP-dependent mitochondrial membrane polarization by safranin O assay in permeabilized cells showed that the proton pumping activity of the ATP synthase in the induced subunit-gRNAi cells is negligibly affected, demonstrating that the monomerized enzyme is catalytically functional. By contrast, RNAi downregulation of subunit-8, ATPTB4 and ATPTB11, and ATPTB1 resulted in a strong decline of the mitochondrial membrane polarization capacity, consistent with the loss of both monomeric and dimeric ATP synthase forms (Fig.\u00a07c). Accordingly, knockdown of the same subunits resulted in inability to produce ATP by oxidative phosphorylation (Fig.\u00a07d). However, upon subunit-g ablation the ATP production was affected only partially, confirming that the monomerized ATP synthase remains catalytically active. The ~50% drop in ATP production of subunit-gRNAi cells can be attributed to the decreased oxidative phosphorylation efficiency due to the impaired cristae morphology. Indeed, when cells were cultured in the absence of glucose, enforcing the need for oxidative phosphorylation, knockdown of subunit-g results in a growth arrest, albeit one to two days later than knockdown of all other tested subunits (Fig.\u00a06a). The data show that dimerization is critical when oxidative phosphorylation is the predominant source of ATP.",
+            "section_image": [
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-022-33588-z/MediaObjects/41467_2022_33588_Fig1_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-022-33588-z/MediaObjects/41467_2022_33588_Fig2_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-022-33588-z/MediaObjects/41467_2022_33588_Fig3_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-022-33588-z/MediaObjects/41467_2022_33588_Fig4_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-022-33588-z/MediaObjects/41467_2022_33588_Fig5_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-022-33588-z/MediaObjects/41467_2022_33588_Fig6_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-022-33588-z/MediaObjects/41467_2022_33588_Fig7_HTML.png"
+            ]
+        },
+        {
+            "section_name": "Discussion",
+            "section_text": "Our structure of the mitochondrial ATP synthase dimer from the mammalian parasite T. brucei offers new insight into the mechanism of membrane shaping, rotary catalysis, and proton transfer. Considering that trypanosomes belong to an evolutionarily divergent group of Kinetoplastida, the ATP synthase dimer has several interesting features that differ from other dimer structures. The subunit-b found in bacterial and other mitochondrial F-type ATP synthases appears to be highly reduced to a single transmembrane helix bH1. The long bH2, which constitutes the central part of the peripheral stalk in other organisms, and is also involved in the composition of the lumenal proton half-channel, is completely absent in T. brucei. Interestingly, the position of bH2 in the proton half channel is occupied by a fully ordered phosphatidylcholine molecule that replaces a well-conserved proteinaceous element in the proton path. However, this replacement is not a common trait of all type-IV ATP synthases, because subunit-b in E. gracilis contains the canonical bH2 but lacks bH110. Thus, while subunit-b is conserved in Euglenozoa, the lineages of T. brucei and E. gracilis retained its different non-overlapping structural elements (Fig.\u00a02c). Lack of bH2 in T. brucei also affects composition of the peripheral stalk in which the divergent subunit-d and subunit-8 binds directly to a C-terminal extension of OSCP, indicating a remodeled peripheral stalk architecture. The peripheral stalk contacts the F1 headpiece at several positions conferring greater conformational flexibility to the ATP synthase.\n\nUsing the structural and functional data, we also identified a conserved structural element of the ATP synthase that is responsible for its multimerization. Particularly, subunit-g is required for the dimerization, but dispensable for the assembly of the F1Fo monomers. Although the monomerized enzyme is catalytically competent, the inability to form dimers results in defective cristae structure, and consequently leads to compromised oxidative phosphorylation and cease of proliferation. The cristae-shaping properties of mitochondrial ATP synthase dimers are critical for sufficient ATP production by oxidative phosphorylation, but not for other mitochondrial functions, as demonstrated by the lack of growth phenotype of subunit-gRNAi cells in the presence of glucose. Thus, trypanosomal subunit-g depletion strain represents an experimental tool to assess the roles of the enzyme\u2019s primary catalytic function and mitochondria-specific membrane-shaping activity, highlighting the importance of the latter for oxidative phosphorylation.\n\nBased on our data and previously published structures, we propose an ancestral state with double rows of ATP synthase monomers connected by e/g modules longitudinally and by other Fo subunits transversally. During the course of evolution, different pairs of adjacent ATP synthase monomer units formed stable dimers in individual lineages (Fig.\u00a08). This gave rise to the highly divergent type-I and type-IV ATP synthase dimers with subunit-e/g modules serving either as oligomerization or dimerization motives, respectively. Because trypanosomes belong to the deep-branching eukaryotic supergroup Discoba, the proposed arrangement might have been present in the last eukaryotic common ancestor. Although sequence similarity of subunit-g is low and restricted to the single transmembrane helix, we found homologues of subunit-g in addition to Opisthokonta and Discoba also in Archaeplastida and Amoebozoa, which represent other eukaryotic supergroups, thus supporting the ancestral role in oligomerization (Supplementary Fig.\u00a08). Taken together, our analysis reveals that mitochondrial ATP synthases that display markedly diverged architecture share the ancestral structural module that promotes oligomerization.\n\nSchematic model of the evolution of type-I and IV ATP synthases. Mitochondrial ATP synthases are derived from a monomeric complex of proteobacterial origin. In a mitochondrial ancestor, acquisition of mitochondria-specific subunits, including the subunit-e/g module resulted in the assembly of ATP synthase double rows, the structural basis for cristae biogenesis. Through divergence, different ATP synthase architectures evolved, with the subunit-e/g module functioning as an oligomerization (type I) or dimerization (type IV) motif, resulting in distinct row assemblies between mitochondrial lineages.",
+            "section_image": [
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-022-33588-z/MediaObjects/41467_2022_33588_Fig8_HTML.png"
+            ]
+        },
+        {
+            "section_name": "Methods",
+            "section_text": "T. brucei procyclic strains were cultured in SDM-79 medium supplemented with 10% (v/v) fetal bovine serum. For growth curves in glucose-free conditions, cells were grown in SDM-80 medium with 10% dialyzed FBS. RNAi cell lines were grown in presence of 2.5 \u03bcg/ml phleomycin and 1 \u03bcg/ml puromycin. For ATP synthase purification, mitochondria were isolated from the Lister strain 427. Typically, 1.5 \u00d7 1011 cells were harvested, washed in 20\u2009mM sodium phosphate buffer pH 7.9 with 150\u2009mM NaCl and 20\u2009mM glucose, resuspended in hypotonic buffer 1\u2009mM Tris-HCl pH 8.0, 1\u2009mM EDTA, and disrupted by 10 strokes in a 40\u2009ml Dounce homogenizer. The lysis was stopped by immediate addition of sucrose to 0.25\u2009M. Crude mitochondria were pelleted (15\u2009min at 16,000\u2009\u00d7\u2009g, 4\u2009\u00b0C), resuspended in 20\u2009mM Tris-HCl pH 8.0, 250\u2009mM sucrose, 5\u2009mM MgCl2, 0.3\u2009mM CaCl2 and treated with 5\u2009\u03bcg/ml DNase I. After 60\u2009min on ice, one volume of the STE buffer (20\u2009mM Tris-HCl pH 8.0, 250\u2009mM sucrose, 2\u2009mM EDTA) was added and mitochondria were pelleted (15\u2009min at 16000\u2009\u00d7\u2009g, 4\u2009\u00b0C). The pellet was resuspended in 60% (v/v) Percoll in STE and loaded on six linear 10\u201335% Percoll gradients in STE in polycarbonate tubes for SW28 rotor (Beckman). Gradients were centrifuged for 1\u2009h at 104000\u2009\u00d7\u2009g, 4\u2009\u00b0C. The middle phase containing mitochondrial vesicles (15\u201320\u2009ml per tube) was collected, washed four times in the STE buffer, and pellets were snap-frozen in liquid nitrogen and stored at \u221280\u2009\u00b0C.\n\nTo downregulate ATP synthase subunits by RNAi, DNA fragments corresponding to individual target sequences were amplified by PCR from Lister 427 strain genomic DNA using forward and reverse primers extended with restriction sites XhoI&KpnI and XbaI&BamHI, respectively (Supplementary Table\u00a03). Each fragment was inserted into the multiple cloning sites 1 and 2 of pAZ0055 vector, derived from pRPHYG-iSL (courtesy of Sam Alsford) by replacement of hygromycine resistance gene with phleomycine resistance gene, with restriction enzymes KpnI/BamHI and XhoI/XbaI, respectively. Resulting constructs with tetracycline inducible T7 polymerase driven RNAi cassettes were linearized with NotI and transfected into a cell line derived from the Lister strain 427 by integration of the SmOx construct for expression of T7 polymerase and the tetracycline repressor38 into the \u03b2-tubulin locus. RNAi was induced in selected semi-clonal populations by addition of 1 \u03bcg/ml tetracycline and the downregulation of target mRNAs was verified by quantitative RT-PCR 2 and 4 days post induction. The total RNA isolated by an RNeasy Mini Kit (Qiagen) was treated with 2 \u03bcg of DNase I, and then reverse transcribed to cDNA with TaqMan Reverse Transcription kit (Applied Biosciences). qPCR reactions were set with Light Cycler 480 SYBR Green I Master mix (Roche), 2 \u03bcl of cDNA, and 0.3 \u03bcM primers (Supplementary Table\u00a03) and run on LightCycler 480 (Roche). Relative expression of target genes was calculated using - \u0394\u0394Ct method with 18S rRNA or \u03b2-tubulin as endogenous reference genes and normalized to noninduced cells.\n\nWhole cell lysates for denaturing sodium dodecyl sulphate polyacrylamide electrophoresis (SDS-PAGE) were prepared from cells resuspended in PBS buffer (10\u2009mM phosphate buffer, 130\u2009mM NaCl, pH 7.3) by addition of 3\u00d7 Laemmli buffer (150\u2009mM Tris pH 6.8, 300\u2009mM 1,4-dithiothreitol, 6% (w/v) SDS, 30% (w/v) glycerol, 0.02% (w/v) bromophenol blue) to final concentration of 1 \u00d7 107 cells in 30\u2009\u03bcl. The lysates were boiled at 97\u2009\u00b0C for 10\u2009min and stored at \u221220\u2009\u00b0C. For immunoblotting, lysates from 3 \u00d7 106 cells were separated on 4\u201320% gradient Tris-glycine polyacrylamide gels (BioRad 4568094), electroblotted onto a PVDF membrane (Pierce 88518), and probed with respective antibodies (Supplementary Table\u00a04). Membranes were incubated with the Clarity Western ECL substrate (BioRad 1705060EM) and chemiluminescence was detected on a ChemiDoc instrument (BioRad). Band intensities were quantified densitometrically using the ImageLab software. The levels of individual subunits were normalized to the signal of mtHsp70.\n\nBlue native PAGE (BN-PAGE) was performed as described earlier12 with following modifications. Crude mitochondrial vesicles from 2.5 \u00d7 108 cells were resuspended in 40\u2009\u00b5l of Solubilization buffer A (2\u2009mM \u03b5-aminocaproic acid (ACA), 1\u2009mM EDTA, 50\u2009mM NaCl, 50\u2009mM Bis-Tris/HCl, pH 7.0) and solubilized with 2% (w/v) dodecylmaltoside (\u03b2-DDM) for 1\u2009h on ice. Lysates were cleared at 16,000\u2009\u00d7\u2009g for 30\u2009min at 4\u2009\u00b0C and their protein concentration was estimated using bicinchoninic acid assay. For each time point, a volume of mitochondrial lysate corresponding to 4\u2009\u03bcg of total protein was mixed with 1.5\u2009\u03bcl of loading dye (500\u2009mM ACA, 5% (w/v) Coomassie Brilliant Blue G-250) and 5% (w/v) glycerol and with 1\u2009M ACA until a final volume of 20\u2009\u03bcl/well and resolved on a native PAGE 3\u201312% Bis-Tris gel (Invitrogen). After the electrophoresis (3\u2009h, 140\u2009V, 4\u2009\u00b0C), proteins were transferred by electroblotting onto a PVDF membrane (2\u2009h, 100\u2009V, 4\u2009\u00b0C, stirring), followed by immunodetection with an appropriate antibody (Supplementary Table\u00a04).\n\nThe capacity to polarize mitochondrial membrane was determined fluorometrically employing safranin O dye (Sigma S2255) in permeabilized cells. For each sample, 2 \u00d7 107 cells were harvested and washed with ANT buffer (8\u2009mM KCl, 110 mM K-gluconate, 10\u2009mM NaCl, 10\u2009mM free-acid Hepes, 10\u2009mM K2HPO4, 0.015\u2009mM EGTA potassium salt, 10\u2009mM mannitol, 0.5\u2009mg/ml fatty acid-free BSA, 1.5\u2009mM MgCl2, pH 7.25). The cells were permeabilized by 8\u2009\u03bcM digitonin in 2\u2009ml of ANT buffer containing 5\u2009\u03bcM safranin O. Fluorescence was recorded for 700\u2009s in a Hitachi F-7100 spectrofluorimeter (Hitachi High Technologies) at a 5\u2009Hz acquisition rate, using 495 and 585\u2009nm excitation and emission wavelengths, respectively. 1\u2009mM ATP (PanReac AppliChem A1348,0025) and 10\u2009\u03bcg/ml oligomycin (Sigma O4876) were added after 230\u2009s and 500\u2009s, respectively. Final addition of the uncoupler SF 6847 (250\u2009nM; Enzo Life Sciences BML-EI215-0050) served as a control for maximal depolarization. All experiments were performed at room temperature and constant stirring.\n\nATP production in digitonin-isolated mitochondria was performed as described previously39. Briefly, 1 \u00d7 108 cells per time point were lysed in SoTE buffer (600\u2009mM sorbitol, 2\u2009mM EDTA, 20\u2009mM Tris-HCl, pH 7.75) containing 0.015% (w/v) digitonin for 5\u2009min on ice. After centrifugation (3\u2009min, 4000\u2009\u00d7\u2009g, 4\u2009\u00b0C), the soluble cytosolic fraction was discarded and the organellar pellet was resuspended in 75\u2009\u03bcl of ATP production assay buffer (600\u2009mM sorbitol, 10\u2009mM MgSO4, 15\u2009mM potassium phosphate buffer pH 7.4, 20\u2009mM Tris-HCl pH 7.4, 2.5\u2009mg/ml fatty acid-free BSA). ATP production was induced by addition of 20\u2009mM DL-glycerol phosphate (sodium salt) and 67\u2009\u03bcM ADP. Control samples were preincubated with the inhibitors potassium cyanide (1\u2009mM) and carboxyatractyloside (6.5\u2009\u03bcM) for 10\u2009min at room temperature. After 30\u2009min at room temperature, the reaction was stopped by the addition of 1.5\u2009\u03bcl of 70% perchloric acid. The concentration of ATP was estimated using the Roche ATP Bioluminescence Assay Kit HS II in a Tecan Spark plate reader. The luminescence values of the RNAi induced samples were normalized to that of the corresponding noninduced sample.\n\nThe samples were centrifuged and pellet was transferred to the specimen carriers which were completed with 20% BSA and immediately frozen using high pressure freezer Leica EM ICE (Leica Microsystems). Freeze substitution was performed in the presence of 2% osmium tetroxide diluted in 100% acetone at \u221290\u2009\u00b0C. After 96\u2009h, specimens were warmed to \u221220\u2009\u00b0C at a slope 5\u2009\u00b0C/h. After the next 24\u2009h, the temperature was increased to 3\u2009\u00b0C (3\u2009\u00b0C/h). At room temperature, samples were washed in acetone and infiltrated with 25, 50, 75% acetone/resin EMbed 812 (EMS) mixture 1\u2009h at each step. Finally, samples were infiltrated in 100% resin and polymerized at 60\u2009\u00b0C for 48\u2009h. Ultrathin sections (70\u2009nm) were cut using a diamond knife, placed on copper grids, and stained with uranyl acetate and lead citrate. TEM micrographs were taken with Mega View III camera (SIS) using a JEOL 1010 TEM operating at an accelerating voltage of 80\u2009kV.\n\nMitochondria from 3 \u00d7 1011 cells were lysed by 1% (w/v) \u03b2-DDM in 60\u2009ml of 20\u2009mM Bis-tris propane pH 8.0 with 10% glycerol and EDTA-free Complete protease inhibitors (Roche) for 20\u2009min at 4\u2009\u00b0C. The lysate was cleared by centrifugation at 30,000\u2009\u00d7\u2009g for 20\u2009min at 4\u2009\u00b0C and adjusted to pH 6.8 by drop-wise addition of 1\u2009M 3-(N-morpholino) propanesulfonic acid pH 5.9. Recombinant TbIF1 without dimerization region, whose affinity to F1-ATPase was increased by N-terminal truncation and substitution of tyrosine 36 with tryptophan20, with a C-terminal glutathione S-transferase (GST) tag (TbIF1(9-64)-Y36W-GST) was added in approximately 10-fold molar excess over the estimated content of ATP synthase. Binding of TbIF1 was facilitated by the addition of neutralized 2\u2009mM ATP with 4\u2009mM magnesium sulphate. After 5\u2009min, sodium chloride was added to 100\u2009mM, the lysate was filtered through a 0.2 \u03bcm syringe filter and immediately loaded on 5\u2009ml GSTrap HP column (Cytiva) equilibrated in 20\u2009mM Bis-Tris-Propane pH 6.8 binding buffer containing 0.1% (w/v) glyco-diosgenin (GDN; Avanti Polar Lipids), 10% (v/v) glycerol, 100\u2009mM sodium chloride, 1\u2009mM tris(2-carboxyethyl)phosphine (TCEP), 1\u2009mM ATP, 2\u2009mM magnesium sulphate, 15\u2009\u03bcg/ml cardiolipin, 50\u2009\u03bcg/ml 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC), 25\u2009\u03bcg/ml 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoethanolamine (POPE) and 10\u2009\u03bcg/ml 1-palmitoyl-2-oleoyl-sn-glycero-3-[phospho-rac-(1-glycerol)] (POPG). All phospholipids were purchased from Avanti Polar Lipids (catalogue numbers 840012C, 850457C, 850757C, and 840757, respectively). ATP synthase was eluted with a gradient of 20\u2009mM reduced glutathione in Tris pH 8.0 buffer containing the same components as the binding buffer. Fractions containing ATP synthase were pooled and concentrated to 150 \u03bcl on Vivaspin centrifugal concentrator with 30\u2009kDa molecular weight cut-off. The sample was fractionated by size exclusion chromatography on a Superose 6 Increase 3.2/300 GL column (Cytiva) equilibrated in a buffer containing 20\u2009mM Tris pH 8.0, 100\u2009mM sodium chloride, 2\u2009mM magnesium chloride, 0.1% (w/v) GDN, 3.75\u2009\u03bcg/ml cardiolipin, 12.5\u2009\u03bcg/ml POPC, 6.25\u2009\u03bcg/ml POPE and 2.5\u2009\u03bcg/ml POPG at 0.03\u2009ml/min. Fractions corresponding to ATP synthase were pooled, supplemented with 0.05% (w/v) \u03b2-DDM that we and others experimentally found to better preserve dimer assemblies in cryo-EM40, and concentrated to 50\u2009\u03bcl.\n\nSamples were vitrified on glow-discharged Quantifoil R1.2/1.3 Au 300-mesh grids after blotting for 3\u2009s, followed by plunging into liquid ethane using a Vitrobot Mark IV. 5199 movies were collected using EPU 1.9 on a Titan Krios (ThermoFisher Scientific) operated at 300\u2009kV at a nominal magnification of 165\u2009kx (0.83\u2009\u00c5/pixel) with a Quantum K2 camera (Gatan) using a slit width of 20\u2009eV. Data was collected with an exposure rate of 3.6 electrons/px/s, a total exposure of 33 electrons/\u00c52, and 20 frames per movie.\n\nImage processing was performed within the Scipion 2 framework41, using RELION-3.0 unless specified otherwise. Movies were motion-corrected using the RELION implementation of the MotionCor2. 294,054 particles were initially picked using reference-based picking in Gautomatch (http://www.mrc-lmb.cam.ac.uk/kzhang/Gautomatch) and Contrast-transfer function parameters were using GCTF42. Subsequent image processing was performed in RELION-3.0 and 2D and 3D classification was used to select 100,605 particles, which were then extracted in an unbinned 560-pixel box (Fig.\u00a0S1). An initial model of the ATP synthase dimer was obtained using de novo 3D model generation. Using masked refinement with applied C2 symmetry, a 2.7\u2009\u00c5 structure of the membrane region was obtained following per-particle CTF refinement and Bayesian polishing. Following C2-symmetry expansion and signal subtraction of one monomer, a 3.7\u2009\u00c5 map of the peripheral stalk was obtained. Using 3D classification (T\u2009=\u2009100) of aligned particles, with a mask on the F1/c-ring region, 10 different rotational substates were then separated, and maps at 3.5\u20134.3\u2009\u00c5 resolution were obtained using 3D refinement. The authors note that the number of classes identified in this study likely reflects the limited number of particles, rather than the complete conformational space of the complex. By combining particles from all states belonging to main rotational state 1, a 3.7\u2009\u00c5 map of the rotor and a 3.2\u2009\u00c5 consensus map of the complete ATP synthase dimer with both rotors in main rotational state 1 were obtained.\n\nAn initial atomic model of the static Fo membrane region was built automatically using Bucaneer43. Subunits were subsequently assigned directly from the cryo-EM map, 15 of them corresponding to previously identified T. brucei ATP synthase subunits21, while three subunits (ATPTB14, ATPEG3, and ATPEG4) were identified herein using BLAST searches. Manual model building was performed in Coot 0.9.544 using the T. brucei F1 PDB 6F5D [https://www.rcsb.org/structure/6F5D] (T. brucei F1)13 and homology models45 of the E. gracilis OSCP and c-ring PDB 6TDU [https://www.rcsb.org/structure/6TDU] (E. gracilis mitochondrial ATP synthase)10 as starting models. Ligands were manually fitted to the map and restraints were generated by the GRADE server (http://grade.globalphasing.org). Cardiolipins were assigned based on the presence of a characteristic elongated density branched on both termini, corresponding to two phosphatidyl groups linked by the central glycerol bridge. Monophosphatidyl lipids were assigned based on their headgroup densities. Characteristic tetrahedral shapes of densities of choline groups served to distinguish phosphatidylcholines from elongated phosphatidylethanolamine head groups (Supplementary Fig.\u00a05g, h). Real-space refinement was performed in PHENIX 1.17.1 using auto-sharpened, local-resolution-filtered maps of the membrane region, peripheral stalk tip, c-ring/central stalk, and F1Fo monomers in different rotational states, respectively, using secondary structure restraints. Model statistics were generated using MolProbity46 and EMRinger47 Finally, the respective refined models were combined into a composite ATP synthase dimer model and real-space refined against the local-resolution-filtered consensus ATP synthase dimer map with both monomers in rotational state 1, applying reference restraints. Figures of the structures were prepared using ChimeraX 0.9148, the proton half-channels were traced using HOLLOW49.\n\nFurther information on research design is available in the\u00a0Nature Research Reporting Summary linked to this article.",
+            "section_image": []
+        },
+        {
+            "section_name": "Data availability",
+            "section_text": "The atomic coordinates generated in this study have been deposited in the Protein Data Bank (PDB) under the accession codes: 8AP7 (membrane-region), 8AP8 (peripheral stalk), 8AP9 (rotor), 8AP6 (F1Fo dimer), 8APA (rotational state 1a), 8APB (rotational state 1b), 8APC (rotational state 1c), 8APD (rotational state 1d), 8APE (rotational state 1e), 8APF (rotational state 2a), 8APG (rotational state 2b), 8APH (rotational state 2c), 8APJ (rotational state 2d), 8APK (rotational state 3). The local resolution filtered cryo-EM maps, half maps, masks and FSC-curves have been deposited in the Electron Microscopy Data Bank under accession codes: EMD-15560 (membrane-region), EMD-15561 (peripheral stalk), EMD-15562 (rotor), EMD-15559 (F1Fo dimer), EMD-15563 (rotational state 1a), EMD-15564 (rotational state 1b), EMD-15565 (rotational state 1c), EMD-15566 (rotational state 1d), EMD-15567 (rotational state 1e), EMD-15568 (rotational state 2a), EMD-15570 (rotational state 2b), EMD-15571 (rotational state 2c), EMD-15572 (rotational state 2d), EMD-15573 (rotational state 3). The TEM micrographs of thin cell sections are available from the authors upon request. All other data are available in the article, Supplementary Information or the Source Data file. Source data are provided with this paper.\n\nThe atomic coordinates that were used in this study: 6TDU (E. gracilis mitochondrial ATP synthase), 6TDV (E. gracilis mitochondrial ATP synthase, membrane region), 6B2Z (S. cerevisiae mitochondrial ATP synthase), 6F5D (T. brucei F1), 6ZNA (S. scrofa mitochondrial ATP synthase)\u00a0Source data are provided with this paper.",
+            "section_image": []
+        },
+        {
+            "section_name": "References",
+            "section_text": "Paumard, P. et al. The ATP synthase is involved in generating mitochondrial cristae morphology. EMBO J. 21, 221\u2013230 (2002).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nDavies, K. M., Anselmi, C., Wittig, I., Faraldo-Gomez, J. D. & Kuhlbrandt, W. Structure of the yeast F1Fo-ATP synthase dimer and its role in shaping the mitochondrial cristae. Proc. Natl Acad. Sci. USA 109, 13602\u201313607 (2012).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nPanek, T., Elias, M., Vancova, M., Lukes, J. & Hashimi, H. Returning to the fold for lessons in mitochondrial crista diversity and evolution. Curr. Biol. 30, R575\u2013R588 (2020).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKuhlbrandt, W. Structure and mechanisms of F-Type ATP synthases. Annu. Rev. Biochem. 88, 515\u2013549 (2019).\n\nArticle\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSpikes, T. E., Montgomery, M. G. & Walker, J. E. Structure of the dimeric ATP synthase from bovine mitochondria. Proc. Natl Acad. Sci. USA 117, 23519\u201323526 (2020).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nPinke, G., Zhou, L. & Sazanov, L. A. Cryo-EM structure of the entire mammalian F-type ATP synthase. Nat. Struct. Mol. Biol. 27, 1077\u20131085 (2020).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGuo, H., Bueler, S. A. & Rubinstein, J. L. Atomic model for the dimeric Fo region of mitochondrial ATP synthase. Science 358, 936\u2013940 (2017).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMurphy, B. J. et al. Rotary substates of mitochondrial ATP synthase reveal the basis of flexible F1-Fo coupling. Science 364, eaaw9128 (2019).\n\nFlygaard, R. K., M\u00fchleip, A., Tobiasson, V. & Amunts, A. Type III ATP synthase is a symmetry-deviated dimer that induces membrane curvature through tetramerization. Nat. Commun. 11, 5342 (2020).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMuhleip, A., McComas, S. E. & Amunts, A. Structure of a mitochondrial ATP synthase with bound native cardiolipin. Elife 8, e51179 (2019).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nM\u00fchleip, A. et al. ATP synthase hexamer assemblies shape cristae of Toxoplasma mitochondria. Nat. Commun. 12, 120 (2021).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGahura, O. et al. The F1-ATPase from Trypanosoma brucei is elaborated by three copies of an additional p18-subunit. FEBS J. 285, 614\u2013628 (2018).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMontgomery, M. G., Gahura, O., Leslie, A. G. W., Zikova, A. & Walker, J. E. ATP synthase from Trypanosoma brucei has an elaborated canonical F1-domain and conventional catalytic sites. Proc. Natl Acad. Sci. USA 115, 2102\u20132107 (2018).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSerricchio, M. et al. Depletion of cardiolipin induces major changes in energy metabolism in Trypanosoma brucei bloodstream forms. FASEB J. 35, 21176 (2020).\n\n\n                    Google Scholar\u00a0\n                \n\nMuhleip, A. W., Dewar, C. E., Schnaufer, A., Kuhlbrandt, W. & Davies, K. M. In situ structure of trypanosomal ATP synthase dimer reveals a unique arrangement of catalytic subunits. Proc. Natl Acad. Sci. USA 114, 992\u2013997 (2017).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSchnaufer, A., Clark-Walker, G. D., Steinberg, A. G. & Stuart, K. The F1-ATP synthase complex in bloodstream stage trypanosomes has an unusual and essential function. EMBO J. 24, 4029\u20134040 (2005).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBrown, S. V., Hosking, P., Li, J. & Williams, N. ATP synthase is responsible for maintaining mitochondrial membrane potential in bloodstream form Trypanosoma brucei. Eukaryot. Cell 5, 45\u201353 (2006).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGahura, O., Hierro-Yap, C. & Zikova, A. Redesigned and reversed: Architectural and functional oddities of the trypanosomal ATP synthase. Parasitology 148, 1151\u20131160 (2021).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHierro-Yap, C. et al. Bioenergetic consequences of FoF1-ATP synthase/ATPase deficiency in two life cycle stages of Trypanosoma brucei. J. Biol. Chem. 296, 100357 (2021).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGahura, O., Panicucci, B., Vachova, H., Walker, J. E. & Zikova, A. Inhibition of F1-ATPase from Trypanosoma brucei by its regulatory protein inhibitor TbIF1. FEBS J. 285, 4413\u20134423 (2018).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nZikova, A., Schnaufer, A., Dalley, R. A., Panigrahi, A. K. & Stuart, K. D. The F(0)F(1)-ATP synthase complex contains novel subunits and is essential for procyclic Trypanosoma brucei. PLoS Pathog. 5, e1000436 (2009).\n\nArticle\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nPerez, E. et al. The mitochondrial respiratory chain of the secondary green alga Euglena gracilis shares many additional subunits with parasitic Trypanosomatidae. Mitochondrion 19, 338\u2013349 (2014).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSathish Yadav, K. N. et al. Atypical composition and structure of the mitochondrial dimeric ATP synthase from Euglena gracilis. Biochim. Biophys. Acta 1858, 267\u2013275 (2017).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nDewar, C.E., Oeljeklaus, S., Wenger, C., Warscheid, B. & Schneider, A. Characterisation of a highly diverged mitochondrial ATP synthase Fo subunit in Trypanosoma brucei. J. Biol. Chem. 298, 101829 (2022)\n\nAphasizheva, I. et al. Lexis and grammar of mitochondrial RNA processing in trypanosomes. Trends Parasitol. 36, 337\u2013355 (2020).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBlum, B., Bakalara, N. & Simpson, L. A model for RNA editing in kinetoplastid mitochondria: \u201cguide\u201d RNA molecules transcribed from maxicircle DNA provide the edited information. Cell 60, 189\u2013198 (1990).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nAdler, B. K., Harris, M. E., Bertrand, K. I. & Hajduk, S. L. Modification of Trypanosoma brucei mitochondrial rRNA by posttranscriptional 3\u2019 polyuridine tail formation. Mol. Cell Biol. 11, 5878\u20135884 (1991).\n\nCAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHofer, A., Steverding, D., Chabes, A., Brun, R. & Thelander, L. Trypanosoma brucei CTP synthetase: A target for the treatment of African sleeping sickness. Proc. Natl Acad. Sci. USA 98, 6412\u20136416 (2001).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSobti, M. et al. Cryo-EM structures provide insight into how E. coli F1Fo ATP synthase accommodates symmetry mismatch. Nat. Commun. 11, 2615, https://doi.org/10.1038/s41467-020-16387-2 (2020).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGupta, K. et al. The role of interfacial lipids in stabilizing membrane protein oligomers. Nature 541, 421\u2013424 (2017).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nArnold, I., Pfeiffer, K., Neupert, W., Stuart, R. A. & Schagger, H. Yeast mitochondrial F1Fo-ATP synthase exists as a dimer: identification of three dimer-specific subunits. EMBO J. 17, 7170\u20137178 (1998).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGu, J. et al. Cryo-EM structure of the mammalian ATP synthase tetramer bound with inhibitory protein IF1. Science 364, 1068\u20131075 (2019).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSpikes, T. E., Montgomery, M. G. & Walker, J. E. Interface mobility between monomers in dimeric bovine ATP synthase participates in the ultrastructure of inner mitochondrial membranes. Proc. Natl Acad. Sci. USA 118, e2021012118 (2021).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nCadena, L. R. et al. Mitochondrial contact site and cristae organization system and F1Fo-ATP synthase crosstalk is a fundamental property of mitochondrial cristae. mSphere 6, e0032721 (2021).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nDavies, K. M. et al. Macromolecular organization of ATP synthase and complex I in whole mitochondria. Proc. Natl Acad. Sci. USA 108, 14121\u201314126 (2011).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBlum, T. B., Hahn, A., Meier, T., Davies, K. M. & K\u00fchlbrandt, W. Dimers of mitochondrial ATP synthase induce membrane curvature and self-assemble into rows. Proc. Natl Acad. Sci. USA 116, 4250\u20134255 (2019).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBochud-Allemann, N. & Schneider, A. Mitochondrial substrate level phosphorylation is essential for growth of procyclic Trypanosoma brucei. J. Biol. Chem. 277, 32849\u201332854 (2002).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nPoon, S. K., Peacock, L., Gibson, W., Gull, K. & Kelly, S. A modular and optimized single marker system for generating Trypanosoma brucei cell lines expressing T7 RNA polymerase and the tetracycline repressor. Open Biol. 2, 110037 (2012).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nAllemann, N. & Schneider, A. ATP production in isolated mitochondria of procyclic Trypanosoma brucei. Mol. Biochem. Parasitol. 111, 87\u201394 (2000).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nAibara, S., Dienemann, C., & Cramer, P. Structure of an inactive RNA polymerase II dimer. Nucleic Acids Res. 49, gkab783 (2021).\n\nde la Rosa-Trevin, J. M. et al. Scipion: A software framework toward integration, reproducibility, and validation in 3D electron microscopy. J. Struct. Biol. 195, 93\u201399 (2016).\n\nArticle\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nZhang, K. Gctf: Real-time CTF determination and correction. J. Struct. Biol. 193, 1\u201312 (2016).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nCowtan, K. The Buccaneer software for automated model building. 1. Tracing protein chains. Acta Crystallogr. D. Biol. Crystallogr. 62, 1002\u20131011 (2006).\n\nArticle\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nEmsley, P., Lohkamp, B., Scott, W. G. & Cowtan, K. Features and development of Coot. Acta Crystallogr. D Biol. Crystallogr. 66, 486\u2013501 (2010).\n\nWaterhouse, A. et al. SWISS-MODEL: Homology modelling of protein structures and complexes. Nucleic Acids Res. 46, W296\u2013W303 (2018).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWilliams, C. J. et al. MolProbity: More and better reference data for improved all\u2010atom structure validation. Protein Sci. 27, 293\u2013315 (2018).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBarad, B. A. et al. EMRinger: Side chain\u2013directed model and map validation for 3D cryo-electron microscopy. Nat. Methods 12, 943\u2013946 (2015).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGoddard, T. D. et al. UCSF ChimeraX: Meeting modern challenges in visualization and analysis. Protein Sci. 27, 14\u201325 (2018).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHo, B. K. & Gruswitz, F. HOLLOW: Generating accurate representations of channel and interior surfaces in molecular structures. BMC Struct. Biol. 8, 49 (2008).\n\nArticle\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nDownload references",
+            "section_image": []
+        },
+        {
+            "section_name": "Acknowledgements",
+            "section_text": "We are grateful to John E. Walker and Martin G. Montgomery for invaluable assistance with ATP synthase purification in the initial stage of the project. We acknowledge cryo-electron microscopy and tomography core facility of CIISB, Instruct-CZ Centre, supported by MEYS CR (LM2018127). This work was supported by the Czech Science Foundation grants number 18-17529S to A.Z. and 20-04150Y to O.G. and by European Regional Development Fund (ERDF) and Ministry of Education, Youth and Sport (MEYS) project CZ.02.1.01/0.0/0.0/16_019/0000759 to A.Z., Swedish Foundation for Strategic Research (FFL15:0325), Ragnar S\u00f6derberg Foundation (M44/16), European Research Council (ERC-2018-StG-805230), Knut and Alice Wallenberg Foundation (2018.0080), and EMBO Young Investigator Programme to A.A.",
+            "section_image": []
+        },
+        {
+            "section_name": "Funding",
+            "section_text": "Open access funding provided by Stockholm University.",
+            "section_image": []
+        },
+        {
+            "section_name": "Author information",
+            "section_text": "These authors contributed equally: Ond\u0159ej Gahura, Alexander M\u00fchleip.\n\nInstitute of Parasitology, Biology Centre, Czech Academy of Sciences, 37005, \u010cesk\u00e9 Bud\u011bjovice, Czech Republic\n\nOnd\u0159ej Gahura,\u00a0Carolina Hierro-Yap,\u00a0Brian Panicucci,\u00a0Minal Jain,\u00a0Martina Slapni\u010dkov\u00e1\u00a0&\u00a0Alena Z\u00edkov\u00e1\n\nScience for Life Laboratory, Department of Biochemistry and Biophysics, Stockholm University, 17165, Solna, Sweden\n\nAlexander M\u00fchleip\u00a0&\u00a0Alexey Amunts\n\nFaculty of Science, University of South Bohemia, 37005, \u010cesk\u00e9 Bud\u011bjovice, Czech Republic\n\nCarolina Hierro-Yap,\u00a0Minal Jain,\u00a0David Hollaus\u00a0&\u00a0Alena Z\u00edkov\u00e1\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nA.Z. and A.A. conceived and designed the work. O.G. prepared the sample for cryo-EM. O.G. and A.M. performed initial screening. A.M. processed the cryo-EM data and built the model. O.G., A.M., and A.A. analyzed the structure. B.P., C.H.Y., M.J., M.S., O.G., D.H., and A.Z. performed biochemical analysis. O.G., A.M., A.A., and A.Z. interpreted the data. O.G., A.M., A.A., and A.Z. wrote and revised the manuscript. All authors contributed to the analysis and approved the final version of the manuscript.\n\nCorrespondence to\n                Alena Z\u00edkov\u00e1 or Alexey Amunts.",
+            "section_image": []
+        },
+        {
+            "section_name": "Ethics declarations",
+            "section_text": "The authors declare no competing interests.",
+            "section_image": []
+        },
+        {
+            "section_name": "Peer review",
+            "section_text": "Nature Communications thanks the anonymous reviewer(s) for their contribution to the peer review of this work.\u00a0Peer reviewer reports are available.",
+            "section_image": []
+        },
+        {
+            "section_name": "Additional information",
+            "section_text": "Publisher\u2019s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.",
+            "section_image": []
+        },
+        {
+            "section_name": "Source data",
+            "section_text": "",
+            "section_image": []
+        },
+        {
+            "section_name": "Rights and permissions",
+            "section_text": "Open Access  This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article\u2019s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article\u2019s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.\n\nReprints and permissions",
+            "section_image": []
+        },
+        {
+            "section_name": "About this article",
+            "section_text": "Gahura, O., M\u00fchleip, A., Hierro-Yap, C. et al. An ancestral interaction module promotes oligomerization in divergent mitochondrial ATP synthases.\n                    Nat Commun 13, 5989 (2022). https://doi.org/10.1038/s41467-022-33588-z\n\nDownload citation\n\nReceived: 22 December 2021\n\nAccepted: 22 September 2022\n\nPublished: 11 October 2022\n\nVersion of record: 11 October 2022\n\nDOI: https://doi.org/10.1038/s41467-022-33588-z\n\nAnyone you share the following link with will be able to read this content:\n\nSorry, a shareable link is not currently available for this article.\n\n\n\n\n                            Provided by the Springer Nature SharedIt content-sharing initiative\n                        ",
+            "section_image": [
+                "https://data:image/svg+xml;base64,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"
+            ]
+        },
+        {
+            "section_name": "This article is cited by",
+            "section_text": "Cell Death & Differentiation (2025)\n\nNature Communications (2024)\n\nCell Death & Differentiation (2023)",
+            "section_image": []
+        }
+    ]
+}

0a8f1ca999301a2ab4d52f4b89787c7ef42bdb88f6c6bbf60d2710384ff32828/preprint/preprint.md

@@ -0,0 +1,429 @@
+Natural Variation of Chalk9 Regulates Grain Chalkiness in Rice
+
+Changjie Yan
+c.jyan@yzu.edu.cn
+Yangzhou University
+Zhi Hu
+Yangzhou University
+Hongchun Liu
+Yangzhou University
+Min Guo
+Yangzhou University
+Xiang Han
+Yangzhou University
+Youguang Li
+Yangzhou University
+Rujia Chen
+Yangzhou University   https://orcid.org/0000-0001-6744-3509
+Yifan Guo
+Yangzhou University
+Yihao Yang
+Yangzhou University
+Shengyuan Sun
+Huazhong Agricultural University
+Yong Zhou
+Yangzhou University
+Minghong Gu
+Yangzhou University
+DOI: https://doi.org/10.21203/rs.3.rs-5850266/v1
+
+License: This work is licensed under a Creative Commons Attribution 4.0 International License.
+Read Full License
+
+Additional Declarations: There is NO Competing Interest.
+
+Version of Record: A version of this preprint was published at Nature Communications on July 19th, 2025. See the published version at https://doi.org/10.1038/s41467-025-61683-4.
+Natural Variation of Chalk9 Regulates Grain Chalkiness in Rice
+
+Zhi Hu1,2#, Hongchun Liu1#, Min Guo1#, Xiang Han1, Youguang Li1, Rujia Chen1, Yifan Guo1, Yihao Yang1, Shengyuan Sun1, Yong Zhou1, Minghong Gu1 and Changjie Yan1,2*
+
+1Jiangsu Key Laboratory of Crop Genomics and Molecular Breeding/Zhongshan Biological Breeding Laboratory/Key Laboratory of Plant Functional Genomics of the Ministry of Education, College of Agriculture, Yangzhou University, Yangzhou 225009 Jiangsu, China
+
+2Co-Innovation Center for Modern Production Technology of Grain Crops of Jiangsu Province, Yangzhou University, Yangzhou 225009 Jiangsu, China
+
+# These authors contribute equally to this work.
+
+*Correspondence to: C.Y. (cjyan@yzu.edu.cn)
+Abstract
+Grain chalkiness is an undesirable trait affecting rice quality, concerning both consumers and breeders. However, the genetic mechanisms underlying rice chalkiness remain largely elusive. Here, we identified Chalk9 as a major gene associated with grain chalkiness in a natural population, explaining 28% of the observed variance. Chalk9 encodes an E3 ubiquitin ligase that targets OsEBP89 for its ubiquitination and degradation during the post-milk stage to balance storage component accumulation in the endosperm. However, low expression of Chalk9 results in excessive accumulation of OsEBP89, disrupting the homeostasis of storage components and leading to the chalkiness phenotype. A 64-bp insertion/deletion in the Chalk9 promoter contributes to its differential transcriptional levels, thus causing chalkiness variation among rice varieties. Moreover, the introgression of the elite allele Chalk9-L into a high-chalkiness rice variety reduced the chalky grain rate by up to 20% and the degree of chalkiness by up to 40%, without compromising yield. Chalk9-L was strongly selected during japonica rice domestication and gradually incorporated into modern indica breeding programs. Our findings reveal novel molecular and genetic mechanisms underlying chalkiness and provide a potential strategy for breeding novel rice variety with improved quality.
+Introduction
+
+Rice (Oryza sativa L.) is a staple food for over half of the global population, but enhancing its grain quality remains a significant challenge as living standards rise^{1,2}. Chalkiness, a major determinant of rice quality, severely reduces the appearance quality of rice and negatively affects milling, eating and cooking, thereby diminishing its commercial value^{3,4}. Chalkiness is an undesirable trait for consumers and marketing^{4}. Preventing grain chalkiness formation is thus a critical goal in rice breeding.
+
+Crop breeding is a dynamic and continuous process that strongly reflects human preferences^{5}. Over the past century, rice breeding efforts have primarily focused on enhancing rice productivity by developing high-yield varieties^{6}. However, these increased yields often come at the cost of poor quality, particularly high chalkiness^{2,7}. Seed storage proteins (SSPs) and starch, the predominant components in rice grains, determine both yield and quality. The negative correlation between yield and quality is likely arises from the disruption of their coordinated synthesis^{8,9}. Breaking this trade-off between yield and quality represents a breakthrough opportunity for rice breeders.
+
+Chalkiness, which refers to opaque regions in the endosperm, is a complex quantitative trait influenced by polygenes and environmental factors, such as high temperature and nutrient availability^{10-12}. Extensive efforts have been made to dissect the genetic basis of chalkiness in rice, and numerous quantitative trait loci (QTLs) related to chalkiness have been identified on all 12 rice chromosomes using biparental mapping and natural populations^{13-19}. Several genes have been functionally cloned and characterized. For example, *Chalk5* influences rice grain chalkiness by regulating pH homeostasis in developing seeds^{20}. Natural variation in *WCR1* regulates redox homeostasis in rice endosperm to affect grain chalkiness^{21}. Recent studies also identified *WBR7* and *LCG1* as regulators of rice chalkiness through their effects on the accumulation of grain storage components^{22,23}. Despite these advancements, the genetic and molecular mechanisms underlying rice grain chalkiness remain unclear.
+E3 ligases are critical components of the ubiquitin–proteasome system determining the substrate specificity of the cascade by covalent attachment of ubiquitin to target proteins\(^{24,25}\). RING-finger proteins, a major family of E3 ligases characterized by a 40-60 residue RING domain, confer substrate specificity through direct interaction with target proteins\(^{26}\). The RING domain, stabilized by zinc ions coordinated by cysteine and histidine residues, is essential for E3 activity. Mutations in these zinc-binding residues can disrupt the domain structure and abolish ligase activity\(^{26}\). E3 ligases play significant roles in plant growth, stress resistance, and signaling\(^{27-29}\); however, their role in regulating grain chalkiness remains unknown.
+
+In this study, we identified *Chalk9* as a major gene controlling chalkiness variation through genome-wide association studies (GWAS) in *indica* rice germplasm and elucidated the molecular mechanism of *Chalk9*-mediated chalkiness regulation. For breeding applications, we identified the elite haplotype *Chalk9*-L, which improves rice appearance quality without yield penalty. Our findings provide novel insight into the molecular mechanisms underlying rice chalkiness and offer promising strategies for breeding rice varieties with high quality.
+
+**Results**
+
+*Chalk9* is a major locus associated with grain chalkiness in *indica* rice
+
+To investigate the genetic basis of grain chalkiness, we collected 175 *indica* rice varieties from a global population with high phenotypic diversity in chalky grain rate (CGR) and degree of chalkiness (DC) (Supplementary Fig. 1a–d and Supplementary Table 1). Whole-genome sequencing of these varieties generated a final set of 2,290,145 high-quality single-nucleotide polymorphisms (SNPs) after filtering. Principal component analysis showed that the score plot of principal components had continuous distribution without any distinct clusters (Supplementary Fig. 1e), indicating that these *indica* varieties did not represent a highly structured population. In addition, the average decay of linkage disequilibrium (LD) distance was estimated about 180 kb in this
+population (\( r^2 = 0.1 \)) (Supplementary Fig. 1f), consistent with the previous estimation in cultivated rice\(^{30}\).
+
+Using a linear mixed model, we identified a major locus on chromosome 9, *Chalk9*, associated with CGR and DC in the 2-year trials through GWAS. This locus explained ~28% of the total phenotypic variation (Extended Data Fig. 1). In the overlapped peak, the top two SNPs associated with CGR were located at 19,506,938 bp (\( P = 8.12 \times 10^{-10} \)) and 19,536,079 bp (\( P = 7.25 \times 10^{-12} \)), while the top two SNPs associated with DC were located at 19,586,699 bp (\( P = 2.65 \times 10^{-10} \)) and 19,536,079 bp (\( P = 4.39 \times 10^{-11} \)) (Fig. 1a). LD analysis delimited the candidate region within an approximately 200-kb block from 19.43 to 19.63 Mb (Fig. 1b). Interestingly, *Chalk9* was located within the previously reported chalkiness-associated QTL regions, such as *qWBR9-1* and *qCR9-1*\(^{17}\).
+
+Using a relatively strict \( P \) value threshold (\( P < 1 \times 10^{-6} \)), we identified 76 SNPs that were significantly associated with chalkiness (Supplementary Table 2). Of these, 11 caused missense mutations, one SNP caused a synonymous mutation were in gene coding regions, 20 were in regulatory regions. These SNPs were assigned to 15 genes (Supplementary Table 2). The others were in the intergenic regions (14 SNPs) or gene introns (30 SNPs). For these 15 genes, three genes were annotated as either transposon-related or expressed proteins, the remaining 12 candidate genes were annotated as putative functional proteins (Supplementary Table 3).
+
+*LOC_Os09g32730* is the candidate of *Chalk9*
+
+To identify the candidate gene for *Chalk9*, we first evaluated SNPs causing amino acid substitutions in the 12 putative functional proteins. Only one SNP affected a functional domain (Supplementary Fig. 2a), but it was not conserved across plant species (Supplementary Fig. 2b), suggesting the missense SNP was unlikely to affect protein function. We then randomly selected eight lines from both high chalky-grain and low chalky-grain varieties to measure the expression levels of these 12 genes in endosperms
+and leaves by quantitative RT-PCR (qRT-PCR). Of the 12 genes, 11 showed no significant differences in the expression levels in the endosperms between the low chalky-grain and high chalky-grain varieties (Fig. 1c). Only gene III (*LOC_Os09g32730*) showed significantly higher expression in low chalky-grain varieties compared to high chalky-grain varieties (Fig. 1c, d). In contrast, these 12 genes exhibited similar expression levels of expression in leaves between high chalky-grain and low chalky-grain varieties (Fig. 1e). Notably, *LOC_Os09g32730* was preferentially expressed in the developing endosperm, compared to the other candidate genes (Supplementary Fig. 2c). Since grain chalkiness is closely associated with endosperm development, *LOC_Os09g32730* was identified as a potential candidate gene for the *Chalk9* locus. Hence, we designated this gene as *Chalk9*.
+
+To further validate *LOC_Os09g32730* as the candidate gene, we generated transgenic lines that either overexpressed *Chalk9* (OE) using the constitutive CaMV 35S promoter or interfered *Chalk9* using RNA interference (RNAi) in the Zhonghua11 (ZH11) background. Two *Chalk9*-overexpression lines (OE1 and OE2) displayed decreased chalkiness with lower CGR and DC values, whereas two *Chalk9*-RNAi lines exhibited increased chalkiness with higher CGR and DC values (Fig. 2a–d). The RANi lines developed in the variety Nipponbare (Nip) or Yangdao 6 (93-11) also displayed increased chalkiness (Supplementary Fig. 3a–h). Additionally, the CRISPR/Cas9 system was used to specifically disrupt the *Chalk9* gene in Nip (Supplementary Fig. 3i–k). Two knockout lines (*chalk9-1* and *chalk9-2*) demonstrated increased chalkiness (Fig. 2e–g). Collectively, these results strongly suggest that *LOC_Os09g32730* is the candidate of *Chalk9*, acting as a negative regulator of grain chalkiness in rice.
+
+**An indel in the *Chalk9* promoter confers grain chalkiness variation**
+
+To address potential limitations in identifying DNA sequence variations in *Chalk9* from low-coverage genome sequencing, we resequenced *Chalk9* and conducted an association analysis with the identified variants (Supplementary Table 4). Two indels (−1331, 64-bp and −791, 1-bp; referred to as v5 and v12) and four SNPs (−1355G>A,
+–817A>G, –749G>A and –634G>C; referred to as v4, v10, v14 and v15) in the promoter region of Chalk9 exhibited stronger associations with grain chalkiness than the top SNP (Extended Data Fig. 2a; Supplementary Table 4). However, a missense SNP in the coding region was not significantly associated with grain chalkiness (Extended Data Fig. 2a).
+
+Based on the identified variants, we classified Chalk9 variations into two haplotypes: one associated with high-chalky varieties [haplotype H (Chalk9-H)] and one with low-chalky varieties [haplotype L (Chalk9-L)] (Fig. 3a–c, Supplementary Table 5). Chalk9-L accessions showed significantly higher Chalk9 expression in the endosperm compared to Chalk9-H accessions (Fig. 3d). We further developed a near-isogenic line (NIL) carrying the Chalk9-H allele from the indica variety Kasalath in the japonica variety Nip, which had the Chalk9-L allele based on the known reference genome (Extended Data Fig. 2b, c). Compared to Nip, NIL^{Chalk9-H} plants exhibited significantly higher grain chalkiness with reduced Chalk9 expression (Fig. 3e–h). These results suggest that the two Chalk9 haplotypes confer different expression levels and variations in grain chalkiness in rice.
+
+To investigate whether the functional differences between the two Chalk9 haplotypes arise from the promoter variants, we created three transgenic constructs (pChalk9-L::Chalk9-L, pChalk9-H::Chalk9-H, and pChalk9-L::Chalk9-H), and used them to generate transgenic plants (see Methods). Compared to wild-type Guichao2 plants (Chalk9-H type), pChalk9-L::Chalk9-L and pChalk9-L::Chalk9-H transgenic lines showed significantly reduced grain chalkiness, with approximately 20% and 40% decreases in CGR and DC values, respectively (Fig. 3i–k). However, pChalk9-H::Chalk9-H transgenic plants showed no significant difference in grain chalkiness relative to wild-type Guichao2 (Fig. 3i–k). Consistent with the phenotypes of these transgenic lines, pChalk9-L::Chalk9-L and pChalk9-L::Chalk9-H transgenic lines showed higher Chalk9 transcript levels than wild-type Guichao2 and pChalk9-H::Chalk9-H transgenic plants (Fig. 3l). These findings indicate that the variants in the
+Chalk9 promoter are responsible for the differences in grain chalkiness.
+
+To further pinpoint the functional variations, we mutated the Chalk9-L promoter by introducing each of six variations individually (v4, v5, v10, v12, v14, and v15) from the Chalk9-H promoter. Transient expression assays showed that the activity of the Chalk9-L promoter was significantly reduced by deleting the 64-bp indel, to a level that was comparable to that of the Chalk9-H promoter (Fig. 3m, Extended Data Fig. 2d). By contrast, none of the other five mutations affected the activity of the Chalk9-L promoter (Fig. 3m, Extended Data Fig. 2d). We also generated gene-edited plants with a deletion in this 64-bp indel region in Nip (Chalk9-L type) (Extended Data Fig. 2e). The Chalk9-L gene-edited (D52) plants exhibited reduced Chalk9 expression and increased chalkiness (Fig. 3n-p, Extended Data Fig. 2f), further confirming that the 64-bp indel in the promoter as the causal variant.
+
+To understand why the 64-bp indel resulted in the different expression, we further analyzed its sequence and identified binding sites for some conserved transcription factors, including AT-Hook, TCR, B3, and ZF-HD families (Supplementary Table 6). Among these, a rice B3 domain transcription factor (OsB3), highly expressed in endosperms and homologous to ABI3 (essential for seed maturation in Arabidopsis\(^{31}\)), was found (Supplementary Fig. 4a, b). We found that OsB3 protein activated the promoter of Chalk9-L (Fig. 3q, Extended Data Fig. 2g). In the absence of the 64-bp sequence of the Chalk9-L promoter, the activation of OsB3 protein was significantly reduced (Fig. 3q, Extended Data Fig. 2g). These results demonstrate that the 64-bp sequence in the Chalk9-L promoter contains the DNA binding elements by the OsB3 protein in rice.
+
+Chalk9 exhibits E3 ubiquitin ligase activity
+
+To investigate the molecular function of Chalk9, we first analyzed its localization and expression pattern. The results showed that Chalk9 was localized in the nucleus and highly expressed in the developing endosperm with gradually increasing during grain
+filling (Fig. 4a, b). Similar results were observed through GUS staining (Supplementary Fig. 5a). Chalk9 is predicted to be a RING-C3HC4 type E3 ubiquitin ligase. To confirm its E3 ligase activity, we produced recombinant MBP-Chalk9 protein in Escherichia coli (E. coli) for in vitro ubiquitination assays. When ubiquitin, ubiquitin-activating enzyme (E1), and ubiquitin-conjugating enzyme (E2) were present, Chalk9 underwent auto-ubiquitination, whereas no ubiquitination was detected when E1, E2, or MBP-Chalk9 was absent (Fig. 4c). We mutated the conserved cysteine at position 189 to serine, creating the MBP-Chalk9C189S mutant (Supplementary Fig. 5b). The self-ubiquitination was abolished by the substitution in the RING finger domain (Fig. 4c), confirming that Chalk9 is a functional RING finger E3 ligase.
+
+Using yeast two-hybrid assays to screen the substrate of Chalk9, we successfully identified OsEBP89, a transcription factor involved in amylose biosynthesis^{32-34}, that interacts with Chalk9 (Fig. 4d). The C-terminal domain of OsEBP89 was found to be critical for this interaction (Supplementary Fig. 6a). This interaction was further validated by in vitro pull-down (Fig. 4e) and coimmunoprecipitation (CoIP) assays in rice protoplasts (Fig. 4f). Co-localization and bimolecular fluorescence complementation (BiFC) assays confirmed that the interaction between Chalk9 and OsEBP89 occurred in the nucleus (Fig. 4g, Supplementary Fig. 6b).
+
+Given that Chalk9 functions as an active RING finger E3 ligase and interacts with OsEBP89 (Fig. 4), we hypothesized that OsEBP89 is the direct substrate of Chalk9. In vitro ubiquitination assay was performed using MBP-Chalk9 and GST-OsEBP89. In the presence of E1, E2, ubiquitin, GST-OsEBP89 was ubiquitinated by MBP-Chalk9 (Fig. 5a). In contrast, no polyubiquitination was observed in the absence of E1, E2, ubiquitin or MBP-Chalk9 (Fig. 5a). Furthermore, replacing MBP-Chalk9 with the MBP-Chalk9C189S mutant failed to ubiquitinate GST-OsEBP89 (Fig. 5a), confirming that Chalk9 targeted OsEBP89 for ubiquitination.
+
+Since ubiquitination often leads to 26S proteasome-dependent degradation of target
+proteins, we tested whether Chalk9 influences the protein stability of OsEBP89 in a rice cell-free system. GST-OsEBP89 protein was expressed in E. coli, and purified protein was incubated in cell-free extracts from Nip and chalk9-1 seedlings. The GST-OsEBP89 protein was found to be more stable in the chalk9-1 mutant extract compared to Nip (Fig. 5b, c). The addition of MG132 significantly inhibited GST-OsEBP89 degradation in both Nip and chalk9-1 extracts (Fig. 5b), indicating that Chalk9 mediated the stability of OsEBP89 in vivo through the 26S proteasome system.
+
+To further determine OsEBP89 abundance in seeds from Nip and chalk9 mutants, we generated a specific antibody against OsEBP89 (Supplementary Fig. 6c). OsEBP89 showed more abundant in chalk9 mutants than in Nip, although the transcript levels of OsEBP89 remained unchanged (Fig. 5d–f). This suggests that the loss of Chalk9 function leads to the accumulation of OsEBP89 protein in rice. We also compared OsEBP89 protein levels between Nip (Chalk9-L type) and NIL^{Chalk9-H} plants. The OsEBP89 protein level in seeds was higher in NIL^{Chalk9-H} plants than in Nip (Fig. 5g, h), while OsEBP89 expression was similar (Fig. 5i), suggesting that Chalk9-L promotes more degradation of OsEBP89 than Chalk9-H.
+
+Chalk9–OsEBP89 module regulates grain chalkiness through regulation of the storage components in endosperm
+
+We observed that the chalk9 mutants produced white-belly endosperms (Supplementary Fig. 7a). Scanning electron microscopy revealed that the chalky endosperm of the chalk9 mutants contained loosely packed spherical starch granules interspersed with large air spaces, whereas the non-chalky endosperm of Nip consisted of densely and regularly packed polyhedral starch granules (Fig. 6a), which is consistent with previous studies^{35,36}. Although the total starch content remained unchanged, the chalk9 mutants exhibited significantly higher amylose content (Fig. 6b, c). Transmission electron microscopy further showed that chalky endosperm cells of the chalk9 mutants contained increased numbers and larger mean areas of spherical protein body I (PBI) and irregularly shaped PBII compared to Nip (Fig. 6d–f). This observation aligned with
+the greatly increased levels of seed storage proteins in chalk9 mutants, including glutelin, prolamin, and albumin (Fig. 6g–k). We further performed a transcriptome deep sequencing (RNA-seq) analysis on seeds from Nip and chalk9-1 (Supplementary Fig. 7b). A total of 2,658 differentially expressed genes were identified in chalk9-1 compared to Nip (Supplementary Fig. 7c, Supplementary Data 1). We found that the Waxy (Wx) gene for amylose synthesis and some genes for seed storage protein (SSP) exhibited significantly higher expression levels in chalk9-1 compared to Nip (Supplementary Fig. 7d and Supplementary Table 7). These findings were further validated by qRT-PCR analysis, which confirmed the increased expression of related genes in chalk9-1 (Supplementary Fig. 8).
+
+To investigate whether OsEBP89 is involved in chalkiness regulation, we generated OsEBP89 knockout plants (osebp89-1 and osebp89-2) using CRISPR/Cas9 (Extended Data Fig. 3a), and OsEBP89 overexpression lines (OsEBP89-OE1 and OsEBP89-OE2) driven by the constitutive CaMV 35S promoter (Extended Data Fig. 3b). Compared to the wild-type Nip, the OsEBP89 knockout plants showed a slight yet significant reduction in both CGR and DC values, whereas the OsEBP89-overexpression lines exhibited markedly increased chalkiness with higher CGR and DC values (Extended Data Fig. 3c–e). These results indicate that OsEBP89 positively regulates chalkiness in rice. Notably, a significant decrease in Wx expression was detected in OsEBP89 knockout mutants, whereas its expression increased in OsEBP89-overexpressing plants (Extended Data Fig. 3f), which is consistent with previous studies showing that OsEBP89 binds to the GCC box and GCC box-like sequences in the Wx promoter, thereby promoting its expression\(^{32-34}\). Several such binding sites were also identified in the promoters of SSP genes (Supplementary Table 8). The expression of SSP genes was greatly repressed in OsEBP89 knockout mutants but upregulated in OsEBP89-overexpressing plants (Extended Data Fig. 3g–l). Furthermore, yeast one-hybrid assays demonstrated that OsEBP89 directly bound to the promoters of SSP genes (Extended Data Fig. 3m).
+We crossed the chalk9-1 mutant with the osebp89-1 mutant to generate double mutant plants (chalk9-1/osebp89-1). While the chalk9-1 mutant showed increased chalkiness, the chalk9-1/osebp89-1 double mutant exhibited a reduced chalkiness, resembling the phenotype of osebp89-1 (Fig. 6l–n). In addition, the chalk9-1/osebp89-1 double mutant showed decreased amylose and total protein similar to osebp89-1 mutants, while the chalk9-1 mutant contained increased levels of both (Fig. 6o, p). Taken together, these results reveal that Chalk9-OsEBP89 module regulated the synthesis of grain storage components by modulating the expression of genes involved in storage components, thereby influencing chalkiness formation in rice.
+
+In addition, based on our whole-genome sequencing data, we observed that OsEBP89 had a single major haplotype in indica rice (Supplementary Table 9). Extending this analysis to 4,726 accessions of cultivated rice^{38-40}, this major haplotype occupied 97.5% of indica (Supplementary Table 10), indicating the strong genetic conservation and unlikely contribution to chalkiness variation in indica varieties. This result clearly demonstrates that Chalk9 plays an important role in determining the grain chalkiness.
+
+*Chalk9-L is artificially selected in cultivated rice during domestication and breeding*
+
+We performed a geographic distribution analysis of haplotypes in 1,424 cultivated varieties from the 3K Rice Genomes Project^{37}. The distribution of rice accessions carrying either Chalk9-L or Chalk9-H was variable across Asia regions relative to other areas (Extended Data Fig. 4a). The frequency of Chalk9-L was nearly 100% in Southeast Asia (e.g., Myanmar, Philippines, Laos, and Thailand), but it was relatively lower in China (71.1%) and South Asia, including Bangladesh (62%), Nepal (68.1%), Pakistan (70%), and India (76.2%) (Extended Data Fig. 4a). We further performed haplotype analysis in 4,726 accessions of cultivated rice^{38-40}. Eight out of 9 unique high-confidence haplotypes belonged to the Chalk9-L group, while only one belonged to the Chalk9-H group (Supplementary Table 11). Chalk9-L was present in 12.3% of Aus, 85.3% of aromatic, 99.9% of japonica, and 80.1% of indica varieties (Supplementary
+Table 12). Within the indica subgroups, its frequency was 40.9% in indica I, 96.6% in indica II, 94% in indica III, and 84% in indica intermediate (Supplementary Table 12).
+In 445 accessions of the wild ancestor Oryza rufipogon (O. rufipogon)38, O. rufipogon had a high frequency of Chalk9-L (89.4%) (Supplementary Table 13). These results suggest that the allele distribution of Chalk9 in different rice subgroups may be corelated to their evolution and selection.
+
+A selective sweep surrounding the Chalk9 locus was observed between japonica and wild rice, with significantly reduced nucleotide diversity in japonica compared to wild rice (Fig. 7a), indicating a strong artificial selection in Chalk9 locus of japonica. Tajima’s D values in the Chalk9 locus was significantly negative in japonica (Fig. 7b), reflecting directional selection across this region. In contrast, no obvious selection was detected in indica because the relative ratio of nucleotide diversity in indica to wild rice was higher than that in japonica to wild rice in Chalk9 locus (Fig. 7a). Further phylogenetic analysis showed that the Chalk9-L haplotype in japonica rice formed a tight cluster, while in indica rice, Chalk9-L was more widely distributed and genetically diverse (Fig. 7c). Haplotype network also showed that Chalk9-L in japonica was closely related to O. rufipogon, with few mutational differences, whereas Chalk9-L in indica exhibited more complex connections and mutational steps (Fig. 7d), suggesting that Chalk9-L in japonica evolved from O. rufipogon through a single lineage, while Chalk9-L in indica had a more complex evolution history with multiple origins.
+
+To trace the selection of Chalk9-L during indica rice breeding, we developed a 64-bp InDel marker in the Chalk9 promoter and genotyped Chalk9 in 127 indica varieties from the 1950s to the 2000s. The frequency of Chalk9-L in varieties prior to 1990 was relatively low, but it increased significantly thereafter (Extended Data Fig. 4b). This trend aligns with the significant reduction of chalkiness observed in indica varieties post-1990 (Extended Data Fig. 4c, d), indicating that Chalk9-L has been artificially selected in modern indica rice breeding programs. All 123 japonica varieties carried Chalk9-L (Extended Data Fig. 4b), consistent with the lower chalkiness observed
+(Extended Data Fig. 4e, f). These findings suggest that Chalk9-L might have been under artificial selection to reduce chalkiness.
+
+Chalk9-L holds the potential for breeding low-chalkiness rice cultivars without yield penalty
+
+We further investigated the effect of Chalk9 on yield. Chalk9 knockout plants displayed no significant differences from Nip in major agronomic traits, including heading date, tiller number, plant height, grain size and weight, as well as yield per plant and yield per plot (Supplementary Fig. 9). These results suggest that Chalk9 has no impact on rice yield. NIL^{Chalk9-H} plants showed no significant differences in grain weight or yield per plant compared to Nip (Chalk9-L type) (Supplementary Fig. 10a, b). Furthermore, introducing the Chalk9-L transgene into the high-yield variety Guichao2 significantly reduced chalkiness without affecting other agronomic traits, particularly yield per plant (Supplementary Fig. 10c–h), demonstrating the potential of Chalk9-L to reduce chalkiness in high-yield rice cultivars without compromising productivity.
+
+Discussion
+
+To date, little progress has been made in understanding the genetic and molecular mechanisms underlying natural variation associated with chalkiness in rice. Here we reported that Chalk9 is the major gene controlling chalkiness variation in indica rice. A 64-bp indel variant in Chalk9 promoter leads to differing expression levels, conferring chalkiness variation among rice varieties. Moreover, we deciphered a Chalk9-OsEBP89-Wx/SSP regulatory module that mediates chalkiness variation (Fig. 7e). These findings deepen our understanding of the genetic and molecular mechanisms underlying grain chalkiness variation in rice.
+
+Developing high-yielding rice with superior quality is challenging for rice breeding due to the trade-off between these traits\(^2\). One notable reason is that many QTLs associated with chalkiness are closely linked to yield-associated genes\(^{12,20}\). Fortunately, Chalk9
+does not exhibit such a linkage drag, as the yield in its near-isogenic lines shows no significant difference compared to the wild type (Supplementary Fig. 10a, b). Chalk9-L as an elite haplotype showed increased Chalk9 expression, conferring reduced chalkiness (Fig. 3a–h). By introducing this favorable allele into a well-known high-yielding indica variety but with high chalkiness, the chalkiness in the new lines was significantly decreased but did not compromise yield (Fig. 3i–k and Supplementary Fig. 10g, h).
+
+The distribution of Chalk9-L in cultivated rice appears to have been influenced by evolution and artificial selection during domestication and breeding. Our evolutionary analysis revealed that Chalk9 originated from wild rice but diverged significantly between japonica and indica rice (Fig. 7b–e). In japonica rice, Chalk9-L is likely derived from a single origin in O. rufipogon, while, in indica rice, Chalk9-L has multiple origins and exhibits greater genetic diversity. Moreover, the increasing incorporation of Chalk9-L in modern indica breeding programs has contributed to a significant reduction of chalkiness. In the light of that approximately 30% of indica varieties lack Chalk9-L and that Chalk9 explains 28% of the variance in chalkiness phenotype, our results strongly indicate that Chalk9-L is a key target for improving rice appearance quality of indica rice.
+
+The accumulated knowledge showed that the regulatory regions of genes involved in starch and storage protein biosynthesis usually share common motifs, which facilitates their co-regulation by common transcription factors, such as OsNAC20 and OsNAC26 in rice41. Similarly, OsEBP89 not only influences Wx expression but also regulates the expression of part of SSP genes, thereby coordinating the synthesis of amylose and storage proteins (Extended Data Fig. 3). In addition, Chalk9 acts as an E3 ubiquitin ligase, targeting OsEBP89 for ubiquitination and subsequent degradation via the 26S proteasome pathway (Figs. 4 and 5). This discovery underscores the critical role of the 26S proteasome in maintaining OsEBP89 protein homeostasis. Notably, recent research showed that OsSK41 phosphorylates OsEBP89, thereby reducing its stability34.
+Whether this phosphorylation is involved in Chalk9-mediated degradation of OsEBP89 remains to be elucidated.
+
+We propose that OsEBP89 is a positive regulator of chalkiness in rice. Genetic analysis demonstrates that Chalk9 operates in an OsEBP89-dependent manner to modulate the expression of genes involved in the biosynthesis of storage substances, thereby influencing chalkiness (Fig. 6, Supplementary Figs. 7 and 8). Notably, OsEBP89 exhibits a single major haplotype in indica varieties, highlighting its high conservation in indica rice. Consequently, the variation in chalkiness observed in indica rice is largely attributed to genetic variation in Chalk9. Moreover, our findings suggest that OsB3 acts as a potential upstream regulator of Chalk9, mediating its differential expression in response to the 64-bp indel. Future studies should aim to elucidate the role of OsB3 in regulating chalkiness and its contribution to chalkiness variation in rice. These efforts will help elucidate the OsB3-Chalk9-OsEBP89-Wx/SSP pathway in chalkiness regulation.
+
+Endosperm development involves the coordinated synthesis and accumulation of storage substances, a process closely associated with chalkiness. This developmental progress begins in the pre-milk stage, peaks during mid-milk, and tapers off in the post-milk stage^{42-44}. Similarly, Wx and SSP genes, which are central to this process, exhibit finely tuned temporal expression patterns that aligns with the synthesis of storage compounds^{45,46}. This coordination is crucial for optimizing grain quality by balancing biosynthetic processes that determine grain texture and appearance. Our findings reveal that Chalk9 expression gradually increases during endosperm development, reaching its peak in the post-milk stage (Fig. 4b), a period when the synthesis of storage substances naturally declines. At this stage, Chalk9 functions as a regulatory “brake”, limiting storage substance accumulation by promoting OsEBP89 degradation. This regulatory mechanism aligns with the natural decline in storage substance synthesis, supporting seed maturation and contributing to the formation of translucent grains. Thus, we propose a model in which the Chalk9-OsEBP89 regulatory module governs
+chalkiness variation in rice (Fig. 7e). In rice varieties carrying the Chalk9-H allele, reduced Chalk9 expression leads to OsEBP89 stabilization, which subsequently upregulates the expression of Wx and SSP genes. This increased synthesis of storage compounds disrupts the natural decline in their accumulation during the post-milk stage, resulting in the formation of chalky endosperm. In contrast, the Chalk9-L allele enhances Chalk9 expression, promoting OsEBP89 degradation. This reduction in OsEBP89 levels downregulates the expression of Wx and SSP genes, reducing storage product synthesis during the post-milk stage, leading to translucent grains and improved grain quality.
+
+Methods
+
+Plant materials and genotyping
+All 175 indica accessions, obtained from germplasm banks and breeders around the world, are listed in Supplementary Table 1. The japonica rice varieties (Nip and ZH11) and the indica rice varieties (93-11 and Guichao2) were used in this study. All rice materials used in this study were cultivated simultaneously during the summer in paddy fields at the experimental station of Yangzhou University, located in Yangzhou, China. The plants were grown under standardized crop management practices.
+
+Total genomic DNA was extracted from the samples and used to generate DNA sequencing libraries. Sequencing was performed, and the resulting libraries were size-checked using an Agilent 2100 Bioanalyzer system. The library preparations were ultimately sequenced on an Illumina Xten platform, producing 150 bp paired-end reads. After removing nucleotide variations with missing rates \( \geq 0.25 \) and minor allele frequency \( < 0.05 \), all nucleotide polymorphisms were categorized based on their location in the reference genome.
+
+Measurements of grain chalkiness and storage components
+Seeds harvested after full maturation were air-dried, stored at room temperature for
+three months. Images of 200-300 polished rice grains, randomly selected from each plant, were captured using a ScanWizard EZ scanner and analyzed with the rice quality TS-G automated analysis system (Hangzhou Shansheng Testing Technology Co., China). For chalkiness traits, the chalky grain rate (CGR) refers to the proportion of chalky grains among all rice grains, while the degree of chalkiness (DC) represents the extent of chalkiness in the rice grains. Total starch, amylose, total protein, and storage protein fractions were measured based on previously published methods47.
+
+Genome-wide association study
+GCR and DC were surveyed in 175 indica varieties over two years (2021 and 2023) and subsequently used for genome-wide association studies (GWAS). The analysis was performed using GEMMA (version 0.941), which fits a linear mixed model48. The \( P \)-value threshold for significance was set at \( 1 \times 10^{-5} \) using the Bonferroni correction method49, and the leading SNP was determined to be the SNP with the minimum \( P \)-value in the associated signal. Linkage disequilibrium (LD), evaluated as \( r^2 \), between SNPs in the 175 varieties was calculated using plink v1.950, and The LD heatmap surrounding the peak region was constructed using the LDBlockShow v1.4051.
+
+Constructs for genetic transformation
+For the Chalk9 RNA-interference vector, Chalk9-specific sequences from the coding region were amplified, and inserted in both sense and antisense orientations into a modified pTAC303-RNAi vector. For the Chalk9 overexpression vectors, the full-length coding sequence of Chalk9 from Nip was inserted into pCAMBIA2300-35S vector to generate the pCAMBIA2300-35S:Chalk9 construct. For the Chalk9 knockout vectors, two small-guide RNA (sgRNA) sequences targeting the Chalk9 coding region were cloned into pYLCRISPR/Cas9-MH vector to generate the Chalk9 CRISPR-Cas9 construct. Additionally, two sgRNA sequences from the Chalk9 promoter surrounding the 64-bp indel were designed and inserted into pYLCRISPR/Cas9-MH vector to generate the Chalk9 promoter-editing construct. For the Chalk9 promoter-GUS vector, a 2-kb genomic upstream region of Chalk9 was amplified and cloned into the
+pCAMBIA1381z vector.
+
+For the pChalk9-L::Chalk9-L vector, the 2,645-bp genomic region including the 2-kb upstream sequence and 645-bp coding sequence was amplified from low chalky-variety IR72 (Chalk9-L type) genomic sequence and cloned into plant binary vector pCAMBIA2300. The construct pChalk9-H::Chalk9-H contains the 2-kb upstream sequence and 645-bp coding sequence from high chalky-variety Guichao2 (Chalk9-H type). The 645-bp coding sequence from Guichao2 was driven by the 2-kb promoter sequence from IR72 to generate the pChalk9-L::Chalk9-H construct.
+
+For the OsEBP89 knockout vector, two sgRNA sequences targeting the OsEBP89 coding region were cloned into pYLCRISPR/Cas9-MH vector to generate the OsEBP89 CRISPR-Cas9 construct. For the OsEBP89 overexpression vector, the full-length coding sequence of OsEBP89 from Nip was inserted into pCAMBIA2300-35S vector to generate the pCAMBIA2300-35S:OsEBP89 construct. Agrobacterium-mediated transformation was used to generate transgenic rice plants. Primer sequences used in this study are listed in Supplementary Table 14.
+
+GUS analysis
+Various rice tissues, including young roots, stems, leaf sheaths, leaves, young panicles, and developing seeds from proChalk9::GUS transgenic plants, were stained with a GUS staining kit (Coolaber Biotech, Beijing, China) at 37°C in the dark for 12 hours, and then decolorized with 100% ethanol and imaged using a microscope (OLYMPUS, MVX10, Japan).
+
+Gene expression analysis
+After synthesizing first-strand cDNA from total RNA extracted from rice samples, quantitative PCR was performed using ChamQ SYBR qPCR Master Mix (Vazyme Biotech, Nanjing, China). Data analysis was conducted from three replicates for each experiment, using the rice OsActin (LOC_Os10g36650) gene as the internal reference.
+Gene-specific primers are listed in Supplementary Table 14.
+
+Transcriptome deep sequencing (RNA-seq) analysis
+Total RNA was isolated from seeds at 20 days after flowering (DAF), and RNA-seq libraries were prepared in triplicate from wild-type Nip and chalk9-1 mutant samples. RNA-seq and gene transcript abundance analysis were performed by the Bioacme Biotechnology Co., Ltd. (Wuhan, China). Differentially expressed genes were identified using DESeq2 with a \( P \)-value < 0.05 and \( |\text{Log}_2\text{FoldChange}| > 1 \). Correlation analysis, heatmap plotting, and volcano plot analysis were performed as previously described\(^{52}\).
+
+Transmission electron microscopy
+Seeds from WT and chalk9-1 mutant plants at 18 DAF were collected and used for transmission electron microscopy (TEM) samples, which were fixed and prepared as previously described\(^{52}\). Micrographs of the endosperm cells were captured on 80-nm ultra-thin sections using a transmission electron microscope.
+
+Scanning electron microscopy
+Brown rice grains were naturally broken from the middle and then coated with gold using an E-100 ion sputter coater. The morphology of starch granules was observed using a scanning electron microscope as previously described\(^{21}\). At least three biological replicates from different mature grains were analyzed.
+
+Subcellular colocalization assay
+The coding regions of IPA1 and Chalk9 were amplified by PCR and individually cloned into the 163-mCherry plasmid and the 163-GFP plasmid, respectively. Protoplasts isolated from 10-day-old Nip rice seedlings were transfected with the constructs as described previously\(^{53}\). GFP and mCherry were excited with 488-nm and 543-nm laser lines, respectively, and all fluorescence signals were detected at 500-580 nm and 565-615 nm using confocal laser-scanning microscopy. Images presented in the figures are
+representative of at least five protoplasts.
+
+Y2H assay
+For the Y2H screening, developing seeds at the reproductive stage were combined to construct a two-hybrid library by Shanghai OE Biotech Company. The coding sequence of Chalk9 was cloned into the pGBK7 vector and used as the bait. The yeast strain Y2H Gold was employed for transformation. To verify the interaction between OsEBP89 and Chalk9 in yeast, the coding sequences of OsEBP89 and mutated OsEBP89 variants (OsEBP89 [1-119], OsEBP89 [120-201], and OsEBP89 [202-326]) were separately cloned into the pGADT7 vector. Y2H assays were performed according to the manufacturer’s instructions (Clontech). Primers used for construction are listed in Supplementary Table 14.
+
+BiFC assay
+The coding regions of OsEBP89 and Chalk9 were amplified and cloned into the pUC-SPYCE and pUC-SPYNE vectors, respectively. The IPA1-mCherry vector served as a nuclear marker. The transfected protoplasts with the indicated constructs were observed using a fluorescence microscope. Yellow fluorescent protein (YFP) and mCherry were excited with 514-nm and 543-nm laser line, respectively, and detected at 522-555 nm and 565-615 nm. Images presented in the figures are representative of at least five protoplasts. Primers used for construction are listed in Supplementary Table 14.
+
+Co-IP assay
+The coding sequences of OsEBP89 and Chalk9 were amplified and cloned into the 163-GFP and pUC35S-HA vectors, respectively, to generate the OsEBP89-GFP and Chalk9-HA constructs. Total proteins for the Co-IP assay were extracted from protoplasts isolated from 10-day-old rice seedlings, which were transfected with the indicated constructs. Chalk9-HA was immunoprecipitated using anti-HA beads at 4 °C for 2 hours. The eluted proteins were analyzed by immunoblotting with anti-HA (1:3000, ab9110, Abcam) and anti-GFP (1:3000, ab290, Abcam) antibodies.
+In vitro GST pull-down assays
+The coding sequences of OsEBP89 and Chalk9 were cloned into the pGEX-5X-1 and pMAL-c5X vectors, respectively, to produce GST-OsEBP89 and MBP-Chalk9. The constructs were transformed into E. coli BL21 and induced with 0.2 mM IPTG for 12 hours at 16 °C to generate GST-OsEBP89 and MBP-Chalk9 recombinant proteins. GST-OsEBP89 and MBP-Chalk9 were purified using glutathione-sepharose resins (CW0190S; CWBIO) and amylose resins (E8021V; NEB), respectively, for Pull-down assays as described previously52. The eluted proteins were analyzed by immunoblotting with anti-GST (CW0084M; CWBIO) and anti-MBP (HT701; Transgene) antibodies.
+
+In vitro self-ubiquitination and substrate ubiquitination analyses
+Recombinant MBP-Chalk9 and its single amino acid substitution mutant (MBP-Chalk9C189S) were expressed in E. coli and purified using amylose resins (E8021V; NEB) for in vitro self-ubiquitination analyses. The ubiquitination assay was performed as previously described, with some modifications54. 400 μg of MBP-Chalk9, MBP-Chalk9C189S, or MBP protein was incubated in a 50-μL reaction mixture containing ubiquitination buffer (50 mM Tris-HCl, pH7.5, 5 mM MgCl2, 2 mM DTT, 4 mM ATP, 15 μg ubiquitin). The reaction was carried out at 30 °C for 2 hours in the presence or absence of 50 ng E1 (Beyotime, Shanghai, China) and 100 ng E2 (Beyotime). The reactions were stopped by the addition of 5×SDS sample buffer and heated at 95 °C for 5 minutes. The reaction products were separated on SDS-PAGE, followed by immunoblot analysis using an anti-MBP antibody (1:5000; HT701; Transgene) and a polyclonal anti-ubiquitin antibody (1:1,000, RM4934; Biodragon).
+
+For in vitro substrate ubiquitination assays, GST-OsEBP89 was used as the target substrate. 300 ng of the GST-OsEBP89 fusion protein was mixed with an equal amount of MBP-Chalk9 or MBP-Chalk9C189S in the presence or absence of the following: 50 ng of E1, 100 ng of E2, and 5 μg of ubiquitin. The reaction was performed in a 50 μL total mixture containing ubiquitination buffer at 30 °C for 3 hours. Ubiquitination levels
+of proteins were determined by Western blotting using a polyclonal anti-ubiquitin antibody (1:1,000, RM4934; Biodragon) and an anti-GST antibody (CW0084M; CWBIO).
+
+Cell-free degradation assays
+The leaf powder, frozen in liquid nitrogen from Nip and chalk9-1 plants, was suspended in extraction buffer (5 mM MgCl2, 40 mM Tris-HCl pH 7.5, 5 mM NaCl, 1 mM DTT, and 10 mM ATP) and vigorously vortexed at 4 °C for 1 hour. After centrifugation at 16,000 g at 4 °C for 30 minutes, the supernatant was collected for the cell-free degradation assay. The GST-OsEBP89 recombinant protein was incubated with the supernatant at 30 °C for different periods, with or without the addition of 50 mM MG132 (Beyotime). The reactions were terminated by adding 5 × SDS sample buffer and then immunoblotted using anti-GST (CW0084M; CWBIO) and anti-Actin (CW0264M; CWBIO) antibodies. The protein levels were quantified using ImageJ software (http://rsb.info.nih.gov/ij/).
+
+Yeast one-hybrid assay
+Yeast one-hybrid (Y1H) assays were performed using the Matchmaker™ Gold Yeast One-Hybrid System (Clontech). The coding sequence of OsEBP89 was fused to the activation domain of the GAL4 protein in the pGADT7 vector, generating the prey construct pGADT7-OsEBP89. 2-kb promoter sequences from GluB1a, GluB2, GluB4, PROLM20, PROLM22, and PROLM23 were individually inserted into the pAbAi vector, generating the bait constructs. The minimal inhibitory concentration of aureobasidin A (AbA) for the bait strains was determined for yeast one-hybrid assay. The prey construct pGADT7-OsEBP89 was then transformed into the recombinant bait-reporter strains. The interaction between the empty pGADT7 and the corresponding bait plasmid was considered a negative control. Yeast cells were grown on SD/-Leu culture media with or without AbA for 3–5 days at 30 °C. Primer sequences are shown in Supplementary Table 14.
+Luciferase activity assay in rice protoplasts
+To investigate the regulatory effect of the Chalk9 promoter on gene expression, approximately 2-kb promoter sequences of Chalk9 were amplified from Nip and Guichao2 and inserted into pGreenII 0800-LUC vector to generate proChalk9-L:LUC and proChalk9-H:LUC, respectively. Six variants were mutated based on proChalk9-L:LUC using a Fast Mutagenesis System (FM111, Transgen Biotech). All the vectors were transformed into protoplasts, respectively. Afterwards, all the protoplasts were incubated in W5 solution for 12 h at 28 °C. Activities of firefly luciferase (LUC) and Renilla luciferase (REN) were examined using a dual luciferase assay kit (Vazyme Biotech, Jiangsu, China). The primers used for PCR amplification and mutation are listed in Supplementary Table 14.
+
+To test the transcriptional activity of OsEBP89 proteins on SSP genes, 2-kb promoter sequences of GluB1a, GluB2, GluB4, PROLM20, PROLM22, and PROLM23 were cloned into the pGreenII 0800-LUC vector to create reporter constructs. The coding sequence of OsEBP89 was cloned into the pGreenII 62-SK vector to generate effector construct. For analyzing the transcriptional activity of the OsB3 protein on the Chalk9 alleles, 2-kb promoter sequences of Chalk9-L and Chalk9-L v5m were cloned into the pGreenII 0800-LUC vector to create reporter constructs, respectively. The coding sequence of OsB3 was cloned into the pGreenII 62-SK vector to generate the effector construct. The empty pGreenII 62-SK vector was used as a negative control. Plasmid combinations were co-transformed into rice protoplasts for transcriptional activity analysis. The transformed cells were incubated in the dark at 28 °C for 12 hours and then used to measure transcriptional activity using a dual luciferase assay kit (Vazyme Biotech, Jiangsu, China). The relevant primers are listed in Supplementary Table 14.
+
+Immunoblot analysis
+Developing seeds were homogenized in protein extraction buffer (2 mM EDTA, 100 mM NaCl, 20 mM Tris-HCl, pH 7.5, 0.1% [v/v] Triton X-100, 1 mM PMSF, and a 1 × proteinase inhibitor cocktail). Total proteins were collected after the homogenate was
+centrifuged at 16,000 g at 4 °C for 20 min. Western blotting was performed as previously described. Briefly, Protein samples were separated by 10% (w/v) SDS-PAGE and transferred to PVDF membranes (Immobilon-P, USA). Protein signals were detected using the eECL Western Blot Kit (CW0049S; CWBIO) after being probed with specific primary antibodies, followed by incubation with the appropriate secondary antibodies.
+
+OsEBP89 polyclonal antibody preparation
+To generate a specific antibody against OsEBP89, we chosen a truncated sequence (residues 1-120) for recombinant protein production. The corresponding coding sequence was amplified and cloned into the pET28a vector with an N-terminal His-tag. The recombinant protein was expressed in E. coli strain BL21 (DE3) transformed with the resulting construct and then purified using a Ni-NTI agarose resin matrix (Qiagen). The purified recombinant protein served as the antigen to raise antibodies in two rabbits, a process conducted by GenScript. The antibody against OsEBP89 was further affinity-purified from serum using immobilized recombinant protein and specifically detected endogenous OsEBP89.
+
+Population genetic and evolutionary analyses
+The geographical information and genomic sequences of 1,424 cultivated varieties were obtained from the 3K Rice Genomes Project\(^{37}\), and marked on map to observe geographic distribution of the two types of Chalk9. Using VCFtools v0.1.16\(^{55}\), the nucleotide diversity (\( \pi \)) and Neutral test (Tajima’s \( D \)) were calculated in 50-kb windows for each japonica, indica, and wild rice population. For all sites in the Chalk9 locus with a minor allele frequency \( \geq 0.01 \), phylogenetic and haplotypes network analyses were constructed following previously established methods\(^{56}\).
+
+The spatiotemporal gene expression and TFBS enrichment analysis
+The spatio-temporal gene expression pattern was analyzed by RiceXPro\(^{57}\). Additionally, the 64-bp sequence of the Chalk9 promoter was examined for transcription factor
+binding site (TFBS) enrichment using PlantPan v4.058.
+
+Statistical analysis
+Prism v.6.0 (GraphPad) software was used for all statistical tests and data visualization. The individual figures and figure legends indicated the sample sizes (\( n \)) and \( P \) values. For two groups, statistical significance was determined using two-tailed paired Student’s \( t \)-test. For more than two groups, statistical significance was determined using one-way analysis of variance (ANOVA) with Tukey’s multiple comparisons test.
+
+Accession numbers
+Sequence data related to this article can be obtained from the Rice Database (https://www.ricedata.cn/gene) under following accession numbers *LOC_Os09g32730* for *Chalk9*, *LOC_Os03g08460* for *OsEBP89*, *LOC_Os06g04200* for *Wx*, *LOC_Os02g15178* for *GluB1a*, *LOC_Os02g15150* for *GluB2*, *LOC_Os02g16830* for *GluB4*, *LOC_Os02g16820* for *GluB5*, *LOC_Os02g14600* for *GluB7*, *LOC_Os02g25640* for *GluC*, *LOC_Os02g15090* for *GluD*, *LOC_Os05g26350* for *PROLM4*, *LOC_Os05g26460* for *PROLM11*, *LOC_Os05g26368* for *PROLM13*, *LOC_Os05g26720* for *PROLM16*, *LOC_Os07g11910* for *PROLM20*, *LOC_Os07g11920* for *PROLM22*, and *LOC_Os06g31060* for *PROLM23*.
+
+Acknowledgements
+We thank Prof. L. Yan (Oklahoma State University, USA) for revising the paper. The pTAC303-RNAi vector was provided by K. Chong; The pYLCRISPR/Cas9-MH vector was provided by Y. Liu; and the pUC-SPYCE and pUC-SPYNE vectors were provided by R. Lin. This work was supported by the grants from the National Natural Sciences Foundation of China (32301828), the Biological Breeding-National Science and Technology Major Project (2023ZD04068), the programs of the Jiangsu Province Government (BE2022335, JBGS [2021]001, and BE2021334-1), and the Project of Zhongshan Biological Breeding Laboratory (ZSBBL-KY2023-01).
+Author contributions
+Z.H., H.L. and M.G. performed experiments and analyzed the data. Z.H. and X.H. collected phenotype data of rice varieties and genetic materials in the field. Y.L. conducted the GWAS analysis. R.C. performed the evolutionary analysis. Y.G., Y.Y., S.S., Y.Z. and M.G. participated in the experiments. Z.H. and C.Y. designed research and wrote the manuscript. C.Y. supervised the project. All authors reviewed the manuscript.
+
+Competing interests
+The authors declare no competing interests.
+
+References
+1. Tian, Z. et al. Allelic diversities in rice starch biosynthesis lead to a diverse array of rice eating and cooking qualities. Proc. Natl Acad. Sci. USA **106**, 21760–21765 (2009).
+2. Zeng, D. et al. Rational design of high-yield and superior-quality rice. *Nat. Plants* **3**, 17031 (2017).
+3. Zhao, D., Zhang, C., Li, Q. & Liu, Q. Genetic control of grain appearance quality in rice. *Biotechnol. Adv.* **60**, 108014 (2022).
+4. Fitzgerald, M. A., McCouch, S. R. & Hall, R. D. Not just a grain of rice: the quest for quality. *Trends Plant Sci.* **14**, 133–139 (2009).
+5. Zhang, J. et al. Releasing a sugar brake generates sweeter tomato without yield penalty. *Nature* **635**, 647–656 (2024).
+6. Sakamoto, T. & Matsuoka, M. Identifying and exploiting grain yield genes in rice. *Curr. Opin. Plant Biol.* **11**, 209–214 (2008).
+7. Zhao, D. et al. *GS9* acts as a transcriptional activator to regulate rice grain shape and appearance quality. *Nat. Commun.* **9**, 1240 (2018).
+8. Cao, S. et al. Orchestrating seed storage protein and starch accumulation toward overcoming yield–quality trade-off in cereal crops. *J. Integr. Plant Biol.* **66**, 468–483 (2024).
+9. She, K. C. et al. A novel factor *FLOURY ENDOSPERM2* is involved in
+regulation of rice grain size and starch quality. Plant Cell **22**, 3280–3294 (2010).
+
+10. Tabassum, R. et al. *FLOWY ENDOSPERM11-2* encodes plastid HSP70-2 involved with the temperature-dependent chalkiness of rice (*Oryza sativa* L.) grains. *Plant J.* **103**, 604–616 (2020).
+
+11. Custodio, M. C. et al. Rice quality: how is it defined by consumers, industry, food scientists, and geneticists? *Trends Food Sci. Tech.* **92**, 122–137 (2019).
+
+12. Sreenivasulu, N. et al. Designing climate-resilient rice with ideal grain quality suited for high-temperature stress. *J. Exp. Bot.* **66**, 1737–1748 (2015).
+
+13. Misra, G. et al. Genome-wide association coupled gene to gene interaction studies unveil novel epistatic targets among major effect loci impacting rice grain chalkiness. *Plant Biotechnol. J.* **19**, 910–925 (2021).
+
+14. Qiu, X. et al. Genetic dissection of rice appearance quality and cooked rice elongation by genome-wide association study. *Crop J.* **9**, 1470–1480 (2021).
+
+15. Gong, J. et al. Dissecting the genetic basis of grain shape and chalkiness traits in hybrid rice using multiple collaborative populations. *Mol. Plant* **10**, 1353–1356 (2017).
+
+16. Misra, G. et al. Dissecting the genome-wide genetic variants of milling and appearance quality traits in rice. *J. Exp. Bot.* **70**, 5115–5130 (2019).
+
+17. Peng, B. et al. Comparative mapping of chalkiness components in rice using five populations across two environments. *BMC Genet.* **15**, 49 (2014).
+
+18. Wan, X. Y. et al. Stability of QTLs for rice grain dimension and endosperm chalkiness characteristics across eight environments. *Theor. Appl. Genet.* **110**, 1334–46 (2005).
+
+19. Zhao, X. et al. Identification of stable QTLs causing chalk in rice grains in nine environments. *Theor. Appl. Genet.* **129**, 141–53 (2016).
+
+20. Li, Y. et al. *Chalk5* encodes a vacuolar H+-translocating pyrophosphatase influencing grain chalkiness in rice. *Nat. Genet.* **46**, 398–404 (2014).
+
+21. Wu, B. et al. Natural variation in *WHITE-CORE RATE 1* regulates redox homeostasis in rice endosperm to affect grain quality. *Plant Cell* **34**, 1912–1932 (2022).
+22. Tu, B. et al. The LCG1-OsBP5/OsEBP89-Wx module regulates the grain chalkiness and taste quality in rice. Plant Biotechnol. J. **23**, 36–50 (2025).
+23. Shi, H. et al. Natural variation of *WBR7* confers rice high yield and quality by modulating sucrose supply in sink organs. Plant Biotechnol. J. **22**, 2985–2999 (2024).
+24. Zheng, N. & Shabek, N. Ubiquitin ligases: structure, function, and regulation. Annu. Rev. Biochem. **86**, 129–157 (2017).
+25. Vierstra, R. D. The ubiquitin/26S proteasome pathway, the complex last chapter in the life of many plant proteins. Trends Plant Sci. **8**, 135–142 (2003).
+26. Deshaies, R. J. & Joazeiro, C. A. RING domain E3 ubiquitin ligases. Annu. Rev. Biochem. **78**, 399–434 (2009).
+27. Lee, H. K. et al. Drought stress-induced Rma1H1, a RING membrane-anchor E3 ubiquitin ligase homolog, regulates aquaporin levels via ubiquitination in transgenic *Arabidopsis* plants. Plant Cell **21**, 622–641 (2009).
+28. Wang, J. et al. Tissue-specific ubiquitination by IPA1 INTERACTING PROTEIN1 modulates IPA1 protein levels to regulate plant architecture in rice. Plant Cell **29**, 697–707 (2017).
+29. Sun, J., Sun, Y., Ahmed, R. I., Ren, A. & Xie, A. M. Research progress on plant RING-finger proteins. Genes **10**, 973 (2019).
+30. Huang, X. et al. Genome-wide association studies of 14 agronomic traits in rice landraces. Nat. Genet. **42**, 961–7 (2010).
+31. Gazzarrini, S. & Song, L. LAFL Factors in Seed Development and Phase Transitions. Annu. Rev. Plant Biol. **75**, 459–488 (2024).
+32. Yang, H. J. et al. The *OsEBP-89* gene of rice encodes a putative EREBP transcription factor and is temporally expressed in developing endosperm and intercalary meristem. Plant Mol. Biol. **50**, 379–391 (2002).
+33. Zhu, Y., Cai, X. L., Wang, Z. Y. & Hong, M. M. An interaction between a MYC protein and an EREBP protein is involved in transcriptional regulation of the rice *Wx* gene. J. Biol. Chem. **278**, 47803–47811 (2003).
+34. Hu, Z. et al. The kinase OsSK41/OsGSK5 negatively regulates amylose content
+in rice endosperm by affecting the interaction between OsEBP89 and OsBP5. J. Integr. Plant Biol. **65**, 1782–1793 (2023).
+
+35. Yan, H. et al. Rice LIKE EARLY STARVATION1 cooperates with FLOURY ENDOSPERM6 to modulate starch biosynthesis and endosperm development. Plant Cell **36**, 1892–1912 (2024).
+
+36. Wang, Y. et al. OsRab5a regulates endomembrane organization and storage protein trafficking in rice endosperm cells. Plant J. **64**, 812–824 (2010).
+
+37. Alexandrov, N. et al. SNP-Seek database of SNPs derived from 3000 rice genomes. Nucleic Acids Res. **43**, D1023–D1027 (2015).
+
+38. Huang, X. et al. A map of rice genome variation reveals the origin of cultivated rice. Nature **490**, 497–501 (2012).
+
+39. Wang, W. et al. Genomic variation in 3,010 diverse accessions of Asian cultivated rice. Nature **557**, 43–49 (2018).
+
+40. Zhao, H. *et al.* RiceVarMap: a comprehensive database of rice genomic variations. Nucleic Acids Res. **43**, D1018–D1022 (2015).
+
+41. Wang, J., Chen, Z., Zhang, Q., Meng, S. & Wei, C. The NAC transcription factors OsNAC20 and OsNAC26 regulate starch and storage protein synthesis. Plant Physiol. **184**, 1775–1791 (2020).
+
+42. Yamagata, H., Sugimoto, T., Tanaka, K. & Kasai, Z. Biosynthesis of storage proteins in developing rice seeds. Plant Physiol. **70**, 1094–1100 (1982).
+
+43. Yang, J. et al. Changes in activities of three enzymes associated with starch synthesis in rice grains during grain filling. Acta Agron. Sin. **27**, 157–164 (2001).
+
+44. Zhong, L. & Cheng, F. Varietal differences in amylose accumulation and activities of major enzymes associated with starch synthesis during grain filling in rice. Acta Agron. Sin. **29**, 452–456 (2003).
+
+45. Duan, M. & Sun, S. S. M. Profiling the expression of genes controlling rice grain quality. Plant Mol. Biol. **59**, 165–178 (2005).
+
+46. Liu, X. et al. Transcriptome analysis of grain-filling caryopses reveals involvement of multiple regulatory pathways in chalky grain formation in rice. BMC Genomics **11**, 730 (2010).
+47. Yang, Y. et al. Natural variation of OsGluA2 is involved in grain protein content regulation in rice. Nat. Commun. **10**, 1949 (2019).
+48. Zhou, X. & Stephens, M. Genome-wide efficient mixed-model analysis for association studies. Nat. Genet. **44**, 821–824 (2012).
+49. Si, L. et al. OsSPL13 controls grain size in cultivated rice. Nat. Genet. **48**, 447–456 (2016).
+50. Purcell, S. et al. PLINK: a tool set for whole-genome association and population-based linkage analyses. Am. J. Hum. Genet. **81**, 559–575 (2007).
+51. Dong, S. S. et al. LDBlockShow: a fast and convenient tool for visualizing linkage disequilibrium and haplotype blocks based on variant call format files. Brief. Bioinform. **22**, bbaa227 (2021).
+52. Hu, Z. et al. Autophagy targets Hd1 for vacuolar degradation to regulate rice flowering. Mol. Plant **15**, 1137–1156 (2022).
+53. Zhang, Y. et al. A highly efficient rice green tissue protoplast system for transient gene expression and studying light/chloroplast-related processes. Plant Methods **7**, 30 (2011).
+54. Huang, L. et al. The LARGE2-APO1/APO2 regulatory module controls panicle size and grain number in rice. Plant Cell **33**, 1212–1228 (2021).
+55. Danecek, P. et al. The variant call format and VCFtools. Bioinformatics **27**, 2156–2158 (2011).
+56. Chen, R. et al. A de novo evolved gene contributes to rice grain shape difference between indica and japonica. Nat. Commun. **14**, 5906 (2023).
+57. Sato, Y. et al. RiceXPro: a platform for monitoring gene expression in japonica rice grown under natural field conditions. Nucleic Acids Res. **39**, D1141–D1148 (2011).
+58. Chow, C. N. et al. PlantPAN 4.0: updated database for identifying conserved non-coding sequences and exploring dynamic transcriptional regulation in plant promoters. Nucleic Acids Res. **52**, D1569–D1578 (2024).
+Figure legends
+
+Fig. 1 GWAS and fine mapping of the major locus that underlies grain chalkiness variation.
+
+Fig. 2 Chalk9 negatively regulates grain chalkiness in rice.
+
+Fig. 3 A 64-bp indel in the Chalk9 promoter confers different grain chalkiness in rice.
+
+Fig. 4 Chalk9 is an E3 ubiquitin ligase that interacts with OsEBP89.
+
+Fig. 5 Chalk9 ubiquitinates OsEBP89 and regulates its stability.
+
+Fig. 6 Chalk9-OsEBP89 module regulates rice grain chalkiness by influencing seed storage substance biosynthesis.
+
+Fig. 7 Geographical distribution, genomic differentiation, and genomic selection of Chalk9 between japonica and indica subspecies.
+
+Extended Data Fig. 1 The genome-wide association study for rice grain chalkiness.
+
+Extended Data Fig. 2 The 64-bp indel in the Chalk9 promoter contributes to grain chalkiness variation.
+
+Extended Data Fig. 3 OsEBP89 positively regulates grain chalkiness in rice.
+
+Extended Data Fig. 4 Temporal and geographic patterns of Chalk9-L distribution and its impact on chalkiness in cultivated rice varieties.
+
+Supplementary Fig. 1 Variations of chalky grain rate and degree of chalkiness in 175 indica varieties.
+
+Supplementary Fig. 2 Functional importance estimation of SNPs located in the coding region.
+
+Supplementary Fig. 3 Identification of Chalk9 RNAi and knockout plants.
+Supplementary Fig. 4 Identification of the candidate genes in transcriptional factors analysis.
+
+Supplementary Fig. 5 The amino acid sequence of RING domain in Chalk9 is highly conserved in plants.
+
+Supplementary Fig. 6 Functional analysis of OsEBP89 and identification of the anti-OsEBP89 antibody.
+
+Supplementary Fig. 7 Transcript levels of storage substance-related genes in seeds of Nip and chalk9-1 plants from RNA-seq data.
+
+Supplementary Fig. 8 Transcript levels of storage substance-related genes in the seeds form Nip and chalk9-1 plants.
+
+Supplementary Fig. 9 Agronomic traits for chalk9 mutants.
+
+Supplementary Fig. 10 Agronomic traits for near-isogenic lines and transgenic plants.
+
+Supplementary Table 1. Chalky grain rate and degree of chalkiness of 175 indica accessions in two years.
+
+Supplementary Table 2. Annotation of significant SNPs associated with the grain chalkiness in the candidate region.
+
+Supplementary Table 3. Annotation of candidate genes on Chromosome 9 associated with the grain chalkiness by MSU Rice Genome Annotation Project Release 7.
+
+Supplementary Table 4. All variations of Chalk9 in 149 indica accessions were identified by re-sequencing based on PCR amplification.
+
+Supplementary Table 5. Major haplotypes of Chalk9 were identified from significant variations in indica accessions.
+
+Supplementary Table 6. Prediction of transcription factors binding to the 64-bp sequence in Chalk9.
+Supplementary Table 7. Expression analysis of starch and SSP genes generated from chalk9-1 vs Nip by RNA-seq.
+
+Supplementary Table 8. Putative cis-regulatory elements identified in the promoters of genes involved in starch and storage protein biosynthesis.
+
+Supplementary Table 9. The SNPs at the OsEBP89 gene region define one major haplotype in indica varieties.
+
+Supplementary Table 10. The SNPs at the OsEBP89 gene region from sequencing data of 4,726 rice accessions.
+
+Supplementary Table 11. The SNPs at the Chalk9 promoter region in 3K Rice Genomes Project.
+
+Supplementary Table 12. Allele frequency of Chalk9-L from sequencing data of 4,726 rice accessions.
+
+Supplementary Table 13. Allele frequency of Chalk9-L in common wild rice (O. rufipogon).
+
+Supplementary Table 14. Primers (5'-3') used in this study.
+
+Supplementary Data 1. 2,658 differentially expressed genes generated from chalk9-1 vs Nip.
+Fig. 1 GWAS and fine mapping of the major locus that underlies grain chalkiness variation. a, The genome-wide association signals for chalky grain rate (CGR) and degree of chalkiness (DC) in the region at 18–21 Mb on chromosome 9 (x axis) across two years. Negative log_{10}-transformed \( P \) values from the linear mixed model are plotted
+on the y-axis. The horizontal dashed line indicates the genome-wide significance threshold (\( P = 1 \times 10^{-6} \)). **b**, Linkage disequilibrium (LD) heatmap of the *Chalk9* locus region. Pairwise linkage disequilibrium was determined by calculating \( r^2 \) (the square of the correlation coefficient between SNPs). **c**, Relative expression level of the 12 candidate genes in the endosperm of eight high-chalk varieties and eight low-chalk varieties at 20 days after flowering (DAF). The 12 predicted genes in the *Chalk9* locus region are labeled by I to XII. Data show means ± SD (\( n = 8 \) varieties). **d**, Relative expression level of the candidate gene III (*Chalk9*) in the endosperm from the selected varieties at 20 DAF. Data show means ± SD (\( n = 3 \) biological replicates). **e**, Relative expression level of the 12 candidate genes in the leaves of high-chalk varieties and low-chalk varieties. Data show means ± SD (\( n = 8 \) varieties). **f**, Expression analysis of the candidate genes from GWAS in various tissues. The result of two genes (II and XI) was not found in RiceXPro database. L, leaf blade; LS, leaf sheath; R, root; S, stem; I, inflorescence; A, anther; P, pistil; L/P, lemma/palea; O, ovary; Em, embryo; En, endosperm. In **c** and **e**, statistical analysis was performed by two-tailed Student’s *t*-test; for *P* values, see Source Data.
+Fig. 2 Chalk9 negatively regulates grain chalkiness in rice. a, Grain chalkiness in ZH11, ZH11-OE1, ZH11-OE2, ZH11-RNAi-1, and ZH11-RNAi-2 plants. Scale bar: 5 mm. b, Expression analysis of Chalk9 in ZH11 and transgenic plants. Data show means ± SD (n = 3 biological replicates). c,d, Chalky grain rate (c) and degree of chalkiness (d) in ZH11 and transgenic plants. Data show means ± SD (n = 16 plants). e, Grain chalkiness in Nip, chalk9-1, and chalk9-2 plants. Scale bar: 5 mm. f,g, Chalky grain rate (f) and degree of chalkiness (g) in Nip, chalk9-1, and chalk9-2 plants. Data show means ± SD (n = 16 plants). In b, c, d, f, and g, different letters indicate significant differences (P < 0.05, one-way ANOVA with Tukey’s multiple comparison test); for P values, see Source Data.
+Fig. 3 A 64-bp indel in the Chalk9 promoter confers different grain chalkiness in rice. a, Major haplotypes of Chalk9. v4, v5, v10, v12, v14 and v15 indicate the variants, and their positions relative to ATG are shown in the table. b,c, The distribution of chalky grain rate (b) and degree of chalkiness (c) in haplotype H (\( n = 45 \) accessions) and
+haplotype L (\( n = 102 \) accessions). **d**, Expression analysis of *Chalk9* in haplotype H (\( n = 24 \) accessions) and haplotype L (\( n = 24 \) accessions) in endosperms. **e**, Grain chalkiness of Nip and NIL*Chalk9-H* plants. Scale bar: 5 mm. **f,g**, Chalky grain rate (**f**) and degree of chalkiness (**g**) of Nip and NIL*Chalk9-H* plants. Data show means ± SD (\( n = 16 \) plants). **h**, Relative *Chalk9* expression levels of Nip and NIL*Chalk9-H* plants in endosperms. Data show means ± SD (\( n = 3 \) biological replicates). **i**, Grain chalkiness of wild-type Guichao2, *pChalk9-H::Chalk9-H*, *pChalk9-L::Chalk9-L*, and *pChalk9-L::Chalk9-H* plants. Scale bar: 6 mm. **j,k**, Chalky grain rate (**j**) and degree of chalkiness (**k**) of Guichao2 and different transgenic plants. Data show means ± SD (\( n = 10 \) plants). **l**, Relative *Chalk9* expression levels of Guichao2 and different transgenic plants in endosperms. Data show means ± SD (\( n = 3 \) biological replicates). **m**, Transient expression assays of the effect of different variations on gene expression, shown by firefly luciferase/Renilla luciferase activity ratio (LUC/REN). v4m, v5m, v10m, v12m, v14m and v15m represent the mutations introduced into the promoter of *Chalk9*-L. Data show means ± SD (\( n = 3 \) biological replicates). **n,o**, Degree of chalkiness (**n**) and chalky grain rate (**o**) in Nip and *D52* plants. Data show means ± SD (\( n = 16 \) plants). **p**, Relative *Chalk9* expression levels of Nip and *D52* plants in endosperms. Data show means ± SD (\( n = 3 \) biological replicates). **q**, Indel of 64 bp resulted in divergent activation of the OsB3 protein to *Chalk9* promoter, as shown by LUC/REN. Data show means ± SD (\( n = 3 \) biologically replicates). In **b-d**, the bars within violin plots represent 25th percentiles, medians, and 75th percentiles. In **b-d, f-h**, and **n-p**, statistical analysis was performed by two-tailed Student’s *t*-test. In **j-m** and **q**, different letters indicate significant differences (\( P < 0.05 \), one-way ANOVA with Tukey’s multiple comparison test); for *P* values, see Source Data.
+Fig. 4 Chalk9 is an E3 ubiquitin ligase that interacts with OsEBP89. a, Subcellular localization of Chalk9-GFP fusion protein in rice protoplasts. IPA1-mCherry was used as a nuclear marker. Scale bars: 5 μm. b, Quantitative PCR with reverse transcription (qRT–PCR)-based transcript abundance analysis of Chalk9 in various tissues. R, root;
+S, stem; L, leaf; LS, leaf sheath; P, panicle; DAF, days after flowering. Data show means ± SD (\( n = 3 \) biological replicates). **c**, Ubiquitin ligase activity of Chalk9. MBP-Chalk9 was expressed in *E. coli* strain BL21, and ubiquitinated proteins were detected using both anti-MBP and anti-ubiquitin (Ub) antibodies. **d**, Yeast two-hybrid (Y2H) assay showing the interaction between Chalk9 and OsEBP89. Strains carrying the indicated constructs were grown on synthetic dropout medium. DDO and QDO/X represent SD/–Trp–Leu and SD/–Trp–Leu–His–Ade + X-Gal selection medium, respectively. **e**, Pull-down assay. GST-OsEBP89 was used as baits, and the pull down of MBP-Chalk9 was detected by the anti-MBP antibody. **f**, Co-immunoprecipitation (Co-IP) assay of rice protoplasts co-expressing Chalk9-HA and GFP-OsEBP89. Total proteins were incubated with magnetic agarose beads conjugated to HA-tag antibody. The immunoprecipitants were probed with antibodies against HA and GFP. IP, immunoprecipitation. **g**, Interaction between the Chalk9 and OsEBP89 demonstrated by bimolecular fluorescence complementation (BiFC) assays in rice protoplasts. N-terminal fragment of YFP (nYFP) fused with Chalk9 and the C-terminal fragment of YFP (cYFP) fused with OsEBP89 were co-expressed in rice protoplasts. IPA1-mCherry was used as a nuclear control.
+Fig. 5 Chalk9 ubiquitinates OsEBP89 and regulates its stability. a, In vitro ubiquitination of OsEBP89 by Chalk9. Ubiquitinated proteins were detected using anti-GST and anti-Ub antibodies. b, Cell-free degradation of GST-OsEBP89 in the protein extracts from Nip and chalk9-1 seedlings. Protein levels of GST-OsEBP89 were detected using anti-GST antibody, and Actin was used as a loading control for total protein extraction. Relative fold changes of GST-OsEBP89 to Actin loading controls
+were quantified by ImageJ and marked on bottom of the lanes. The protein level at time point 0 min was marked as 1. c, GST-OsEBP89 degradation rate in Nip and chalk9-1 seedlings. d, Detection of OsEBP89 protein abundance in Nip and chalk9-1 plants. Total proteins were extracted from seeds at 18 DAF. OsEBP89 protein abundance was determined by immunoblotting. e, Relative quantification of OsEBP89 protein abundance in Nip, chalk9-1, and chalk9-2 plants. OsEBP89 protein levels were quantified relative to Actin by ImageJ. The protein level at time point 0 min was set as 1. f, OsEBP89 mRNA level in Nip, chalk9-1, and chalk9-2 plants. g, Detection of OsEBP89 protein abundance in Nip and NILChalk9-H plants. Total proteins were extracted from seeds at 15 DAF. OsEBP89 protein abundance was determined by immunoblotting. h, Relative quantification of OsEBP89 protein abundance in the Nip and NILChalk9-H plants. i, OsEBP89 mRNA level in Nip and NILChalk9-H plants. In c, e, f, h, and i, data show means ± SD (n = 3 biological replicates). In c, h, and i, statistical analysis was performed by two-tailed Student’s t-test (**P < 0.01); for P values, see Source Data.. In e and f, different letters indicate significant differences (P < 0.05, one-way ANOVA with Tukey’s multiple comparison test); for P values, see Source Data.
+Fig. 6 Chalk9-OsEBP89 module regulates rice grain chalkiness by influencing seed
+storage substance biosynthesis. a, The scanning electron microscopy observation of transverse sections of mature seeds from Nip, chalk9-1, and chalk9-2 plants. Scale bars: 0.8 mm (upper), 5 μm (down). b,c, Starch (b) and amylose (c) contents of Nip, chalk9-1, and chalk9-2 plants. d, Transmission electron microscopy analysis of the endosperm cells from Nip, chalk9-1, and chalk9-2 plants at 18 DAF. Scale bars: 5 μm (upper), 2 μm (down). White asterisk indicates PBI; red asterisk indicates PBII. e,f, Number (per 400 μm^2) of protein bodies (e) and mean area of protein bodies (f) in the endosperms from Nip, chalk9-1, and chalk9-2 plants. g-k, Total protein (g), glutelin (h), prolamin (i), albumin (j), and globulin (k) contents of Nip, chalk9-1, and chalk9-2 plants. l, Grain chalkiness of Nip, chalk9-1, osebp89-1 and chalk9-1/osebp89-1 plants. Scale bar: 5 mm. m-p, Degree of chalkiness (m), chalky grain rate (n), amylose (o), and total protein (p) in Nip, chalk9-1, osebp89-1, and chalk9-1/osebp89-1 plants. In b, c, g-k, and m-p, data show means ± SD (\( n = 9 \) plants); different letters indicate significant differences (\( P < 0.05 \), one-way ANOVA with Tukey’s multiple comparison test); for \( P \) values, see Source Data. In e and f, data show means ± SD (\( n = 3 \)); statistical analysis was performed by two-tailed Student’s \( t \)-test (**\( P < 0.01 \); *\( P < 0.05 \)***); for \( P \) values, see Source Data.
+Fig. 7 Geographical distribution, genomic differentiation, and genomic selection of Chalk9 between japonica and indica subspecies. a,b, The relative ratio of nucleotide diversity (a) and Tajima’s D (b) analyses in the whole chromosome 9 of cultivated and wild rice. Red dashed line indicates the Chalk9 locus. c,d Phylogeny (c) and haplotype networks (d) generated from the genomic sequences of Chalk9 in both cultivated and wild rice varieties. Outer circle of the tree indicates various rice populations. Circle size of the network is proportional to the number of samples for each haplotype. Black spots on the lines indicate mutational steps between two
+haplotypes. e, A proposed model for the Chalk9–OsEBP89 module in the regulation of grain chalkiness. In rice varieties with the Chalk9-H allele, Chalk9 expression in the endosperm is relatively lower, which reduces the degradation of OsEBP89. This accumulation of OsEBP89 leads to the upregulation of \( Wx \) and \( SSP \) genes, resulting in increased levels of amylose and storage protein in the endosperm. This elevated synthesis of storage compound during the post-milk stage contributes to the formation of chalky grains. Conversely, in rice varieties with the Chalk9-L allele, Chalk9 is highly expressed, which accelerates OsEBP89 degradation. The reduction in OsEBP89 levels leads to the downregulation of \( Wx \) and \( SSP \) genes in endosperm, resulting in decreased synthesis of storage substances during the post-milk stage and the formation of translucent grains.
+Extended Data Fig. 1 The genome-wide association study for rice grain chalkiness.
+
+a,b, Manhattan plots and Quantile-quantile plots for chalky grain rate (a) and degree of chalkiness (b) in 175 indica varieties in 2021. c,d, Manhattan plots and Quantile-quantile plots for chalky grain rate (c) and degree of chalkiness (d) in 175 indica varieties in 2023. The horizontal dash-dot line indicates the genome-wide significant threshold (\( P = 1 \times 10^{-5} \)).
+Extended Data Fig. 2 The 64-bp indel in the Chalk9 promoter contributes to grain chalkiness variation. a, Structure of Chalk9 and association mapping with 28 variants. Red dots connected with the dashed lines indicate the six variants that are significantly associated with chalkiness. x axis, position relative to ATG (0 bp). b, Identification and diagram of the NIL Chalk9-H line. The numbering above the line represents the molecular markers used for the construction of NIL Chalk9-H. The green bar indicates the genomic
+region from Kasalath (Chalk9-H type). The double-headed arrow shows the length of the substitution segment. **c**, Representative photographs of Nip and NIL*Chalk9-H* plants at the mature stage in the field. Scale bar: 10 cm. **d**, Schematic representation of the reporter constructs for the luciferase assay in **Fig. 3m**. v4m, v5m, v10m, v12m, v14m and v15m represent the mutations introduced into the promoter of *Chalk9*-L. **e**, Diagram of the *Chalk9* promoter sequences of Nip and *D52* plants. Red box indicates the position of the 64-bp indel. **f**, Grain chalkiness of Nip and *D52* plants. Scale bar: 5 mm. **g**, Schematic diagrams of the effector and reporter plasmids used in the luciferase assays from **Fig. 3q**.
+Extended Data Fig. 3 OsEBP89 positively regulates grain chalkiness in rice. a,
+Loss-of-function mutants (osebp89-1 and osebp89-2) of OsEBP89 generated using
+CRISPR/Cas9 on the wild-type Nip (WT). The 20-bp target sequence for CRISPR/Cas9-mediated editing is underlined. **b**, Expression analysis of *OsEBP89* in WT, *OsEBP89-OE1*, and *OsEBP89-OE2* plants. Data show means ± SD (*n* = 3 biological replicates). **c**, Grain chalkiness of WT, *osebp89-1*, *osebp89-2*, *OsEBP89-OE1*, and *OsEBP89-OE2* plants. Scale bar: 5 mm. **d,e**, Chalky grain rate (**d**) and degree of chalkiness (**e**) in WT, *osebp89-1*, *osebp89-2*, *OsEBP89-OE1*, and *OsEBP89-OE2* plants. Data show means ± SD (*n* = 16 plants). **f-l**, Expression analysis of *Wx* (**f**), *GluB1a* (**g**), *GluB2* (**h**), *GluB4* (**i**), *PROLM20* (**j**), *PROLM22* (**k**), *and PROLM23* (**l**) genes from WT, *osebp89-1*, *osebp89-2*, *OsEBP89-OE1*, and *OsEBP89-OE2* plants. Data show means ± SD (*n* = 3 biological replicates). **m**, Y1H assay demonstrated the interaction between OsEBP89 and the promoters of *GluB1a*, *GluB2*, *GluB4*, *PROLM20*, *PROLM22*, and *PROLM23* genes. In **b** and **d-l**, different letters indicate significant differences (*P* < 0.05, one-way ANOVA with Tukey’s multiple comparison test); for *P* values, see Source Data.
+Extended Data Fig. 4 Temporal and geographic patterns of Chalk9-L distribution and its impact on chalkiness in cultivated rice varieties. a, Geographic distributions of 1,424 cultivated rice varieties. blue and orange circles indicate the Chalk9-L and Chalk9-H type, respectively. b, The frequency of Chalk9-L in the cultivated indica varieties from different years (~1960-2000s) and the cultivated japonica varieties. c, d, The Chalky grain rate (c) and degree of chalkiness (d) in indica varieties from two periods: the 1980s and earlier versus the 1990s and later. e,f, Chalky grain rate (e) and degree of chalkiness (f) in japonica varieties compared with indica varieties.
+Supplementary Files
+
+This is a list of supplementary files associated with this preprint. Click to download.
+
+• SupplementaryTable.xlsx
+• SupplementaryData.xlsx
+• SupplementaryFigure.pdf

0acd819abe48c47a84cbc21e7054100bd0d1363bbc114db43e3926604a5e536e/preprint/preprint.md

@@ -0,0 +1,182 @@
+Electrically Driven Spin Resonance of 4f Electrons in a Single Atom on a Surface
+
+Yujeong Bae
+bae.yu.jeong@qns.science
+
+IBS Center for Quantum Nanoscience   https://orcid.org/0000-0002-9983-8529
+Stefano Reale
+IBS Center for Quantum Nanoscience
+Jiyoon Hwang
+IBS Center for Quantum Nanoscience
+Jeongmin Oh
+IBS Center for Quantum Nanoscience
+Harald Brune
+Ecole Polytechnique Fédérale de Lausanne (EPFL)   https://orcid.org/0000-0003-4459-3111
+Andreas Heinrich
+Institute for Basic Science at Ewha Womans University   https://orcid.org/0000-0001-6204-471X
+Fabio Donati
+IBS Center for Quantum Nanoscience   https://orcid.org/0000-0002-3932-2889
+
+Article
+
+Keywords:
+
+Posted Date: November 7th, 2023
+
+DOI: https://doi.org/10.21203/rs.3.rs-3385164/v1
+
+License: This work is licensed under a Creative Commons Attribution 4.0 International License.
+Read Full License
+
+Additional Declarations: There is NO Competing Interest.
+
+Version of Record: A version of this preprint was published at Nature Communications on June 20th, 2024. See the published version at https://doi.org/10.1038/s41467-024-49447-y.
+Electrically Driven Spin Resonance of 4f Electrons in a Single Atom on a Surface
+
+Stefano Reale1,2,3, Jiyoon Hwang1,4, Jeongmin Oh1,4, Harald Brune5, Andreas J. Heinrich1,4, Fabio Donati1,4*, and Yujeong Bae1,4*
+
+1 Center for Quantum Nanoscience (QNS), Institute for Basic Science (IBS), Seoul 03760, Republic of Korea
+2 Ewha Womans University, Seoul 03760, Republic of Korea
+3 Department of Energy, Politecnico di Milano, Milano 20133, Italy
+4 Department of Physics, Ewha Womans University, Seoul 03760, Republic of Korea
+5 Institute of Physics, Ecole Polytechnique Fédérale de Lausanne, 1015 Lausanne, Switzerland
+
+*Corresponding authors: F.D. (donati.fabio@qns.science), Y.B. (bae.yujeong@qns.science)
+
+A pivotal challenge in quantum technologies lies in reconciling long coherence times with efficient manipulation of the quantum states of a system. Lanthanide atoms, with their well-localized 4f electrons, emerge as a promising solution to this dilemma if provided with a rational design for manipulation and detection. Here we construct tailored spin structures to perform electron spin resonance on a single lanthanide atom using a scanning tunneling microscope. A magnetically coupled structure made of an erbium and a titanium atom enables us to both drive the erbium’s 4f electron spins and indirectly probe them through the titanium’s 3d electrons. In this coupled configuration, the erbium spin states exhibit a five-fold increase in the spin relaxation time and a two-fold increase in the driving efficiency compared to the 3d electron counterparts. Our work provides a new approach to accessing highly protected spin states, enabling their coherent control in an all-electric fashion.
+
+The last two decades have witnessed a rising focus on the control and application of quantum coherent effects, marking the advent of the so-called “second quantum revolution”. Utilizing quantum coherent functionalities of materials for novel technologies, such as imaging, information processing, and communications, requires robustness of their quantum coherence, addressability, and scalability1. However, these requirements often clash since decoupling the quantum states from the environment prolongs the quantum coherent properties but hinders the possibility of efficient state manipulation.
+Lanthanide atoms represent a promising platform to tackle this dilemma. Their well-localized \(4f\) electrons show long spin relaxation \(T_1^{2,3}\) and coherence times \(T_2^{4,5}\). In addition, their strong hyperfine interaction facilitates the read-out of nuclear spins\(^{6,7}\). In bulk insulators, exceedingly long \(T_1\) and \(T_2\) have been demonstrated using optical control and detection\(^{8-11}\) down to the single atom level\(^{12,13}\). While hybrid optical-electrical approaches have been developed to access individual lanthanide atom’s spins embedded in a silicon transistor\(^{14}\), it is still challenging to achieve efficient control of the quantum states using electrical transport methods. This necessitates the rational design of a quantum platform capable of tackling both control and detection schemes, along with their interactions with local environments. In this context, single crystal surfaces constitute an advantageous framework both for building atomically engineered nanostructures and addressing individual spin centers, in particular using probe techniques\(^{15-18}\). However, coherent manipulation and detection of surface-adsorbed lanthanide atoms have so far remained elusive.
+
+In this work, we demonstrate the control and detection of \(4f\) electron spins by building atomic-scale structures on a surface using a scanning tunneling microscope (STM) with electron spin resonance (ESR) capabilities\(^{19-22}\). The atomic structures are composed of an erbium (Er) atom as the target spin system and a magnetically coupled titanium (Ti) atom as the sensor spin. This architecture allows us to drive ESR transitions on the Er \(4f\) electrons with a projected angular momentum of \(½^{23}\) and to probe them indirectly through Ti. We observed an Er \(T_1\) of close to \(1\ \mu s\), 5 times longer than what was previously measured in \(3d\) electrons with spin \(½\) on the same surface\(^{18}\). This novel platform allows for the ESR driving and read-out of the well-screened \(4f\) electron spin states, paving the way to integrate lanthanide atoms in quantum architectures.
+
+Sensing Er Spin States through a Ti Atom
+
+Erbium atoms on a few monolayer-thick MgO(100) on Ag(100) present a \(4f^{11}\) configuration with no unpaired electrons in the \(5d\) and \(6s\) shells\(^{23}\). The atomic-like spin and orbital momenta are coupled through the large spin-orbit interaction into a total angular momentum \(J_{\text{Er}}\) with magnitude of \(15\hbar/2^{23}\). When adsorbed on the oxygen site of MgO (Fig. 1a), the crystal field leads to a strong hard-axis magneto-crystalline anisotropy that stabilizes a doubly-degenerate ground state with an out-of-plane component of the angular momentum \(±\hbar/2^{23}\), which splits into two singlets when an external magnetic field (\(\mathbf{B}\)) is applied. As found in a previous work\(^{23}\), the component of \(J_{\text{Er}}\) along the magnetic field direction (z), defined as \(J_z\), increases from \(±\hbar/2\) to \(±4\hbar\) by rotating \(\mathbf{B}\) from the out-of-plane (\(\theta = 0^\circ\)) to the in-plane (\(\theta = 90^\circ\)) direction
+(Fig. 1b), while retaining a large probability for spin dipole transitions. Given these properties, Er can be regarded as a highly tunable two-level system allowing for efficient ESR driving. To characterize the magnetic states and anisotropy of Er, we utilized the dipole field sensing technique\(^{24}\) with a Ti atom on the bridge binding site of MgO as a well-known spin sensor. On this binding site, Ti has a spin \( S_{Ti} \) of magnitude \( \hbar/2 \) and a relatively weak g-factor anisotropy\(^{25}\) with respect to the oxygen binding site\(^{26}\).
+
+We deposited Er and Ti at cryogenic temperatures (\(~10\) K) on 2 monolayers of MgO grown on Ag(100) (Methods and Fig. S1a). Their binding sites on the surface can be changed by atom manipulation (Supplementary Section 2). Figure 1c shows the ESR spectra obtained on Ti in an Er-Ti dimer with the atomic separation of 0.928 nm (Fig. S2b). For \( \theta = 8^\circ \), we observed one ESR peak at the resonance frequency of Ti which splits into two peaks separated by \( \Delta f = 334 \pm 3 \) MHz when rotating \( \mathbf{B} \) close to the in-plane direction (\( \theta = 68^\circ \)). The two ESR peaks stem from the magnetic interaction with the Er spin fluctuating between two states\(^{24}\) during the measurement, with the relative peak intensity being proportional to the time-averaged population of the Er states. The pronounced difference in the relative intensity of the ESR peaks indicates a large imbalance in the Er state occupation even at \( B = 0.3 \) T and 1.3 K, which reflects the large \( J_z \) of Er at \( \theta = 68^\circ \) (Fig. 1b). The sign of this asymmetry depends on the character of the magnetic interactions between the two atoms. In Fig. 1c, the peak at the lower frequency is less intense than the one at the higher frequency and, hence, the interaction can be regarded as ferromagnetic\(^{27}\).
+
+The angle dependence of \( \Delta f \) (Fig. 1d) gives a direct measurement of the Er-Ti interaction energy and of its anisotropy\(^{24,27}\). To interpret it, we model the system through a spin-Hamiltonian including both the single atom Zeeman and anisotropy terms, as well as the interaction between the two spins:
+
+\[
+H = \mu_B g_{Er} \mathbf{B} \cdot \mathbf{J}_{Er} + DJ_\perp^2 + \mu_B \mathbf{B} \cdot \mathbf{g}_{Ti} \cdot \mathbf{S}_{Ti} + H_{\text{dip}} + H_{\text{exc}}.
+\]
+
+Here, \( \mu_B \) is the Bohr magneton, \( J_\perp \) is the out-of-plane component of \( \mathbf{J}_{Er} \), \( g_{Er} = 1.2 \) is the Er g-factor , and \( \mathbf{g}_{Ti} \) is the Ti anisotropic g-tensor\(^{25}\). We use a magnetic anisotropy parameter \( D = 2.4 \) meV to match the Er energy splitting found in a previous study\(^{23}\). The magnetic coupling consists of dipolar (\( H_{\text{dip}} \)) and Heisenberg exchange interactions (\( H_{\text{exc}} \)):
+
+\[
+H_{\text{dip}} = \frac{\mu_0 \mu_B^2}{4 \pi \hbar^2 r^3} [g_{Er} J_{Er} \cdot \mathbf{g}_{Ti} \cdot \mathbf{S}_{Ti} - 3 (\hat{r} \cdot g_{Er} J_{Er}) (\hat{r} \cdot \mathbf{g}_{Ti} \cdot \mathbf{S}_{Ti})],
+\]
+\[
+H_{\text{exc}} = \frac{J_{\text{exc}}}{\hbar^2} (\mathbf{J}_{Er} \cdot \mathbf{S}_{Ti}),
+\]
+where \( \mu_0 \) is the vacuum permittivity, \( r \) the separation between the two atoms, \( \hat{r} \) the unit vector connecting them\(^{24}\), and \( J_{exc} \) the exchange interaction energy expressed in terms of \( J_{Er}^{28} \). In our model, \( J_{exc} \) is the only free parameter for the fit. As shown in Fig. 1d, our model accurately reproduces the data for \( J_{exc}/h = -48 \) MHz, where the negative sign indicates an antiferromagnetic coupling. This value is more than 20 times smaller than that observed for a Ti-Ti dimer at the same distance (−1.16 GHz)\(^{29}\). We ascribe the smaller Er-Ti coupling to the localization of the 4f orbitals near the atom’s core, which limits the overlap between Er and Ti orbitals when compared to the Ti-Ti case.
+
+The strong angle dependence of \( \Delta f \) can be understood by considering the large magneto-crystalline anisotropy of \( J_{Er} \). At \( \theta = 90^\circ \), \( J_z \) is largest (4ħ) and the angular momenta of both atoms are parallel to \( \hat{r} \) (Fig. 1e), which maximizes the contribution of the dipolar coupling with a positive sign (ferromagnetic). When rotating \( \mathbf{B} \) away from the in-plane direction, \( S_{Ti} \) follows the direction of \( \mathbf{B} \), while the anisotropy of Er preserves a large component of \( J_{Er} \) mainly aligned along the in-plane direction (Fig. 1f). This misalignment between the two angular momenta reduces the dipolar interaction. Finally, as \( \mathbf{B} \) approaches the surface normal (Fig. 1g), \( J_{Er} \) turns towards the out-of-plane direction with a much smaller value of \( J_z = \hbar/2 \). With the two momenta being perpendicular to \( \hat{r} \), the dipolar contribution is minimal and negative (antiferromagnetic). Conversely, the mutual projection of \( S_{Ti} \) and \( J_{Er} \) is the only factor modulating the exchange interaction term, which remains negative (antiferromagnetic) at all angles.
+Fig. 1 | Probing Er 4f electron spins through a Ti spin sensor. a, Schematic of the experimental set-up for ESR-STM measurement of an Er-Ti dimer built on MgO/(100)/Ag(100). The Ti atom (purple) is positioned close to the Er atom (orange) and located under a spin-polarized (SP) STM tip. The external magnetic field (\( \mathbf{B} \)) defines the z-direction and is applied at an angle \( \theta \) from the out-of-plane direction. A dc voltage \( V_{dc} \) is applied to the STM junction while the radio-frequency (rf) voltage is applied to the tip or to the antenna with an amplitude \( V_{rf} \). b, The projected total angular momentum of Er (\( J_z \)) onto the \( \mathbf{B} \) field direction as a function of \( \theta \). The strong magnetic anisotropy favors an in-plane alignment of \( J_z \). c, ESR spectra of the Ti atom placed 0.928 nm apart from the Er atom at different \( \theta \). At \( \theta = 8^\circ \), a single ESR peak is visible (pink) while, at \( \theta = 68^\circ \) (purple), the two ESR peaks are separated due to the magnetic interactions between the Er and Ti (set-point: \( V_{dc} = 50 \) mV, \( I_{dc} = 20 \) pA, \( V_{rf} = 12 \) mV, \( B = 0.3 \) T). The spectrum at \( \theta = 8^\circ \) (pink) was normalized at its maximum intensity while the spectrum at \( \theta = 68^\circ \) (purple) was normalized to the sum of the intensities of its two peaks. The frequency detuning is defined with respect to 9.1 GHz (8.1 GHz) for the spectrum at \( \theta = 8^\circ \) (\( \theta = 68^\circ \)). d, ESR peak separation, \( \Delta f \), as a function of \( \theta \). The experimental points (black dots) were acquired at different set-points (\( V_{dc}
+= 50 mV, \( I_{dc} = 12\text{-}30 \) pA, \( V_{rt} = 12\text{-}20 \) mV, \( B = 0.3 \) T). The total interaction (solid purple line) calculated by the model Hamiltonian is composed of a dipolar contribution (dashed blue line) and an exchange contribution (dashed pink line). e–g, Schematic of the angular momenta of Er and Ti on MgO/Ag(100). The dipolar fields induced by Er are depicted as black curved arrows. When \( \mathbf{B} \) is applied along the in-plane direction (\( \theta = 90^\circ \)), the \( J_z \) is maximum and aligned with the spin of Ti giving the largest ferromagnetic interaction. When \( \mathbf{B} \) is rotated, the spin of Ti follows the direction of \( \mathbf{B} \) while the total angular momentum of Er is aligned preferentially in-plane (f). In the out-of-plane direction (\( \theta = 0^\circ \)), \( J_z \) is minimum and aligned with the spin of Ti (g) giving a small antiferromagnetic interaction.
+
+Spin Resonance of Er 4f Electrons
+
+The direct drive of ESR in STM requires positioning the tip directly on top of the target atom\(^{19}\). However, we observed no ESR when positioning the tip over an Er atom (Fig. S3b), which we attribute to the small polarization of the 5\( d \) and 6\( s \) shells of Er and to the weak interaction between the 4\( f \) and tunneling electrons. These factors were found to limit the tunneling magnetoresistance at the STM junction in other lanthanide atoms\(^{30,31}\), possibly hindering both the ESR drive and detection\(^{23}\).
+
+To overcome this limitation, we built a strongly interacting Er-Ti dimer by positioning Ti at 0.72 nm from Er through atom manipulation (Fig. 2a and Supplementary Section 2). Similar to the isolated atom, we observed no ESR peaks at the Er position in the dimer (yellow curve in Fig. 2b). However, when the tip was positioned on Ti, we observed up to 5 peaks (pink curve in Fig. 2b). The first two peaks below 10 GHz with \( \Delta f = 2.70 \pm 0.01 \) GHz correspond to the ESR transitions of Ti that were similarly found in the dimer with larger atomic separations (Fig. 1c). Hence, we label them as \( f_1^{\mathrm{Ti}} \) and \( f_2^{\mathrm{Ti}} \), respectively. In this dimer, we observed that \( f_1^{\mathrm{Ti}} \) shows a higher intensity than \( f_2^{\mathrm{Ti}} \), indicating an antiferromagnetic exchange interaction\(^{27}\) dominating over the dipolar coupling at this atomic separation. At higher frequencies, we further observed two peaks that are significantly blue-shifted when rotating \( \mathbf{B} \) from \( \theta = 52^\circ \) (pink curve in Fig. 2b) to 97° (purple). The higher resonance frequencies and pronounced angle dependence indicate that those transitions involve the large and anisotropic angular momentum of Er, and, thus, we label them as \( f_3^{\mathrm{Er}} \) and \( f_4^{\mathrm{Er}} \). In addition, their frequency separation exactly matches the one between \( f_1^{\mathrm{Ti}} \) and \( f_2^{\mathrm{Ti}} \), reflecting the same Er-Ti interaction. On the other hand, \( f_3^{\mathrm{Er}} \) and \( f_4^{\mathrm{Er}} \) are approximately equal in intensity, indicating that Ti fluctuates between two spin states with almost equal occupations. The comparable Ti states’ occupation stems from the scattering with tunneling electrons and from the Zeeman splitting of Ti (~7 GHz) being smaller than the thermal energy at the measurement temperature of 1.3 K (~27 GHz). With \( \mathbf{B} \) at \( \theta = 52^\circ \), we observed one more peak at even higher frequencies. Its frequency exactly matches the sum of \( f_1^{\mathrm{Ti}} \) and \( f_4^{\mathrm{Er}} \) (or equivalently \( f_2^{\mathrm{Ti}} \) and \( f_3^{\mathrm{Er}} \)), which suggests an ESR transition
+involving both Ti and Er spins. We label this peak as \( f_5^{\mathrm{TiEr}} \). Remarkably, the sign of \( f_3^{\mathrm{Er}} \), \( f_4^{\mathrm{Er}} \) and \( f_5^{\mathrm{TiEr}} \) is opposite to that of \( f_1^{\mathrm{Ti}} \) and \( f_2^{\mathrm{Ti}} \), indicating a different detection mechanism for the transitions involving the Er spin, which will be discussed below. Finally, we observed an energy level crossing between Er and Ti transitions at \( \theta \sim 12^\circ \), with the Er resonance frequencies further shifting below the Ti transitions at \( \theta \sim 0^\circ \) (Fig. 2c and Fig. S4). This peculiar behavior is a consequence of the large difference in magnetic anisotropy between Er and Ti\(^{23}\).
+
+As shown in Fig. 2c, the angular dependence of the ESR frequencies is well reproduced by using Eq. 1 with \( J_{\mathrm{exc}}/\hbar = -326 \) MHz. We observed small deviations for \( f_1^{\mathrm{Ti}} \), \( f_2^{\mathrm{Ti}} \) and \( f_5^{\mathrm{TiEr}} \), which we ascribe to different experimental conditions and magnetic interaction of Ti with the tip, which is not included in our model. Diagonalizing Eq. 1 allows us to analyze the quantum states of the Er-Ti dimer in terms of individual Er and Ti spin states. For an in-plane \( B = 0.3 \) T, the energy detuning between the Er and Ti spins (30 GHz) is much larger than the interaction energy (about 3 GHz). Therefore, the Er-Ti dimer can be modeled with the 4 Zeeman product states of the Er and Ti spins. Following this picture, we can support the assignment of \( f_{1,2}^{\mathrm{Ti}} \) as Ti spin transitions occurring with no changes in the Er state, while \( f_{3,4}^{\mathrm{Er}} \) correspond to Er spin transitions without altering Ti. Finally, we attribute \( f_5^{\mathrm{TiEr}} \) to a double-flip transition involving both Er and Ti spins. Even though a \( |\Delta m| > 1\hbar \) process is generally forbidden to first order, anisotropic terms in the magnetic interaction can give rise to higher order matrix elements connecting states with \( \Delta m = \pm 2\hbar^{32} \).
+
+When the field is oriented at \( \theta = 0^\circ \), both \( J_{\mathrm{Er}} \) and \( S_{\mathrm{Ti}} \) show an expectation value of \( \hbar/2 \), but a detuning still occurs due to the difference between the g-factors, \( g_{\mathrm{Er}} = 1.2 \) and \( g_{\mathrm{Ti}} = 1.989^{25} \). This detuning is comparable to their interaction energy and, thus, the two middle levels are no longer described by Zeeman product states (Fig. 2e). Finally, at the level crossing angle (\( \theta \sim 12^\circ \)), the two Er and Ti middle levels become singlet and triplet states\(^{29}\). However, measuring ESR spectra under these conditions becomes challenging (Fig. S5), possibly due to the limitation in our detection as discussed in the following.
+Fig. 2 | Measurement of Er ESR transitions through a strongly coupled Ti atom. a, Constant-current STM image of the engineered Er-Ti dimer with the atomic separation of 0.72 nm. The intersection of grids represents the oxygen sites of MgO. The Er atom (circled in yellow) is adsorbed on the oxygen site of MgO, while the Ti atom (circled in purple) is adsorbed on the bridge site (set-point: \( V_{dc} = 100 \) mV, \( I_{dc} = 20 \) pA). b, ESR spectra of the dimer given in a. When the STM tip is located on top of Er, no peaks are observed (yellow) (set-point: \( V_{dc} = 50 \) mV, \( I_{dc} = 20 \) pA, \( V_{rt} = 20 \) mV, \( B = 0.28 \) T, \( \theta = 97^\circ \)). When the STM tip is located on top of Ti, 5 ESR peaks are detected (\( f_{1,2}^{Ti} \), \( f_{3,4}^{Er} \) and \( f_5^{TiEr} \)) with \( \theta = 52^\circ \) (pink), while 4 ESR peaks are detected (\( f_{1,2}^{Ti} \), and \( f_{3,4}^{Er} \)) with \( \theta = 97^\circ \) (purple) (set-point: \( V_{dc} = 70, 60 \) mV, \( I_{dc} = 30, 40 \) pA, \( V_{rt} = 20, 15 \) mV, \( B = 0.3 \) T). The spectra measured on Ti at \( \theta = 52^\circ \) and at \( \theta = 97^\circ \) were normalized at their respective maxima while the spectrum measured on top of Er was rescaled by the same amount used for the spectrum measured on Ti at \( \theta = 97^\circ \). The spectra measured on Ti at \( \theta = 52^\circ \) and on Er are offset for clarity. c, ESR resonance frequencies as a function of \( \theta \) at \( B = 0.32 \) T. The ESR frequencies obtained from each measurement are given as black dots alongside the transition energies predicted from the model Hamiltonian for \( f_1^{Ti} \) (blue line), \( f_2^{Ti} \) (light blue line), \( f_3^{Er} \) (red line), \( f_4^{Er} \) (orange line), \( f_5^{TiEr} \) (green line) and flip-flop transition (dashed gray line). The experimental points were obtained at different set-points (\( V_{dc} = 60–70 \) mV, \( I_{dc} = 12–40 \) pA, \( V_{rt} = 15–25 \) mV, \( B = 0.28–0.8 \) T); the resonance frequencies were rescaled by 0.32 T/B. d,e, Four-level schemes corresponding to the energies of the 4 spin states of the Er-Ti dimer and the corresponding transitions depicted as colored arrows at \( B = 0.32 \) T with different \( \theta \) (90° and 0°, respectively). At \( \theta = 90^\circ \) (d) the spin states are given by the Zeeman products states, while at \( \theta = 0^\circ \) (e), a linear combination of the Zeeman product states is needed to describe the levels.
+
+Erbium ESR Detection and Driving Mechanisms
+
+The detection of ESR peaks exclusively occurs when the tip is positioned on top of Ti. Moving the tip from Ti to Er, the intensities of \( f_3^{Er} \) and \( f_4^{Er} \) gradually decrease and eventually vanish at ~0.3 nm from the Ti
+center (Fig. S6). This behavior indicates that driving an ESR transition on Er must induce a change in the Ti state occupation, subsequently modifying the spin polarization of the tunnel junction. In addition, Er ESR signals differ depending on specific tip conditions, i.e., different tips show positive or negative sign for \( f_{3,4}^{\text{Er}} \) (Fig. 3a).
+
+To further delve into the driving and detection mechanisms of the Er spin, we measured the intensities of \( f_1^{\text{Ti}} \) and \( f_3^{\text{Er}} \) as a function of \( V_{rt} \) using a tip that shows negative Er peaks (Fig. 3b). While \( f_1^{\text{Ti}} \) exhibits a continuous increase in intensity with increasing \( V_{rt} \), \( f_3^{\text{Er}} \) reaches saturation at \( V_{rt} \sim 20 \) mV. The result for \( f_1^{\text{Ti}} \) aligns with previous measurements on Ti\(^{29}\), while the low-power saturation of Er is comparable to that of Fe, which might reflect a long \( T_1 \) and/or a high Rabi rate (\( \Omega \))\(^{33}\). To understand this \( V_{rt} \)-dependence as well as the signs of ESR signals, we developed a rate equation model (Supplementary Section 7) based on the four-level scheme depicted in Fig. 3c. When driving \( f_3^{\text{Er}} \), the populations of the initial and final states involved in the transition tend to equalize through a population transfer\(^{34}\). The changes in population are counteracted by the relaxation rates of each state (\( \Gamma_{1,2}^{\text{Ti}} \) and \( \Gamma_{3,4}^{\text{Er}} \)), which tend to repopulate the depleted states. These rates are inversely proportional to the \( T_1 \) of the atom involved in the spin flip. Since Ti located under the tip is strongly influenced by tunneling electrons, relaxation events occur on a much shorter timescale than for Er\(^{35}\), providing a more efficient pathway to attain the steady state. In addition, to account for the tip-dependent sign and intensity of Er ESR signals, we included a spin-pumping term originating from the spin-polarized current that can shift the Ti spin occupation (Fig. 3c for a negatively polarized tip)\(^{17,36}\). The proposed detection scheme based on the change of Ti state population accurately describes the \( V_{rt} \)-dependence (Fig. 3b) and the tip-dependent sign variations of the ESR signals (Fig. S7).
+
+Finally, to identify the ESR driving source of the Er spin, we follow the relative peak intensity (\( \Delta I/I_{dc} \)) at different tip heights, as controlled by \( l_{dc} \). As shown in Fig. 3d, \( \Delta I/I_{dc} \) of \( f_1^{\text{Ti}} \) increases with reducing the tip-sample distance, indicating that the main driving term for Ti arises from the exchange interaction with the spin-polarized tip\(^{37,38}\). On the other hand, \( \Delta I/I_{dc} \) for \( f_4^{\text{Ti}} \) remains independent of \( l_{dc} \), which identifies the modulation of the magnetic interaction with Ti as the ESR driving source of Er\(^{39}\). The modulation of the magnetic coupling\(^{40}\), in combination with anisotropic interaction terms\(^{32}\), additionally explains the drive of the double-flip transition \( f_5^{\text{TiEr}} \).
+Fig. 3 | Detection and driving mechanisms of Er ESR transitions. a, ESR spectra showing \( f_{3,4}^{\mathrm{Er}} \) for two different STM tips: negative peaks related to negative spin-pumping (yellow line) and positive peaks related to positive spin-pumping (orange line) (set-point: \( I_{dc} = 12, 20 \) pA, \( V_{dc} = 70 \) mV, \( V_{rt} = 25 \) mV, \( B = 0.28, 0.32 \) T, \( \theta = 67^\circ \)). b, ESR peak intensities as a function of \( V_{rt} \). The measured values for \( f_1^{\mathrm{Ti}} \) and \( f_3^{\mathrm{Er}} \) are given by black dots while the intensities predicted from the rate equation model for \( f_1^{\mathrm{Ti}} \) and \( f_3^{\mathrm{Er}} \) are given as blue, light blue, red solid lines and an orange dashed line, respectively (set-point: \( I_{dc} = 40 \) pA, \( V_{dc} = 70 \) mV, \( B = 0.28 \) T, \( \theta = 97^\circ \)). c, Four-level scheme explaining the rate equation model while driving \( f_3^{\mathrm{Er}} \) (red arrow). The Ti’s spin relaxation rates \( \Gamma_1^{\mathrm{Ti}} \) and \( \Gamma_2^{\mathrm{Ti}} \) are depicted as purple arrows while the Er’s spin relaxation rates \( \Gamma_3^{\mathrm{Er}} \) and \( \Gamma_4^{\mathrm{Er}} \) are given as dashed yellow arrows. The negative spin pumping effect is represented as blue double arrows. d, Normalized ESR peak intensities (\( \Delta I/I_{dc} \)) for \( f_1^{\mathrm{Ti}} \) (blue circles) and for \( f_4^{\mathrm{Er}} \) (orange circles) at different tip heights. Here, the tip height is controlled by the set-point current \( I_{dc} \) (set-point: \( V_{dc} = 70 \) mV, \( V_{rt} = 10 \) mV, \( B = 0.28 \) T, \( \theta = 97^\circ \)). The blue and the orange lines serve as guides for the eye. The insets show two different tip-Ti distances: larger for low \( I_{dc} \) and smaller for higher \( I_{dc} \).
+Relaxation Time Measurement through Electron-Electron Double Resonance
+
+By applying an additional rf voltage (\( V_{rf2} \)), Ti and Er spins can be simultaneously driven in the so-called “electron-electron double resonance” scheme\(^{41}\). In a single-frequency ESR sweep, the relative intensities of \( f_1^{\mathrm{Ti}} \) and \( f_2^{\mathrm{Ti}} \) (Fig. 4a) reflect the thermal population of the Er spin. Instead, in double resonance the relative intensities of \( f_1^{\mathrm{Ti}} \) and \( f_2^{\mathrm{Ti}} \) are equalized when \( f_3^{\mathrm{Er}} \) is simultaneously driven (Fig. 4b). As shown in Fig. 4c, the intensity ratio of \( f_1^{\mathrm{Ti}} \) and \( f_2^{\mathrm{Ti}} \) (\( \Delta I_{f_2}^{\mathrm{Ti}} / \Delta I_{f_1}^{\mathrm{Ti}} \)) increases with increasing \( V_{rf} \) only when \( V_{rf2} \) is applied at the resonance frequency of \( f_3^{\mathrm{Er}} \) or \( f_4^{\mathrm{Er}} \), enabling selective modulation of the Er states to an out-of-equilibrium configuration.
+
+Taking advantage of this selective driving mechanism, we implemented an inversion recovery measurement to estimate the spin relaxation time of Er (\( T_1^{\mathrm{Er}} \)) in a pump-probe scheme (Fig. 4d). After exciting \( f_3^{\mathrm{Er}} \) with a pumping rf pulse of 200 ns duration that equalized the Er population, we applied a probe pulse of 500 ns for \( f_1^{\mathrm{Ti}} \) after a delay time \( \Delta t \). Using this sequence, we monitored the time evolution of the intensity of \( f_1 \) as a function of \( \Delta t \) from the out-of-equilibrium to the thermal state (Fig. 4e). The fit to an exponential function (Fig. 4e) gives \( T_1^{\mathrm{Er}} = 0.818 \pm 0.115 \) \( \mu s \), which is five times longer than what previously measured in Fe-Ti dimers in the absence of tunnel current\(^{18}\). We attribute this enhancement to the efficient decoupling of 4f electrons from the environment, which reduces the relaxation events arising from the scattering with substrate electrons.
+
+The large \( T_1^{\mathrm{Er}} \) measured through Ti indicates that the rapid spin fluctuations of Ti occurring on the timescale of a few ns\(^{35} \) do not significantly perturb the stability of the Er states. This property partially originates from the large energy detuning between Er and Ti levels, which prevents the energy exchange required for spin-flip events. Using the experimentally obtained value of \( T_1^{\mathrm{Er}} \) in the rate equation model, we extract a driving term \( W = \Omega^2 T_2/2 \) for Er that is two times larger than for Ti in the same dimer (Supplementary Section 7). Despite the long spin lifetime and large driving term, attempts to drive Er Rabi oscillations through Ti do not yield a complete cycle (Fig. S8b), preventing a direct measure of the Er \( T_2 \). This is most likely due to a relatively low Rabi rate \( \Omega \) provided by the moderate Er-Ti exchange coupling, which is about 2–3 times smaller than in the Fe-Ti dimer\(^{39}\). In turn, a low value of \( \Omega \) together with a large driving term \( W \) would imply much longer \( T_2 \) for Er than previous 3d elements, highlighting the potential of 4f electrons to realize higher performance atomic-scale qubits.
+Fig. 4 | Determination of Er spin relaxation time. a,b, Double resonance spectra in the frequency range covering Ti ESR transitions \( f_{1,2}^{T_1} \) (a) without and (b) with simultaneous driving of Er at the ESR frequency of \( f_3^{Er} \). The peak intensities of \( f_{1,2}^{T_1} \) are related to the relative population of the Er spin states (insets). The spectra were normalized to the sum of their peak intensity. c, ESR intensity ratios between \( \Delta f_{1}^{T_1} \) and \( \Delta f_{2}^{T_1} \) as a function of the driving strength \( V_{rfz} \) at different Er ESR transition states (red, orange, and grey circles for \( f_3^{Er} \), \( f_4^{Er} \), and off-resonance, respectively). The solid curves show the correspondent simulation results by the rate equation model for \( f_3^{Er} \) (red line), \( f_4^{Er} \) (orange line) and at an off-resonance frequency (grey line). Set-point: \( I_{dc} = 15 \) pA, \( V_{dc} = 70 \) mV, \( V_{rf} = 30 \) mV, \( V_{rfz} = 1, 30 \) mV, \( B = 0.28 \) T, \( \theta = 97^\circ \). d, Schematics of the inversion recovery measurement in a pump-probe pulse scheme to determine the Er spin relaxation time \( T_1^{Er} \). Each sequence is composed of a pump pulse at the resonance frequency of \( f_3^{Er} \) (red box) and a probe pulse at the resonance frequency of \( f_1^{T_1} \) (blue box). The probe pulse follows the pump pulse after a delay time \( \Delta t \). The population of the Er states after the pump pulse relaxes back to the thermal state following its \( T_1 \). e, The experimental data for the inversion recovery measurement (blue circles) show the intensity of the ESR signal at the probe pulse \( f_1 \) as a function of the delay time. The black line shows the fit using an exponential function with \( T_1^{Er} \) of about 1 \( \mu \)s. Set-point: \( I_{dc} = 50 \) pA, \( V_{dc} = 70 \) mV, \( V_{rf,pump} = 60 \) mV, \( V_{rf,probe} = 100 \) mV, \( B = 0.28 \) T, \( \theta = 97^\circ \).
+
+Conclusions
+
+We demonstrated a new experimental approach to electrically drive ESR on the elusive 4f electrons in a surface-adsorbed lanthanide atom with long spin relaxation time. Given the reduced scattering with the substrate electrons, it is reasonable to anticipate an enhancement in the coherence time of Er in comparison to 3d elements. We expect that, by employing a similar approach in different atomic structures, the ESR driving on the 4f electrons can be amplified, enabling the use of lanthanide atoms as surface spin qubits with superior properties compared to the routinely adopted 3d elements.
+Methods
+
+STM measurements
+
+Our experiment was performed in a home-built STM operating at the cryogenic temperature of ~1.3 K in an ultrahigh vacuum environment (< 1 × 10^{-9} Torr)^{42}. Using a two-axis vector magnet (6 T in-plane/4 T out-of-plane), the magnetic fields were varied from 0.28 T to 0.9 T at different angles from the surface^{42}. To allow atom deposition on the sample kept in the STM stage, the sample is slightly tilted from the axis of the magnet by ~7° as estimated from the fit to the data shown in Fig. 1d. Considering this misalignment, all our experimental \( \theta \) were offset by that amount accordingly. The magnetic tips used in our measurements were prepared by picking up ~4–9 Fe atoms from the MgO surface until the tips presented good ESR signals on isolated Ti atoms.
+
+ESR measurements
+
+We used two different schemes to apply \( V_{rf} \) to the STM junction: one through the tip and one through an antenna (rf generators: Keysight E8257D and E8267D)^{42}. In all our measurement involving a single rf sweep, we applied the \( V_{rf} \) using an antenna located near the STM tip except for the data in Fig. 3b, where the \( V_{rf} \) was combined with the dc bias voltage \( V_{dc} \) using a diplexer at room temperature and then applied to the STM tip. The data in Fig. 4a–c were acquired by applying \( V_{rf1} \) to the tip and simultaneously \( V_{rf2} \) to the antenna. For the measurements reported in Fig. 4e and Fig. S8, the two rf voltages (\( V_{rf1} \) and \( V_{rf2} \)) were combined through a power splitter (minicircuits ZC2PD-K0244+) and applied to the STM tip. For these measurements, both rf generators were gated by an arbitrary waveform generator (Tektronix, AWG 70002B).
+
+Sample preparation
+
+The surface of a Ag(100) substrate was cleaned by repeated cycles of Ar+ sputtering and annealing (700 K). We grew atomically thin layers of MgO(100) on the Ag(100) following a procedure described in a previous work^{43}. We deposited Fe, Ti and Er atoms (< 1% of monolayer) from high purity rods (>99%) using an e-beam evaporator. During the deposition the sample was held at ~10 K in order to have well-isolated single atoms on the surface.
+
+Analysis of ESR spectra
+We fit the ESR spectra using a model given in\(^{29}\) in order to extract the resonance frequency, peak intensity, and peak width for the data shown in Fig. 1d, Fig. 2c, Fig. 3b,d and Fig. 4c.
+
+Acknowledgements
+
+We thank Taehong Ahn and Leonard Edens for their support at the initial stage of the experiment and Yi Chen, Arzhang Ardavan, and Joaquín Fernández-Rossier for fruitful discussions. We acknowledge support from the Institute for Basic Science (IBS-R027-D1). Y.B. acknowledges support from Asian Office of Aerospace Research and Development (FA2386-20-1-4052). H.B. acknowledges funding from the SNSF AdG (TMAG-2_209266).
+References
+
+1 Heinrich, A. J. et al. Quantum-coherent nanoscience. Nature Nanotechnology **16**, 1318-1329 (2021). https://doi.org/10.1038/s41565-021-00994-1
+2 Donati, F. et al. Magnetic remanence in single atoms. Science **352**, 318-321 (2016). https://doi.org/10.1126/science.aad9898
+3 Dey, A., Kalita, P. & Chandrasekhar, V. Lanthanide(III)-Based Single-Ion Magnets. ACS Omega **3**, 9462-9475 (2018). https://doi.org/10.1021/acsomega.8b01204
+4 Pedersen, K. S. et al. Toward Molecular 4f Single-Ion Magnet Qubits. Journal of the American Chemical Society **138**, 5801-5804 (2016). https://doi.org/10.1021/jacs.6b02702
+5 Shiddiq, M. et al. Enhancing coherence in molecular spin qubits via atomic clock transitions. Nature **531**, 348-+ (2016). https://doi.org/10.1038/nature16984
+6 Vincent, R., Klyatskaya, S., Ruben, M., Wernsdorfer, W. & Balestro, F. Electronic read-out of a single nuclear spin using a molecular spin transistor. Nature **488**, 357-360 (2012). https://doi.org/10.1038/nature11341
+7 Thiele, S. et al. Electrically driven nuclear spin resonance in single-molecule magnets. Science **344**, 1135-1138 (2014). https://doi.org/10.1126/science.1249802
+8 Macfarlane, R. M. High-resolution laser spectroscopy of rare-earth doped insulators: a personal perspective. J. Lumin. **100**, 1-20 (2002). https://doi.org:https://doi.org/10.1016/S0022-2313(02)00450-7
+9 Serrano, D. et al. Ultra-narrow optical linewidths in rare-earth molecular crystals. Nature **603**, 241-246 (2022). https://doi.org/10.1038/s41586-021-04316-2
+10 Rančić, M., Hedges, M. P., Ahlefeldt, R. L. & Sellars, M. J. Coherence time of over a second in a telecom-compatible quantum memory storage material. Nature Physics **14**, 50-54 (2017). https://doi.org/10.1038/nphys4254
+11 Zhong, T. et al. Optically Addressing Single Rare-Earth Ions in a Nanophotonic Cavity. Phys. Rev. Lett. **121**, 183603 (2018). https://doi.org/10.1103/PhysRevLett.121.183603
+12 Kolesov, R. et al. Optical detection of a single rare-earth ion in a crystal. Nature Communications **3**, 1029 (2012). https://doi.org/10.1038/ncomms2034
+13 Kindem, J. M. et al. Control and single-shot readout of an ion embedded in a nanophotonic cavity. Nature **580**, 201-204 (2020). https://doi.org/10.1038/s41586-020-2160-9
+14 Yin, C. et al. Optical addressing of an individual erbium ion in silicon. Nature **497**, 91-94 (2013). https://doi.org/10.1038/nature12081
+15 Meier, F., Zhou, L., Wiebe, J. & Wiesendanger, R. Revealing Magnetic Interactions from Single-Atom Magnetization Curves. Science **320**, 82-86 (2008). https://doi.org:doi:10.1126/science.1154415
+16 Franke, K. J., Schulze, G. & Pascual, J. I. Competition of Superconducting Phenomena and Kondo Screening at the Nanoscale. Science **332**, 940-944 (2011). https://doi.org:doi:10.1126/science.1202204
+17 Khajetoorians, A. A. et al. Current-Driven Spin Dynamics of Artificially Constructed Quantum Magnets. Science **339**, 55-59 (2013). https://doi.org:doi:10.1126/science.1228519
+18 Wang, Y. et al. An electron-spin qubit platform assembled atom-by-atom on a surface. arXiv:2108.09880 (2021). <https://arxiv.org/abs/2309.02348>.
+19 Baumann, S. et al. Electron paramagnetic resonance of individual atoms on a surface. Science **350**, 417-420 (2015). https://doi.org/10.1126/science.aac8703
+20 Chen, Y., Bae, Y. & Heinrich, A. J. Harnessing the Quantum Behavior of Spins on Surfaces. Advanced Materials **35**, 2107534 (2023). https://doi.org:https://doi.org/10.1002/adma.202107534
+21 Drost, R. et al. Combining electron spin resonance spectroscopy with scanning tunneling microscopy at high magnetic fields. Rev. Sci. Instrum. **93** (2022). https://doi.org/10.1063/5.0078137
+22 Seifert, T. S., Kovarik, S., Gambardella, P. & Stepanow, S. Accurate measurement of atomic magnetic moments by minimizing the tip magnetic field in STM-based electron paramagnetic resonance. Physical Review Research **3**, 043185 (2021). https://doi.org:10.1103/PhysRevResearch.3.043185
+23 Reale, S. et al. Erbium and thulium on MgO(100)/Ag(100) as candidates for single atom qubits. Physical Review B **107**, 045427 (2023). https://doi.org:10.1103/PhysRevB.107.045427
+24 Choi, T. et al. Atomic-scale sensing of the magnetic dipolar field from single atoms. Nature Nanotechnology **12**, 420-424 (2017). https://doi.org/10.1038/nnano.2017.18
+25 Kim, J. et al. Spin resonance amplitude and frequency of a single atom on a surface in a vector magnetic field. Physical Review B **104**, 174408 (2021). https://doi.org/10.1103/PhysRevB.104.174408
+26 Steinbrecher, M. et al. Quantifying the interplay between fine structure and geometry of an individual molecule on a surface. Physical Review B **103**, 155405 (2021). https://doi.org/10.1103/PhysRevB.103.155405
+27 Yang, K. et al. Engineering the Eigenstates of Coupled Spin-$\frac{1}{2}$ Atoms on a Surface. Phys. Rev. Lett. **119**, 227206 (2017). https://doi.org/10.1103/PhysRevLett.119.227206
+28 Vieru, V., Iwahara, N., Ungur, L. & Chibotaru, L. F. Giant exchange interaction in mixed lanthanides. Scientific Reports **6**, 24046 (2016). https://doi.org/10.1038/srep24046
+29 Bae, Y. et al. Enhanced quantum coherence in exchange coupled spins via singlet-triplet transitions. Science Advances **4**, eaau4159 (2018). https://doi.org/doi:10.1126/sciadv.aau4159
+30 Singha, A. et al. Engineering atomic-scale magnetic fields by dysprosium single atom magnets. Nature Communications **12**, 4179 (2021). https://doi.org:10.1038/s41467-021-24465-2
+31 Natterer, F. D. et al. Reading and writing single-atom magnets. Nature **543**, 226–+ (2017). https://doi.org:10.1038/nature21371
+32 Sadiek, G., Lashin, E. I. & Abdalla, M. S. Entanglement of a two-qubit system with anisotropic XYZ exchange coupling in a nonuniform time-dependent external magnetic field. Physica B: Condensed Matter **404**, 1719-1728 (2009). https://doi.org:https://doi.org/10.1016/j.physb.2009.02.011
+33 Willke, P. et al. Tuning Single-Atom Electron Spin Resonance in a Vector Magnetic Field. Nano Lett. **19**, 8201-8206 (2019). https://doi.org:10.1021/acsnanolett.9b03559
+34 Willke, P. et al. Probing quantum coherence in single-atom electron spin resonance. Science Advances **4**, eaqq1543 (2018). https://doi.org:doi:10.1126/sciadv.aqq1543
+35 Seifert, T. S. et al. Longitudinal and transverse electron paramagnetic resonance in a scanning tunneling microscope. Science Advances **6**, eabc5511 (2020). https://doi.org:doi:10.1126/sciadv.abc5511
+36 Loth, S. et al. Controlling the state of quantum spins with electric currents. Nature Physics **6**, 340-344 (2010). https://doi.org:10.1038/nphys1616
+37 Lado, J. L., Ferrón, A. & Fernández-Rossier, J. Exchange mechanism for electron paramagnetic resonance of individual adatoms. Physical Review B **96**, 205420 (2017). https://doi.org:10.1103/PhysRevB.96.205420
+38 Reina-Gálvez, J., Wolf, C. & Lorente, N. Many-body nonequilibrium effects in all-electric electron spin resonance. Physical Review B **107**, 235404 (2023). https://doi.org:10.1103/PhysRevB.107.235404
+39 Phark, S.-h. et al. Electric-Field-Driven Spin Resonance by On-Surface Exchange Coupling to a Single-Atom Magnet. Advanced Science **n/a**, 2302033 (2023). https://doi.org:https://doi.org/10.1002/advs.202302033
+40 Niskanen, A. O. et al. Quantum Coherent Tunable Coupling of Superconducting Qubits. Science **316**, 723-726 (2007). https://doi.org:10.1126/science.1141324
+41 Phark, S.-h. et al. Double-Resonance Spectroscopy of Coupled Electron Spins on a Surface. ACS Nano **17**, 14144-14151 (2023). https://doi.org:10.1021/acsnano.3c04754
+42 Hwang, J. et al. Development of a scanning tunneling microscope for variable temperature electron spin resonance. Rev. Sci. Instrum. **93** (2022). https://doi.org:10.1063/5.0096081
+43 Paul, W. et al. Control of the millisecond spin lifetime of an electrically probed atom. Nature Physics **13**, 403-407 (2017). https://doi.org:10.1038/nphys3965
+Supplementary Files
+
+This is a list of supplementary files associated with this preprint. Click to download.
+
+• ErMgOESRSIsubmission.docx

0ae7f8f6cbad98f78d95595f0fd34157385af5879d5ea6b07e3d14e42afd8221/peer_review/peer_review.md

@@ -0,0 +1,417 @@
+Peer Review File
+
+Constructing synthetic nuclear architectures via transcriptional condensates in a DNA protonucleus
+
+Corresponding Author: Professor Andreas Walther
+
+This file contains all reviewer reports in order by version, followed by all author rebuttals in order by version.
+
+Version 0:
+
+Reviewer comments:
+
+Reviewer #1
+
+(Remarks to the Author)
+This paper reports that nuclear-like DNA-based condensates that can transcribe RNAs. The DNA condensates were formed from three types of long single-stranded DNAs with a promoter sequence of T7 RNA polymerase. A transcribed RNA can fold into a three-branched structure with hairpin loops for a kissing-loop interaction. The transcribed RNAs were self-assembled into RNA condensates. This study revealed the behaviors and features of RNA condensates transcribed from DNA condensates, such as the localization of a formed RNA condensate in a DNA condensate. The experiments are well-organized, and the results are clear and interesting. This study is excellent research that will boost the DNA-based synthetic condensate field. The reviewer’s questions and comments are below.
+
+1) Figs. 2j, 3b, and 4a,c,f. What is the ‘sponge-like’ structure of RNA condensates? How did the structure form? Is it like bubbling caused by RNA degradation (doi.org/10.1073/pnas.2001654117)? Or is it a kind of phase separation (doi.org/10.1002/admi.202300898)?
+
+2) Fig. 4: Why was the RNA condensate in Fig. 4f not spherical, different from the spherical RNA condensates in other results such as Fig. 4a and c? What decides the formed RNA condensate structures?
+
+3) Fig. 5 and p.12, l.353-356 (This suggests a preferred interaction between KL1-R1 and PN matrix, retaining the KL1-R1 condensate inside the PN, whereas KL2-R2 gets obviously expelled. KL1-R1 dominates the interaction with the PN matrix in this competitive system, whereas pure KL2-R2-PN would form a single central condensate (Supplementary Fig. 9.).): This phenomenon is very interesting. What decides the preference of the interaction between KL1-R1 and PN matrix and between KL2-R2 and PN matrix?
+
+4) Fig. 5b: Why did the expelled KL2-R2 condensates attach on the outer surface of the PN, not inside the PN?
+
+5) Fig. 5 and p.12, l.356-360 (At 15 mM Mg2+, the transcriptional KL1-R1 occupies the major PN space, while KL2-R2 forms multiple condensates in the PNs (Fig. 5e-g). This can be attributed to weakened interactions between KL1-R1 and the PN at low salinity, allowing KL2-R2 to occupy some of the available volume in the PN to form condensates.): Why did the stable structure (KL2-R2) have a weaker interaction with the PN matrix? What is the mechanism of this phenomenon?
+
+6) Fig. 5e and p.12, l.356-360: Why did KL2-R2 form multiple condensates in the PNs? This behavior is different from KL1-R1. Can the authors explain this phenomenon based on the surface tension difference in both RNA condensates? Or is there another hypothesis?
+
+7) Fig. 5: Are the asymmetric behaviors in the interaction of KL1-R1/KL2-R2 with PN related to not only the stability of KL1-
+R1/KL2-R2 condensates but also the kinetics in the formation process of both the condensates?
+
+8) Fig. 5: Does the asymmetric behaviors change if the ratio of the concentrations of the KL1-R1/KL2-R2 template DNAs in the PN?
+
+9) Fig. 5: If only the KL1-R1 condensate is degraded after the formation of Fig. 5b structures, does the KL2-R2 condensate enter the PN and re-form a single condensate in the PN like Supplementary Fig. 9b?
+
+Reviewer #2
+
+(Remarks to the Author)
+I have read with great interest the manuscript by Xie, Chen et al. The report combines the DNA based protocells developed by the Walther group with recently introduced RNA nanostructures capable of forming condensates upon in vitro transcription. The team uses these technologies to study co-transcriptional condensation in the crowded microenvironments of the DNA protocells that, arguably, shares many similarities with condensate formation in eukaryotic nuclei.
+The authors perform a wide array of experimental assays and controls and observe a number of intriguing effects including the key role played by interactions between the RNA nanostars and the surrounding DNA matrix in determining the characteristics (and presence) of the condensates.
+Overall, I find this manuscript well written, and the results relevant to the interdisciplinary readership of the journal, particularly those interested in biomolecular condensates, synthetic cells and DNA/RNA nanotechnology. The document is well written and clear, and the figures are of high quality. I am happy to recommend publications after the authors have addressed the following minor comments:
+1) In page 5, line 151, the authors introduce [NTP] as “defined as the maximum amount of KL1 transcripts that can be produced per template”. Could the authors clarify better this definition? Is this the overall NTP concentration divided by the number of nucleotides in one nanostar construct?
+
+2) In page 5, line 158, the authors state “This comparison demonstrates that the spatial transcription of the KL1 in PNs leads to locally high concentrations sufficient for condensation, similar to the enrichment mechanism in natural nuclear condensates”, referring to the observation that under the relevant conditions KL1 condensates form in PNs but not in bulk. However, we find out later in the manuscript that the effect is likely due to affinity between the nanostars and poly(A20-o), so this statement is not really correct. Perhaps the authors could simply refer the reader to the later parts of the manuscript for a rationalisation of this effect?
+
+3) Similarly, when discussing the difference in FRAP recovery between PNs and bulk (page 5, line 166), the authors could refer to the subsequent discussion on interactions between the KL1 nanostars and the DNA matrix, which is probably behind the slower recovery seen in PNs.
+
+4) In Figs 2 c and k the authors convert a fluorescent signal from a molecular beacon into RNA concentrations. Could they please provide information on how this is done?
+
+5) The authors use a single-stranded DNA template for transcribing the RNA nanostars. This will probably fold into a DNA nanostar similar to the RNA version, which may hinder transcription. Have the authors tested transcription using a double-stranded template? Does this change transcription efficiency in PNs or the bulk? Have they included GU wobble pairs in the RNA nanostar designs to destabilise the secondary structure of the single-stranded template?
+
+6) Following up from 5: the authors ascribe the preferential condensation of KL1 stars within PNs to the affinity with the o domain in poly(A20-o), and they show good evidence for it. However, 10% of the scaffold is DNA in the PNs poly(A20-p), hybridised to single-stranded DNA templates that are likely to bin the transcription products (particularly when transcription ends and the polymerase loses activity). The authors could perhaps test or discuss this possibility?
+
+Reviewer #3
+
+(Remarks to the Author)
+Within a core-shell DNA coacervates, the authors demonstrated that specific RNAs (kissing loop sequence, KLs) can be transcribed in-situ to form another phase of condensate inside, and this platform is termed as protonuclei (PNs). To construct this, they designed polyA and polyT containing ssDNA by incorporating the T7 promotor sequence or other barcode DNAs, and mixed with T7 DNA ligase. They provided the mechanism of co-condensationof transcriptional KL1 with the PN DNA matrix. This platform is capable of transcribing two KLs simultaneously that produces different patterns of multiphasic condensates by changes in magnesium concentration.
+
+This work is original because the authors provided the first demonstration of in-situ co-transcription condensation KLs in PNs that can form multiphasic condensates. This work provides a clear observation of the KLs transcribed in PNs to
+undergo condensation with DNA matrix. This work provides the new design strategy to construct the synthetic system that can perform the in-situ transcription with modality. This is potentially an interesting system for artificial cell research and synthetic biology.
+
+While they provide thorough experimental designs to support their claims, there are a couple of remaining questions regarding their interpretation and observations:
+
+1. If this core-shell DNA coacervates have a highly concentrated DNA-enriched core, why does the KLs transcription start from the inner-interface of PN shell (Fig 3b)? Is this a combined effect of NTPs diffusion and T7 polymerase from outside of DNA PNs? Perhaps, the T7 polymerase is enriched at the shell?
+2. In this context, the authors describe the slower rate by T7 RNA polymerase in PNs caused by the slow diffusion of NTPs. Supposedly, this highly concentrated ssDNAs at the interface could suggest more dense networks of DNAs. Then, I would expect that T7 RNA polymerase diffusion would be slow as well, or even limited, because it is bigger than the NTPs. If this transcription is in a kinetic regime where the enzyme concentration doesn’t affect the overall rate of reaction, the authors point of view is valid. Otherwise, some discussion regarding the T7 RNA polymerase encapsulation in PNs needs to be addressed.
+3. Interpretation about the dsDNA exclusion from KL1 condensates in Figure 4 needs clarification.
+a. If KL1 sequence does not have any specific interactions with the PN matrix, why would the hybridization of ssDNA with o* make KL excluded from the condensates? What is this interaction between KL1 and PN DNA matrix? Nonspecific interactions between KL and PN DNA matrix could be the charge interaction mediated by magnesium. Complementary sequence for PN DNA matrix would likely have partial dsDNA structure, which can yield in the increased charge density overall. If so, wouldn’t it make these nonspecific interactions between KL and the polyA matrix stronger? Perhaps, magnesium is depleted to mediate the nonspecific charge interactions between PN matrix and KL1 that leads to the exclusion of the KL1? This could partially explain why the condensates inside of PNs become dominant by DNA matrix. I wonder higher magnesium concentration will make the co-condensates to be persisted upon adding o* invader.
+b. Why does the KL1 condensate morphology by the prehybridization of DNA matrix in Fig 4f differ from Fig 4a? Is it because of different time point of imaging? Maybe the morphology in Fig 4a is kinetically arrested?
+c. dsDNA can have greater persistence length that leads to exclusion from protein membraneless organelles in-vitro (Nott, T., Cragg, T. & Baldwin, A. Membraneless organelles can melt nucleic acid duplexes and act as biomolecular filters. Nature Chem 8, 569–575 (2016), some discussion about dsDNA vs ssDNA in Jeffrey R. Vieregg et al. Journal of the American Chemical Society 2018 140 (5), 1632-1638). I wonder whether this dsDNA material properties affect its partitioning trends in Fig 4 g & h.
+4. It is still unclear how these two different KLs localize at the different locations under different magnesium concentration (Fig 5).
+a. KL2-R2 condensates still have little amount of KL1 in both cases of magnesium concentrations, suggesting that the coexistence of KL1-R1 condensate and KL2-R2 condensate is more than sum of each individual condensates. Likely, DNA matrix for PN is involved. Please address that.
+b. The reasoning of different morphology of KL condensates based on its material properties seem valid. This is a minor question, but do KL2 get transcribed inside of PNs and get transported to be outside of shell? Or does the KL2 get transcribed outside of the DNA PNs at 30 mM Mg2+?
+5. In line 428, what does it mean by helper proteins?
+
+Additionally, the manuscript does not specify the number of independent replicates. Some of fluorescence intensity curves have shaded areas representing standard deviations, but it is unclear where this standard deviation is coming from. Clarifying this information is important to improve the transparency and reproducibility of the study. The details of methods seem enough.
+
+Version 1:
+
+Reviewer comments:
+
+Reviewer #1
+
+(Remarks to the Author)
+The authors have carefully revised their manuscript in response to the reviewers’ questions and comments. New experimental data have been added, making it scientifically very interesting. Since the revised paper is very well structured, the reviewer recommends its publication in Nature Communications.
+
+Reviewer #2
+
+(Remarks to the Author)
+The authors have addressed my minor concerns and I am happy to recommend publication.
+
+Reviewer #3
+
+(Remarks to the Author)
+I am very pleased with the revised manuscript and appreciate its comprehensiveness and the authors' efforts in addressing the kinetics of RNA polymerase, diffusion of molecules, and condensation dynamics. I strongly recommend this manuscript for publication in Nature Communications.
+Open Access This Peer Review File is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.
+In cases where reviewers are anonymous, credit should be given to 'Anonymous Referee' and the source.
+The images or other third party material in this Peer Review File are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
+To view a copy of this license, visit https://creativecommons.org/licenses/by/4.0/
+We submit the revised version of our manuscript titled “Constructing synthetic nuclear architectures via transcriptional condensates in a DNA protonucleus”.
+
+We thank the reviewers for their positive and constructive comments on the manuscript, and we appreciate the opportunity to discuss the reviewers’ comments and revise our manuscript accordingly.
+
+We have considered all the comments and added requisite experiments and explanations to underscore specific points raised by the reviewers.
+
+Summary of major additional experiments:
+• Simulated interactions between orthogonal KLs transcripts and PN matrix by NUPACK to deepen the understanding on biphasic transcriptional condensates in PN.
+• Effect of ssDNA or dsDNA templates on the formation of transcriptional KL1 condensates, both in solution and in PN.
+• Effect of T7 RNAP diffusion on the transcription efficiency and the formation process of transcriptional KL1 condensate in PN by quantitative plate reader and visible CLSM.
+• Confocal microscopy experiments demonstrate that increasing [Mg^{2+}] from 30 mM to 100 mM slows down the disassembly of the co-condensate in PN during o* invasion.
+• Investigation of impact of transcription kinetics on the asymmetric morphology of orthogonal KL1-R1 and KL2-R2 condensates formed in PN.
+• CLSM experiment verifies that in-solution transcribed KL2-R2 could enter the pristine PN to some extent, whereas spongy co-condensate can only be formed by in-PN co-transcriptional condensation.
+
+We trust that our revisions comprehensively address the points raised by the reviewers and the newly added experiments fall in line with expectations and broaden the scope of our study.
+Reviewer #1 (Remarks to the Author):
+This paper reports that nuclear-like DNA-based condensates that can transcribe RNAs. The DNA condensates were formed from three types of long single-stranded DNAs with a promoter sequence of T7 RNA polymerase. A transcribed RNA can fold into a three-branched structure with hairpin loops for a kissing-loop interaction. The transcribed RNAs were self-assembled into RNA condensates. This study revealed the behaviors and features of RNA condensates transcribed from DNA condensates, such as the localization of a formed RNA condensate in a DNA condensate. The experiments are well-organized, and the results are clear and interesting. This study is excellent research that will boost the DNA-based synthetic condensate field. The reviewer’s questions and comments are below.
+
+Thank you for your highly positive evaluation of our work.
+
+1) Figs. 2j, 3b, and 4a,c,f. What is the ‘sponge-like’ structure of RNA condensates? How did the structure form? Is it like bubbling caused by RNA degradation (doi.org/10.1073/pnas.2001654117)? Or is it a kind of phase separation (doi.org/10.1002/admi.202300898)?
+
+Response: Thank you for your insightful question. The sponge-like structure of KL-PN co-condensate originates either from spinodal or from viscoelastic phase separation that can appear in phase-segregating systems of high concentration of components and for situations of low dynamics, respectively. Different from the binodal phase separation via a nucleation and growth process for in-solution KL condensates formed at comparably high dilution\(^1\), in-PN KL transcripts continuously compact and co-phase-segregation with PN matrix, ultimately forming interconnected networks, where the KL transcripts and polyA matrix, as slow-diffusion components, continuously expel the solvent molecule as the fast-diffusion component through spinodal or viscoelastic phase separation\(^2\). Due to the distinct phase separation processes, the KL-PN co-condensate displays a rather rough surface (see single z-plane images in Fig. 2h) and loose structure (see the higher intensity of pure KL condensate on the outside PN shell and the lower intensity of co-condensate within PN in Fig. 4f), compared with pure KL condensates.
+
+We add additional text to clarify this:
+
+The sponge-like, co-continuous structures of KL-PN co-condensates arises from spinodal or viscoelastic phase separation\(^3\) of both polyA and KL transcripts in PN, that can appear in phase-segregating systems of high concentration of components and for situations of low dynamics. Different from the binodal phase separation via a nucleation and growth process for in-solution KL condensate formation at comparably high dilution\(^1\), KL transcripts transcribed inside PN continuously compact and co-phase-segregate with the PN matrix, ultimately forming interconnected networks. KL transcripts co-phase-segregating with the polyA matrix constitute slowly diffusing components with viscoelastic properties. Solvent molecules as the quickly
+diffusing component are expelled in the process, leading to interconnected network-like phase segregation that ultimately collapses into a spherical domain by interfacial energy minimization².
+
+2) Fig. 4: Why was the RNA condensate in Fig. 4f not spherical, different from the spherical RNA condensates in other results such as Fig. 4a and c? What decides the formed RNA condensate structures?
+
+Response: Thank you for your question. In Fig. 4a and c, RNA condensates are first formed in PN, which are NOT hybridized with o*-Atto647 (o*-Atto647 has the function to reduce affinity between KL condensates and the PN interior). o*-Atto647 is later added as an invader after RNA condensates were formed. That is why a spherical RNA condensate gradually dissolves from the outer surface to the center of PN.
+
+In contrast, in Fig. 4f, KL transcription and condensation is conducted in PN prehybridized with different amounts of Atto647-o*. This hybridization weakens the KL-to-PN affinity, making RNA condensation within the PN difficult at 10% prehybridization and leading to ill-defined RNA condensates. One can also see the brighter edges of these condensates. These arise from rather pure KL condensates, whereas the less intense core is a co-condensate. The pure KL layer even appears to dewet from the central condensate as seen by the undulations/non-perfect bright shell. At higher prehybridization levels, the internal PN environment becomes more repulsive, which prevents RNA condensation inside PN. Therefore, small condensates form at the outer shell of PN due to diffusion of KL transcripts from the PN to the surrounding.
+
+We find that the experimental description is insufficient in the original manuscript for showing the differences between experimental settings for Fig. 4a-c and Fig. 4f. Therefore, we now provide additional text and explanation for Fig. 4a-c as follows:
+
+To do so, we added different stoichiometric amounts of o*-Atto647 into solutions of PN containing already formed KL1-poly(A_{20}-o)_n co-condensates and investigated how the invasion by o*-Atto647 would alter the pre-existing KL1-poly(A_{20}-o)_n co-condensates. A gradual invasion of o*-Atto647 into the KL1-poly(A_{20}-o)_n co-condensates takes place as the amount of o*-Atto647 increases (Fig. 4a, b). A continuous surface erosion of the co-condensates occurs because the o/o*-Atto647 hybridization reduces the affinity between KL condensate and PN interior by introducing stronger electrostatic and steric repulsion inside PN due to increased negative charge density and increasing persistence length of the formed dsDNA parts⁴,⁵ (Fig. 4c-e and Supplementary Video 3).
+
+For Fig. 4f, we have also modified the original text and believe the relevant discussion (highlighted) now can stand alone and make a clear distinction to the discussion about Fig. 4a-c:
+“Seeing such a profound impact, we then investigated KL1 transcription in PN with a poly(A_{20}-o)_n matrix pre-hybridized by different amounts of o*-Atto647 (from 0% – 300%) to provide weakened affinity between PN matrix and KL1 transcripts. In analogy with the above result, single KL1 condensates form in pristine PN (Fig. 4f). When applying 10% o*-Atto647, the KL1 transcripts form
+single condensates with irregular secondary nucleation on its surface inside the PN, along with multiple tiny nuclei outside the PN shell (Fig. 4f). The brighter green parts are condensates purely enriched with KL1 transcripts that remain inside the PN due to relatively sufficient affinity. The marked difference – heterogeneously structured RNA condensates in Fig. 4f compared to the rather spherical and homogenous structures formed after invasion of pre-formed condensate in Fig. 4a, c – can be attributed to their different formation processes. Increasing the content of pre-hybridized o*-Atto647 domains from 10% to 300% gradually prevents KL1 condensate formation inside the PN due to weakened PN-KL1 interaction, which likely becomes even repulsive at higher pre-hybridization degrees because the ssDNA to dsDNA transition leads to higher negative charge density and persistence length, allowing for stronger electrostatic and steric repulsion45, respectively, inside the PN. As a result, the KL1 transcripts formed inside the PN do not yield condensates insides the PN, instead, multiple small transcriptional KL1 condensates form in the PN surroundings.”
+
+3) Fig. 5 and p.12, I.353-356 (This suggests a preferred interaction between KL1-R1 and PN matrix, retaining the KL1-R1 condensate inside the PN, whereas KL2-R2 gets obviously expelled. KL1-R1 dominates the interaction with the PN matrix in this competitive system, whereas pure KL2-R2-PN would form a single central condensate (Supplementary Fig. 9.).): This phenomenon is very interesting. What decides the preference of the interaction between KL1-R1 and PN matrix and between KL2-R2 and PN matrix?
+
+Response: Thank you for the question. In the original manuscript, we had experimentally verified the preferred affinity of KL1-R1/PN over KL2-R2/PN in original Supplementary Fig. 10 (now Supplementary Fig. 16 in revised manuscript). As KL1-R1 and KL2-R2 share the same stem sequence but have different tails (R1 and R2) for binding fluorescent labels, we hypothesize that the difference in the KL-PN affinity interaction might be rooted in the different tails. To further investigate this, we performed NUPACK simulations (rna06 model) at 30 °C with the following sequences (tail regions are underlined):
+
+repeating unit of PN matrix p(A20-o)n (400 μM)
+AAAAAAAAAAAAAAAAAAAAAGCATTGAGGTATTGTTGCTCCCA
+
+RNA sequence of KL1 (200 μM):
+GGGAUGGCACAUUAGAGUCGCUCUAUCGCGAAAGAGCGGCCUCUAGUGUGCUCGCGUGCCUCAGAGGAC
+AUCGCGAAAGGUCCUUUGAGGUACGCGUCACUCGUAGCAUUGUGCAUCGCGAAAGCACAGUGCUAUG
+AGUGCAUCUGAACGAGUAAGGACCCCA
+
+RNA sequence of KL2 (200 μM):
+GGGAUGGCACAUUAGAGUCGCUCAGUCGACAAGAGCGGCCUCUAGUGUGCUCGCGUGCCUCAGAGGAC
+AGUGCAAGGUCCUUUGAGGUACGCGUCACUCGUAGCAUUGUGCAUCGACAAGCACAGUGCUAUG
+AGUGGGUGGCUUUUUUACAGGCGUUAG
+As shown below, we find that at this condition, KL1-R1 indeed has a higher level of binding to the p(A_{20-o})_n than KL2-R2. And the origin of this binding interaction comes from the tail of the KL structure, i.e., the R1 sequence. In contrast, KL2-R2 only shows a minor binding at a different position. We have added this new result to Supplementary Figure 16.
+
+![Bar graph showing simulated concentration distribution of complex sequences in a pool of A_{20-o} (400 μM), KL1-R1 (200 μM), and KL2-R2 (200 μM) at 30 °C based on mao6 model. Salt: 1 M Na+. Predicted minimum of free energy (MFE) proxy structure at 30 °C of A_{20-o} + KL1-R1 and A_{20-o} + KL2-R2. Tails of KL1-R1 and KL2-R2 are highlighted in orange. The binding between A_{20-o} and KLS are indicated by arrows.](page_180_370_1087_370.png)
+
+Supplementary Figure 16. Binding interaction between A_{20-o}, KL1-R1, and KL2-R2.
+a, Simulated concentration distribution of complex sequences in a pool of A_{20-o} (400 μM), KL1-R1 (200 μM), and KL2-R2 (200 μM) at 30 °C based on mao6 model. Salt: 1 M Na+. b, Predicted minimum of free energy (MFE) proxy structure at 30 °C of A_{20-o} + KL1-R1 and A_{20-o} + KL2-R2. Tails of KL1-R1 and KL2-R2 are highlighted in orange. The binding between A_{20-o} and KLS are indicated by arrows.
+
+We added additional discussion in the text:
+
+NUPACK simulations further indicate that the origin of the preferred interaction between KL1-R1 and PN matrix is rooted in the tail structure (Supplementary Fig. 16).
+
+4) Fig. 5b: Why did the expelled KL2-R2 condensates attach on the outer surface of the PN, not inside the PN?
+
+Response: Thank you for your question. As KL1-R1 dominates the affinity interaction with the PN matrix, it co-condenses with the PN. The co-condensation between KL1-R1 and PN on one hand screens the affinity interaction between KL2-R2 and the PN matrix, and on the other hand induces electrostatic repulsion of KL2-R2 due to the highly dense and negatively charged properties. Therefore, the transcribed KL2-R2 will get expelled from the PN interior to the outside. Since KL2-R2 has a local high concentration just outside of the PN due to the in-PN transcription, KL2-R2 condensates form on the outer surface of the PN. We believe that this is now sufficiently and clearly explained in context with the new data provided (Updated Supplementary Fig. 12 and new Supplementary Fig. 16) and with the explanation existent in the text (see around page 13).
+
+5) Fig. 5 and p.12, l.356-360 (At 15 mM Mg2+, the transcriptional KL1-R1 occupies the major PN space, while KL2-R2 forms multiple condensates in the PN (Fig. 5e-g). This can be attributed to weakened interactions between KL1-R1 and the PN at low salinity, allowing KL2-R2 to occupy some of the available volume in the PN to form condensates.):
+Why did the stable structure (KL2-R2) have a weaker interaction with the PN matrix? What is the mechanism of this phenomenon?
+
+Response: As we also answered to your comment #3, KL1-R1 and KL2-R2 share the same stem sequence but have different tail sequences for binding fluorescent labels. As discussed above, we found that KL1-R1 indeed has a higher level of binding to the p(A_{20-o})_n than KL2-R2. The origin of this binding interaction comes from the tail of the KL structure, i.e., the R1 sequence. We have added this new result in Supplementary Figure 16.
+
+![Bar graph showing concentration distribution of complex sequences and schematic diagrams of KL1-R1 and KL2-R2 tails binding to A_{20-o}](page_184_579_1082_388.png)
+
+Supplementary Figure 16. Binding interaction between A_{20-o}, KL1-R1, and KL2-R2.
+a, Simulated concentration distribution of complex sequences in a pool of A_{20-o} (400 μM), KL1-R1 (200 μM), and KL2-R2 (200 μM) at 30 °C based on rna06 model. Salt: 1 M Na+. b, Predicted minimum of free energy (MFE) proxy structure at 30 °C of A_{20-o} + KL1-R1 and A_{20-o} + KL2-R2. Tails of KL1-R1 and KL2-R2 are highlighted in orange. The binding between A_{20-o} and KLS are indicated by arrows.
+
+We added additional discussion in the text:
+
+NUPACK simulations further indicate that the origin of the preferred interaction between KL1-R1 and PN matrix is rooted in the tail structure (Supplementary Fig. 16).
+
+6) Fig. 5e and p.12, l.356-360: Why did KL2-R2 form multiple condensates in the PN? This behavior is different from KL1-R1. Can the authors explain this phenomenon based on the surface tension difference in both RNA condensates? Or is there another hypothesis?
+
+Response: Thank you for your very inspiring question. The suggestion from the surface tension perspective is very insightful and helpful for the interpretation of the data presented in Fig. 5e. We have deeply searched in the relevant literature in biomolecular condensate research and found similar structure explained from surface tension perspective^{6,7}. And we found that such mechanism is also applicable to our system:
+
+As KL2-R2 has a higher melting point, it means that the binding strength among KL2-R2 in condensates is stronger and therefore KL2-R2 condensates possess higher surface tension than KL1-R1 condensates. Besides, as we have newly determined in the Supplementary Fig. 16 (See also our answers to your questions #3 and #5), KL1-R1 has stronger binding interaction with the
+PN matrix than KL2-R2. In a ternary system of PN, KL1-R1, and KL2-R2, the surface tension (\( \gamma \)) will follow the relationship: \( \gamma_{PN/KL2-R2} > \gamma_{PN/KL1-R1} \), so that KL2-R2 condensate will always be enveloped by KL1-R1 condensate to minimize its contact with the PN matrix and the whole system can minimize the free energy. There is an additional requirement that \( \gamma_{KL1-R1/KL2-R2} \) should not be too high, which is applicable to our system, as KL1-R1 and KL2-R2 have very similar sequences and structures.
+
+We added this discussion in the revised manuscript and cited the relevant papers:
+
+This also helps to explain the overall architecture of the multiphase structures in Figure 5b, e from the perspective of surface tension (\( \gamma \)) of the ternary system of PN, KL1-R1, and KL2-R2\(^{6,7}\). Since KL2-R2 has a higher \( T_m \) regarding the KL interactions, the binding strength among KL2-R2 in condensates is stronger and therefore KL2-R2 condensates possess higher surface tension than KL1-R1 condensates. In addition, KL1-R1 has a stronger binding interaction with the PN matrix than KL2-R2. This means \( \gamma_{PN/KL2-R2} > \gamma_{PN/KL1-R1} \) applies, so that KL2-R2 condensate are preferably enveloped by KL1-R1 or expelled from PN to minimize its contact with the PN matrix.
+
+7) Fig. 5: Are the asymmetric behaviors in the interaction of KL1-R1/KL2-R2 with PN related to not only the stability of KL1-R1/KL2-R2 condensates but also the kinetics in the formation process of both the condensates?
+
+Response: Thank you for your inspiring question regarding the kinetics aspects. To investigate the influence of transcription kinetics on the asymmetric behavior of KL1-R1/KL2-R2 with PN matrix, we compared in-solution transcription kinetics of KL1-R1 and KL2-R2 to assess the sequence effect on the transcription efficiency. To quantify the transcribed KLs, we employed a strand displacement reaction to react with the dangling strands (R1 reporter with R1 in KL1-R1, R2 reporter with R2 in KL2-R2 in Supplementary Fig. 13a; see sequence in revised Supplementary Table 2). Overall, both structures show a similar transcription efficiency with a slight trend for a faster transcription of KL2-R2. However, since orthogonal KL transcription in PN favors a co-condensate of KL1-R1 with PN matrix (Fig. 5b), these results collectively underpin that KL-PN interaction are the dominant cause for co-condensate formation, and not faster transcription kinetics.
+Supplementary Figure 13. Transcription kinetics of KL1-R1 and KL2-R2 in solution.
+
+a, Scheme showing transcription of KL1-R1 (top) or KL2-R2 (down) in solution characterized by plate reader. To quantify the transcribed KLs RNA, dsDNA reporters with fluorophore-quencher pairs are present in solutions to react with dangling strand of transcribed KLs (R1 Reporter for KL1-R1, R2 Reporter for KL2-R2) by SDR, generating fluorescent signals. b, Transcription kinetics for KL1-R1 and KL2-R2 in solution with free promoter oligonucleotide, monitored by plate reader ([NTP] : [Rx Reporter] : [T_{KL-R}] : [p] = 100 : 10 : 1 : 1, 30 °C, 6 mM Mg^{2+}, 2.5 U/μL T7 RNAP, x = 1 or 2). N = 2 independent experiments. Error areas represent standard deviation.
+
+We added experiment and discussion in the revised manuscript:
+
+To ensure that this bias in structural development does not arise solely from different transcription efficiency, we conducted three control experiments: First, we compared the in-solution transcription efficiency of KL1-R1 and KL2-R2, finding out comparable transcription efficiency of both KL2-R2 and KL1-R1 (Supplementary Fig. 13).
+
+8) Fig. 5: Does the asymmetric behaviors change if the ratio of the concentrations of the KL1-R1/KL2-R2 template DNAs in the PN?
+
+Response: Thank you for your question. In addition to the experiment presented in the main MS with an equal template ratio of KL1-R1 to KL2-R2, we have now conducted in-PN KLs transcription experiments with altered templates ratios: (1) KL1-R1/KL2-R2 = 1/3 and (2) KL1-R1/KL2-R2 = 3/1 to investigate the transcription speed on the structure of the formed condensate at 30 mM Mg^{2+}. We added this new result in Supplementary Fig. 14.
+
+Most importantly, even when the KL2-R2 is present in 3-fold excess (KL1-R1/KL2-R2 = 1/3), the result is qualitatively similar to equal stoichiometry (KL1-R1/KL2-R2 = 1/1, present in the MS). Hence, even when favoring KL2-R2 transcription (top row), the preferred interactions of KL1-R1 with the PN matrix dominate the structure formation. KL2-R2 condensates form on the surface of the PN. As a control (bottom row), we also inverted the ratio to have a higher concentration of KL1-R1 template inside the PN (KL1-R1/KL2-R2 = 3/1), where only KL1-R1 condensates can be found. Overall, these results suggest that KL-to-PN interactions, rather than transcription speed, dominates the co-condensate formation.
+Supplementary Figure 14. Effect of asymmetric template concentrations for orthogonal KLs on the formation of condensates in PN.
+
+Representative CLSM images of orthogonal KLs transcription in PN with 1/3 ratio (top) and 3/1 ratio (bottom) of T_{KL1-R1}/T_{KL2-R2} at 30 mM Mg^{2+} ([NTP] : [R1*] : [R2*] : [T_{KL1-R1}] : [T_{KL2-R2}] : [p] = 3.6 : 1.8 : 1.8 : 0.75 : 0.25 : 1, 30 °C, 2.5 U/μL T7 RNAP). Green channel: KL1-R1; Magenta channel: KL2-R2. Scale bar: 10 μm.
+
+We added the experiment and relative discussion in the revised manuscript:
+
+Second, we further screened whether biasing the in-PN transcription towards KL2-R2 by increasing its template concentration to a 3-fold excess over the KL1-R1 template would change the structure formation. However, even under such conditions, the co-condensate morphology does not invert, but KL1 condensates remain inside the PN and KL2 condensates form on the surface of the PN (Supplementary Fig. 14). Taken together, these results provide mutually reinforcing evidence supporting that the interaction between different KL and PN dominates the formation of the co-condensates and not the pure transcription efficiency.
+
+9) Fig. 5: If only the KL1-R1 condensate is degraded after the formation of Fig. 5b structures, does the KL2-R2 condensate enter the PN and re-form a single condensate in the PN like Supplementary Fig. 9b?
+
+Response: Thank you for the very interesting question. Since it is not feasible to specifically degrade KL1-R1 without degrading KL2-R2, we provide an additional experiment based on some feasible assumptions to address this issue. We believe the question actually focuses on whether KL2-R2, existing outside of PN, would enter the pristine PN and form a co-condensate solely through KL-PN and KL interactions.
+
+To investigate whether the KL-PN and KL interactions are strong enough to recruit KL2-R2 from outside of PN to form co-condensate, KL2-R2 was first transcribed in solution with 3.6 × NTP for 18 h. This low concentration does not enable the formation of KL condensate in solution (Fig. 2e).
+Then, PN was added and incubated with existing KL2-R2 for another 18 h at 30 °C with shaking (Supplementary Fig. 12c; NEW). The KL2-R2 transcripts are recruited into the PN, however, the structure is clearly different from co-transcriptional condensation that shows a spongy phase segregation (original Supplementary Fig. 9b, now in Supplementary Fig. 12b). Here, KL2-R2 recruitment into PN does occur, but no evident condensation/shrinkage of the structure occurs. This means that interactions between KL2-R2 and PN are able to partition some amount of transcript into the PN, but KL2/KL2 interactions are not abundant enough to initiate the same level of co-condensation as for in-PN transcription. This new result has now been added to Supplementary Fig. 12c.
+
+![Scheme and representative CLSM image showing the formation of single condensates in PN by localized transcription of KL1-R1. The condensate is labeled by R1*-Atto647 (green channel). b, Scheme and representative CLSM image showing the formation of single condensates in PN by localized transcription of KL2-R2. The condensate is labeled by R2*-Atto488 (magenta channel). 2.5 U/μL T7 RNAP, 30 mM Mg^{2+}, 30 °C, [NTP] : [R1* or R2*] : [T_{KL1,R1} or T_{KL2,R2}] : [p] = 3.6 : 3.6 : 1 : 1, 18 h reaction for both (a) and (b). c, Scheme and representative CLSM images for the transcription of KL2-R2 in solution, which gets recruited into the PN. KL2-R2 was first transcribed in solution with UTP-Atto488 feeding ([NTP] : [T_{KL2,R2}] = 3.6 : 1, 1 mol% UTP-Atto488, 30 °C, 2.5 U/μL T7 RNAP, 30 mM Mg^{2+}). Pristine PN was added and incubated with the existing KL2-R2 for 18 h for recruitment at 30 °C with 30 mM Mg^{2+}. Green channel: PN shell labeled with o*-Atto647; Magenta channel: KL2-R2-Atto488. Scale bars are 5 μm for (a) and (b), and 10 μm for (c).](page_186_624_1077_377.png)
+
+Supplementary Figure 12. Formation of KL1-R1 and KL2-R2 condensates in PNs versus KL2-R2 transcribed in solution and recruited into PN.
+
+a, Scheme and representative CLSM image showing the formation of single condensates in PN by localized transcription of KL1-R1. The condensate is labeled by R1*-Atto647 (green channel). b, Scheme and representative CLSM image showing the formation of single condensates in PN by localized transcription of KL2-R2. The condensate is labeled by R2*-Atto488 (magenta channel). 2.5 U/μL T7 RNAP, 30 mM Mg^{2+}, 30 °C, [NTP] : [R1* or R2*] : [T_{KL1,R1} or T_{KL2,R2}] : [p] = 3.6 : 3.6 : 1 : 1, 18 h reaction for both (a) and (b). c, Scheme and representative CLSM images for the transcription of KL2-R2 in solution, which gets recruited into the PN. KL2-R2 was first transcribed in solution with UTP-Atto488 feeding ([NTP] : [T_{KL2,R2}] = 3.6 : 1, 1 mol% UTP-Atto488, 30 °C, 2.5 U/μL T7 RNAP, 30 mM Mg^{2+}). Pristine PN was added and incubated with the existing KL2-R2 for 18 h for recruitment at 30 °C with 30 mM Mg^{2+}. Green channel: PN shell labeled with o*-Atto647; Magenta channel: KL2-R2-Atto488. Scale bars are 5 μm for (a) and (b), and 10 μm for (c).
+
+We added the experiment and relative discussion in the revised manuscript:
+
+Third, we also probed whether KL2-R2 transcribed in solution can be enriched into pristine PN to check for interactions. Here, distinct differences are visible when comparing the structures formed by in-PN transcription of KL2-R2 versus KL2-R2 transcripts recruited from solution (Supplementary Fig. 12c). The in-PN transcription clearly induces phase segregation by KL2/KL2 interactions, whereas the latter rather points to some KL2-R2/PN interactions that enable a certain level of recruitment. Clear phase segregation via contraction of a spongy co-condensate phase is not
+visible for the latter. Overall, the last control also emphasizes the critical role of in-PN transcription to facilitate co-condensate formation.
+
+Reviewer #2 (Remarks to the Author):
+I have read with great interest the manuscript by Xie, Chen et al. The report combines the DNA based protocells developed by the Walther group with recently introduced RNA nanostructures capable of forming condensates upon in vitro transcription. The team uses these technologies to study co-transcriptional condensation in the crowded microenvironments of the DNA protocells that, arguably, shares many similarities with condensate formation in eukaryotic nuclei.
+
+The authors perform a wide array of experimental assays and controls and observe a number of intriguing effects including the key role played by interactions between the RNA nanostars and the surrounding DNA matrix in determining the characteristics (and presence) of the condensates.
+
+Overall, I find this manuscript well written, and the results relevant to the interdisciplinary readership of the journal, particularly those interested in biomolecular condensates, synthetic cells and DNA/RNA nanotechnology. The document is well written and clear, and the figures are of high quality. I am happy to recommend publications after the authors have addressed the following minor comments:
+Thank you for your very positive evaluation of our work!
+
+1) In page 5, line 151, the authors introduce [NTP] as “defined as the maximum amount of KL1 transcripts that can be produced per template”. Could the authors clarify better this definition? Is this the overall NTP concentration divided by the number of nucleotides in one nanostar construct?
+
+Response: Thank you for the comment. The [NTP] is calculated by the maximum amount of transcript [Transcript] divided by template concentration in the solution. [Transcript] is calculated by the concentration of the corresponding monomers divided by the number of the most abundant nucleotide in the transcript. For example, if U is the most abundant nucleotide in the transcript, we will calculate the [transcript] by taking the concentration of UTP divided by the number of U in transcript. To improve clarity, we have now added a new method section “[NTP] calculation” and refereed it in page 5 line 154.
+
+Added text in method section:
+
+[NTP] calculation
+
+[NTP] is defined as the maximum amount of transcripts that can be produced per template given the nucleotide concentrations in the NTP mix in relation to the sequence of the transcript. For example for [NTP] : [template] = 5 : 1, the [NTP] concentration is set in a way to at least allow for 5 full transcripts from 1 template based on the most abundant nucleotide in the transcript. Other NTPs will be in a slight excess as the NTP mix has equal stoichiometry for all four needed NTPs.
+2) In page 5, line 158, the authors state “This comparison demonstrates that the spatial transcription of the KL1 in PN leads to locally high concentrations sufficient for condensation, similar to the enrichment mechanism in natural nuclear condensates”, referring to the observation that under the relevant conditions KL1 condensates form in PN but not in bulk. However, we find out later in the manuscript that the effect is likely due to affinity between the nanostars and poly(A20-o), so this statement is not really correct. Perhaps the authors could simply refer the reader to the later parts of the manuscript for a rationalisation of this effect?
+
+Response: Thank you for your very helpful suggestion.
+
+To improve the clarity of the manuscript, we have added the following:
+
+In addition, as we will demonstrate below, the KL-PN affinity also plays an important role in the condensation process.
+
+3) Similarly, when discussing the difference in FRAP recovery between PN and bulk (page 5, line 166), the authors could refer to the subsequent discussion on interactions between the KL1 nanostars and the DNA matrix, which is probably behind the slower recovery seen in PN.
+
+Response: Thank you for your helpful suggestion. We had mentioned this in the original manuscript in page 5, line 170-173, and now in page 6, line 174-177 in the revised manuscript, which might have slipped your eyes.
+
+Text in original manuscript: “In contrast, half-bleached KL1 condensates in PN show less recovery and lack the bright edge, likely due to their restricted dynamics in a DNA-crowded environment and interactions between PN matrix and the KL1 transcripts, as we will further discuss below.”
+
+4) In Figs 2 c and k the authors convert a fluorescent signal from a molecular beacon into RNA concentrations. Could they please provide information on how this is done?
+
+Response: Thank you for your question. We assume that the transcribed RNA concentration is equal to the concentration of the activated reporter, which is reflected by the measured fluorescence intensity. And in parallel, we have a pure DNA-fluorophore conjugate sample at 1000 nM, which serves as reference. By calculating the intensity ratio between the individual samples and the reference, we can obtain the RNA concentration in individual samples. We have now included a more detailed description for calculating RNA concentration in the Method section in revised manuscript.
+
+For each plate reader experiment, we have included a reference sample containing 1000 nM DNA-fluorophore conjugate. We calculate the intensity ratio between individual samples and the reference sample to yield the transcribed RNA concentration by following equation:
+
+\[
+[\text{RNA}] = 1000 \text{ nM} \times \frac{\text{Intensity (sample)}}{\text{Intensity (reference)}}
+\]
+5) The authors use a single-stranded DNA template for transcribing the RNA nanostars. This will probably fold into a DNA nanostar similar to the RNA version, which may hinder transcription. Have the authors tested transcription using a double-stranded template? Does this change transcription efficiency in PN or the bulk? Have they included GU wobble pairs in the RNA nanostar designs to destabilise the secondary structure of the single-stranded template?
+
+Response: Thank you for your question. For KL (RNA nanostars) transcription, we actually used dsDNA as template, as we had the same concern that the ssDNA template might hinder the RNA transcription due to the internal hairpin structure. There is only use of ssDNA template for linear transcript in Figure 2 c, k. We did not add GU wobble pairs in the KL designs, as we used dsDNA, which avoids the hairpin structure formation inside the template. We have noticed that the scheme in the original Figure 1 was not clear enough. We have now modified the Figure 1 with a dsDNA template to improve the accuracy of the scheme.
+
+In addition, in the revision, we tested whether ssDNA templates enable proper transcription. As expected, transcription is very inefficient. Condensate formation is absent after 24 h for both in-solution and in-PN transcription (Supplementary Fig. 2).
+
+We added this result to the new Supplementary Fig. 2, and referenced it in the text:
+
+As a proof of concept, we first focus on a three-armed singled-stranded RNA (ssRNA) nanostar with a wildtype palindromic KL sequence1,8,9 at the tip of each arm (KL1 in Fig. 2d). We used a dsDNA template (T_{KL}/T_{KL}') because ssDNA templates (T_{KL}) alone do not allow for efficient transcription and condensate formation on account of intramolecular folding of such ssDNA template (Supplementary Fig. 2). T_{KL} contains a p* ssDNA sequence for hybridization to poly(A_{20}-p), inside the PN to initiate transcription (sequences in Supplementary Table 2).
+
+![Representative CLSM images of transcriptional KL condensate in solution and PN](page_370_1042_1002_246.png)
+
+Supplementary Figure 2. ssDNA template hinders the formation of transcriptional KL condensates in both solution and PN.
+
+a, Representative CLSM images of transcriptional KL condensate in solution with [NTP] : [T_{KL}] = 7.2, transcribed by promoter ssDNA, which is hybridized with T_{KL} as a ssDNA template or T_{KL}/T_{KL}' as a dsDNA template ([T_{KL}] : [p] = 1 : 1, 30 °C, 30 mM Mg^{2+}, 2.5 U/μL T7 RNAP). Note that [NTP] is set to a concentration where condensation in solution appears (see Supplementary Figure 3). b, Representative CLSM images of transcriptional KL in PN with ssDNA
+template or dsDNA template for transcription ([NTP] : [T_{KL}] = 3.6, 30 °C, 30 mM Mg^{2+}, 2.5 U/μL T7 RNAP). Scale bar: 10 μm.
+
+6) Following up from 5: the authors ascribe the preferential condensation of KL1 stars within PN to the affinity with the o domain in poly(A20-o), and they show good evidence for it. However, 10% of the scaffold is DNA in the PN poly(A20-p), hybridized to single-stranded DNA templates that are likely to bind the transcription products (particularly when transcription ends and the polymerase loses activity). The authors could perhaps test or discuss this possibility?
+
+Response: Thank you for your question.
+As our answer to your question #5, we actually used dsDNA templates for transcription of KL nanostars throughout the whole paper. Therefore, strong interaction between the transcribed products with the template cannot be the major cause.
+
+Reviewer #3 (Remarks to the Author):
+Within a core-shell DNA coacervates, the authors demonstrated that specific RNAs (kissing loop sequence, KLs) can be transcribed in-situ to form another phase of condensate inside, and this platform is termed as protonuclei (PN). To construct this, they designed polyA and polyT containing ssDNA by incorporating the T7 promotor sequence or other barcode DNAs, and mixed with T7 DNA ligase. They provided the mechanism of co-condensation of transcriptional KL1 with the PN DNA matrix. This platform is capable of transcribing two KLs simultaneously that produces different patterns of multiphasic condensates by changes in magnesium concentration.
+This work is original because the authors provided the first demonstration of in-situ co-transcription condensation KLs in PN that can form multiphasic condensates. This work provides a clear observation of the KLs transcribed in PN to undergo condensation with DNA matrix. This work provides the new design strategy to construct the synthetic system that can perform the in-situ transcription with modality. This is potentially an interesting system for artificial cell research and synthetic biology.
+Thank you for your very positive evaluation of our work!
+
+While they provide thorough experimental designs to support their claims, there are a couple of remaining questions regarding their interpretation and observations:
+1. If this core-shell DNA coacervates have a highly concentrated DNA-enriched core, why does the KLs transcription start from the inner-interface of PN shell (Fig 3b)? Is this a combined effect of NTPs diffusion and T7 polymerase from outside of DNA PN? Perhaps, the T7 polymerase is enriched at the shell?
+Response: Thank you for your comment. We agree that the phenomenon of the initial transcription at the inner interface of PN shell is a result of the diffusion of both NTPs and T7 RNAP.
+Although NTPs are relatively small molecules (~500 Da), they need to be transported from outside to the PN interior, and they get continuously consumed by T7 RNAP during transcription. Therefore, the outer part has the first access to the NTPs and can have higher transcription rate of KLs there. Besides, T7 RNAP is relatively large, with a molecular weight of around 100 kDa with slower diffusion. As the T7 RNAP gets into the PN through diffusion and binding to the template, the inner interface of the PN shell will have the first interaction and binding with the T7 RNAP. Therefore, the T7 RNAP will get firstly enriched at that part and induces transcription from there at the beginning of the process, which also leads to a local high transcription rate of KLs. Over time, this transcription kinetics difference throughout the PN will disappear, as the T7 RNAP become homogenous inside the PN by diffusion.
+
+To confirm our hypothesis, we performed KL1 transcription in PN pre-equilibrated with T7 RNAP to eliminate the effect of T7 RNAP diffusion. To do so, we charged the dsDNA template and T7 RNAP into PN 2 h prior to NTP addition. Upon the NTP addition to initiate transcription, KL1 intensity shows a more uniform increase in PN during the first 6 h transcription due to the continuous NTP consumption and the slow diffusion of NTP. This result implies that the diffusion of both T7 RNAP and NTP plays a role in the KL1 transcription in PN. The co-transcriptional phase segregation is however very similar. We have added this new result to Supplementary Figure 6, as also shown below.
+Supplementary Figure 6. Pre-equilibrium of T7 RNAP enables a more homogenous transcription inside PN.
+
+Rationale: To investigate if the diffusion of T7 RNAP into PN has an impact on the peripheral transcription, we performed transcription of KL1 inside PN equilibrated with T7 RNAP prior the addition of NTPs. a, scheme and representative CLSM images of KL1 transcription in PN pre-equilibrated with T7 RNAP for 2 h, before the addition of NTPs to trigger transcription. The plot shows cross-sectional line profile along the white line in the CLSM image at 6 h, demonstrating a more homogenous transcription inside the PN. b, scheme and representative CLSM images of KL1 transcription in PN with simultaneous addition of T7 RNAP and NTPs (standard conditions used in the main text). The plot shows cross-sectional line profile along the white line in the CLSM image at 6 h, demonstrating a non-homogenous and peripheral transcription inside the PN. Green channel: KL1 condensate labeled by UTP-Atto488; Magenta channel: PN shell (poly(T_{20-k}), labeled with k^+-Atto647). [NTP]: [T_{KL1}] = 3.6 : 1, 1 mol% UTP-Atto488, 30 °C, 2.5 U/μL T7 RNAP, 30 mM Mg^{2+}. Note that we only focus on the first 6h of co-transcriptional condensation, because afterwards it’s rather the maturation of the structure/co-condensate. Scale bar: 10 μm.
+
+We added new text to discuss this point in the revised manuscript:
+
+In the first 12 h, transcription takes place from the edge of the PN to their center due to the continuous consumption of NTPs as well as diffusive uptake of T7 RNAP and NTPs. The KL1 intensity gradient can be diminished if the PN are pre-equilibrated with T7 RNAP (for 2 h) prior to the addition of the NTPs (Supplementary Fig. 6). The overall structure formation proceeds however in a very similar fashion.
+2. In this context, the authors describe the slower rate by T7 RNA polymerase in PN caused by the slow diffusion of NTPs. Supposedly, this highly concentrated ssDNAs at the interface could suggest more dense networks of DNAs. Then, I would expect that T7 RNA polymerase diffusion would be slow as well, or even limited, because it is bigger than the NTPs. If this transcription is in a kinetic regime where the enzyme concentration doesn’t affect the overall rate of reaction, the authors point of view is valid. Otherwise, some discussion regarding the T7 RNA polymerase encapsulation in PN needs to be addressed.
+
+Response: Thank you for your insightful and inspiring question. You are right that the T7 RNAP has much higher molecular weight (~ 100 kDa) than NTPs (~ 500 Da), so that the diffusion of T7 RNAP could potentially also influence the transcription inside the PN. Indeed, for the result shown in Fig. 2c in the original manuscript, the experiment was performed by directly adding T7 RNAP and NTPs into the solution to trigger the transcription. In this case, T7 RNAP needs certain time to diffuse into the PN and get encapsulated, which likely hinders the transcription kinetics. According to our new result shown in Supplementary Fig. 6 (for your comment just above), pre-equilibrating PN with T7 RNAP diminishes the gradient intensity of KL1 during its transcription, which confirms the encapsulation of T7 RNAP within PN (see detailed answer to your question #1). In this case, different concentrations of T7 RNAP could affect its diffusion rate into PN, potentially influencing transcription kinetics.
+
+To investigate whether the diffusion of T7 RNAP with applied concentration in the paper (2 U/uL or 2.5 U/uL) affects the overall transcription kinetics, we now performed new experiments by using various concentrations of T7 RNAP (ranging from 2-10 U/uL) without pre-incubation. The result shows that the RNA polymerase concentration has a slight influence on the transcription rate. We just mention these results here to not clutter the manuscript and the SI. According to the transparent review process, this information will still be available for the reader.
+
+![In-PN transcription kinetics with various T7 RNAP concentrations.](page_489_1042_484_312.png)
+
+Figure for Review 1. In-PN transcription kinetics with various T7 RNAP concentrations.
+
+Higher concentrations of T7 RNAP facilitate faster diffusion and encapsulation into PN, thereby enhancing transcription kinetics. The experiment was monitored by SDR of the R in a plate reader ([NTP] : [R] : [TRep*] : [p] = 2000 : 10 : 1 : 1, 30 °C, 6 mM Mg^{2+}, 2 – 10 U/μL T7 RNAP). N = 3. Shaded areas represent the standard deviation. Note
+that the transcription kinetics differ from the data shown in Fig. 2c in the original manuscript due to different enzymatic activity from different batches.
+
+We also modified the main text to discuss the influence of diffusion of T7 RNAP on the slower transcription kinetics in PN:
+
+The slightly lower activity can be understood considering constraints of the diffusion of T7 RNAP and NTPs into the PN, and RNA strands out of the PN.
+
+3. Interpretation about the dsDNA exclusion from KL1 condensates in Figure 4 needs clarification.
+a. If KL1 sequence does not have any specific interactions with the PN matrix, why would the hybridization of ssDNA with o* make KL excluded from the condensates? What is this interaction between KL1 and PN DNA matrix? Nonspecific interactions between KL and PN DNA matrix could be the charge interaction mediated by magnesium. Complementary sequence for PN DNA matrix would likely have partial dsDNA structure, which can yield in the increased charge density overall. If so, wouldn’t it make these nonspecific interactions between KL and the polyA matrix stronger? Perhaps, magnesium is depleted to mediate the nonspecific charge interactions between PN matrix and KL1 that leads to the exclusion of the KL1? This could partially explain why the condensates inside of PN become dominant by DNA matrix. I wonder higher magnesium concentration will make the co-condensates to be persisted upon adding o* invader.
+
+Response: We believe that KL (no R1/R2 tail) and PN have non-specific binding interaction, which comes from the hydrogen-bond between individual nucleobases. The Mg^{2+} also induces some extent of charge interaction for the binding between phosphate backbones. However, when hybridizing the o* in the PN matrix, the overall net negative charge density largely increases, which assists to expel the KL1, which is also negatively charged, due to electrostatic repulsion. Besides, hybridization also increases the persistence length of matrix strands in PN. This enhances the steric repulsion between PN matrix and the KL1. Therefore, the exclusion of the KL1 is a result of both electrostatic repulsion and steric repulsion, induced by o/o* hybridization.
+
+High magnesium concentration helps to stabilize the co-condensates, by screening the electrostatic repulsion. To verify this, we have now performed new experiments by firstly forming KL-PN co-condensates at 30 mM Mg^{2+} and then increasing the overall Mg^{2+} in solution to 100 mM, before adding o*-Atto647 invader strand. While KL-PN co-condenses at 30 mM Mg^{2+} is completely disassembled 1 h after o*-Atto647 addition (Fig. 4c), the co-condensate still exists after 2 h of o*-Atto647 invasion (Supplementary Fig. 11). After 24 h of o*-Atto647 addition, PN-KL co-condensate is disassembled, resulting in the occupation of poly(A_{20}-o)/o*-Atto647 in the majority of PN, with very few KLS remaining inside PN. This result implies that a higher Mg^{2+} concentration indeed helps stabilize KL interactions as well as the PN-KL co-condensates, thereby slowing down the disassembly of the co-condensate. However, charge screening with a higher Mg^{2+} concentration could not completely preserve the integrity of KL-PN co-condensate.
+Supplementary Figure 11. Invasion of o*-Atto647 at 100 mM Mg^{2+} shows slower disassembly of KL-PN co-condensate compared to 30 mM Mg^{2+}.
+
+a, Representative CLSM images of KL-PN co-condensate formed by KL1 transcription in PN with 30 mM Mg^{2+} (KL1-Atto488, green channel). The overall Mg^{2+} concentration was then increased to 100 mM, and the solution was incubated for 24 h before the addition of o*-Atto647 to hybridize with poly(A_{20}-o)_{n} in PN (magenta channel). b, Representative CLSM images of KL-PN co-condensate at 2 or 24 h after o*-Atto647 addition at 100 mM Mg^{2+}. Scale bar: 10 μm.
+
+We added new texts to discuss this result:
+
+Moreover, an invasion process of o*-Atto647 at 100 mM Mg^{2+} shows a slower disassembly of the co-condensate compared to 30 mM Mg^{2+} (Fig. 4c), implying the critical role of Mg^{2+} in stabilizing interactions of KLs and KL-PN (Supplementary Fig. 11).
+
+b. Why does the KL1 condensate morphology by the prehybridization of DNA matrix in Fig 4f differ from Fig 4a? Is it because of different time point of imaging? Maybe the morphology in Fig 4a is kinetically arrested?
+
+Response: Thank you for your question. This question is similar to reviewer 1 - question #2. In Fig. 4a and c, RNA condensates are first formed in PN, which are NOT hybridized with o*-Atto647 (o*-Atto647 has the function to reduce affinity between KL condensates and the PN interior). o*-Atto647 is later added as an invader after RNA condensates were formed. That is why a spherical RNA condensate gradually dissolves from the outer surface to the center of PN.
+
+In contrast, in Fig. 4f, KL transcription and condensation is conducted in PN prehybridized with different amounts of Atto647-o*. This hybridization weakens the KL-to-PN affinity, making RNA condensation within the PN difficult at 10% prehybridization and leading to ill-defined RNA condensates. One can also see the brighter edges of these condensates. These arise from rather pure KL condensates, whereas the less intense core is a co-condensate. The pure KL layer even
+appears to dewet from the central condensate as seen by the undulations/non-perfect bright shell. At higher prehybridization levels, the internal PN environment becomes more repulsive, which prevents RNA condensation inside PN. Therefore, small condensates form at the outer shell of PN due to diffusion of KL transcripts from the PN to the surrounding.
+
+We find that the experimental description is insufficient in the original manuscript for showing the differences between experimental settings for Fig. 4a-c and Fig. 4f. Therefore, we now provide additional text and explanation for Fig. 4a-c as follows:
+
+To do so, we added different stoichiometric amounts of o*-Atto647 into solutions of PN containing already formed KL1-poly(A_{20-o})_n co-condensates and investigated how the invasion by o*-Atto647 would alter the pre-existing KL1-poly(A_{20-o})_n co-condensates. A gradual invasion of o*-Atto647 into the KL1-poly(A_{20-o})_n co-condensates takes place as the amount of o*-Atto647 increases (Fig. 4a, b). A continuous surface erosion of the co-condensates occurs because the o/o*-Atto647 hybridization reduces the affinity between KL condensate and PN interior by introducing stronger electrostatic and steric repulsion inside PN due to increased negative charge density and increasing persistence length of the formed dsDNA parts^{4,5} (Fig. 4c–e and Supplementary Video 3).
+
+For Fig. 4f, we have also modified the original text and believe the relevant discussion (highlighted) now can stand alone and make a clear distinction to the discussion about Fig. 4a-c:
+
+“Seeing such a profound impact, we then investigated KL1 transcription in PN with a poly(A_{20-o})_n matrix pre-hybridized by different amounts of o*-Atto647 (from 0% – 300%) to provide weakened affinity between PN matrix and KL1 transcripts. In analogy with the above result, single KL1 condensates form in pristine PN (Fig. 4f). When applying 10% o*-Atto647, the KL1 transcripts form single condensates with irregular secondary nucleation on its surface inside the PN, along with multiple tiny nuclei outside the PN shell (Fig. 4f). The brighter green parts are condensates purely enriched with KL1 transcripts that remain inside the PN due to relatively sufficient affinity. The marked difference – heterogeneously structured RNA condensates in Fig. 4f compared to the rather spherical and homogenous structures formed after invasion of pre-formed condensate in Fig. 4a, c – can be attributed to their different formation processes. Increasing the content of pre-hybridized o*-Atto647 domains from 10% to 300% gradually prevents KL1 condensate formation inside the PN due to weakened PN-KL1 interaction, which likely becomes even repulsive at higher pre-hybridization degrees because the ssDNA to dsDNA transition leads to higher negative charge density and persistence length, allowing for stronger electrostatic and steric repulsion^{4,5}, respectively, inside the PN. As a result, the KL1 transcripts formed inside the PN do not yield condensates insides the PN, instead, multiple small transcriptional KL1 condensates form in the PN surroundings.”
+
+c. dsDNA can have greater persistence length that leads to exclusion from protein membraneless organelles in-vitro (Nott, T., Craggs, T. & Baldwin, A. Membraneless organelles can melt nucleic acid duplexes and act as biomolecular filters. Nature Chem 8, 569–575 (2016), some discussion
+about dsDNA vs ssDNA in Jeffrey R. Vieregg et al, Journal of the American Chemical Society 2018 140 (5), 1632-1638). I wonder whether this dsDNA material properties affect its partitioning trends in Fig 4 g & h.
+
+Response: Thank you for the insightful suggestion. Yes, indeed, the greater persistence length of dsDNA is relevant for interpreting the partitioning trends in Fig. 4g, h. As reported in the suggested two literatures, dsDNA has higher charge density and is less flexible (JACS, 140, 1632-1638 (2018)), and the different properties of dsDNA and ssDNA have been shown to affect their partition into the membraneless organelles (Nat. Chem. 8, 569–575 (2016)). We believe that these are important references to demonstrate the two critical properties of dsDNA, i.e., higher charge density and greater persistence length for stronger electrostatic repulsion and steric repulsion, which helps to explain the partitioning trend observed in Fig. 4g, h, that dsDNA is excluded while ssDNA is enriched in the transcriptional KL condensates in solution. In addition, we believe the persistence length aspect of dsDNA also plays an important role in the invasion experiments shown in Fig. 4a-f.
+
+We have now cited the referred papers and added relevant texts to discuss the persistence length and steric repulsion of dsDNA for the relevant results:
+
+Increasing the content of pre-hybridized o*-Atto647 domains from 10% to 300% gradually prevents KL1 condensate formation inside the PN due to weakened PN-KL1 interaction, which likely becomes even repulsive at higher pre-hybridization degrees because the ssDNA to dsDNA transition leads to higher negative charge density and persistence length, allowing for stronger electrostatic and steric repulsion\(^{4,5}\), respectively, inside the PN.
+
+Furthermore, electrostatic repulsion from increased negative charge density, and steric repulsion from higher persistence length after dsDNA formation also play important roles\(^{4,5}\), as in analogy to re-entrant phenomena in living cells.
+
+4. It is still unclear how these two different KLS localize at the different locations under different magnesium concentration (Fig 5).
+
+a. KL2-R2 condensates still have little amount of KL1 in both cases of magnesium concentrations, suggesting that the coexistence of KL1-R1 condensate and KL2-R2 condensate is more than sum of each individual condensates. Likely, DNA matrix for PN is involved. Please address that.
+
+Response: Thank you for your interesting question. In Fig. 5a-d, for the two KLS transcriptions at 30 mM Mg\(^{2+}\), the budding KL2-R2 condensates do not contain PN matrix, as the PN matrix is trapped within the PN shell. Therefore, the PN matrix is not involved in the case of 30 mM Mg\(^{2+}\). While it is true that we observe the existing intensity of KL1-R1 at the KL2-R2 condensate. But the intensity is lower than KL1-R1 at the PN shell. We assume that the remaining intensity is due to the very few KL1-R1 arrested by slow diffusion kinetics at the KL2-R2 condensate.
+In the case of 15 mM Mg^{2+} transcription, there is indeed a minor distribution of KL1-R1 in KL2-R2, or KL2-R2 in KL1-R1 (Fig. 5e, f). Both KL1-R1 and KL2-R2 are transcribed in PN, where KL1-R1 or KL2-R2 phase separates individually. The remaining intensity is likely due to the slow diffusion of the transcripts in the crowded environments.
+
+We added text to explain this:
+
+Note that a minor distribution of KL1-R1 in KL2-R2, or KL2-R2 in KL1-R1 (Fig. 5e, f), can be observed, possibly due to the slow diffusion of the produced KLs in the crowded condensates.
+
+b. The reasoning of different morphology of KL condensates based on its material properties seem valid. This is a minor question, but do KL2 get transcribed inside of PN and get transported to be outside of shell? Or does the KL2 get transcribed outside of the DNA PN at 30 mM Mg2+?
+
+Response: As we show in Supplementary Fig. 1, promoter sequences on poly(A_{20-p})_n are enclosed in the PN after phase separation, so that the KL2-R2 is transcribed inside PN and transported to the outside later by the repulsion from KL1-R1.
+
+5. In line 428, what does it mean by helper proteins?
+
+Response: Thank you for your question. The helper proteins are proteins that are involved in transcription elongation and RNA processing, such as P-TEFb as discussed for instance in literature^{10}.
+
+We modified the text in the revised manuscript:
+
+While we focus on a rather artificial and well controllable system of KL condensates, this work lays an important cornerstone to study more sophisticated phase separation processes, such as in case of polymerase II that forms rich condensate architectures with helper proteins, e.g., P-TEFb that are involved in transcription elongation, and those which are implicated in disease and ageing^{10,11}.
+
+Additionally, the manuscript does not specify the number of independent replicates. Some of fluorescence intensity curves have shaded areas representing standard deviations, but it is unclear where this standard deviation is coming from. Clarifying this information is important to improve the transparency and reproducibility of the study. The details of methods seem enough.
+
+Response: Thank you for your comment. We have now included the number of independent repeats in the figure captions in revised manuscript and supporting information.
+
+References
+
+1. Fabrini, G. et al. Co-transcriptional production of programmable RNA condensates and synthetic organelles. Nat. Nanotechnol. **19**, 1665-1673 (2024).
+2. Morita, M. et al. Liquid DNA Coacervates form Porous Capsular Hydrogels via Viscoelastic Phase Separation on Microdroplet Interface. Adv. Mater. Interfaces **11** (2024).
+3. Liu, W., Lupfer, C., Samanta, A., Sarkar, A. & Walther, A. Switchable Hydrophobic Pockets in DNA Protocells Enhance Chemical Conversion. *J. Am. Chem. Soc.* **145**, 7090-7094 (2023).
+4. Nott, T.J., Craggs, T.D. & Baldwin, A.J. Membraneless organelles can melt nucleic acid duplexes and act as biomolecular filters. *Nat. Chem.* **8**, 569-575 (2016).
+5. Vieregg, J.R. et al. Oligonucleotide-Peptide Complexes: Phase Control by Hybridization. *J. Am. Chem. Soc.* **140**, 1632-1638 (2018).
+6. Feric, M. et al. Coexisting Liquid Phases Underlie Nucleolar Subcompartments. *Cell* **165**, 1686-1697 (2016).
+7. Shin, Y. & Brangwynne, C.P. Liquid phase condensation in cell physiology and disease. *Science* **357**, eaaf4382 (2017).
+8. Stewart, J.M. et al. Modular RNA motifs for orthogonal phase separated compartments. *Nat. Commun.* **15**, 6244 (2024).
+9. Udono, H. et al. Programmable Computational RNA Droplets Assembled via Kissing-Loop Interaction. *Acs Nano* **18**, 15477-15486 (2024).
+10. Changiarath, A. et al. Promoter and Gene-Body RNA-Polymerase II co-exist in partial demixed condensates. Preprint at 10.1101/2024.03.16.585180 (2024).
+11. Pei, G.F., Lyons, H., Li, P.L. & Sabari, B.R. Transcription regulation by biomolecular condensates. *Nat. Rev. Mol. Cell Bio.* **26**, 213–236 (2025).

0ae7f8f6cbad98f78d95595f0fd34157385af5879d5ea6b07e3d14e42afd8221/preprint/preprint.md

@@ -0,0 +1,209 @@
+Constructing synthetic nuclear architectures via transcriptional condensates in a DNA protonucleus
+
+Andreas Walther
+andreas.walther@uni-mainz.de
+
+University of Mainz   https://orcid.org/0000-0003-2170-3306
+Miao Xie
+University of Mainz
+Weixiang Chen
+University of Mainz   https://orcid.org/0009-0000-5518-0799
+Maria Roy
+University of Mainz
+
+Article
+
+Keywords:
+
+Posted Date: February 7th, 2025
+
+DOI: https://doi.org/10.21203/rs.3.rs-5959823/v1
+
+License: This work is licensed under a Creative Commons Attribution 4.0 International License.
+Read Full License
+
+Additional Declarations: There is NO Competing Interest.
+
+Version of Record: A version of this preprint was published at Nature Communications on September 10th, 2025. See the published version at https://doi.org/10.1038/s41467-025-63445-8.
+Constructing synthetic nuclear architectures via transcriptional condensates in a DNA protonucleus
+
+Miao Xie1,2,#, Weixiang Chen1,2,#, Maria de Roy1, Andreas Walther1,2*
+
+Affiliations
+1Life-Like Materials and Systems, University of Mainz, Duesbergweg 10-14, 55128 Mainz, Germany.
+2Max Planck Institute for Polymer Research, 55128 Mainz, Germany.
+*Corresponding author. Email: andreas.walther@uni-mainz.de
+#These authors contributed equally.
+
+Abstract
+Nuclear biomolecular condensates are essential sub-compartments within the cell nucleus and play key roles in transcription and RNA processing. Bottom-up construction of nuclear architectures in synthetic settings is non-trivial but vital for understanding the mechanisms of condensates in real cellular systems. Here, we present a facile and versatile synthetic DNA protonucleus (PN) platform that facilitates localized transcription of branched RNA motifs with kissing loops (KLs) for subsequent condensation into complex condensate architectures. We identify salinity, monomer feeding, and KL-PN interactions as key parameters to control co-transcriptional condensation of these KLs into diverse artificial nuclear patterns, including single and multiple condensates, interface condensates, and biphasic condensates. Over time, KL transcripts co-condense with the PN matrix, with the final architecture determined by their interactions, which can be precisely modulated using a short DNA invader strand that outcompetes these interactions. Our findings deepen the understanding of RNA condensation in nuclear environments and provide new strategies for designing functional nucleus-mimetic systems with precise architectural control.
+Introduction
+
+In eukaryotic cells, the nucleus provides a compartment for essential processes such as transcription, mRNA pre-splicing, and ribosome assembly\(^1\). To ensure precise spatial and temporal regulation of these biochemical processes\(^2\), membrane-less organelles such as nucleolus, Cajal bodies, and nuclear speckles form sub-compartments within the nucleus, which are biomolecular condensates that concentrate specific nucleic acids, enzymes, and metabolites\(^3-6\). Beyond regulating these crucial processes, unique nuclear patterns formed by biomolecular condensates vary across cell types, adapting to specific demands and functional cell states\(^7\). Importantly, dysfunctions in nuclear condensates have been implicated in diseases such as cancer, ribosomopathy, and neurodegeneration\(^6,\ 8,\ 9\). Thus, understanding and reconstructing nuclear biomolecular condensates is not only essential for uncovering their mechanisms but also holds significant potential for therapeutic applications.
+
+Despite considerable advances in studying natural biomolecular condensates and attempts to engineer transcriptional condensates within the nucleus\(^8,\ 10-12\) based on specific or non-specific interactions of protein-protein, protein-nucleic acid, and RNA-RNA pairs\(^2,\ 13,\ 14\), much still remains unknown about their formation mechanisms and the involved kinetic processes. Specifically, the mechanisms by which these condensates concentrate molecules, maintain structural integrity, regulate composition, and modulate internal biochemical activities remain elusive, largely due to the complexity of *in vivo* environments. In contrast, *in vitro* models of biomolecular condensates allow for precise control over composition in a simplified setting\(^11\), enabling detailed mechanism assessment through experiments and computational modeling\(^15\). Here, studies presently however rely on plain solutions that are far from the conditions in a nucleus.
+
+Transcriptional RNAs with specific sequences have been identified to play a key role in many biomolecular condensation processes\(^15\). However, achieving control in synthetic nuclear architectures and functions requires more advanced RNA designs capable of forming higher-order structures. In nature, the self-complementary kissing loop sequence in type 1 human immunodeficiency virus (HIV-1) virions has been identified as framework for systematically manipulating genomic dimerization\(^16\). Similar kissing loop interactions have been shown to facilitate condensation in bacterial riboswitches\(^13,\ 17\). Inspired by the sequence-dependent interaction of kissing loops, which enables specific pairing between internally folded RNAs\(^18,\ 19\), the groups of Takinoue\(^20\), di Michele\(^21\), and Franco\(^22\) have recently introduced programmable condensates in solution formed by nanostar-like RNA motifs. The latter two groups have further shown that RNA nanostars with kissing loops at the end of each arm (KLs) could co-transcriptionally condense into condensates with controlled size, number, morphology, and composition either in solution or confined within water-in-oil emulsions\(^21,\ 22\). Through integration of RNA aptamers into KLs, such condensates can mimic natural membrane-less organelles capable of selective capture of client molecules with biofunctions\(^21\). However, it remains unexplored whether RNA condensates can form in crowded conditions and how they may interact with DNA-rich environments resembling the cellular nucleus, where intricate RNA-DNA interactions occur.
+How such DNA environments influence the organizational principles of such designer condensates is unknown.
+
+We have recently introduced core-shell DNA coacervates, formed by single-stranded DNA (ssDNA) polymers, with a highly concentrated DNA-enriched core\(^{23}\), that can flexibly recruit molecules and proteins for enzymatic functions\(^{24},\ ^{25}\) and chemical reactions\(^{26}\). These DNA coacervates closely resemble the crowded environment of the cellular nucleus, making them an ideal platform for constructing nucleus mimics\(^{27}\). Therefore, we term them protonuclei (PNs) in this study. As the internal composition of the PNs can be flexibly tuned based on the ssDNA polymer selection, we incorporate T7 promoter sequences into the DNA core to recruit transcription templates and facilitate localized *in-protonucleo* transcription. We demonstrate that KL can be transcribed within these PNs, leading to the formation of co-transcriptional KL condensates with various morphologies. We demonstrate a range of synthetic nuclear architectures, including single condensates, multiple condensates, interfacial condensates formed through secondary nucleation, and biphasic condensates of orthogonal KLs, all controlled by salinity, PN-KL affinity, and competing PN-KL interactions, respectively. Given the design flexibility of transcriptional KLs and the tunable condensate patterns in our crowded PN system, we believe this artificial nucleus platform will significantly advance the field of synthetic biology, in particular synthetic cells, providing a powerful toolkit for designing and constructing synthetic nuclear architectures with unprecedented control and precision.
+
+Results
+
+Figure 1 shows an overview of our entire approach. It consists of constructing a modular PNs platform using DNA nanoscience approaches, followed by immobilization of short KL templates to initiate transcription therein. The transcribed KLs are designed to undergo phase separation by complementary interactions. By precisely controlling KL-PN interactions and environmental conditions, we study structure formation and response in detail through easily accessible pathways. In more detail, the DNA PNs are derived from our previous work on DNA protocells\(^{23},\ ^{24}\), where we have identified that temperature ramps of mixtures of long poly(A\(_{20-m}\)) ssDNA and long poly(T\(_{20-k}\)) ssDNA form micron-sized core-shell coacervates with an adenine-rich ssDNA polymer (polyA) core and a thymine-rich ssDNA polymer (polyT) shell\(^{23-25},\ ^{28}\). This process features a selective liquid-liquid phase separation (LLPS) of polyA during heating, forming polyA droplets at high temperature, which are then stabilized by polyT with A\(_{20}/\)T\(_{20}\) hybridization during cooling, forming a thin and crosslinked hydrogel shell. This ultimately furnishes a highly concentrated polyA core of around 10 g/L\(^{29}\). The dynamic properties of the PNs can be regulated from an arrested state to a liquid-like state by tuning the salinity. Additional ssDNA barcode sequences (o, p, k) can be modularly incorporated into the ssDNA polymers for integrating functionalities into the core and the shell (Fig. 1).
+Fig. 1 Transcriptional kissing loop (KL) condensates form different synthetic nuclear patterns in DNA protonuclei (PNs). PolyA strands with barcode p (T7 promoter sequence), polyA strands with dummy barcode (o), and polyT with k barcode are used for LLPS process to form PNs with an incorporated promoter region. The promoter barcodes inside the PNs recruit DNA templates, T7 RNA Polymerase (T7 RNAP), and nucleotide triphosphate (NTP) monomers to induce a localized transcription and enrichment of KL sequences, forming distinct nuclear patterns via different nucleation and condensation processes.
+
+We synthesized several ssDNA polymers using rolling circle amplification (details in Supplementary Table 1), including poly(A20-p)n, poly(A20-o)n, and poly(T20-k)n with n ranging roughly from 10 to 60 repeating units23. The barcodes p, o, and k serve specific functions. The most critical part is the p barcode in poly(A20-p)n, which is the T7 RNA polymerase (T7 RNAP) promoter sequence that allows for the flexible integration of ssDNA templates (short genes) amenable to transcription of RNA in the PNs through simple addition of the templates after formation of the PNs. Poly(A20-o)n serves to homogeneously dilute the p barcode and provides an addressable matrix barcode to tune properties and (as we will see below) adjust the affinity to the transcribed RNA, which regulates the subsequent growth of the transcriptional condensates. Following our established protocols23-25, 28, we prepared a set of core-shell PNs by mixing poly(A20-p)n and poly(A20-o)n for the core, and poly(T20-k)n for the shell, using a temperature ramp in TE buffer at 50 mM Mg^{2+}. Functionalization of the p and o barcodes with complementary dye-appended ssDNA confirms a homogeneous integration of both polyAs in the PN cores (confocal laser scanning microscopy (CLSM) images in Supplementary Fig. 1). The PNs can be conditioned to different salinity after preparation. We focus on a PN system where 10% of promoter sequences (poly(A20-p)n) are diluted with 90% of a matrix (poly(A20-o)n).
+
+To verify transcription to occur inside the PNs, we hybridized a transcription template T_x* containing p* for hybridization with the promoter sequence p and an active transcription region x* at stochiometric ratio into the PNs (sequences in Supplementary Table 2). x* codes for a simple RNA not amenable to undergo condensation. Subsequent addition of T7 RNAP and a nucleotide triphosphate (NTP) monomer mix containing 1% fluorescent monomer (UTP-Atto488) induces transcription with local formation of fluorescent RNA strands (Fig. 2a, b, and Supplementary Video 1).
+
+To better quantify the transcription efficiency and kinetics inside the PNs and compare it to free
+transcription in solution, we further designed a reporter (R) containing a fluorophore-quencher pair, which is a partially complementary double-stranded DNA (dsDNA; Rep/Rep’ sequences in Supplementary Table 2; ’ denotes a partially complementary sequence with a toehold). The transcribed Rep* from template T_{Rep*} will trigger a strand displacement reaction (SDR) with the R by fully hybridizing with the Rep strand, generating a fluorescent signal, which can be monitored by fluorescence measurements using a plate reader. In more detail, we compared the transcription kinetics between PNs with embedded promoter sequence of poly(A_{20}-p)_{n}, pure poly(A_{20}-p)_{n} ssDNA in solution, and short p ssDNA in solution – all at identical p concentration and otherwise identical transcription conditions (Fig. 2a, c). All systems show relatively similar kinetic profiles, with the free promoter in solution being the most active transcription system, and the PN showing a slightly lower activity compared to the free poly(A_{20}-p)_{n} in solution. The slightly lower activity can be understood considering constraints on the diffusion of NTPs into the PNs and RNA strands out of the PNs.
+
+After confirming successful transcription in the PNs, we turn to KL-condensate formation by transcriptional control in the PNs versus in solution. As a proof of concept, we first focus on a three-armed singled-stranded RNA (ssRNA) nanostar with a wildtype palindromic KL sequence^{20-22} at the tip of each arm (KL1 in Fig. 2d; Template = hybridized T_{KL1}/T_{KL1}’, where T_{KL1} contains a p* ssDNA sequence for hybridization to poly(A_{20}-p)_{n} inside the PNs, Supplementary Table 2). We compared differences in KL1 condensate formation at low [NTP] = 3.6 × [T_{KL1}] after 18 h transcription ([NTP] is defined as the maximum amount of KL1 transcripts that can be produced per template). The PN system clearly shows a single KL1 condensate in every PN with an average diameter of approximately 4 μm for PNs with an average diameter of around 6.7 μm (Fig. 2e, f). In striking contrast, no KL1 condensates can be found in solution due to the limited concentration of RNA transcripts (Fig. 2e, f). Transcriptional KL1 condensates in solution start to appear with diameter of ~ 4.5 μm at increased [NTP] ([NTP] = 14.4 ×; Fig. 2e, f). The size of the transcriptional KL1 condensate in solution increases with [NTP] due to the increased amount of RNA transcripts (Supplementary Fig. 2). This comparison demonstrates that the spatial transcription of the KL1 in PNs leads to locally high concentrations sufficient for condensation, similar to the enrichment mechanism in natural nuclear condensates^{2}.
+
+Interestingly, one single KL1 condensate forms in each PN, confirming sufficient dynamics within the PN to follow energy minimization constraints to yield a minimum surface area (Fig. 2e and Supplementary Fig. 3). We further performed fluorescence recovery after photobleaching (FRAP) experiments on the KL1 condensates in PNs and in solution to study their dynamic properties. Strikingly, their fluorescence recovery kinetics differ substantially. Whereas KL1 condensates in solution show near full recovery overnight, KL1 condensates in PNs only show limited recovery, highlighting much better diffusion dynamics of KL1 condensate in solution than in PNs (Fig. 2g, h). A complementary half-bleaching experiment shows a bright edge of transcriptional KL1 condensates in solution during recovery, indicating a dynamic exchange of soluble KL1 transcripts from the solution with the condensate phase (Supplementary Fig. 4). In contrast, half-bleached KL1 condensates in PNs show less recovery and lack the bright edge, likely due to their restricted
+dynamics in a DNA-crowded environment and interactions between PN matrix and the KL1 transcripts, as we will further discuss below.
+
+Next, we discuss the effects of [NTP] and [Mg^{2+}] on transcriptional KL1 condensate formation inside the PNs. KL1 transcription with [NTP] varying from 2.4 × to 3.6 × show a morphological transition from peripheral localization of KL1 transcripts to reorganization and compaction into a single condensate (Fig. 2i). The formation of peripheral KL1 transcripts at low [NTP] shows that incoming NTPs are converted to RNA as they reach the embedded transcription templates in the outer PN parts. The lack of a centrally compacted condensate points to the fact that, at this low concentration of KL1 transcripts, phase segregation is at least not very pronounced. The remaining ring indicates an interaction between the PN matrix and the KL1 transcripts. At higher [NTP], KL1 transcripts are homogeneously produced throughout the PN, and phase segregation drives the formation of the KL1 condensate.
+
+[Mg^{2+}] shows a profound impact on nucleation and condensate morphology. Multiple small condensates can be observed at 15 mM Mg^{2+}, whereas [Mg^{2+}] > 20 mM leads to the formation of a single condensate droplet. 20 mM Mg^{2+} corresponds to a transition point. Interestingly, a transition in the condensate formation process is visible. Whereas isolated nucleation events dominate at 15 mM Mg^{2+}, co-continuous phase separation is visible above 25 mM Mg^{2+} with a sponge-like structure. At 40 mM Mg^{2+}, condensate formation in PNs is no longer visible (Fig. 2j). Such distinct condensate formation in PNs is associated with multiple influences of Mg^{2+} on the system: First, higher [Mg^{2+}] leads to reduced dynamics in the crowded environment of PNs, as previously studied by us in detail^{23, 24}. Second, higher [Mg^{2+}] also assists in tighter condensation of the KL condensates and potentially increases non-specific interactions between the KL condensates and the PN matrix^{22}. Third, increasing [Mg^{2+}] leads to a continuous decrease of the transcription efficiency as depicted in Fig. 2k. Thus, multiple isolated nucleation events and binodal phase separation occur at 15 mM Mg^{2+}, driven by the high dynamics of the PN core and the high transcription efficiency. In contrast, spinodal or viscoelastic phase separation^{26, 30, 31} is favored at high Mg^{2+} concentrations, where the dynamics of the PNs become more arrested. Further, charge screening increases the propensity for non-specific interactions between nucleic acids (RNA and DNA). The transcription is strongly suppressed at 40 mM Mg^{2+} with limited KL1 transcripts so that condensation of KL1 cannot take place (Fig. 2j, k).
+Fig. 2 Transcriptional KL condensates in PNs show different nuclear patterns. a, Scheme showing transcription in PNs characterized by CLSM and plate reader. For CLSM experiment, UTP-Atto488 is added to the NTP mix for transcript labeling. For plate reader experiments, dsDNA reporters with fluorophore-quencher pairs (R) are present in solution to react with transcribed RNA by strand displacement reaction (SDR), generating fluorescent signals. b, Representative CLSM images showing the localized transcription of fluorescent x* (green, labeled by UTP-Atto488 during transcription) inside PNs at different times ([NTP] : [T_x*] : [p] = 250 : 1 : 1, 30 °C, 30 mM Mg^{2+}, 2.5 U/μL T7 RNAP). The whole process is recorded in Supplementary Video 1. c, Transcription kinetics inside PNs with embedded poly(A_{20}-p)_n in solution with free poly(A_{20}-p)_n, and in solution with free promoter (p) oligonucleotide, monitored by SDR of the R in a plate reader ([NTP] : [R] : [T_{Rep^*}] : [p] = 200 : 10 : 1 : 1, 30 °C, 6 mM Mg^{2+}, 2.5 U/μL T7 RNAP). d, Scheme for the formation of transcriptional KL1 condensates in PNs and in solution, with KL1 labeled by UTP-Atto488 during transcription. e, CLSM images with maximum intensity projection of z-stacked images showing the formation of transcriptional KL1 condensates in PNs containing embedded poly(A_{20}-p)_n ([NTP] : [T_{KL1}] = 3.6), and in solution with free promoter (p) oligonucleotide ([NTP] : [T_{KL1}] = 3.6 : 1 or 14.4 : 1, 30 °C, 30 mM Mg^{2+}, 2.5 U/μL T7 RNAP, 18 h for in-PN and in-solution transcription). f, Diameter distributions of KL1 condensates in PNs and in solution at different [NTP]. g, Normalized fluorescence recovery kinetics in the bleached areas in (h) during FRAP experiments on
+KL1 condensates in PNs and in solution at 30 °C. Intensity values were normalized to pre-bleached levels. n = 3. h, Time-lapse CLSM images for FRAP experiments on KL1 condensates in PNs and in solution at 30 °C. White dashed circles indicate the bleached regions. i, Representative CLSM images showing the effect of [NTP] on transcriptional KL1 condensate formation in PNs after 18 h at 30 °C. Note that the left half of the 3.6 × [NTP] sample is a maximum intensity projection of z-stacked images. Green channel: KL1-Atto488; Magenta channel: K*-Atto647/poly(A20-k)h. j, Representative CLSM images showing the effects of [Mg^{2+}] on transcriptional KL1 condensate formation in PNs after 18 and 48 h reaction. Note that a hyperstack image generated from z-stacked images is used for KL1 condensates in PNs at 15 mM Mg^{2+} after 48 hours of transcription to visualize condensates formed in different planes with a corresponding color-coded z-scale. k, Effects of [Mg^{2+}] on transcription efficiency in PNs, monitored via RNA-triggered SDR of the R. ([NTP] : [R] : [T_{rep}] : [p] = 200 : 10 : 1 : 1, 30 °C, 2.5 U/μL T7 RNAP at indicated [Mg^{2+}]), measured by a plate reader. In the box plot (f), the central line marks the median, the box represents the interquartile range (IQR) from Q1 (first quartile) to Q3 (third quartile), and the whiskers enclose all data points from the minimum to the maximum. This applies to all box plots shown in this paper. Error areas represent standard deviation. Scale bars are 10 μm for (b) and (e), 5 μm for (h), (i), and (j).
+
+To get a deeper understanding of the morphological development of single KL condensates in the PNs at 30 mM Mg^{2+}, we monitored the whole process over 24 h through CLSM (Fig. 3a, b). Two distinct stages occur. In the first 12 h, transcription takes place from the edge of the PNs to their center, and the entire structures reach maximum fluorescence intensities at 12 h (Fig. 3b-d). Spongy structures of KL1 condensates during phase separation start to appear at ca. 8-10 h, whereas significant coarsening and compaction into single spherical condensates follows in the later 12-24 h (Fig. 3b-e). Interestingly, we can observe a relatively slow and continuous increase of the PN dimensions, as facilitated by the relaxation of polyA/polyT shell as a result of the increasing negative charge density inside the PNs during localized RNA production and condensation (Fig. 3c, e).
+
+For further probing the universality of this single condensate formation phenomena for various KL structures, we adapted a KL1 condensate with an RNA light-up broccoli aptamer (BrA) as one of the arms, termed KL1-BrA (NUPACK-simulated structure shown in Supplementary Fig. 5a). After 12-24 h transcription, single condensates are formed in each PN. In contrast, KL1-BrA only forms irregular aggregates in solution. Here, the interaction between the PN matrix and the KL1-BrA condensate may facilitate better relaxation and stabilization of KL1-BrA condensate within the PNs (Supplementary Fig. 5). Taken together, KL condensation in solution and in-PN differ profoundly in both the kinetic formation process and the formed final structures at what could be considered closer to the thermodynamic equilibrium. The system can be easily tuned by adjusting the NTP and Mg^{2+} concentrations and is robust to changes in the KL components.
+Fig. 3 Mechanism of co-condensation of transcriptional KL1 with the PN matrix. a, Scheme for formation of single co-condensates of transcriptional KL1 and polyA in PNs through peripheral initiation of transcription, re-organization, and co-condensation. b, Representative CLSM images of PNs conducting KL1 transcription over 24 h ([NTP] : [TKL1] = 3.6 : 1, 1 mol% UTP-Atto488, 30 °C, 2.5 U/μL T7 RNAP, 30 mM Mg^{2+}). Note a slight slow-down of co-condensation kinetics of KL1 compared with results shown in Fig. 2e, j after 18 h, due to the interruption of shaking during incubation for the CLSM imaging. Green channel: KL1 condensate labeled by UTP-Atto488; Magenta channel: PN shell (poly(T_{20}-k)_{n} labeled with k*-Atto647). c, Space-time plot analysis corresponding to the two dashed lines in (b) over 24 h shows the KL1 transcription, condensation, and reorganization process. d, Normalized fluorescence intensity change in the KL1 condensate channel over 24 h in the two white dashed circle in (b), normalized to the intensity at 24 h. e, Normalized radius change of KL1 condensate and PN as measured from (b) over 24 h. f, g Representative CLSM images of fluorescent PNs (magenta) containing KL1 transcripts (green) at 12 h (f) and 24 h (g). The right plots correspond to the line segment analysis of the white line in the CLSM images, showing fluorescence distribution for KL1-Atto488 and poly(A_{20}-o)_{n}-Atto643. The KL1 transcripts show a peripheral distribution in the PN matrix at 12 h (f), while colocalization and co-condensation between KL1 and PN matrix occur after 24 h (g). Error bars and error areas represent standard deviation. Scale bars are all 5 μm.
+
+To study the behavior and aforementioned interactions of the PN DNA matrix with the KL1 RNA condensates, we covalently labeled poly(A_{20}-o)_{n} with Atto643 to prepare fluorescent PNs and used these new PNs to initiate localized KL1 transcription. As expected, the initial production and localization of KL1 transcripts occur at the periphery of the PNs (Fig. 3f). Unexpectedly,
+co-condensation of the PN matrix with the KL1 condensates occurs over time. These co-condensates deposit at the bottom of PNs after 24 h in the imaging chamber, highlighting their higher density and compactness (Fig. 3g, Supplementary Fig. 6, and Supplementary Video 2). FRAP experiments reveal a better recovery for the KL1 components compared to the PN matrix, corresponding to higher dynamics for the KL1 condensate part composed of small RNAs than the PN matrix composed of long ssDNA polymers (Supplementary Fig. 7). This demonstrates the molecular level diffusivity of the RNA nanostars in this co-condensate structure.
+
+Overall, this co-condensation between PN and KL1 condensate comes unexpectedly because the KL1 condensate was not designed to have any specific interactions with the PN matrix. Indeed, a NUPACK simulation suggests no specific hybridization between the A_{20-o} repeats and the KL1 sequence (Supplementary Fig. 8). Experimentally, we probed interactions between mature KL1 condensates and poly(A_{20-o})_n inside PNs by adding different quantities of o*-Atto647 (from 10% - 300%) that can bind to the majority phase of poly(A_{20-o})_n in the PNs. We hypothesized that the hybridization between o/o* may break non-specific KL1-PN interactions (Fig. 4a, b). A gradual invasion of o*-Atto647 into the KL1-poly(A_{20-o})_n co-condensates occurs as the amount of o*-Atto647 increases. This process leads to continuous surface erosion of the co-condensates (Fig. 4c-e and Supplementary Video 3). A sharp interface defined by a bright ring of o*-Atto647 with a locally high concentration appears. The poly(A_{20-o})_n/o*-Atto647 thereafter occupies the space within the entire PN, whereas the KL1 transcripts are squeezed to the PN periphery and eventually dissolve into solution to equilibrate to their low concentration there. This process verifies that the interaction between KL1 and poly(A_{20-o})_n PN matrix promotes the formation of the KL1-poly(A_{20-o})_n co-condensate.
+
+Seeing such a profound impact, we then investigated KL1 transcription in PNs with a poly(A_{20-o})_n matrix pre-hybridized by different amounts of o*-Atto647 (from 0% - 300%) to provide weakened affinity between PN matrix and KL1 transcripts. In analogy with the above result, single KL1 condensates form in pristine PNs (Fig. 4f). When applying 10% o*-Atto647, the KL1 transcripts form single condensates with irregular secondary nucleation on its surface inside the PN, along with multiple tiny nuclei outside the PN shell (Fig. 4f). The brighter green parts are condensates purely enriched with KL1 transcripts that remain inside the PN due to relatively sufficient affinity. Increasing the content of pre-hybridized o*-Atto647 domains from 10% to 300% gradually prevents KL1 condensate formation inside the PNs due to weakened PN-KL1 interaction, which likely becomes even repulsive at higher pre-hybridization degrees. As a result, the KL1 transcripts formed inside the PNs do not yield condensates inside the PNs, instead, multiple small transcriptional KL1 condensates form in the PN surroundings. These results highlight the importance of the interaction between the DNA matrix of the PNs and the KL1 transcripts in both the formation and the maintenance of the condensates within the PNs. Hence, modulating the DNA-RNA interaction is a way for regulating nucleus condensate architectures.
+
+To directly study the affinity between polyA sequence of the PN matrix and KL1 condensate, we prepared transcriptional KL1 condensates in solution and added A_{20-o}-Atto647 ssDNA, and
+A20-o/o*-Atto565 dsDNA. Such pure KL1 condensates sequester A20-o-Atto647 whereas A20-o/o*-Atto565 is excluded (Fig. 4g, h). Such marked differences among interactions between KL1-to-ssDNA versus KL1-to-dsDNA confirm some level of unspecific interaction between the KL1 transcript and the o region, which is removed through hybridization into o/o*. Furthermore, electrostatic repulsion from increased negative charge density after dsDNA formation could also play a role, as in analogy to re-entrant phenomena in living cells, where transcriptional condensate formation is promoted at low rates of RNA synthesis up to a point of charge imbalance, beyond which higher rates of RNA synthesis disfavors condensate formation11,32.
+
+![Representative CLSM images of transcriptional KL1 condensates (Atto488, green channel) in PNs (Atto647, magenta channel) 60 min after adding 10%, 50%, 80% and 300% o*-Atto647 as an invader strand. Schematic diagrams illustrating the hybridization of the polyA matrix of PNs induces disassembly of KL1-PN co-condensate.](page_184_393_1080_1442.png)
+
+Fig. 4 Hybridization of the polyA matrix of PNs induces disassembly of KL1-PN co-condensate.
+a, Representative CLSM images of transcriptional KL1 condensates (Atto488, green channel) in PNs (Atto647, magenta channel) 60 min after adding 10%, 50%, 80% and 300% o*-Atto647 as an invader strand. b, Schematic
+illustration of the o*-Atto647 invasion process. Hybridization of o*-Atto647 with poly(A20-o)h starts from the edge of the KL1-polyA co-condensate with a bright and sharp invading front to final dissolution of the whole co-condensate. e, Representative CLSM images showing the process of co-condensate (Atto488, green channel) dissolution by adding 300% o*-Atto647 (magenta channel) to hybridize to the poly(A20-o)h of the PNs. See also Supplementary Video 3. d, Space-time plot analysis along the white dashed line in (c) shows the gradual dissolution of the condensate. e, Normalized fluorescence intensities of KL1 condensates (KL1-Atto488) and invader strand (o*-Atto647) measured in the white dashed circle in (c) during the invasion process. f, Scheme and representative CLSM images of KL1 transcription and condensation (Atto488, green channel) after 18 h in pre-hybridized PNs with 0%, 10%, 50%, 80%, and 300% o*-Atto647 (magenta channel). g, Scheme shows the attractive interaction between KL1 condensate and ssDNA A20-o-Atto647, and repulsive interaction between KL1 condensate and dsDNA A20-o-Atto647/o*-Atto565. h, Representative CLSM images of pure transcriptional KL1 condensates (Atto488, green channel) prepared in solution, with the addition of (top) ssDNA A20-o-Atto647 (magenta channel) for 1 h, showing preferential partitioning, or (bottom) dsDNA A20-o-Atto647/o*-Atto565 (red channel) showing rejection. Shaded areas represent standard deviations. Scale bars are 5 μm for (a), (c), and (f), and 10 μm for (h).
+
+Finally, we attempted to integrate orthogonal KL transcription systems into PNs for constructing more complex structures to mimic multiple RNA condensates in the crowded environment of natural cell nuclei. We adapted two KL nanostars (KL1-R1 and KL2-R2) with orthogonal kissing loop sequences at the end of their arms, and distinct tail regions (R1 and R2) for specific labeling by R1*-Atto488 and R2*-Atto647, respectively (Fig. 5a). We firstly confirmed the transcription and the formation of centrally located condensates for both KL1-R1 or KL2-R2 inside PNs (Supplementary Fig. 9). Hence, both systems form similar condensate structure as the original KL1 system and the KL1-BrA system studied above.
+
+Since [Mg^{2+}] can control the condensate morphology (Fig. 2j), we conducted co-transcription of both KLS in the same PN at 15 and 30 mM Mg^{2+}, respectively (Fig. 5a). At 30 mM Mg^{2+}, KL1-R1 assembles to a large single condensate (~0.7-fold the diameter of the host) at the PN center, while KL2-R2 forms small condensates, budding at the PN shell, with diameters less than 0.2-fold of the host PN (Fig. 5b-d). This suggests a preferred interaction between KL1-R1 and PN matrix, retaining the KL1-R1 condensate inside the PN, whereas KL2-R2 gets obviously expelled. KL1-R1 dominates the interaction with the PN matrix in this competitive system, whereas pure KL2-R2-PN would form a single central condensate (Supplementary Fig. 9). At 15 mM Mg^{2+}, the transcriptional KL1-R1 occupies the major PN space, while KL2-R2 forms multiple condensates in the PNs (Fig. 5e-g). This can be attributed to weakened interactions between KL1-R1 and the PN at low salinity, allowing KL2-R2 to occupy some of the available volume in the PN to form condensates. Hence, the combined effect of Mg^{2+} on changing the viscoelastic properties and modulating KL interactions as well as KL-to-PN interactions again shows a profound effect. We can conclude that phase segregation of KL1-R1 is energetically favored to be retained in the PNs. A mixing of both KL phases does not occur.
+
+To verify the competitive interaction between KL1-R1 and KL2-R2 with the polyA in the PN matrix, we performed competitive partition experiments of A20-o or A20-o/o* with pure transcriptional KL1-R1 and KL2-R2 condensates grown in solution. The results show preferential partitioning of A20-o into the KL1-R1 condensates, whereas A20-o/o* is excluded by both condensates (Supplementary Fig. 10). This confirms a higher affinity of KL1-R1 condensates to
+the PN matrix and explains the different condensate architectures formed in the PNs. Additionally, transcriptional KL2-R2 condensates show a less spherical structure compared with transcriptional KL1-R1 ones (Supplementary Fig. 10), suggesting stronger condensation interactions for KL2-R2 than KL1-R1, consistent with the higher melting temperature of KL2 interactions than KL1 interactions provided in literature21. This helps to explain that the KL2-R2 could still form condensates, whether expelled from PN at 30 mM Mg^{2+} or remained in PN at 15 mM Mg^{2+}. In summary, these results reveal that, in addition to salinity effect, subtle variations in RNA composition and sequence modulate their interaction with the DNA matrix of PN in a competitive environment, leading to profoundly different condensation processes and resulting in distinct multi-phase co-condensate architectures in PNs.
+
+![Scheme showing orthogonal transcription and condensation of KL1-R1 and KL2-R2 in the same PN at different salinity, and representative single-plane CLSM images and maximum intensity projections of z-stacked CLSM image showing orthogonal transcriptional condensates of KL1-R1 and KL2-R2 in PNs at 30 mM Mg^{2+}](page_184_670_1080_670.png)
+
+Fig. 5 Formation of orthogonal transcriptional KL condensates in PNs. a, Scheme showing orthogonal transcription and condensation of KL1-R1 and KL2-R2 in the same PN at different salinity. KL1-R1 is identical to KL1 in its nanostar framework, but with an additional recognition tail (R1) for R1*-Atto647 labelling. KL2-R2 shares the same stem sequence as KL1 but has orthogonal kissing loop sequences and a distinct recognition tail (R2) for R2*-Atto488 labeling. R1*-Atto647 and R2*-Atto488 are added during transcription. b, Representative single-plane CLSM image and maximum intensity projection of z-stacked CLSM image showing orthogonal transcriptional condensates of KL1-R1 (green channel) and KL2-R2 (magenta channel) in PNs at 30 mM Mg^{2+} ([NTP] : [R1*] : [R2*] : [T_{KL1-R1}] : [T_{KL2-R2}] : [p] = 3.6 : 1.8 : 1.8 : 0.5 : 0.5 : 1, 30 °C, 30
+mM Mg^{2+}, 2.5 U/μL T7 RNAP). c, Normalized intensity profiles of line segment analyses corresponding to the white line in (b) for both channels. d, Diameter of formed orthogonal condensates at 30 mM Mg^{2+}, normalized to the diameter of the host PNs. e, Representative single-plane CLSM image and maximum intensity projection of z-stacked CLSM image showing orthogonal transcriptional condensates of KL1-R1 and KL2-R2 in PNs at 15 mM Mg^{2+} ([NTP] : [R1*] : [R2*] : [TKL1-R1] : [TKL2-R2] : [p] = 3.6 : 1.8 : 1.8 : 0.5 : 0.5 : 1, 2.5 U/μL T7 RNAP, 15 mM Mg^{2+}, 30 °C, 18 h reaction). f, Normalized intensity profiles of line segment analyses corresponding to the white line in (e) for both channels. g, Diameter of formed KL2-R2 condensates at 15 mM Mg^{2+}, normalized to the diameter of host PNs. Note that the diameter of KL1-R1 condensates cannot be quantified due to their hollow shape. Scale bars are all 5 μm.
+
+Discussion
+
+We have introduced a versatile nucleus-mimicking DNA condensate platform – a protonucleus – that enables localized transcription and the study of phase-separation of transcribed RNA nanostars in crowded and highly concentrated DNA environments. Since the strategy builds on our previous work on all-DNA synthetic cells^{23-25, 28}, our approach shows how specific components from a completely different area of research, that is synthetic artificial cell research, can be effectively repurposed into new application domains. These protonuclei offer a highly programmable platform for introducing short genes for transcription while also enabling control over properties such as gene density and the dynamic behavior of the matrix^{23, 24}. Transcription inside these crowded PNs proceeds with satisfying efficiency up to high salt concentration. To study transcriptional folding and phase segregation in the crowded, nuclear-mimetic environment, we focused on KL condensates formed by ssRNA nanostars. We identified that ionic strength is one key parameter for cross-regulating transcription efficiency, viscoelasticity of the PNs, and KL-PN affinity. These effects in turn affect the nucleation of condensates from binodal to spinodal or viscoelastic phase separation^{26, 30, 31}, resulting in tunable artificial nuclear architectures inside the PNs. The non-specific interactions between KL and PN matrix turned out to be crucial for retaining KL transcripts inside PNs via KL-PN co-condensation. We showed how such interactions can be efficiently modulated using DNA nanoscience approaches in such synthetic settings, ultimately leading to a repulsion and exclusion of the KL condensates from the PNs.
+
+We further studied co-transcription and condensation of orthogonal KLs systems within the same PN, which resulted in distinct structures arising from competitive interactions between different RNA nanostars and the PN matrix. This highlights the potential of using our PN platform to study subtle interactions between RNA and DNA, as well as competitive interactions among RNAs in a DNA-enriched environment. Finally, at proper conditions, multiphase condensate structures can be built, which are further regulated by salinity through the cross-regulation of the viscoelastic environment, transcription efficiency, and competitive KL-PN interaction.
+
+Looking into the future, our work opens new perspectives for constructing artificial nuclear architectures in synthetic model systems with DNA nanoscience tools. While we focused on a rather artificial and well controllable system of KL condensates, this work lays an important cornerstone to study more sophisticated phase separation processes, such as in case of polymerase II that forms rich condensate architectures with helper proteins, and those which are implicated in
+disease and ageing\(^{33,34}\). In addition, the modulable KL-PN interactions within protonucleus could serve as simplified models for transcriptional condensates in living cells, which are dynamically forming and dissolving, and essential for transcription regulations\(^{11,32}\). Moreover, from the perspectives of molecular systems engineering, synthetic biology, and artificial cell research, we have identified important pathways to transcriptionally regulate structure formation processes towards multiscale condensates that can be selectively addressed in their compartments. We anticipate that this system will serve as a valuable platform and toolkit for DNA nanoscience and synthetic biology.
+
+Methods
+Materials
+ssDNA were purchased from Biomers and Integrated DNA Technologies (IDT). Supplementary Table 1 and 2 summarize all sequences used in this study. T4 DNA Ligase (2 U/μL), Exonuclease I (40 U/μL), Exonuclease III (200 U/μL), and Φ29 polymerase (10 U/μL) were purchased from Lucigen. Thermostable Inorganic Pyrophosphatase (2 U/μL), T7 polymerase (50000 U/mL) and nuclease-free water were bought from New England BioLabs (NEB). Deoxynucleotide triphosphate (dATP, dTTP, dGTP and dCTP) (100 mM), Aminoallyl-dUTP-XX-ATTO-643 (1 mM), Aminoallyl-UTP-Atto488 (1 mM), and Aminoallyl-UTP-Atto630 (1 mM) were purchased from Jena Bioscience. Hexadecane, sodium chloride, magnesium chloride, Tris(hydroxymethyl)-aminomethane hydrochloride (Tris-HCl), Trizma base, acetic acid and Ethylenediaminetetraacetic acid disodium salt dihydrate (EDTA), were purchased (as bioreagent grade if available) from Sigma-Aldrich. RNase Inhibitor (40 U/μL), RNase-free TE buffer (Invitrogen, 10 mM Tris and 1 mM EDTA, pH 8.0, 500 mL), 384-well high-content imaging glass bottom microplates were purchased from Corning.
+
+Instruments
+All thermal annealing and heating ramps were performed on a TPpersonal Thermocycler (Analytik Jena). Incubation with shaking was carried out on an Eppendorf ThermoMixer C with heated lid. DNA concentration was determined by a DS-11 Spectrophotometer (DeNovix). Confocal laser scanning microscopy (CLSM) was performed on a Leica StellaRis 5.
+
+Synthesis of circular ssDNA templates and long ssDNA polymers
+The synthesis of the circular DNA template and its corresponding ssDNA polymer can be found in our previous reports\(^{23}\). In short, the linear ssDNA template and the corresponding ligation strand were firstly mixed at concentration of 1 μM in 100 μL TE buffer containing 100 mM NaCl. The solution was heated to 85 °C for 5 min before cooling to 25 °C with a cooling rate of 0.01 °C/s for hybridization. Afterwards, 20 μL of 10× Ligase buffer (500 mM Tris-HCl, 100 mM MgCl₂, 50 mM dithiothreitol and 10 mM ATP (Lucigen)), 70 μL of nuclease-free water and 10 μL of T4 DNA Ligase (2 U/μL (Lucigen)) were introduced into the reaction mixture at room temperature for 3 h reaction. The solution was then heated to 70 °C for 20 min to deactivate the enzyme. Then, 10 μL of Exonuclease I (40 U/μL (Lucigen)) and 10 μL of Exonuclease III (200 U/μL (Lucigen)) were added into the reaction mixture for further overnight reaction at 37 °C to remove the ligation strands and any non-circularized templates in solution. Afterwards, the reaction mixture was heated to 80 °C for 40 min to deactivate the enzymes. To obtain the final circular ssDNA templates,
+the reaction mixture was washed with 400 μL TE buffer and filtrated using Amicon Ultra-centrifugal filters with a 10 kDa cut-off (Merck Millipore) for three times. The concentrations of the collected circular ssDNA templates were measured by the DS-11 Spectrophotometer (DeNovix), and the templates were stored in TE buffer at -20 °C.
+
+For the synthesis of the long ssDNA polymers, we used rolling circle amplification (RCA). 5 μL of circular template (1 μM in TE buffer) and 1 μL of exonuclease resistant primer (10 μM in TE buffer) were mixed with 76 μL nuclease-free water, 10 μL of commercial 10× polymerase buffer (500 mM Tris-HCl, 100 mM (NH4)2SO4, 40 mM dithiothreitol, 100 mM MgCl2 (Lucigen)), 2 μL of Φ29 DNA polymerase (10 U/μL (Lucigen)), 1 μL of thermal stable inorganic pyrophosphatase (2 U/μL (NEB)) and 5 μL of adjusted deoxyribose nucleoside 5'-triphosphate mix (100 mM, the mix contains pure dATP, dTTP, dCTP, and dGTP solutions mixed in corresponding proportions of the exact composition of the desired ssDNA polymer repeating units (Jena Bioscience)). Note that for the synthesis of ssDNA polymers with in-chain fluorophores of Atto643, we replaced 2 mol% of the dTTP in the mix with Aminallyl-dUTP-XX-ATTO-643 for random insertion of the dye into the ssDNA chains during RCA. The reaction mixture was kept at 30 °C for 50 h before thermal cleavage at 95 °C for 15 min to shorten the ultrahigh molecular weight of the synthesized DNA polymer23. The final products were purified by rinsing with 400 μL TE buffer and filtration in Amicon Ultra-centrifugal filters with 30 kDa cut-off (Merck Millipore) three times. The concentrations of the collected final ssDNA polymers were measured using the DS-11 Spectrophotometer (DeNovix), and the DNA polymers were stored in TE buffer at -20 °C.
+
+Preparation of all-DNA PNs embedded with T7 promoter sequence.
+The preparation of the PNs is adapted from our previous reports23 with modifications for the formation of PNs containing T7 promoter sequence. Adenine-rich DNA polymers (poly(A20-p)n + poly(A20-o)n in a ratio of 1:9) (0.5556 g/L) and poly(T20-k)n (0.0694 g/L) were mixed in TE buffer without any salt at a final volume of 9 μL. The solution mixture was heated at 95 °C for 15 min for thermal cleavage to further reduce the chain length of the ssDNA polymers. Afterwards, 1 μL of TE buffer containing 500 mM MgCl2 was introduced into the reaction mixture. The solution containing finally 0.5 g/L mixture of polyA and 0.0625 g/L poly(T20-k)n with 50 mM MgCl2 was heated to 95 °C for 20 min (3 °C/s) and cooled down to room temperature (3 °C/s), yielding core-shell PNs. Finally, the 10 μL solution containing the PNs was diluted 5 times by adding 40 μL TE buffer containing various amounts of MgCl2 to reach desired salinity. The obtained 50 μL DNA condensates solution (as 5× diluted) has 0.1 g/L polyA mixture and 0.0125 g/L poly(T20-k)n, corresponding to ca. 0.8 μM p barcode, ca. 7.2 μM o barcode and ca. 1 μM k barcode, respectively, in total solution. The solution was then stored in a fridge at 4 °C for 1 week for equilibration before usage.
+
+Spatially controlled transcription assay in PN.
+For transcription in PNs monitored by plate reader, 3.125 μL of 5× diluted PNs (90% o barcode + 10% p barcode) is further diluted into 25 μL solution containing 1× RNA polymerase buffer (40 mM Tris-HCl, 6 mM MgCl2, 1 mM DTT, 2 mM spermidine), 100 nM template (TRep*), 1 μM prehybridized fluorophore-quencher reporter (Rep/Rep’ dsDNA), 2.5 U/μL T7 RNAP, 0.02 U/μL Thermostable Inorganic Pyrophosphatase, 1 U/μL RNase Inhibitor. MgCl2 concentration was adjusted in different settings as noted in each figure caption. At the end, 2 μL of NTP mix (to reach 2 mM of ATP, GTP, CTP, and UTP each) was added into the solution to trigger the transcription reaction at different temperatures ranging from 25-30 °C. The final promoter sequence
+concentration in the solution is at 100 nM. As for control, transcription with pure promoter ssDNA (p) and poly(A20-p)n ssDNA polymer was also performed to compare the transcription efficiency. For kinetic experiments under CLSM, Tx* is loaded into the PNs at a final concentration of 100 nM. Reporter is not used, instead, we further added 0.0833 mM Aminoallyl-UTP-Atto488 so that the transcribed RNA can be fluorescently labeled and observed under CLSM.
+
+Transcriptional KLs condensates formation.
+2.5 μL of 5× diluted PNs (90% o barcode + 10% p barcode) is further diluted into 20 μL solution containing 1× RNA polymerase buffer (40 mM Tris-HCl, 6 mM MgCl2, 1 mM DTT, 2 mM spermidine), 100 nM dsDNA template (Tkl1/Tkl1*, Tkl1-BrA/Tkl1-BrA*, Tkl1-R1/Tkl1-R1*, or Tkl2-R2/Tkl2-R2*; or 50 nM Tkl1-R1/Tkl1-R1* + 50 nM Tkl2-R2/Tkl2-R2*), 2.5 U/μL T7 RNAP, 0.02 U/μL Thermostable Inorganic Pyrophosphatase, 1 U/μL RNase Inhibitor, and 0.048 mM NTP mix (0.048 mM of ATP, GTP, CTP, and UTP each at [NTP] : [Tkl1] = 3.6 : 1, for maximum amount of KL1 produced, which is 3.6-fold of Tkl1, adjusted in different settings as noted in figure captions). MgCl2 concentration is adjusted in different settings as noted in each figure caption. The mixture is incubated with shaking at 30 °C for 18 h reaction. The final promoter sequence concentration in the solution is at 100 nM. As control, transcription of KLs with pure promoter oligo was also performed with corresponding NTP concentration. For transcription of KLs with covalent label, 1 mol% of UTP is replaced by either Aminoallyl-UTP-Atto488, or Aminoallyl-UTP-Atto630 in transcription system. For KL1-BrA transcription, 0.05 mM DFHBI is added to the solution. For KL1-R1 or KL2-R2 transcription, 360 nM R1*-Atto647 or R2*-Atto488 sequence is added to the system, respectively. For transcriptional KL1 condensate formed in solution as a control, 100 nM promoter ssDNA is added to system, instead of PNs.
+
+Fluorescence recovery after photobleaching (FRAP) experiments.
+FRAP experiments were performed by applying 3 times bleaching in a small circular region of interest (ROI) with diameter of 2 μm by 100 % laser intensity. Post-bleaching images were recorded over different periods. The intensities within the circular ROI (\( I_{ROI} \)), and intensities in a circular region of the same size away from bleached region within the condensates (\( I_{ref} \)), in pre- and post-bleaching images were measured in ImageJ for performing double normalization in bleached regions by:
+
+\[
+I_{Norm}(t) = \frac{I_{ROI}(t)}{I_{ROI}(t_0)} \times \frac{I_{ref}(t_0)}{I_{ref}(t)}
+\]
+
+(1)
+
+to quantify the recovery kinetics over time. Note that \( I(t_0) \) represents the intensity measured in the first image before bleaching.
+
+Data availability
+Additional supporting data are available from the corresponding author upon request. Source data are provided with this paper.
+
+Acknowledgments
+We would like to thank Dr. Siyu Song and Tao Xu for their helpful discussion about data analysis and plotting. M.X. acknowledges the support of the Alexander von Humboldt Foundation. W.C. acknowledges support from the Max Planck Graduate Center with the Johannes Gutenberg
+University of Mainz (MPGC), and the RTG 2516 “Structure Formation of Soft Matter at Interfaces”. This research was funded by the German Research Foundation (DFG) in the framework of the CRC 1552; Project No. 465145163. A.W. acknowledges funding from the Gutenberg Research Council Mainz underpinning his Life-Like Materials Program, the German Research Foundation grant WA 3084/19-1, the Max Planck Fellowship, and from the EU in the framework of the ERC Consolidator Grant to AW – M3ALI (101001638).
+
+Author contributions
+M.X. and A.W. conceived the project. M.X. and W.C. designed and performed all the experiments. M.R. helped with the initial transcription experiments. M.X. prepared the draft manuscript. M.X., W.C., and A.W. reviewed and edited the manuscript. A.W. supervised the project. M.X. and W.C. contributed equally.
+
+Competing interest
+The authors declare no competing interests.
+
+Additional information
+Supplementary Information is available for this paper.
+Correspondence and requests for materials should be addressed to Andreas Walther.
+
+Reference
+1. Lamond, A.I. & Earnshaw, W.C. Structure and function in the nucleus. Science **280**, 547-553 (1998).
+2. Sabari, B.R., Dall'Agnese, A. & Young, R.A. Biomolecular Condensates in the Nucleus. Trends Biochem. Sci. **45**, 961-977 (2020).
+3. Sawyer, I.A., Sturgill, D. & Dundr, M. Membraneless nuclear organelles and the search for phases within phases. Wiley Interdiscip. Rev. RNA **10**, e1514 (2019).
+4. Feric, M. et al. Coexisting Liquid Phases Underlie Nucleolar Subcompartments. Cell **165**, 1686-1697 (2016).
+5. Lyon, A.S., Peeples, W.B. & Rosen, M.K. A framework for understanding the functions of biomolecular condensates across scales. Nat. Rev. Mol. Cell Biol. **22**, 215-235 (2021).
+6. Lafontaine, D.L.J., Riback, J.A., Bascetin, R. & Brangwynne, C.P. The nucleolus as a multiphase liquid condensate. Nat. Rev. Mol. Cell Bio. **22**, 165-182 (2021).
+7. Solovei, I. et al. Nuclear architecture of rod photoreceptor cells adapts to vision in mammalian evolution. Cell **137**, 356-368 (2009).
+8. Jain, A. & Vale, R.D. RNA phase transitions in repeat expansion disorders. Nature **546**, 243-247 (2017).
+9. Patel, A. et al. A Liquid-to-Solid Phase Transition of the ALS Protein FUS Accelerated by Disease Mutation. Cell **162**, 1066-1077 (2015).
+10. Wu, J.H. et al. Self-Assembly of Intracellular Multivalent RNA Complexes Using Dimeric Corn and Beetroot Aptamers. J. Am. Chem. Soc. **144**, 5471-5477 (2022).
+11. Henninger, J.E. et al. RNA-Mediated Feedback Control of Transcriptional Condensates. Cell **184**, 207-225 (2021).
+12. Mao, Y.S., Sunwoo, H., Zhang, B. & Spector, D.L. Direct visualization of the co-transcriptional assembly of a nuclear body by noncoding RNAs. Nat. Cell Biol. 13, 95-101 (2011).
+13. Bevilacqua, P.C., Williams, A.M., Chou, H.-L. & Assmann, S.M. RNA multimerization as an organizing force for liquid–liquid phase separation. RNA 28, 16-26 (2022).
+14. Dai, Y., You, L. & Chilkoti, A. Engineering synthetic biomolecular condensates. Nat. Rev. Bioeng. 1, 466-480 (2023).
+15. Roden, C. & Gladfelter, A.S. RNA contributions to the form and function of biomolecular condensates. Nat. Rev. Mol. Cell Biol. 22, 183-195 (2021).
+16. Clever, J.L., Wong, M.L. & Parslow, T.G. Requirements for kissing-loop-mediated dimerization of human immunodeficiency virus RNA. J. Virol. 70, 5902-5908 (1996).
+17. Poudyal, R.R., Sieg, J.P., Portz, B., Keating, C.D. & Bevilacqua, P.C. RNA sequence and structure control assembly and function of RNA condensates. RNA 27, 1589-1601 (2021).
+18. Guo, P. The emerging field of RNA nanotechnology. Nat. Nanotechnol. 5, 833-842 (2010).
+19. Geary, C., Rothemund, P.W. & Andersen, E.S. A single-stranded architecture for cotranscriptional folding of RNA nanostructures. Science 345, 799-804 (2014).
+20. Udoño, H. et al. Programmable Computational RNA Droplets Assembled via Kissing-Loop Interaction. Acs Nano 18, 15477-15486 (2024).
+21. Fabrini, G. et al. Co-transcriptional production of programmable RNA condensates and synthetic organelles. Nat. Nanotechnol. 19, 1665-1673 (2024).
+22. Stewart, J.M. et al. Modular RNA motifs for orthogonal phase separated compartments. Nat. Commun. 15, 6244 (2024).
+23. Merindol, R., Loescher, S., Samanta, A. & Walther, A. Pathway-controlled formation of mesostructured all-DNA colloids and superstructures. Nat. Nanotechnol. 13, 730-738 (2018).
+24. Samanta, A., Sabatino, V., Ward, T.R. & Walther, A. Functional and morphological adaptation in DNA protocells via signal processing prompted by artificial metalloenzymes. Nat. Nanotechnol. 15, 914-921 (2020).
+25. Samanta, A., Horner, M., Liu, W., Weber, W. & Walther, A. Signal-processing and adaptive prototissue formation in metabolic DNA protocells. Nat. Commun. 13, 3968 (2022).
+26. Liu, W., Lupfer, C., Samanta, A., Sarkar, A. & Walther, A. Switchable Hydrophobic Pockets in DNA Protocells Enhance Chemical Conversion. J. Am. Chem. Soc. 145, 7090-7094 (2023).
+27. Samanta, A., Pellejero, L.B., Masukawa, M. & Walther, A. DNA-empowered synthetic cells as minimalistic life forms. Nat. Rev. Chem. 8, 454-470 (2024).
+28. Liu, W., Samanta, A., Deng, J., Akintayo, C.O. & Walther, A. Mechanistic Insights into the Phase Separation Behavior and Pathway-Directed Information Exchange in all-DNA Droplets. Angew. Chem. Int. Ed. 61, e202208951 (2022).
+29. Walther, A. et al. Growing functional artificial cytoskeletons in the viscoelastic confinement of DNA synthetic cells. Preprint at 10.21203/rs.3.rs-5048001/v1 (2024).
+30. Alberti, S., Gladfelter, A. & Mittag, T. Considerations and Challenges in Studying Liquid-Liquid Phase Separation and Biomolecular Condensates. Cell 176, 419-434 (2019).
+31. André, A.A.M. & Spruijt, E. Liquid-Liquid Phase Separation in Crowded Environments. Int. J. Mol. Sci. 21, 5908 (2020).
+32. Pei, G.F., Lyons, H., Li, P.L. & Sabari, B.R. Transcription regulation by biomolecular condensates. Nat. Rev. Mol. Cell Bio. (2024).
+33. Changiarath, A. et al. Promoter and Gene-Body RNA-Polymerase II co-exist in partial demixed condensates. Preprint at 10.1101/2024.03.16.585180 (2024).
+34. Delaney, C.E. et al. SETDB1-like MET-2 promotes transcriptional silencing and development independently of its H3K9me-associated catalytic activity. Nat. Struct. Mol. Biol. 29, 85-96 (2022).
+Supplementary Files
+
+This is a list of supplementary files associated with this preprint. Click to download.
+
+• SupplementaryInformation.pdf
+• MovieS1.mp4
+• MovieS2.mp4
+• MovieS3.mp4

0b483a72e0c58a9950db9f28aced8eb28f2145605ee9fc9e2c0f9dd377ab37bd/metadata.json

@@ -0,0 +1,171 @@
+{
+    "title": "Bayesian log-normal deconvolution for enhanced in silico microdissection of bulk gene expression data",
+    "pre_title": "BLADE: Bayesian Log-normAl DEconvolution for enhanced in silico microdissection of bulk gene expression data",
+    "journal": "Nature Communications",
+    "published": "20 October 2021",
+    "supplementary_0": [
+        {
+            "label": "Supplementary Information",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-021-26328-2/MediaObjects/41467_2021_26328_MOESM1_ESM.pdf"
+        },
+        {
+            "label": "Peer Review File",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-021-26328-2/MediaObjects/41467_2021_26328_MOESM2_ESM.pdf"
+        },
+        {
+            "label": "Description of Additional Supplementary Files",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-021-26328-2/MediaObjects/41467_2021_26328_MOESM3_ESM.pdf"
+        },
+        {
+            "label": "Supplementary Data 1",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-021-26328-2/MediaObjects/41467_2021_26328_MOESM4_ESM.xlsx"
+        },
+        {
+            "label": "Reporting Summary",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-021-26328-2/MediaObjects/41467_2021_26328_MOESM5_ESM.pdf"
+        }
+    ],
+    "supplementary_1": NaN,
+    "supplementary_2": NaN,
+    "source_data": [
+        "https://support.10xgenomics.com/single-cell-gene-expression/datasets/3.0.2/5k_pbmc_protein_v3",
+        "https://support.10xgenomics.com/single-cell-gene-expression/datasets/3.0.0/pbmc_10k_protein_v3",
+        "https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE107572",
+        "https://ngdc.cncb.ac.cn/gsa/browse/CRA001160"
+    ],
+    "code": [
+        "https://github.com/tgac-vumc/BLADE",
+        "/articles/s41467-021-26328-2#ref-CR41"
+    ],
+    "subject": [
+        "Computational models",
+        "Machine learning",
+        "Statistical methods"
+    ],
+    "license": "http://creativecommons.org/licenses/by/4.0/",
+    "preprint_pdf": "https://www.researchsquare.com/article/rs-123595/v1.pdf?c=1637613073000",
+    "research_square_link": "https://www.researchsquare.com//article/rs-123595/v1",
+    "nature_pdf": "https://www.nature.com/articles/s41467-021-26328-2.pdf",
+    "preprint_posted": "15 Dec, 2020",
+    "research_square_content": [
+        {
+            "section_name": "Abstract",
+            "section_text": "High-resolution deconvolution of bulk gene expression profiles is pivotal to characterize the complex cellular make-up of tissues, such as tumor microenvironment. Single-cell RNA-seq provides reliable prior knowledge for deconvolution, however, a comprehensive statistical model is required for efficient utilization due to the inherently variable nature of gene expression. We introduce BLADE (Bayesian Log-normAl Deconvolution), a comprehensive probabilistic framework to estimate both cellular make-up and gene expression profiles of each cell type in each sample. Unlike previous comprehensive statistical approaches, BLADE can handle >20 cell types thanks to the efficient variational inference. Throughout an intensive evaluation using >700 datasets, BLADE showed enhanced robustness against gene expression variability and better completeness than conventional methods, in particular to reconstruct gene expression profiles of each cell type. All-in-all, BLADE is a powerful tool to unravel heterogeneous cellular activity in complex biological systems based on standard bulk gene expression data.Computational Biologybulk gene expressionBayesian Log-normAl Deconvolution",
+            "section_image": []
+        },
+        {
+            "section_name": "Figures",
+            "section_text": "Figure 1Figure 2Figure 3Figure 4Figure 5Figure 6",
+            "section_image": [
+                "https://assets-eu.researchsquare.com/files/rs-123595/v1/7adbc1aab5268475b1ba7d9a.png",
+                "https://assets-eu.researchsquare.com/files/rs-123595/v1/5cd2e6e964499b70d53449b4.png",
+                "https://assets-eu.researchsquare.com/files/rs-123595/v1/2184027981b8b33fcea2a686.png",
+                "https://assets-eu.researchsquare.com/files/rs-123595/v1/6b010e1b760f816456abb2ec.png",
+                "https://assets-eu.researchsquare.com/files/rs-123595/v1/d0aa7027da0c559a8cf5e427.png",
+                "https://assets-eu.researchsquare.com/files/rs-123595/v1/fb4de1166172f918d5e3715b.png"
+            ]
+        },
+        {
+            "section_name": "Full Text",
+            "section_text": "Due to technical limitations, full-text HTML conversion of this manuscript could not be completed. However, the latest manuscript can be downloaded and  accessed as a PDF.",
+            "section_image": []
+        },
+        {
+            "section_name": "Additional Declarations",
+            "section_text": "There is NO Competing Interest.",
+            "section_image": []
+        },
+        {
+            "section_name": "Supplementary Files",
+            "section_text": "SupplementaryFigures.pdfSupplementaryNotes.pdf",
+            "section_image": []
+        }
+    ],
+    "nature_content": [
+        {
+            "section_name": "Abstract",
+            "section_text": "Deconvolution of bulk gene expression profiles into the cellular components is pivotal to portraying tissue\u2019s complex cellular make-up, such as the tumor microenvironment. However, the inherently variable nature of gene expression requires a comprehensive statistical model and reliable prior knowledge of individual cell types that can be obtained from single-cell RNA sequencing. We introduce BLADE (Bayesian Log-normAl Deconvolution), a unified Bayesian framework to estimate both cellular composition and gene expression profiles for each cell type. Unlike previous comprehensive statistical approaches, BLADE can handle > 20 types of cells due to the efficient variational inference. Throughout an intensive evaluation with > 700 simulated and real datasets, BLADE demonstrated enhanced robustness against gene expression variability and better completeness than conventional methods, in particular, to reconstruct gene expression profiles of each cell type. In summary, BLADE is a powerful tool to unravel heterogeneous cellular activity in complex biological systems from standard bulk gene expression data.",
+            "section_image": []
+        },
+        {
+            "section_name": "Introduction",
+            "section_text": "Over the past decade, gene expression profiling has been applied to elucidate the complexity of transcriptional regulation in diverse biological contexts, such as cancer1,2. Conventional gene expression profiling, either by RNA sequencing (RNA-seq) or microarrays, captures cumulative gene expression levels of many cells combined. Therefore, it is often referred to as bulk gene expression profiling to distinguish it from the recent single-cell gene expression profiling technologies3. In oncology, single-cell RNA sequencing (scRNA-seq) is employed to study cellular heterogeneity within a tumor, composed of malignant (tumor) and non-malignant cells4,5,6,7,8,9,10. However, scRNA-seq has severe limitations, including technical challenges such as drop-out11,12 and high cost, which hinder its application to large series and translation to clinical applications.\n\nSeveral computational deconvolution methods have been developed to predict cellular composition from bulk RNA-seq data by employing a signature of pre-determined cell type-specific gene expression profiles. Initially, these signatures were constructed by sorting each cell type followed by gene expression profiling13, whereas recent methods such as CIBERSORTx14 and MuSiC15 employed scRNA-seq data for this purpose. Most approaches perform linear regression to reconstruct the bulk gene expression profiles using the gene expression signatures, where the regression coefficients correspond to the cellular composition. However, the standard regression approach does not account for variability in gene expression within the same cell type and may render biased results.\n\nTo the best of our knowledge, no deconvolution method can adequately and efficiently account for the gene expression variability within the same cell type. Modeling gene expression variability is challenging specifically for deconvolution due to the incompatibility of the log-normalization16, which significantly stabilizes gene expression variability. Without the log-normalization (i.e., in linear-scale), gene expression data has a heavily skewed distribution, which is not adequately modeled by the standard linear regression approaches, such as non-negative least square (NNLS) used in EPIC17. Currently, few probabilistic deconvolution approaches take skewed variability into account. However, these methods handle only a restricted number of cell types due to optimization difficulties (e.g., three cell types in DeClust18 and Demix/DemixT19,20).\n\nRecently, several linear-regression deconvolution approaches have been introduced that consider gene expression variability. MuSiC is a variant of NNLS that prioritizes genes for deconvolution by their variability obtained from the multi-subject single-cell RNA-seq data. CIBERSORTx introduced a two-step approach to address variable cell-type-specific gene expression profiles across the samples: first estimate cellular fraction (deconvolution) and then reconstruct gene expression per cell type in each sample (purification). However, the purification step of CIBERSORTx can handle only a part of genes because of the underdetermination problem where too many parameters need to be inferred. In terms of cellular fraction estimation, both MuSiC and CIBERSORTx outperformed the standard linear regression methods, though they are also linear regression approaches.\n\nHere, we introduce BLADE (Bayesian Log-normAl DEconvolution), a Bayesian method that jointly performs deconvolution and purification in a single-step, taking into account prior knowledge of cell type-specific gene expression profiles obtained from scRNA-seq data. BLADE takes a Bayesian framework that integrates two signatures of mean and variability of gene expression per-cell type using a log-normal probability model. The unified probabilistic model for both deconvolution and purification of BLADE can leverage the prior knowledge for purification, which can remedy the underdetermination issue. Furthermore, an efficient variational inference algorithm was developed, for which we show that it can handle at least 20 cell types. Through a comprehensive evaluation based on more than 700 simulated and real bulk gene expression data sets, we demonstrate a robust performance of BLADE regardless of gene expression variability. In particular, BLADE achieves high accuracy and completeness in gene expression purification, underpinning the power of the unified Bayesian framework for both tasks.",
+            "section_image": []
+        },
+        {
+            "section_name": "Results",
+            "section_text": "We first assessed gene expression variability within a cell type using publicly available Peripheral Blood Mononuclear Cell (PBMC) CITE-seq (Cellular Indexing of Transcriptomes and Epitopes by Sequencing) data from 10x Genomics. Based on the integration and clustering analysis followed by phenotyping of 9439 cells, we identified fifteen immune cell types, among which nine are in common, with distinct cell-surface markers and gene expression profiles (Fig.\u00a01a; see \u201cMethods\u201d and Supplementary Figs.\u00a0S1\u20132). The size of cell populations ranges from 38 regulatory T cells (0.36%) to 2518 classical monocytes (24%). We then identified differentially expressed genes (DEGs) for each cell type. Subsequently, the standard deviation of gene expression levels per gene and per cell type was measured to assess gene expression variability among the same cell types. We identified high gene expression variability among the same cell populations, especially for DEGs without log-transformation (i.e., linear-scale; Fig.\u00a01b, c). The variability further increased when cells from the two scRNA-seq datasets were combined, indicating the presence of more variability between individuals (Fig.\u00a01d; P\u2009<\u20092.2\u2009\u00d7\u200910e\u221216 from a one-tailed paired t-test of within-sample and between-sample variability).\n\na t-SNE plots show the similarities in Pearson correlation coefficients among gene expression profiles of individual cells in two single-cell PBMC RNA-seq data, respectively, on the left and right. Cell type* is denoted by color. b, c Comparison of gene expression variability measured in standard deviation (y-axis) per gene and cell type pair in log-scale (b) and linear-scale (c) for both datasets (x-axis). The genes were split by differentially expressed genes (DEGs; n\u2009=\u20092876 gene and cell type pairs; red) and non-differentially expressed genes (non-DEGs; n\u2009=\u2009145,305 gene and cell type pairs; blue). The standard boxplot notation was used (lower/upper hinges\u2014first/third quartiles; whiskers extend from the hinges to the largest/lowest values no further than 1.5 * inter-quartile ranges). d. Comparison of within-sample (x-axis) and between-sample variability (y-axis) in gene expression levels per cell type, split by DEGs (n\u2009=\u20092876) and non-DEGs (n\u2009=\u2009145,305) per cell type. Standard deviation is measured for each gene and cell type first separately in two PBMC single-cell datasets followed by taking the average (x-axis), then also in merged PBMC data set (y-axis). Only the nine cell types commonly detected in two data sets were used in the analysis. *(CMCD4T: central memory CD4+ T cell; CMonocytes: classical monocytes; EMCD4T: effector memory CD4+ T cell; mDC: myeloid dendritic cell; MemoryB: memory B cell; MemoryCD8T: memory CD8+ T cell; NaiveB: naive B cell; NaiveCD4T: naive CD4+ T cell; NaiveCD8T: naive CD8+ T cell; NKcells: natural killer cell; NKT: natural killer T cell; Nmonocyte: non-classical monocyte; pDC: plasmacytoid dendritic cell; TRegs: regulatory T cell).\n\nTo properly account for variation in gene expression, we examined multiple probability distributions. We evaluated normal distribution, negative binomial distribution, and log-normal distribution to fit the expression level of each gene per cell type without log-normalization. The normal distribution is the standard variability model in many deconvolution algorithms, including CIBERSORTx14, EPIC17, and ABIS21, while the negative binomial distribution is frequently used for handling count data such as RNA-seq data22. Note that Poisson distribution was also introduced for modeling count data23,24, but it is a special case of negative binomial. The log-normal distribution is identical to the normal distribution but includes an exponential function, assuming gene expression data is normally distributed on a log scale but not on a linear scale. To evaluate the performance of these probability distributions on gene expression variability, we assessed (1) the maximum likelihood of fitting gene expression profiles and (2) the difference between estimated and empirical modes (i.e., the most probable gene expression level; Fig.\u00a02a\u2013c). The log-normal distribution, in general, shows the best performance in per-gene maximum likelihood, followed by the negative binomial and normal distributions (Fig.\u00a02a, c). In particular, we noted a biased fit of the normal distribution toward outlier observations, which led to low accuracy in identifying modes (Fig.\u00a02b; see four example genes with a biased fit with normal distribution in Fig.\u00a02d). In mode estimation, log-normal and negative binomial appears to be fairly competitive, where the log-normal had a somewhat worse median but a better third quartile (Fig.\u00a02b).\n\na A bar chart of average log-likelihood of the three types of distribution fitted to PBMC single-cell RNA-seq data. The genes were split by DEGs (red; n\u2009=\u20091723) and non-DEGs (blue; n\u2009=\u20091496). b Comparison of the distance of the estimated mode to the true mode (y-axis) per distribution type (x-axis). The standard boxplot notation was used (lower/upper hinges\u2014 first/third quartiles; whiskers extend from the hinges to the largest/lowest values no further than 1.5 * inter-quartile ranges). c Pairwise comparison of per-gene log-likelihood of log-normal distribution (y-axis) and that of normal (x-axis; top) and negative binomial distribution (x-axis; bottom). The genes were split into non-DEGs (left) and DEGs (right). d Density plots for raw-counts (red) and optimized log-normal (green), normal (blue), and negative binomial distribution (purple) for four example genes (gene name at the top) with low maximum log-likelihood for normal distribution. e, f Maximum log-likelihood values (e) and root mean squared error (root MSE: f) of each gene for log-normal (y-axis) and negative binomial (x-axis) convolutions of T\u2009=\u20098 cell types, applied to TCGA-MESO (left) and TCGA-SARC (right) data.\n\nWe further evaluated the performance of the log-normal and negative binomial distributions in the context of deconvolution. To this end, we constructed a generic statistical deconvolution method that can model gene expression profiles with various probabilistic assumptions given known cellular fractions. The method approximates the convolution of random variables with an arbitrary distribution using a probabilistic generating function, for which both negative binomial and log-normal random variables can be accurately approximated (see \u201cMethods\u201d, Supplementary Note\u00a01, and Supplementary Fig.\u00a0S3). Based on this method, we evaluated the performance of negative binomial and log-normal distribution in fitting the gene expression profiles per cell type using RNA-seq data from TCGA25. First, we obtained TCGA RNA-seq data of mesothelioma (TCGA-MESO; n\u2009=\u200984) and sarcoma (TCGA-SARC; n\u2009=\u2009256), from which we estimated the fraction of eight cell types using EPIC17, a deconvolution method previously applied to the TCGA. Then, we applied the flexible deconvolution method with two different probabilistic assumptions, log-normal and negative binomial, to estimate expression profiles per cell type of 200 random genes. In terms of log-likelihood and root mean square error (RMSE) measured per gene, log-normal and negative binomial deconvolutions performed equally well for most of the genes, except for a few genes (Fig.\u00a02e, f). Cumulatively, we concluded that the log-normal distribution is an attractive probabilistic distribution to model the gene expression variability of each cell type.\n\nWe constructed a Bayesian Log-normal Deconvolution method, BLADE, by emulating bulk gene expression profiles through convolution of gene expression profiles per cell type (Fig.\u00a03a). The bulk gene expression level of each gene \\(j\\) in sample \\(i\\) was modeled by \\({y}_{ij}={\\sum }_{t}{f}_{i}^{t}{x}_{ij}^{t}+{{\\in }}_{ij}\\). Here, the hidden variables \\({f}_{i}^{t}\\) and \\({{x}_{{ij}}}^{t}\\) denote the fraction of cell type \\(t\\) for sample \\(i\\) and the purified expression level of gene \\(j\\) of cell type \\(t\\) for sample \\(i\\). These hidden variables \\({f}_{i}^{t}\\) and \\({{x}_{{ij}}}^{t}\\) are, respectively, endowed with the Dirichlet distribution and the log-normal distribution. To incorporate prior knowledge from scRNA-seq data, we take a hierarchical approach to model \\({{x}_{{ij}}}^{t}\\) by taking a conjugate prior of log-normal distribution with hyperparameters \\({{\\mu }^{t}}_{0j},{{\\kappa }^{t}}_{0j},{{\\alpha }^{t}}_{0j}\\), and \\({{\\beta }^{t}}_{0j}\\) (Fig.\u00a03b). The hyperparameters are chosen based on the mean and standard deviation of each gene per cell type from the scRNA-seq data. By inferring the hidden variables, we can jointly estimate the fraction of cell types, captured by \\({f}^{t}_{i}\\), and purified gene expression profiles of each cell type in each sample, captured by \\({{x}^{t}}_{{ij}}\\). For inference, we employed a collapsed variational inference that maximize efficiency by integrating out a subset of hidden variables with a conjugate prior in advance. Furthermore, we employed the L-BFGS algorithm in conjunction with machine-code translated Python code for gradient and objective function calculations instead of native Python code. The compilation of native Python code by the Numba package26 significantly accelerates gradient and objective functions that are executed thousands of times during the L-BFGS optimization (Supplementary Fig.\u00a0S4). See \u201cMethods\u201d and Supplementary Note\u00a02 for further details of the framework. As a result, BLADE can handle many cell types (>20 cell types); unlike the previous log-normal-based deconvolution that can account for a maximum of three cell types20.\n\na To construct a prior knowledge for BLADE, we used CITE-seq data that contains gene expression and cell surface marker profiles of each cell. Cells are then subject to phenotyping, clustering, and differential gene expression analysis. Then, for each cell type, we retrieved average expression profiles (red cross and top heat map) and standard deviation per gene as the variability (blue circle and bottom heatmap). This prior knowledge is then used in the hierarchical Bayesian model (bottom right) to deconvolute bulk transcriptome profiles. b A graphical model of BLADE represents random variables, observed and hidden variables, respectively, in blue and gray nodes, and their dependency associations (arrows). See the text for the details of the model.\n\nWe assessed the robustness of BLADE, CIBERSORTx, and non-negative least squares (NNLS) against gene expression variability by applying them to model-based simulation data. The simulation data was created to have diverse but controlled variability levels of gene expression profiles (standard deviation of 0.1\u20131.5) as well as different numbers of cell types (5\u201320 cell types), marker genes (100\u20131000 genes), and samples (5\u2013100 samples; in total 700 training data sets). Note that NNLS is a regularized linear regression, a type of constrained linear regression used in many deconvolution methods, including MuSiC15, EPIC17, TIMER27, ABIS21, and also in the purification step of CIBERSORTx14. The simulation data variability levels were selected to recapitulate the observed range in the scRNA-seq data (up to standard deviation of 1.5 in log scale; Fig.\u00a01b, c). In general, all three methods could accurately estimate cellular fractions in case of a high number of genes, a low number of cell types, and a low variability level. In contrast, the performance decreased when a smaller number of genes are presented, and the number of cell types is increased (Fig.\u00a04a; Supplementary Figs.\u00a0S5\u20137). However, BLADE was the most robust against gene expression variability. In particular, in the range of observed expression variability of DEGs in the PBMC scRNA-seq data (on average standard deviation of > 0.5; Fig.\u00a01b), BLADE significantly outperformed CIBERSORTx and NNLS.\n\na Performances (Pearson correlation coefficient; y-axis) of BLADE (orange), CIBERSORTx (blue), and NNLS (dark red) to predict the cellular fraction of a subset of simulation data with ten cell types, 1000 genes, and various variability levels (standard deviation of 0.1\u20131.5; x-axis; n\u2009=\u200950 per variability level; five independent data set with ten cell types each). The standard boxplot notation was used (lower/upper hinges\u2014first/third quartiles; whiskers extend from the hinges to the largest/lowest values no further than 1.5 * inter-quartile ranges). b, c Performances (Pearson correlation coefficient; y-axis) of BLADE (orange) and CIBERSORTx (blue) to predict gene expression profiles per cell type for all samples jointly (group mode; b) and for each sample separately (high-resolution mode; c) using the same simulation data (n\u2009=\u200950 per variability level; five independent data set with ten cell types each). The standard boxplot notation was used. d Fractions of purified genes in the simulation data with two extreme levels of gene expression variability (left and right panels) by CIBERSORTx in group mode (top) and high-resolution mode (bottom). x- and y-axis represent the number of cell types and samples in the simulation data, respectively.\n\nWe then compared the performance of BLADE and CIBERSORTx in estimating gene expression profiles per cell type. In this comparison, NNLS is not included because of redundancy since the purification step of CIBERSORTx is based on NNLS. There are two modes of purification in CIBERSORTx, both of which were compared with BLADE: (1) estimating average profile per cell type across the samples (group mode purification), and (2) estimating the profile per cell type for each sample (high-resolution mode purification). For the data set with low variability levels, both BLADE and CIBERSORTx accurately reconstructed gene expression profiles per cell type (Fig.\u00a04b, c; Supplementary Figs.\u00a0S8\u20139). However, unlike BLADE, the performance of CIBERSORTx decreased rapidly as the RNA expression variability within a cell type increased. Furthermore, CIBERSORTx often excludes genes for purification, especially in high-resolution mode, when: (1) the number of cell types is larger than or equal to the number of samples, and (2) the variability in gene expression is high (Fig.\u00a04d; Supplementary Figs.\u00a0S10, S11). BLADE could accurately estimate the gene expression profiles of each cell type in both group mode and high-resolution mode, regardless of the number of cell types and samples, without any filtering (Fig.\u00a04b, c; Supplementary Figs.\u00a0S8\u20139).\n\nWe constructed realistic bulk gene expression data by in silico mixing the scRNA-seq data from PBMC samples without any model assumption to further evaluate our method. To this end, we randomly sample 100 cells 20 times from the 9439 cells from the two PBMC scRNA-seq data sets. We chose to use 100 cells since more cells get selected commonly in multiple samples as we sample more, making the simulated bulk gene expression data lose variability between the samples. In order to make the simulation data as realistic as possible, a cumulative sum of raw counts of 100 cells was obtained, followed by a standard normalization. The resulting simulation data recapitulate the gene expression variability of 15 cell types (Fig.\u00a05a; Supplementary Fig.\u00a0S12). We constructed signature matrices that capture the true mean and the standard deviation of 1007 genes selected and measured using all of 9439 cells (top 200 DEGs with FDR\u2009<\u20090.2 per cell type, combined). We also generated three extra data sets with a coarse classification of the 15 cell types by four (level 1; 441 genes selected), eight (level 2; 604 genes), and 12 cell types (level 3; 880 genes) in the same manner to diversify the difficulty levels for deconvolution (see Supplementary Data\u00a01 for the details of classifications). The increase of cell type often lowers the fraction of each cell type and the number of genes that can classify each cell type (Supplementary Figs.\u00a0S13\u201314). In particular, the fraction of T cells in level 1 is 0.47 on average, which gets much lower for their subtypes in level 4 (0.01 and 0.094 on average for naive and memory CD8+ T cells; Supplementary Fig.\u00a0S13). Furthermore, although more genes selected in the higher levels, there are 25 unique DEGs for T cells (DEGs only identified for T cells) in level 1, whereas there are only 16 and 3 unique DEGs for naive and memory CD8+ T cells in level 4 (Supplementary Fig.\u00a0S14). Collectively, deconvolution gets more challenging as the number of cell types increases from level 1 to level 4.\n\na A t-SNE plot represents the similarities in Pearson correlation coefficients among gene expression profiles of 15 cell types* (denoted by label) in 20 simulated bulk PBMC data. b Performances (Pearson correlation coefficient; y-axis) of BLADE (orange), CIBERSORTx (blue), NNLS (dark red), and MuSiC (light yellow) in predicting cellular fractions of the 20 simulated PBMC bulk RNA-seq data with diverse levels (n\u2009=\u20094, 7, 12, and 15 cell types, respectively, in levels 1\u20134; x-axis). The standard boxplot notation was used (lower/upper hinges \u2014first/third quartiles; whiskers extend from the hinges to the largest/lowest values no further than 1.5 * inter-quartile ranges). c Comparison of performance in estimating the cellular fractions per cell type of BLADE (y-axis) with CIBERSORTx, NNLS, and MuSiC (x-axis) at level 4. The fraction of each cell type is indicated by the size of the point. Pearson correlation coefficient and two-tailed test P-values are indicated at the top left in each panel. d Performance of BLADE (indicated by color) and its association to the number of unique DEGs per cell type (x-axis) and the respective fraction in the simulated data (y-axis). e Performance in Pearson correlation coefficient of BLADE (orange), CIBERSORTx (blue) for group mode purification of four levels of PBMC simulation data (n\u2009=\u20094, 7, 12, and 15 cell types, respectively, in levels 1\u20134; x-axis). The standard boxplot notation was used. f Performance (Pearson correlation coefficient; y-axis) of BLADE (orange) and CIBERSORTx (blue) in estimating gene expression profiles per cell type (x-axis) and per sample in level 4 (n\u2009=\u200920 samples per cell type; left). Fraction of genes in silico purified in high-resolution mode by CIBERSORTx at all levels of PBMC simulation data (n\u2009=\u200920 samples with 4, 7, 12, and 15 cell types, respectively, in levels 1\u20134; x-axis; right). The standard boxplot notation was used. *(CMCD4T: central memory CD4+ T cell; CMonocytes: classical monocytes; EMCD4T: effector memory CD4+ T cell; mDC: myeloid dendritic cell; MemoryB: memory B cell; MemoryCD8T: memory CD8+ T cell; NaiveB: naive B cell; NaiveCD4T: naive CD4+ T cell; NaiveCD8T: naive CD8+ T cell; NKcells: natural killer cell; NKT: natural killer T cell; Nmonocyte: non-classical monocyte; pDC: plasmacytoid dendritic cell; TRegs: regulatory T cell).\n\nUsing the bulk PBMC data generated above, we evaluated BLADE taking CIBERSORTx, NNLS, and also MuSiC as the baseline. We used the same list of genes and signatures for the baseline methods for a fair comparison. In general, the accuracy of estimated cell type fractions gets lower as the number of cell types gets higher, as expected (Fig.\u00a05b, see also Spearman correlation coefficients and RMSE in Supplementary Fig.\u00a0S15). All algorithms reached > 0.5 Pearson correlation coefficient for almost all cell types at level 1, where many cell types failed to reach as high performance as the number of cell types increased. Interestingly, the performance was sometimes higher in level 3 than level 2, especially for MuSiC, possibly because the advantage of having more genes overcomes the complexity due to the increased number of cell types (e.g., 880 genes in level 3, compared to 604 genes in level 2). At level 4, BLADE outperformed CIBERSORTx (P-value of 0.0087; a one-tailed paired t-test) and NNLS (P-value of 0.021; a one-tailed paired t-test) and performed comparably to MuSiC (P-value of 0.46; one-tailed paired t-test). The performance of the four methods are significantly correlated (P-value < 0.05 from Pearson correlation test), especially in pairs of MuSiC and BLADE (Pearson correlation coefficient\u2009=\u20090.82; P-value\u2009=\u20091.9e\u221204), and NNLS and CIBERSORTx (Pearson correlation coefficient\u2009=\u20090.87; P-value\u2009=\u20093.0e\u221205; Fig.\u00a05c; Supplementary Fig.\u00a0S16 for the comparison in the levels 1\u20133). Among the 15 cell types, plasmablasts, classical monocytes, natural killer (NK) cells were the best predicted by all four methods, which commonly failed to predict the composition of regulatory T cells (Tregs), naive CD8+ T cells (NaiveCD8T), and central memory CD4+ T cells (CMCD4T). These cell types are commonly low abundant (fraction of < 7% on average), and only a few unique DEGs were identified for each cell type (< 50 unique DEGs; Fig.\u00a05d; see Supplementary Fig.\u00a0S17 for other levels). In contrast, we noted a decent predictive performance of all methods for the abundant cell types (> 10%) with a high number of DEGs (> 50 unique DEGs).\n\nBLADE significantly outperformed CIBERSORTx in estimating gene expression profiles per cell type in both group mode and high-resolution mode across all the levels (Fig.\u00a05e, f and Supplementary Fig.\u00a0S18). For group mode purification, CIBERSORTx performed comparably to BLADE at level 1, which, however, gets lower at the higher level. Here, BLADE\u2019s performance was near-perfect, as expected, since BLADE integrates cell-type-specific gene expression profiles for purification (Fig.\u00a05e). CIBERSORTx did not estimate expression levels of most genes in high-resolution mode, and essentially no genes were purified for 11 cell types at level 4 (right panel of Fig.\u00a05f; Supplementary Fig.\u00a0S19). Furthermore, estimated expression profiles by CIBERSORTx are in general less accurate than BLADE in all levels, except for few cell types (e.g., central memory CD4+ T cells and naive CD4+ T cells at level 4; Fig.\u00a05f). The performance of BLADE in high-resolution mode purification is consistently accurate (> 0.7 Pearson correlation coefficient) across all cell types in all levels (Supplementary Fig.\u00a0S20). Cumulatively, Bayesian simultaneous deconvolution and in silico purification by BLADE significantly outperformed CIBERSORTx in reconstructing gene expression profiles per cell type.\n\nWe further challenged BLADE and other deconvolution algorithms using the standard bulk RNA-seq data of PBMC immune cell mixtures for which the composition of eight immune cell types was determined by flow cytometry28 (Fig.\u00a06a). Of these eight cell types, neutrophils were not identified in our PBMC scRNA-seq data. Furthermore, there are undetermined cells by the flow cytometry analysis that still contributed to the bulk RNA-seq data. Therefore, there is only limited prior knowledge available on cell-type-specific gene expression profiles, which is the case for most applications of deconvolution. We applied BLADE and other baseline methods using the gene expression signatures consisting of 532 genes that can distinguish seven cell types derived in the same manner as in the previous section (see Supplementary Data\u00a01 for the cell type classification). BLADE was able to reconstruct fractions of the seven cell types rather accurately, except for myeloid dendritic cells (mDC; Fig.\u00a06b, c and Supplementary Fig.\u00a0S21). We confirmed a low concordance of mDC signature compared to the previously determined signature using a large number of RNA-seq data28 (53 samples; Fig.\u00a06d). In fact, mDC signature has a higher correlation with previous B cell and monocyte signatures (Fig.\u00a06e), which makes the signatures less informative and the deconvolution extra challenging. Other baseline methods estimated compositions of monocytes accurately, but they failed to do the same for the majority of the other cell types including mDC (Fig.\u00a06b). In fact, they often failed to detect some cell types, particularly Tregs are commonly missed (Fig.\u00a06c). Instead, BLADE over and underestimated the fractions of Tregs and CD8+ T cells, respectively, by absorbing CD8+ T cell fractions to Tregs. However, BLADE was still able to rank samples accurately by their fractions. Cumulatively, BLADE was the most robust method for estimating cell type fractions when available prior knowledge was incomplete.\n\na Cell types fractions (y-axis) determined by flow cytometry in nine samples (x-axis). All cell types* have a color associated as shown in the legend. b Performances of BLADE (orange), CIBERSORTx (blue), NNLS (dark red) and MuSiC (light yellow), measured by Pearson correlation (y-axis) of the estimated sample-specific cell type (x-axis) fractions with those determined by flow cytometry. c Estimated cell fractions (y-axis) per sample (x-axis) by BLADE (top-left), NNLS (top-right), MuSiC (bottom-left) and CIBERSORTx (bottom-right). d, e Pearson correlation (y-axis in d and color gradient in e) of the signature per pair of cell types determined by Finotello et al. and two PBMC scRNA-seq data used in this study. *(TRegs: regulatory T cells; NKcells: natural killer cells; mDC: myeloid dendritic cells).\n\nWe further evaluated our method using scRNA-seq data from tumor samples. First, we obtained scRNA-seq data for 35 pancreas samples (CRP000653; Genome Sequence Archive), of which 24 are tumors while the other 11 are normal. The scRNA-seq data contains 57,530 cells classified into 10 cell types29 (Fig.\u00a07a; Supplementary Fig.\u00a0S22). For a fair evaluation of deconvolution algorithms, the 35 samples and their cells were split into auxiliary (six samples, of which four are tumors) and main samples (29 samples, of which 20 are tumors; Supplementary Fig.\u00a0S23). From the auxiliary samples, we obtained the mean and standard deviation of 818 genes that can classify ten cell types reliably (top 100 DEGs with FDR\u2009<\u20090.1 per cell types). For the main samples, we generated bulk gene expression profiles by calculating a cumulative sum of the raw count of all cells, followed by the standard log-normalization. For predicting the fraction of 10 cell types, MuSiC performed the best, followed by BLADE and CIBERSORTx (Fig.\u00a07b; see Spearman correlation coefficients and RMSE in Supplementary Fig.\u00a0S24). Interestingly, the performance of BLADE correlates the most with MuSiC (Pearson correlation coefficient of 0.62; P-value of 0.056), whereas it is less so with CIBERSORTx (Pearson correlation coefficient of 0.39; P-value of 0.27) and NNLS (Pearson correlation coefficient of \u22120.18; P-value of 0.62; Fig.\u00a07c). BLADE outperformed MuSiC for predicting the fraction of B cells but was worse for predicting endocrine cell fractions. Most cell types achieved high performance (> 0.5 of Pearson correlation coefficient) in all methods, except for B cells (in MuSiC and CIBERSORTx), T cells (in CIBERSORTx and NNLS), and Stellate cells (in NNLS). These cell types are often less dominant (< 5%) and with a small number of DEGs (less than 40 unique DEGs; Fig.\u00a07d). For reconstructing cell-type-specific gene expression profiles, both BLADE and CIBERSORTx achieved high performance for all cell types (> 0.8 of Pearson correlation coefficient; > 0.9 mostly for BLADE; Fig.\u00a07e, f). However, the purification by BLADE is without any filtering, unlike CIBERSORTx, which purified around 30% of genes per cell type on average in high-resolution mode (Fig.\u00a07g). Cumulatively, BLADE is a reliable deconvolution method especially to reconstruct cell-type-specific gene expression profiles in the tumor context.\n\na t-SNE plot showing the variability of the cell populations in the PDAC single-cell RNA-seq data. b Performances (Pearson correlation coefficient; y-axis) of BLADE (orange), CIBERSORTx (blue), NNLS (dark red), and MuSiC (light yellow) in predicting cellular fractions of the PDAC bulk RNA-seq data (n\u2009=\u200929 main samples). The standard boxplot notation was used (lower/upper hinges\u2014first/third quartiles; whiskers extend from the hinges to the largest/lowest values no further than 1.5 * inter-quartile ranges). c Comparison of performance in estimating the cellular fractions of each cell type of BLADE (y-axis) with CIBERSORTx (left), NNLS (middle), and MuSiC (right; x-axis). The fraction of each cell type is indicated by the point size. Pearson correlation coefficient and two-tailed test P-values are indicated at the top left in each panel. d The number of unique DEGs per cell type (x-axis) and the respective fraction in the PDAC data (y-axis). e Performance in Pearson correlation of BLADE (orange), CIBERSORTx (blue) for group mode purification (n\u2009=\u200910 cell types). The standard boxplot notation was used. f Performance (Pearson correlation coefficient; y-axis) of BLADE (orange) and CIBERSORTx (blue) in estimating gene expression profiles per cell type (x-axis) and per sample (n\u2009=\u200929 samples for each cell type; right). The standard boxplot notation was used. g Fraction of genes in silico purified in group mode (blue) and high-resolution mode (blue) by CIBERSORTx.",
+            "section_image": [
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-021-26328-2/MediaObjects/41467_2021_26328_Fig1_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-021-26328-2/MediaObjects/41467_2021_26328_Fig2_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-021-26328-2/MediaObjects/41467_2021_26328_Fig3_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-021-26328-2/MediaObjects/41467_2021_26328_Fig4_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-021-26328-2/MediaObjects/41467_2021_26328_Fig5_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-021-26328-2/MediaObjects/41467_2021_26328_Fig6_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-021-26328-2/MediaObjects/41467_2021_26328_Fig7_HTML.png"
+            ]
+        },
+        {
+            "section_name": "Discussion",
+            "section_text": "One of the major challenges in the deconvolution of bulk RNAseq data is the adequate and efficient handling of gene expression variability, mainly since stabilization of variability by log-normalization is inapplicable. Most of the previous algorithms implicitly or explicitly assumed normal distribution, as otherwise, the inference is highly challenging and limits the number of cell types that can be handled maximally (three cell types in Demix19). However, the normal distribution often renders a biased fit for gene expression variability (Fig.\u00a02a\u2013d), leading to a suboptimal outcome of deconvolution algorithms. Consequently, the performance of the standard regression technique, NNLS, was consistently inferior, especially when there is a realistic level of gene expression variability (Figs.\u00a04\u20137).\n\nCIBERSORTx and MuSiC are also linear-regression approaches that partially alleviate the issue by prioritizing genes for deconvolution. Support vector regression, the core algorithm of CIBERSORTx, depends on a subset of genes with high reconstruction errors. On the contrary, MuSiC explicitly learns gene weights from the single-cell RNA-seq data and prioritizes genes with low variability, for which the normal distribution can fit accurately due to the low skewness. We noted a consistently superior performance for fraction estimations of these algorithms over NNLS (Figs.\u00a04a, 5b, 6b, 7b). MuSiC outperformed BLADE in some cases, indicating normal distribution-based deconvolution can also be accurate when genes are prioritized based on the gene expression variability. However, performance of MuSiC compared unfavorably to that of BLADE when prior knowledge was incomplete (Fig.\u00a06). Furthermore, the strategy of prioritizing genes reduces the completeness of the purification results (Figs.\u00a04d, 5f, 7g). We observed a lower performance of linear regression-based purification by CIBERSORTx, particularly in high-resolution mode, which may be due to the inefficient variability model and a large number of variables to be estimated (Figs.\u00a04b, c, 5e, f, 7e, f).\n\nBLADE is a hierarchical Bayesian model that simultaneously performs deconvolution and estimation of gene expression profiles per cell type. The log-normal convolution model efficiently accounts for variability in gene expression and also for prior knowledge of gene expression profiles per cell type derived from scRNA-seq data (Fig.\u00a03). Notably, thanks to the unified probabilistic model used in BLADE, the prior knowledge contributes to both deconvolution and gene expression purification. This prior knowledge significantly reduces the search space of solutions for both tasks, which leads to enhanced accuracy and completeness, especially for gene expression purification. The efficient variational inference of BLADE allowed it to handle many cell types while accurately modeling the gene expression variability. Furthermore, the hierarchical approach of BLADE makes it robust against the quality of prior knowledge, as demonstrated in Fig.\u00a06. Finally, unlike MuSiC and CIBERSORTx, the Bayesian framework of BLADE provide the uncertainties of estimates, which may be valuable to evaluate the quality of the results and for further downstream analysis.\n\nEnhanced in silico microdissection by BLADE opens up the possibility to molecularly characterize individual cell types in tissue based on the standard RNA-seq data. For instance, we demonstrated that BLADE could be applied to estimate each cell type\u2019s gene expression profiles that make up the tumor microenvironment (TME). This allows us to characterize pathway activity in each immune cell type and possibly to recognize additional cell (sub-)types. Furthermore, BLADE can aid previously established gene expression subtypes (e.g., PDAC30,31) by characterizing the subtypes with distinct TME profiles. Finally, the detailed profiling of the TME, particularly immune TME profiles, may lead to a clinically applicable biomarker strategy for immunotherapy based on the standard bulk gene expression profiling. In conclusion, BLADE is a powerful tool that can significantly contribute to unravel cellular heterogeneity in complex biological systems.",
+            "section_image": []
+        },
+        {
+            "section_name": "Methods",
+            "section_text": "Two public peripheral blood mononuclear cell (PBMC) CITE-seq (cellular indexing of transcriptomes and epitopes by sequencing) datasets of healthy donors were downloaded from 10x Genomics datasets database [https://support.10xgenomics.com/single-cell-gene-expression/datasets/3.0.2/5k_pbmc_protein_v3] [https://support.10xgenomics.com/single-cell-gene-expression/datasets/3.0.0/pbmc_10k_protein_v3]. Genes and cells were filtered based on the following criterions: percentage of mitochondrial genes <10% and number of genes per cell between 200 and 4000. After the filtering, raw count data was normalized and scaled, using SCTransform, which performs normalization and variance stabilization using regularized negative binomial regression. Dimensionality reduction was done using principal component analysis (PCA) and t-distributed stochastic neighbor embedding (t-SNE). Following that, k-nearest neighbors (knn) of each cell using 25 dimensions of PCA were determined. This knn graph was used to construct the Shared Nearest Neighbor (SNN) graph by calculating the neighborhood overlap (Jaccard index) between every cell and its 20 nearest neighbors. Cluster determination was done by SNN graph modularity optimization based on the Louvain algorithm with the resolution of 1. Cells were phenotyped separately in both datasets, using primarily cell surface markers and then gene expression levels in case of lack of usable cell surface markers (Supplementary Figs.\u00a0S1\u20132). The two datasets were individually normalized, followed by selecting variable genes. The two data set were then integrated, and batch corrected using the common variable genes. The same analysis as described above was performed on the merged data set, including PCA, SNN, and cluster determination32. Finally, the top 200 differentially expressed genes per cell type were identified using a two-sided Wilcoxon Rank sum test by taking a contrast between one cell type versus the rest with an FDR cutoff of 0.2.\n\nTo evaluate log-normal, normal, and negative binomial distribution in fitting gene expression profiles, we retrieved raw counts per gene and per cell type and fit the three distribution types using the maximum-likelihood method available in fitdistrplus R package. For each cell type, genes with a standard deviation of lower than 0.5 were filtered out as they are mostly not expressed in that cell type. Finally, the log-likelihoods of the optimized distributions were obtained per gene and per cell type for comparison. As an alternative measure, we also identified the mode (i.e., the peak of the probability distribution) in each of the optimized distributions and assessed its accuracy by comparing it to the mode of the empirical distribution for each gene and cell type pair.\n\nFor a fair comparison of log-normal and negative binomial distribution for deconvolution, we developed a simple, generic maximum-likelihood-based convolution model. Formally it is assumed that there are \\(i=1,...,I\\) samples in which \\(t=1,...,T\\) cell types jointly contribute to expression profiles of \\(j=1,...,J\\) genes. For each sample \\(i\\) and gene \\(j\\), a bulk expression level is given, indicated by \\({y}_{{ij}}\\). Then, two hidden variables were introduced that jointly makeup \\({y}_{{ij}}\\): (1) expression level of the gene per cell type and sample, xtij; and (2) cellular composition for each cell type t, fti, where \\({\\forall} f^{t}_{i}\\ge 0\\) and \\({\\sum }_{t}f_{i}^{t}=1\\). An important strength of our method here is that it applies to any underlying parametric distribution for \\({{x}^{t}}_{{ij}}\\). \\({y}_{{ij}}\\) is a (weighted) convolution:\n\nwhich implies, with \\({{\\hat{x}}^{t}}_{{ij}}={f}^{t}_{i}{x}^{t}_{{ij}}\\),\n\nBy assuming \\({{x}^{t}}_{{ij}}\\) follows log-normal distribution (i.e., \\({x}^{t}_{{ij}}\\sim {LN}({{\\mu }^{t}_{j},({\\sigma}^{t}_{j})}^{2})\\)) and thus \\({\\hat{x}}^{t}_{{ij}} \\sim {LN}({\\mu }^{t}_{j}+{{{{{\\rm{log }}}}}}{{f}^{t}}_{i},({\\sigma }^{t}_{j})^{2})\\), \\({y}_{{ij}}\\) is a convolution of \\(T\\) log-normal random variables. The interest lies in estimating parameters \\({\\theta }_{j}=({\\mu }_{j}^{t},{\\sigma }_{j}^{t})\\) by maximum likelihood.\n\nWhile numerical evaluation of (2) may still be efficient for \\(T=2\\)28, however, the extension to \\(T \\; > \\; 2\\) is not straightforward to a \\(T-1\\) dimensional integral. To this end, the log-normal density \\({{g}_{t}=g}_{{{\\hat{x}}^{t}}_{{ij}}}\\) is approximated by a probability generating function (PGF). See Supplementary Note\u00a01 for the details of PGF approximation. The PGF-based approximation of \\({g}_{t}\\) showed higher accuracy than an alternative approximation method, Fenton-Wilkinson (FW) approximation33, which was also included as a benchmark (see Supplementary Note 1 and Supplementary Fig.\u00a0S3).\n\nThe aforementioned generic deconvolution was used to evaluate LN and NB for deconvolution. For this, two RNA-seq data sets are retrieved from The Cancer Genome Atlas (https://tcga-data.nci.nih.gov/tcga/) using TCGAbiolinks34. We considered all complete samples from the following tumor types: Mesothelioma (MESO35, n\u2009=\u200984; and Sarcoma (SARC36, n\u2009=\u2009256. We retrieved the upper quartile normalized RSEM (RNASeq by expectation-maximization) TPM (transcript per million) gene expression values (R package curatedTCGAData), merged replicated measurements (R package MultiAssayExperiment), and extracted the sample definitions from the barcodes (R package TCGAutils). We retained genes with mean count larger or equal to 5. For visualizing results, 200 genes were sampled randomly from this set37. The comparison procedure for LN and NB distributions is:\n\n1. Apply a non-statistical method, EPIC17, to estimate cell type fractions for bulk RNA-seq data using cell type-specific reference signatures. It has shown that EPIC provides a reliable estimate of cellular fractions of \\(T=8\\) cell types38, and it provides fractions that add up to 1.\n\n2. Fix the cellular fractions and fit generic deconvolution models with \\(T=8\\) LN or NB components using maximum likelihood.\n\n3. Compare the maximum likelihood values of the LN and NB models for of \\(J\\) genes.\n\nThe above procedure was done for 200 randomly selected genes with mean count per million larger or equal to 5 to exclude lowly expressed genes. Note that the comparison of the maximum likelihood values is fair, because the number of parameters used in the LN and NB components is the same, of \\(2T=16\\) per gene. As an alternative metric, we also measured the accuracy in reconstructing bulk gene expression levels based on deconvolution. Taking actual and predicted bulk gene expression level in LN or NB deconvolution model, root-mean-squared error (RMSE) was evaluated per gene and per model.\n\nBLADE is a hierarchical Bayesian model for log-normal convolution while accounting for the prior knowledge of per cell-type gene expression profiles (see Overview at Fig.\u00a03a). Formally, we assume \\({y}_{ij}={\\sum }_{t}{f}_{i}^{t}{x}_{ij}^{t}+{{\\in }}_{ij}\\), where \\({\\epsilon }_{{ij}}\\) is a log-normal error with mean parameter 0 and variance parameter \\({\\gamma }_{j}\\). Then, \\({{x}^{t}}_{{ij}}\\) follows a log-normal distribution: \\({{x}^{t}}_{{ij}}\\sim {LN}({{\\mu }^{t}}_{j},\\frac{1}{{\\lambda }_{j}})\\), where \\({{\\mu }^{t}}_{{j} }\\) and \\({{\\lambda }^{t}}_{j}\\) are expected value and precision in log-scale. Note that the parameters \\({{\\mu }^{t}}_{j}\\) and \\({{\\lambda }^{t}}_{j}\\) are shared across the samples. To incorporate prior knowledge on gene expression profiles per cell type, a hierarchical Bayesian approach was taken: \\({{\\mu }^{t}}_{j}\\) and \\({{\\lambda }^{t}}_{j}\\) are endowed with normal-gamma priors with hyperparameters \\({{{\\mu }^{t}}_{0j}} ,{{{\\kappa }^{t}}_{0j}} ,{{{\\alpha }^{t}}_{0j}}\\), and \\({{{\\beta }^{t}}_{0j}}\\): \\(({{\\mu }^{t}}_{j},{{\\lambda }^{t}}_{j})\\sim {NG}({{{\\mu }^{t}}_{0j}},{{{\\kappa }^{t}}_{0j}} ,{{{\\alpha }^{t}}_{0j}} ,{{{\\beta }^{t}}_{0j}} )\\). Note that the normal-gamma distribution is a conjugate prior of log-normal distribution, based on which marginal distribution of \\({{x}^{t}}_{{ij}}\\) given the hyperparameters \\({{{\\mu }^{t}}_{0j}} ,{{{\\kappa }^{t}}_{0j}} ,{{{\\alpha }^{t}}_{0j}}\\), and \\({{{\\beta }^{t}}_{0j}}\\) is analytically tractable. The other hidden variable, \\({{f}^{t}}_{i}\\), was endowed with Dirichlet distribution: \\(({{f}^{1}}_{i},...,{{f}^{T}}_{i})\\sim D({{\\alpha }^1}_{i},...,{{\\alpha }^T}_{i})\\).\n\nFor the inference, a collapsed variational inference was employed to handle analytically intractable posterior distribution of hidden variables given observed variables39. In the framework, the random variables with conjugate prior distribution, which are \\({{\\mu }^{t}}_{j}\\) and \\({{\\lambda }^{t}}_{j}\\), were integrated out, which allows us to find a fully Bayesian estimation of \\({{x}^{t}}_{{ij}}\\) instead of estimation of the single most probable \\({{\\mu }^{t}}_{j}\\) and \\({{\\lambda }^{t}}_{j}\\)39. By defining the variational distribution for the hidden variables, \\({{x}^{t}}_{{ij}}\\) and \\({f}_{i}^{t}\\), the objective function is to minimize the dissimilarity between the variational distribution and probability distribution, measured by Kullback-Leibler divergence (see Supplementary Note\u00a02 for the detailed derivation). The minimization was done by the Limited-memory Broyden-Fletcher-Goldfarb-Shanno (L-BFGS) algorithm available in SciPy Python library with the constraints of \\({{f}_{{i} }}^{t}\\)\u2009\u2265\u2009\\(0\\) and \\({\\sum }_{t}{f}_{i}^{t}=1\\). Numba-compiled objective function and gradients were used for the acceleration.\n\nBLADE has multiple hyperparameters for the hidden variables \\({{x}^{t}}_{{ij}}\\) and \\({{f}_{{i} }}^{t}\\), and also for observed variable \\({y}_{{ij}}\\). For \\({{f}_{{i} }}^{t}\\), a hyperparameter \\({{\\alpha }^{t}}_{i}\\) for Dirichlet distribution needs to be set. A user-defined value is assigned to \\({{\\alpha }^{t}}_{i}\\) for all \\(t\\) since we do not have prior information on cellular composition. For \\({y}_{{ij}}\\), we need to specify a precision of each gene, \\({\\gamma }_{j}\\), which we chose \\(\\frac{1}{{\\mathbb{V}}(\\,{{{{{\\rm{log }}}}}}{y}_{{ij}})s}\\), where \\(s\\) and \\({\\mathbb{V}}({{{{{\\rm{log }}}}}}{y}_{{ij}})\\) are a user-defined scale factor and a variance in log-scale measured per gene, respectively. For hyperparameters of \\({{x}^{t}}_{{ij}}\\), \\({{{\\mu }^{t}}_{0j}} ,{{{\\kappa }^{t}}_{0j}} ,{{{\\alpha }^{t}}_{0j}}\\), and \\({{{\\beta }^{t}}_{0j},}\\) we incorporated prior knowledge of gene expression profiles per cell type obtained from the scRNA-seq data. Given log-normal likelihood and normal-gamma priors, average expression level and standard deviation of \\({{x}^{t}}_{{ij}}\\) are: \\({\\mathbb{E}}({{{{{\\rm{log }}}}}}{{x}^{t}}_{{ij}})={{{\\mu }^{t}}_{0j}}\\) and \\({\\mathbb{V}}({{{{{\\rm{log }}}}}}{{x}^{t}}_{{ij}})={\\frac{{{\\beta }^{t}}_{0j}}{{{\\alpha }^{t}}_{0j}}}\\), respectively. To make use of the prior knowledge, we obtained the sample estimates of \\({\\mathbb{E}}({{{{{\\rm{log }}}}}}{{x}^{t}}_{{ij}})\\) and \\({\\mathbb{V}}({{{{{\\rm{log }}}}}}{{x}^{t}}_{{ij}})\\) from the scRNA-seq data, denoted by \\({{\\mu }^{t}}_{j}\\) and \\({({{\\sigma }^{t}}_{j})}^{2}\\). Then, we assigned \\({{{\\mu }^{t}}_{0j}={{\\mu }^{t}}_{j}}\\) whereas \\(\\,{{{\\alpha }^{t}}_{0j}}\\) is set by users followed by deriving: \\({{{\\beta }^{t}}_{0j}={\\alpha }^{t}}_{0j}{({{\\sigma }^{t}}_{j})}^{2}\\). Here, \\({{\\alpha }^{t}}_{0j}\\) allows to adapt to how much information the single cell data carries for the bulk RNA-seq data. The other hyperparameter \\({{{\\kappa }^{t}}_{0j}}\\) is also user-defined, which serve as a scale factor for variance of \\({{\\mu }^{t}}_{j}\\) (see also Supplementary Note\u00a02).\n\nAn empirical Bayes approach was employed to select the best set of user-defined parameters40. For each configuration of parameters, a maximum likelihood estimate of variational parameters is obtained using a subset of samples. Then, the hyperparameter configuration with the highest likelihood is selected, followed by performing deconvolution using the entire data set. Only a subset of samples is used in the empirical Bayes step, not only to gain computational efficiency but also to avoid overfitting. Throughout the manuscript, we considered a total of 90 different parameter configurations that cover all possible combinations of \\({{\\alpha }^{t}}_{i}\\in \\left\\{{{{{\\mathrm{1,10}}}}}\\right\\}\\), \\({{\\alpha }^{t}}_{0j}\\in \\left\\{0.1,0.5,1,5,10\\right\\}\\), \\({{\\kappa }^{t}}_{0j}\\in \\left\\{1,0.5,0.1\\right\\}\\), and \\(s\\in \\left\\{1,0.3,0.5\\right\\}\\).\n\nWe constructed simulation data sets of bulk gene expression profiles with known cellular fraction, gene expression profiles per cell type, and a diverse number of cell types and samples. To this end, given a number of cell types and genes, we first randomly sample an expected gene expression level \\({{\\mu }^{t}}_{j}\\) for gene \\(j\\) and cell type \\(t\\) from a normal distribution with 0 mean and standard deviation of 1.5: \\({{\\mu }^{t}}_{j}\\sim N(0,2)\\). Then, we sample gene expression levels per sample and per cell type, \\({{x}^{t}}_{{ij}}\\) from a log-normal distribution with mean \\({{\\mu }^{t}}_{j}\\) and standard deviation of \\(\\sigma\\) (\\({{x}^{t}}_{{ij}}\\sim {LN}({{\\mu }^{t}}_{j},\\sigma )\\)), where \\(\\sigma\\) is the parameter to control the variability in gene expression per cell type of each simulation data set. Fraction of cell types are sampled from a Dirichlet distribution with uninformative prior: \\({f}_{i}^{t}\\sim ({\\forall }_{i}{\\alpha }_{i}^{t})\\), where \\({\\alpha }_{i}^{t}=1\\). Then, the bulk gene expression profiles are generated by \\({y}_{ij}={\\sum }_{t}{f}_{i}^{t}{{x}_{ij}}^{t}\\). We constructed a total of 700 training data sets with the following settings: (1) number of samples\u2009=\u2009[5,10,20,50,100]; (2) number of genes\u2009=\u2009[100,200,500,1000]; (3) number of cell types\u2009=\u2009[2,3,5,10,20]; and (4) level of variability in gene expression profiles per cell type: \\(\\sigma\\)\u2009=\u2009[0.1,0.2,0.5,0.75,1,1.25,1.5].\n\nTo construct realistic simulation data, 20 bulk gene expression data sets were generated by randomly sampling and merging a subset of 9439 cells from the two PBMC scRNA-seq datasets. For each sample, the cellular fraction was first sampled from a Dirichlet distribution. The actual fractions of the 15 cell types were used as the parameter of the Dirichlet distribution so that the sampled fraction is similar to the total fraction. The fraction was then converted into the count of each cell type, with the following constraints: (1) the total number of cells is 100, and (2) the minimum number of cells per type is one. Then, the given number of cells were sampled with replacement, followed by obtaining the raw counts per cell type as the cumulative sum of raw counts of the sampled cells. Up to three distinct cells per type were allowed to be sampled since otherwise, gene expression variability was over-stabilized due to the averaging. Finally, the simulated bulk raw counts were obtained by taking the cumulative sum of the raw counts per cell type among 15 cell types. The bulk gene expression data was log-normalized using the Seurat package32.\n\nThe raw counts of RNAseq data and immune cell fractions determined by flow cytometry were obtained from the GEO databases with accession GSE107572. The raw counts was log-normalized using the Seurat package32.\n\nPDAC single-cell RNAseq data were obtained from the Genome Sequence Archive database under the accession code CRP00065329. A total of 57,530 cells from 35 pancreas samples (11 normal pancreas and 24 PDAC samples) were previously classified into ten cell types. For auxiliary data, we selected 17,266 cells (30% of cells) from six samples, of which two are normal and four are PDAC samples with the most cells. The rest of the 29 samples were used as the main data for evaluation. The signature genes were selected by the top 100 DEGs from each of the ten cell types (FDR\u2009<\u20090.1; 818 genes in total), followed by obtaining mean and standard deviation from the reference data. Note that we used more stringent criteria to select DEGs than for the PBMC data, because a sufficient number of DEGs (>500 DEGs) still satisfies these. For main data, a cumulative sum of the raw count of all cells was obtained from each sample. The standard log-normalization was then applied to the raw count. For the evaluation, the true cell type fractions and cell-type-specific gene expression profiles were obtained per main sample.\n\nThe original implementation of CIBERSORTx, NNLS, and MuSiC were obtained from https://cibersortx.stanford.edu/ (docker image), SciPy Python library, and https://github.com/xuranw/MuSiC (R package), respectively. For all four methods, the same set of genes were consistently used for a fair comparison. For the simulation data sets with the controlled gene expression variability level, true mean \\({{\\mu }^{t}}_{j}\\) and variability \\(\\sigma\\) per cell type of all genes were retrieved. For the PBMC and PDAC bulk transcriptome data, average and standard deviations of the union of DEGs of, respectively, 15 and 10 cell types were obtained from the scRNAseq data. These DEGs were selected using a FDR cutoff of 0.2 for PBMC data (in total 1007 genes) and a FDR cutoff of 0.1 for PDAC data (in total 818 genes). CIBERSORTx and NNLS require average expression profiles per gene and cell type, and BLADE requires both mean and standard deviation. MuSiC internally calculates the gene weight using the raw counts from scRNA-seq data, which was only available in PBMC and PDAC evaluation data set. The Pearson correlation coefficient, Spearman correlation coefficient, and root mean squared error (RMSE) were measured using the predicted and true fraction of each cell type across the samples to evaluate the deconvolution performance. Likewise, the Pearson correlation coefficient was measured between true and estimated gene expression profiles per cell type for group mode purification and per cell type and per sample for the high-resolution mode purification. The performance evaluation for purification was done only for CIBERSORTx and BLADE as NNLS and MuSiC only estimate cellular fractions.\n\nFurther information on research design is available in the\u00a0Nature Research Reporting Summary linked to this article.",
+            "section_image": []
+        },
+        {
+            "section_name": "Data availability",
+            "section_text": "The data used in this study is from public sources. The two PBMC CITE-seq datasets of healthy donors were downloaded from 10x Genomics datasets database [https://support.10xgenomics.com/single-cell-gene-expression/datasets/3.0.2/5k_pbmc_protein_v3] [https://support.10xgenomics.com/single-cell-gene-expression/datasets/3.0.0/pbmc_10k_protein_v3]. The TCGA data is retrievable using the TCGA-biolinks R package. The PBMC data is available from GEO under the accession code GSE107572. The single-cell RNA-seq data of the PDAC cohort is available from the Genome Sequence Archive under the accession code CRP000653.",
+            "section_image": []
+        },
+        {
+            "section_name": "Code availability",
+            "section_text": "BLADE python software along with a user-friendly demo is available and maintained at https://github.com/tgac-vumc/BLADE41.",
+            "section_image": []
+        },
+        {
+            "section_name": "References",
+            "section_text": "Angelova, M. et al. Characterization of the immunophenotypes and antigenomes of colorectal cancers reveals distinct tumor escape mechanisms and novel targets for immunotherapy. Genome Biol. 16, 1\u201317 (2015).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nTirosh, I. et al. Dissecting the multicellular ecosystem of metastatic melanoma by single-cell RNA-seq. Science 352, 189\u2013196 (2016).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nZheng, Z. et al. Single-cell transcriptomic analysis. Compr. Physiol. 10, 767\u2013783 (2020).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nPottier, C. et al. The importance of the tumor microenvironment in the therapeutic management of cancer. Expert Rev. Anticancer Ther. 15, 943\u2013954 (2015).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKumar, M. P. et al. Analysis of single-cell RNA-Seq identifies cell-cell communication associated with tumor characteristics. Cell Rep. 25, 1458\u20131468.e4 (2018).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nXiao, Z., Dai, Z. & Locasale, J. W. Metabolic landscape of the tumor microenvironment at single cell resolution. Nat. Commun. 10, 1\u201312 (2019).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSathe, A. et al. Single-cell genomic characterization reveals the cellular reprogramming of the gastric tumor microenvironment. Clin. Cancer Res. 26, 2640\u20132653 (2020).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nXiao, Z., Locasale, J. W. & Dai, Z. Metabolism in the tumor microenvironment: insights from single-cell analysis. Oncoimmunology 9, 1726556 (2020).\n\nLee, H. W. et al. Single-cell RNA sequencing reveals the tumor microenvironment and facilitates strategic choices to circumvent treatment failure in a chemorefractory bladder cancer patient. Genome Med. 12, 1\u201321 (2020).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGonz\u00e1lez-Silva, L., Quevedo, L. & Varela, I. Tumor functional heterogeneity unraveled by scRNA-seq technologies. Trends Cancer 6, 13\u201319 (2020).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nAngerer, P. et al. Single cells make big data: New challenges and opportunities in transcriptomics. Curr. Opin. Syst. Biol. 4, 85\u201391 (2017).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nL\u00e4hnemann, D. et al. Eleven grand challenges in single-cell data science. Genome Biol. 21. https://doi.org/10.1186/s13059-020-1926-6 (2020).\n\nNewman, A. M. et al. Robust enumeration of cell subsets from tissue expression profiles. Nat. Methods 12, 453\u2013457 (2015).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nNewman, A. M. et al. Determining cell type abundance and expression from bulk tissues with digital cytometry. Nat. Biotechnol. 37, 773\u2013782 (2019).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWang, X., Park, J., Susztak, K., Zhang, N. R. & Li, M. Bulk tissue cell type deconvolution with multi-subject single-cell expression reference. Nat. Commun. 10, 380 (2019).\n\nZhong, Y. & Liu, Z. Gene expression deconvolution in linear space. Nat. Methods 9, 8\u20139 (2012).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nRacle, J., de Jonge, K., Baumgaertner, P., Speiser, D. E. & Gfeller, D. Simultaneous enumeration of cancer and immune cell types from bulk tumor gene expression data. Elife 6, 1\u201325 (2017).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nAoto, Y. et al. DEclust: a statistical approach for obtaining differential expression profiles of multiple conditions. PLoS ONE 12, 1\u201315 (2017).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nAhn, J. et al. DeMix: deconvolution for mixed cancer transcriptomes using raw measured data. Bioinformatics 29, 1865\u20131871 (2013).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWang, Z. et al. Transcriptome deconvolution of heterogeneous tumor samples with immune infiltration. iScience 9, 451\u2013460 (2018).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMonaco, G. et al. RNA-Seq signatures normalized by mRNA abundance allow absolute deconvolution of human immune cell types. Cell Rep. 26, 1627\u20131640.e7 (2019).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nAnders, S. & Huber, W. Differential expression analysis for sequence count data. Genome Biol. 11, R106 (2010).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGr\u00fcn, D., Kester, L., & Van Oudenaarden, A. Validation of noise models for single-cell transcriptomics. Nat. Methods 11, 637\u2013640 (2014).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKlein, A. M. et al. Droplet barcoding for single-cell transcriptomics applied to embryonic stem cells. Cell 161, 1187\u20131201 (2015).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nTCGA Research Network: https://www.cancer.gov/tcga.\n\nLam, S. K., Pitrou, A. & Seibert, S. Numba: a LLVM-based Python JIT compiler. In Proc. Second Work. LLVM Compil. Infrastruct. HPC - LLVM \u201915 1\u20136. https://doi.org/10.1145/2833157.2833162 (2015).\n\nLi, T. et al. TIMER: a web server for comprehensive analysis of tumor-infiltrating immune cells. Cancer Res. 77, e108\u2013e110 (2017).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nFinotello, F. et al. Molecular and pharmacological modulators of the tumor immune contexture revealed by deconvolution of RNA-seq data. Genome Med. 11, 34 (2019).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nPeng, J. et al. Single-cell RNA-seq highlights intra-tumoral heterogeneity and malignant progression in pancreatic ductal adenocarcinoma. Cell Res. 29, 725\u2013738 (2019).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nDijk, F. et al. Unsupervised class discovery in pancreatic ductal adenocarcinoma reveals cell-intrinsic mesenchymal features and high concordance between existing classification systems. Sci. Rep. 10, 1\u201312 (2020).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBailey, P. et al. Genomic analyses identify molecular subtypes of pancreatic cancer. Nature 531, 47\u201352 (2016).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nStuart, T. et al. Comprehensive Integration of single-cell data. Cell 177, 1888\u20131902.e21 (2019).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nFenton, L. F. The sum of log-normal probability distributions in scatter transmission systems. IRE Trans. Commun. Syst. 8, 57\u201367 (1960).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nColaprico, A. et al. TCGAbiolinks: an R/Bioconductor package for integrative analysis of TCGA data. Nucleic Acids Res. 44, e71 (2016).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHmeljak, J. et al. Integrative molecular characterization of malignant pleural mesothelioma. Cancer Discov. 8, 1549\u20131565 (2018).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nAbeshouse, A. et al. Comprehensive and integrated genomic characterization of adult soft tissue sarcomas. Cell 171, 950\u2013965.e28 (2017).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nRauschenberger, A., Cioc\u0103nea-Teodorescu, I., Jonker, M. A., Menezes, R. X. & van de Wiel, M. A. Sparse classification with paired covariates. Adv. Data Anal. Classif. 14, 571\u2013588 (2020).\n\nArticle\u00a0\n    MathSciNet\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSturm, G. et al. Comprehensive evaluation of transcriptome-based cell-type quantification methods for immuno-oncology. Bioinformatics 35, i436\u2013i445 (2019).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nTeh, Y. W., Newman, D. & Welling, M. A collapsed variational Bayesian inference algorithm for latent Dirichlet allocation. Adv. Neural Inf. Process. Syst. 1353\u20131360. https://doi.org/10.7551/mitpress/7503.003.0174 (2007).\n\nCarlin, B. P. & Louis, T. A. Empirical bayes: past, present and future. J. Am. Stat. Assoc. 95, 1286\u20131289 (2000).\n\nArticle\u00a0\n    MathSciNet\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nAndrade Barbosa, B., van de Wiel, M. A. & Kim, Y. Bayesian log-normal deconvolution for enhanced in silico microdissection of bulk gene expression data. https://doi.org/10.5281/zenodo.5337105 (2021).\n\nDownload references",
+            "section_image": []
+        },
+        {
+            "section_name": "Acknowledgements",
+            "section_text": "The authors thank Kai Ruan for his careful review of the derivation of the BLADE algorithm. This project was supported by stichting Cancer Center Amsterdam (CCA2019-9-62).",
+            "section_image": []
+        },
+        {
+            "section_name": "Author information",
+            "section_text": "Department of Pathology, Cancer Center Amsterdam, Amsterdam UMC, Vrije Universiteit Amsterdam, Amsterdam, the Netherlands\n\nB\u00e1rbara Andrade Barbosa,\u00a0Saskia D. van Asten,\u00a0Bauke Ylstra\u00a0&\u00a0Yongsoo Kim\n\nDepartment of Molecular Cell Biology & Immunology, Amsterdam UMC, Vrije Universiteit Amsterdam, Amsterdam Infection and Immunity Institute, Amsterdam, the Netherlands\n\nSaskia D. van Asten\u00a0&\u00a0Juan J. Garcia Vallejo\n\nDepartment of Anatomy, School of Medicine, Kyungpook National University, Daegu, South Korea\n\nJi Won Oh\n\nBio-Medical Research Institute, Kyungpook National University Hospital, Daegu, South Korea\n\nJi Won Oh\n\nDepartment of Pathology, Amsterdam UMC, University of Amsterdam, Amsterdam, the Netherlands\n\nArantza Farina-Sarasqueta,\u00a0Joanne Verheij\u00a0&\u00a0Frederike Dijk\n\nDepartment of Medical Oncology, Cancer Center Amsterdam, Amsterdam UMC, University of Amsterdam, Amsterdam, the Netherlands\n\nHanneke W. M. van Laarhoven\n\nDepartment of Epidemiology and Data Science, Amsterdam Public Health research institute, Amsterdam UMC, Vrije Universiteit Amsterdam, Amsterdam, the Netherlands\n\nMark A. van de Wiel\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nY.K. and M.W. conceived the ideas, and designed the algorithm. Y.K. developed the python software. B.A.B. and S.D.A. analyzed PDAC single-cell RNAseq and PBMC CITE-seq data. B.A.B., S.D.A., and J.G.V. classified immune cell types in the CITE-seq data. Biological interpretation of the outcome is done by S.D.A., J.O., A.F.S., J.V., F.D., H.L. B.Y., and J.G.V. Evaluation of the algorithm performance is designed and performed by Y.K. and B.A.B. All authors discussed the results and contributed to the writing.\n\nCorrespondence to\n                Mark A. van de Wiel or Yongsoo Kim.",
+            "section_image": []
+        },
+        {
+            "section_name": "Ethics declarations",
+            "section_text": "The authors declare no competing interests.",
+            "section_image": []
+        },
+        {
+            "section_name": "Additional information",
+            "section_text": "Peer review information Nature Communications thanks the anonymous reviewers for their contribution to the peer review of this work. Peer reviewer reports are available.\n\nPublisher\u2019s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.",
+            "section_image": []
+        },
+        {
+            "section_name": "Rights and permissions",
+            "section_text": "Open Access  This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article\u2019s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article\u2019s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.\n\nReprints and permissions",
+            "section_image": []
+        },
+        {
+            "section_name": "About this article",
+            "section_text": "Andrade Barbosa, B., van Asten, S.D., Oh, J.W. et al. Bayesian log-normal deconvolution for enhanced in silico microdissection of bulk gene expression data.\n                    Nat Commun 12, 6106 (2021). https://doi.org/10.1038/s41467-021-26328-2\n\nDownload citation\n\nReceived: 07 December 2020\n\nAccepted: 27 September 2021\n\nPublished: 20 October 2021\n\nVersion of record: 20 October 2021\n\nDOI: https://doi.org/10.1038/s41467-021-26328-2\n\nAnyone you share the following link with will be able to read this content:\n\nSorry, a shareable link is not currently available for this article.\n\n\n\n\n                            Provided by the Springer Nature SharedIt content-sharing initiative\n                        ",
+            "section_image": [
+                "https://data:image/svg+xml;base64,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"
+            ]
+        },
+        {
+            "section_name": "This article is cited by",
+            "section_text": "Nature Communications (2025)\n\nGenome Biology (2024)\n\nBMC Genomics (2024)\n\nNature Communications (2022)",
+            "section_image": []
+        }
+    ]
+}

0b7608bc251c8024ae493244379344108add8b6da42c8277a7031cc03e69c3fc/peer_review/peer_review.md

@@ -0,0 +1,222 @@
+Peer Review File
+
+The negligible role of carbon offsetting in corporate climate strategies
+
+Corresponding Author: Mr Niklas Stolz
+
+This file contains all reviewer reports in order by version, followed by all author rebuttals in order by version.
+
+Version 0:
+
+Reviewer comments:
+
+Reviewer #1
+
+(Remarks to the Author)
+This study sets out to understand to what extent companies’ purchases of carbon credits impact their broader climate change strategy, for example, related to achieved emission reductions and future reduction targets. Although conflicting existing literature has both indicated (or hypothesized) a positive impact and a negative impact, the authors find no substantial impact, except potentially in the case of a handful of airlines (and other major purchasers of carbon credits). The research question is relevant and timely, the findings are interesting, the underlying methods appear appropriate and the study is well-written. I therefore support the publication of the study.
+
+I mainly have minor comments:
+• Page 1, abstract: consider replacing “we find no significant difference between companies that purchased credits and those that did not” by “we find no significant difference between the climate strategies of companies that purchased credits and those that did not.”
+• Figure 1 and its captions: for someone like me that is not very familiar with the OSL regression, it would be nice if you could “hold the readers’s hand” and explain basic things, like: 1) “positive values indicate slower decarbonisation over the study time” compared to companies that do not buy credits? 2) “positive values indicate more ambitious targets” compared to companies that do not buy credits? 3) what is the meaning of the “sectoral categorical variables” and the “geographic categorical variables”? (I do not find the explanation in Methods) 4) Why is “number of retired credits” indicated in red? 5) what is the meaning of emission share in mid-term target? (I do not find the explanation in Methods).
+• Page 8, line 9: with the wording “more meaningful decarbonization initiatives”, you appear to assume that carbon credits do not reduce emissions as efficiently as “investments in operations or value-chain decarbonisation”. However, from my understanding, your results do not say anything about, for example, the quantity of reduced emissions per dollar spent on carbon credits vs. “investments in operations or value-chain decarbonisation”? Check if an adjustment in wording is needed.
+• Page 8, line 30: Discussion section: consider comparing your findings to broader literature on “moral hazard” of market-based instruments in carbon accounting, for example “Ascuı, F., Brander, M., Cojoianu, T. and Li, Q., 2021. Moral hazard and the market-based method: Does using REAs affect corporate emissions performance?. In Academy of Management Proceedings (Vol. 2021, No. 1, p. 15686). Briarcliff Manor, NY 10510: Academy of Management.”.
+• Page 10, line 29: the section on “Climate Target Ambition” is clearly written and seems methodologically sound. However, I encourage the authors to reference existing literature that also has the purpose of establishing comparability between the emission reduction targets of different companies, potentially in the light of the Paris Agreement temperature ceiling, and then highlight and justify any differences that your approach might have. Here are a handful of key studies that I am aware of:
+o Dietz, S., Gardiner, D., Jahn, V. and Noels, J., 2023. Carbon Performance assessment of oil & gas producers: note on methodology. Transition Pathway Initiative Centre, London School of Economics and Political Science.
+o Bolay, A.F., Bjørn, A., Weber, O. and Margni, M., 2022. Prospective sectoral GHG benchmarks based on corporate climate mitigation targets. Journal of Cleaner Production, 376, p.134220.
+o Bjørn, A., Lloyd, S. and Matthews, D., 2021. From the Paris Agreement to corporate climate commitments: evaluation of seven methods for setting ‘science-based’ emission targets. Environmental Research Letters, 16(5), p.054019.
+
+(Remarks on code availability)
+Reviewer #2
+
+(Remarks to the Author)
+Manuscript Title: The negligible role of carbon offsetting in corporate climate strategies
+
+I express sincere gratitude to the editorial board for offering to review the captioned paper. The manuscript caters to a critical and timely research topic and presents robust and useful findings. The results suggest that carbon credit retirement does not affect the climate strategy of the sample firms as well as the internal decarbonization (in most cases). Provided the quality of the work, I would recommend authors consider the below points to enrich the manuscript value.
+
+• The introduction section fairly describes the need for this research. Yet, the authors should consider adding a paragraph of review of relevant past studies to further strengthen the contributions to the body of knowledge.
+
+• Under the Methods section, the author should describe the sample selection process leading to selection of 89 firms.
+
+• The findings have huge potential to provide critical policy implications for regulators. Authors should incorporate a section for the same and discuss.
+
+(Remarks on code availability)
+NA
+
+Version 1:
+
+Reviewer comments:
+
+Reviewer #1
+
+(Remarks to the Author)
+The authors have done an excellent job at responding to my original review comments, and the comments of the other reviewer. The paper has been substantially improved as a result and I have no further comments.
+
+(Remarks on code availability)
+
+Reviewer #2
+
+(Remarks to the Author)
+I have reviewed the revised manuscript and found that authors have sufficiently addressed the comments provided earlier. I recommend the article for publication.
+
+(Remarks on code availability)
+NA
+Open Access This Peer Review File is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.
+In cases where reviewers are anonymous, credit should be given to 'Anonymous Referee' and the source.
+The images or other third party material in this Peer Review File are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
+To view a copy of this license, visit https://creativecommons.org/licenses/by/4.0/.
+Point-by-point response to the referees
+
+Dear referees,
+Thank you for your constructive reviews on our submitted manuscript, "The negligible role of carbon offsetting in corporate climate strategies". We welcome the opportunity to submit a thoroughly revised version of our manuscript together with this point-by-point revision letter.
+
+We highly appreciate the extremely helpful feedback from the reviewers. In the following, we would like to summarise the most important changes that we incorporated during the revision. We are confident that these changes address all raised concerns. In addition, we strongly believe that they led to significant improvements.
+
+1. Refinement of integration in literature: As suggested by reviewer 2, we have improved the integration of our study into the existing literature. To achieve this, we added a paragraph in the Introduction that highlights key findings from previous studies on the relationship between corporate carbon management practices and environmental performance. This addition strengthens the foundation of our research and clarifies its contribution to the field.
+2. Discussion on moral hazard: Based on the suggestion of reviewer 1, we have refined our study to include a discussion around moral hazard associated with emission offsetting. To address this, we introduced the concept of moral hazard and relevant findings for other market-based carbon accounting instruments to the Introduction. Additionally, we incorporated a dedicated section in the Discussion that examines the potential risks associated with moral hazard.
+3. Improved robustness of Results: We ran additional tests to ensure the robustness of our regression results. After evaluating the influence of single observations on the regression analysis (Figure 1), we decided to exclude INPEX Corporation from the emission analysis and to re-run the regression on the change in scope 1 emissions (Fig. 1a) since it is an outlier that strongly influences the results. The conclusions of the analysis remain the same, and the messages are reinforced by additional robustness checks presented in the Supplementary Information section S1.
+4. Refinement of Methodology section: Based on the suggestions from Reviewers 1 and 2, we have refined the methodology section to improve its clarity and better integrate it with the existing literature. Specifically, we added a section detailing the selection of companies for the study and enhanced the section on quantifying corporate climate targets to make the connections to previous studies more explicit. Additionally, we revised the caption of Figure 1 to provide clearer guidance for readers.
+5. Recommendation for policymakers: As suggested by Reviewer 2 we added a section on recommendations for policymakers to the discussion. We highlight that policymakers should not view voluntary carbon markets as an alternative to compliance mechanisms. Further, our work strengthens the rationale of policies like the European Union's Green Claims Directive that prevent companies from making excessive environmental claims
+based on emission offsetting. Finally, we warn policymakers from expanding the use-case of voluntary carbon credits without improving market mechanisms.
+
+We believe our work on the role of carbon credits in corporate climate strategies is highly relevant to ensure effective Net Zero transitions. Given the prominent role of carbon credits in the public discourse on decarbonisation and some companies' communication on their climate strategies, we believe that our study will help to guide the debate on voluntary emission offsetting, which stakeholders from the industry currently dominate.
+
+Based on the reviewers' feedback, we have substantially improved the transparency of the method, the integration with the literature, and the implications of our findings. We are confident that the manuscript has improved considerably.
+
+Reviewer 1:
+
+Reviewer Comment:
+This study sets out to understand to what extent companies’ purchases of carbon credits impact their broader climate change strategy, for example, related to achieved emission reductions and future reduction targets. Although conflicting existing literature has both indicated (or hypothesized) a positive impact and a negative impact, the authors find no substantial impact, except potentially in the case of a handful of airlines (and other major purchasers of carbon credits). The research question is relevant and timely, the findings are interesting, the underlying methods appear appropriate and the study is well-written. I therefore support the publication of the study.
+
+Author Response:
+We thank the reviewer for the positive and constructive feedback. We indicate additions to the manuscript in blue and deleted text in red.
+
+Reviewer Comment 1:
+• Page 1, abstract: consider replacing “we find no significant difference between companies that purchased credits and those that did not” by “we find no significant difference between the climate strategies of companies that purchased credits and those that did not.”
+
+Author Response:
+We thank the reviewer for this suggestion. We changed the abstract accordingly (p. 1, l.9).
+
+Reviewer Comment 2:
+• Figure 1 and its captions: for someone like me that is not very familiar with the OSL regression, it would be nice if you could “hold the readers’s hand” and explain basic things, like: 1) “positive values indicate slower decarbonisation over the study time” compared to companies
+that do not buy credits? 2) “positive values indicate more ambitious targets” compared to companies that do not buy credits? 3) what is the meaning of the “sectoral categorical variables” and the “geographic categorical variables”? (I do not find the explanation in Methods) 4) Why is “number of retired credits” indicated in red? 5) what is the meaning of emission share in mid-term target? (I do not find the explanation in Methods).
+
+Author Response:
+We thank the reviewer for this helpful comment. We added clarification to the caption of Figure 1 (p. 3) and the Methods section (p. 11, l. 1-5; p. 12 l. 5-11). Additionally, we provide a brief explanation of the unclear points:
+(1) & (2): The regression coefficients reflect not only the number of carbon credits a company retired but also the influence of all other explanatory variables in the model. These coefficients indicate whether explanatory variables are significantly correlated with the outcome variables (a) change in scope 1 emissions over time and (b) climate target ambition. The coefficients do not compare companies that retire carbon credits to those that do not. Instead, they show whether an increase in the number of carbon credits retired (starting from zero for companies that do not retire credits) is significantly associated with the decarbonisation speed and the ambition of climate targets. We now make this explicit in the caption of Figure 1 by stating (p.3): “In (a), positive regression coefficients indicate a negative relationship between the explanatory variables (on y-axis) and decarbonisation speed (x-axis), suggesting that as the explanatory variables increase, we observe a decreased decarbonisation speed. In (b), positive regression coefficients indicate a positive relationship between the explanatory variables (y-axis) and climate target ambition (x-axis), suggesting that as the explanatory variables increase, we observe an increased climate target ambition. The sectoral categorical variables are relative to the aviation sector, and the geographic categorical variables are relative to headquarters in Asia.”
+
+(3): We clarified the meaning of sectoral variables in the Methods section (p. 11, l.4-9):
+
+“We convert the categorical variables for the sector (automobile, oil and gas, and airlines) and the continent of headquarters (Asia, Europe, Latin America, and North America) into binary indicator columns (i.e. one-hot encoding). That means in the regression, each categorical variable equals 1 if a company belongs to a specific sector or is headquartered in a particular region and 0 otherwise. To avoid multicollinearity, we excluded the sectoral category "Airlines" and the geographical category "Asia" from the regression. These omitted categories serve as the reference groups against which the effects of other categories are compared.”
+
+(4): We highlighted the number of retired carbon credits to help the reader identify the outcome variable of interest for the study. To avoid confusion, we changed the colour of the errorbar to blue and explicitly stated in the caption of Figure 1 (p. 3): “The label of retired carbon credits is written in bold as it represents the primary outcome variable of interest in the study.”
+(5): We clarified the reason for including the share of intermediate targets in the Methods section (p. 12, l. 10-14):
+    “We include the share of emissions covered by intermediate targets (i.e. emission targets that are no net zero targets) to avoid systematically favouring companies that set only long-term net zero targets without intermediate goals. For instance, a company with a net zero target for 2050 but no intermediate targets would appear to have higher target ambition than the exemplary company depicted in Figure 6, as the red area decreases without Target 1..”
+
+Reviewer Comment 3:
+• Page 8, line 9: with the wording “more meaningful decarbonization initiatives”, you appear to assume that carbon credits do not reduce emissions as efficiently as “investments in operations or value-chain decarbonisation”. However, from my understanding, your results do not say anything about, for example, the quantity of reduced emissions per dollar spent on carbon credits vs. “investments in operations or value-chain decarbonisation”? Check if an adjustment in wording is needed.
+
+Author Response:
+We thank the reviewer for this valuable input. We changed the wording according to the reviewer’s suggestion (p.8, l. 9) to “investments in operations or value chain decarbonisation”. Initially, we based our wording on recent findings by Probst et al. (2024) that less than 16% of carbon credits constitute real emission reductions. However, we do not evaluate this in our study, and based on our data and analysis, as you rightfully remarked, we cannot conclude that investments in internal and value-chain decarbonisation are more effective at reducing emissions per dollar spent than carbon credits.
+
+Reviewer Comment 4:
+• Page 8, line 30: Discussion section: consider comparing your findings to broader literature on “moral hazard” of market-based instruments in carbon accounting, for example “Ascui, F., Brander, M., Cojoianu, T. and Li, Q., 2021. Moral hazard and the market-based method: Does using REAs affect corporate emissions performance?. In Academy of Management Proceedings (Vol. 2021, No. 1, p. 15686). Briarcliff Manor, NY 10510: Academy of Management.”.
+
+Author Response:
+
+We thank the reviewer for this valuable addition. We added a section to introduce the moral hazard concept (p.1, l. 24-37) and the literature on the moral hazard of renewable energy certificates (p. 2, l. 9-15) to the Introduction. Further, we added a section about how moral hazard of carbon credit usage is linked to renewable energy certificates to the Discussion (p. 9, l. 43-51).
+We now write in the introduction (p.1, l. 34-38): “Purchasing carbon credits instead of pursuing – potentially more effective – internal decarbonisation can be conceptualised as a moral hazard. Moral hazard occurs when actors take on higher risks or engage in socially suboptimal behaviour because they are shielded from the consequences of their actions (Baker, 1996) . Therefore, emission offsetting might lead to moral hazard when companies neglect internal and value chain emission reductions because the improved public perception achieved through emission offsetting shields them from the risk of reputational damage, public scrutiny, or governmental regulation.” and (p. 2., l.10-17): “While research on companies' use of carbon credits is still nascent, research on renewable energy attributes (REAs), another market-based carbon accounting tool, is more advanced. REAs allow companies to verify and claim the purchase of renewable energy, directly reducing market-based scope 2 emissions under the Greenhouse Gas Protocol and Science-Based Targets initiative (SBTi)(The Greenhouse Gas Protocol, 2015) . Unlike voluntary carbon credits, REAs can be counted towards SBTi goals. Ascui et al. (Ascui et al., 2021) show that companies using REAs tend to increase their scope 1 and 2 emissions without improving energy efficiency compared to peers who do not use them, which indicates their potential to induce moral hazard (Ascui et al., 2021). Additionally, setting targets for market-based scope 2 emissions and achieving them by purchasing REAs might undermine the integrity of SBTi because these certificates do not lead to real emission reductions (Bjørn et al., 2022; Brander et al., 2018; Gillenwater et al., 2014).”
+
+In the Discussion, we added a paragraph on potential moral hazard, linking it to REAs. We now write (p.9, l.30-39): “The lack of significant correlation between the number of retired carbon credits and environmental performance suggests that voluntary carbon offsetting has historically not been associated with moral hazard. However, there is qualitative evidence that some oil and gas companies, like Shell, Eni, and Inpex, plan to integrate voluntary emission offsetting into their climate targets. This approach could result in moral hazard if companies purchase carbon credits that do not entail promised emission reductions rather than actively reduce their emissions. In such a scenario, the use of carbon credits could reflect the same type of moral hazard previously observed in the context of renewable energy attributes, where the claimed benefits fall short of real emission reductions. Lastly, some compliance pricing mechanisms, such as the Colombian and South African carbon taxes or the Korean and Californian ETSs, permit companies to use carbon credits sourced from voluntary carbon markets to meet part of their obligations (UNFCCC, 2023). Hence, if companies reduce their compliance emission costs by retiring carbon credits, this can be associated with moral hazard.”
+
+Reviewer Comment 5:
+• Page 10, line 29: the section on “Climate Target Ambition” is clearly written and seems methodologically sound. However, I encourage the authors to reference existing literature that also has the purpose of establishing comparability between the emission reduction targets of
+different companies, potentially in the light of the Paris Agreement temperature ceiling, and then highlight and justify any differences that your approach might have. Here are a handful of key studies that I am aware of:
+o Dietz, S., Gardiner, D., Jahn, V. and Noels, J., 2023. Carbon Performance assessment of oil & gas producers: note on methodology. Transition Pathway Initiative Centre, London School of Economics and Political Science.
+o Bolay, A.F., Bjørn, A., Weber, O. and Margni, M., 2022. Prospective sectoral GHG benchmarks based on corporate climate mitigation targets. Journal of Cleaner Production, 376, p.134220.
+o Bjørn, A., Lloyd, S. and Matthews, D., 2021. From the Paris Agreement to corporate climate commitments: evaluation of seven methods for setting ‘science-based’emission targets. Environmental Research Letters, 16(5), p.054019.
+
+Author Response:
+We thank the reviewer for raising this critical point. Our approach is very similar to those used in the studies listed by the reviewer. We significantly revised the subsection “Climate Target Ambition” of the Methods to more clearly discuss the similarities and differences between the approaches..
+
+The main difference between the approaches presented in the proposed literature and our analysis is how intensity targets are handled. When companies reported an intensity target in the CDP survey, Bjørn et al. (2021) used the expected change in absolute emissions that companies must report. However, we decided against using this metric since the data quality of this survey field is low, and some large companies like Shell did not report this number. The low data quality is especially concerning since the expected change in absolute emissions is not usually disclosed in other sources like websites or sustainability reports and can not be verified. Consequently, we decided to treat absolute and intensity targets equally and add a sensitivity check to the Supplementary Information (Table S5) to ensure the presence of intensity targets does not alter our main results (Table S6, p. 21). The sensitivity check shows that the presence of an intensity target is not significantly correlated with companies' climate target ambitions.
+
+The methodology section on quantifying climate target ambition now reads (p. 11, l. 11-44):
+
+“It is difficult to compare climate target ambition due to differences in scope, base years and target years. Therefore, corporate climate targets must be harmonised before they can be directly compared across companies (Bolay et al., 2022). Here we calculate the ambition for each emission scope (scope 1, 2, and 3) and then add the ambitions weighted with the relative importance of that scope for a specific industry (relative importance = avg. share of total emissions for scope n in industry X).
+We use the following assumptions and simplifications to construct the target emission trajectories between 2020 and 2050 in line with previous studies that compared climate targets across organisations:
+• Between intermediate targets, emission trajectories are linear (Bolay et al., 2022)
+• Emissions that are not covered by any target remain unchanged (Bolay et al., 2022; Simon Dietz et al., 2021)
+• We only consider company-wide targets, not product targets - e.g. when a company sells oil and gas but only has a target for their oil operation, we do not regard it since a reduction in oil emissions could be compensated by increased gas production. The only exception is when there are product-level targets for all main products of a company (e.g. separate targets for oil and gas operation). This assumption helps to avoid the difficulties of aggregating product-level emission-intensity targets for integrated energy companies that often sell different energy and non-energy products (Simon Dietz et al., 2021) by constructing a company-wide intensity metric.
+• In line with SBTi's net zero standard (SBTi, 2023), we accept both absolute and intensity targets. If, for the same year, intensity and absolute targets are given, we use the absolute target. While other studies also accept absolute and intensity targets (Bolay et al., 2022; Simon Dietz et al., 2021), in contrast to Bolay et al. (2022), we do not use the associated expected change in absolute emissions for intensity targets that companies need to report to CDP. Instead, we directly use the targeted change in emission intensity to construct a company's emission trajectory. The deviation from previous studies is due to low data quality for the expected change in absolute emissions for intensity targets. Often, it is unclear if companies correctly use positive and negative numbers to indicate expected increases or decreases in absolute emissions. Further, it is not possible to verify the data since this data is typically not reported in annual or sustainability reports. To verify that companies that use intensity targets do not systematically set more ambitious climate targets, we show in the Supplementary Information that the share of a company's target that is covered by intensity targets is not significantly correlated with its target ambition (Tab. S5) and, therefore, does not bias the result reported in the regression analysis.
+
+We assign an ambition score for each subtarget by comparing planned emission trajectories with emission reductions in line with the annual average of all 1.5 degree Celsius warming emission scenarios from IPCC's Sixth Assessment Report (Byers et al., 2022). Studies that evaluated climate targets compare target ambitions either to emission trajectories (Bjørn et al., 2021) or to targeted average annual change in emissions (Bolay et al., 2022; Simon Dietz et al., 2021). We choose to compare companies based on the targeted cumulative emissions until 2050 instead of targeted average annual changes in emissions since not only the average change in emissions but also the shape of the emission trajectory is directly linked to global warming. The choice of the reference emission trajectory is not relevant for the regression results since it is the same constant that is deducted from each observation and, hence, does not influence the regression coefficients. Further, we do not use different reference emission trajectories per sector or geography to avoid biasing our regression results. Instead, we explicitly control for geography and sector in the regression.
+Reviewer #2 (Remarks to the Author):
+
+Manuscript Title: The negligible role of carbon offsetting in corporate climate strategies
+
+Reviewer Comment:
+I express sincere gratitude to the editorial board for offering to review the captioned paper. The manuscript caters to a critical and timely research topic and presents robust and useful findings. The results suggest that carbon credit retirement does not affect the climate strategy of the sample firms as well as the internal decarbonization (in most cases). Provided the quality of the work, I would recommend authors consider the below points to enrich the manuscript value.
+
+Author Response:
+We thank the reviewer for the very positive and constructive feedback.
+
+Reviewer Comment:
+• The introduction section fairly describes the need for this research. Yet, the authors should consider adding a paragraph of review of relevant past studies to further strengthen the contributions to the body of knowledge.
+
+Author Response:
+We thank the reviewer for this constructive feedback. We added a paragraph on past studies on the relationship between corporate carbon management and emission performance and evidence around the influence of renewable energy certificates, another market-based carbon accounting tool, on corporate climate performance (p. 2, l. 1-17). We now write:
+“There is mixed evidence regarding the relationship between corporate carbon management practices and subsequent emission reductions. While some studies demonstrate a positive relationship between corporate carbon disclosure and emission reductions (Downar et al., 2021; Qian & Schaltegger, 2017), others find this link only among emission-intensive companies (Hsueh, 2019). Conversely, to our knowledge, no study to date has established a significant relationship between adopting reporting guidelines, such as the Global Reporting Initiative (GRI) and improved corporate emission performance (Belkhir et al., 2017; Haque & Ntim, 2020). Further, the impact of corporate climate strategies on emission reductions remains ambiguous. For example, there is no relationship between the mere presence of corporate climate targets and subsequent decarbonisation (Dahlmann et al., 2019), though more ambitious targets are associated with greater emission reductions (Dahlmann et al., 2019; Ioannou et al., 2016). Recent findings suggest that only a comprehensive mix of corporate climate instruments (e.g., absolute emission targets, internal carbon prices, value chain engagement) is linked to absolute emission reductions (Klaassen & Stoll, 2021)
+
+While research on companies' use of carbon credits is still nascent, research on renewable energy attributes (REAs), another market-based carbon accounting tool, is more advanced. REAs allow companies to verify and claim the purchase of renewable energy, directly reducing market-based scope 2 emissions under the Greenhouse Gas Protocol and Science-Based Targets initiative (SBTi) (The Greenhouse Gas Protocol, 2015). Unlike voluntary carbon credits, REAs can be counted towards SBTi goals. Ascui et al. show that companies using REAs tend to increase their scope 1 and 2 emissions without improving energy efficiency compared to peers who do not use them, which indicates their potential to induce moral hazard (Ascui et al., 2021). Additionally, setting targets for market-based scope 2 emissions and achieving them by purchasing REAs might undermine the integrity of SBTi because these certificates do not lead to real emission reductions (Bjørn et al., 2022; Brander et al., 2018; Gillenwater et al., 2014).”
+
+Reviewer Comment:
+• Under the Methods section, the author should describe the sample selection process leading to selection of 89 firms.
+
+Author Response:
+We thank the reviewer for this important clarification question. We added a paragraph to the methods outlining the sample selection (p. 10, l. 13-20):
+
+“The study includes 89 oil and gas, airlines, and automobile manufacturing companies. These sectors are well-suited for investigating the role of voluntary emission offsetting, as they are characterised by high emissions and include some of the largest emission offsetting companies. In contrast, other hard-to-decarbonise industries, such as steel, cement, and maritime shipping,
+rarely engage in voluntary offsetting (CDP, 2023). The sample includes all passenger airlines and automobile manufacturers that disclose their emissions to CDP. We selected the 40 companies with the highest scope 1 emissions for the oil and gas sector due to the large number of companies in the CDP dataset. PJSC Lukoil was excluded from the analysis due to significant structural changes in the Russian gas industry during the study period. Consequently, the final sample consists of 39 oil and gas companies, 27 passenger airlines, and 23 automobile manufacturers.”
+
+Reviewer Comment:
+• The findings have huge potential to provide critical policy implications for regulators. Authors should incorporate a section for the same and discuss.
+
+Author Response:
+
+We thank the reviewer for this valuable addition. We have added policy recommendations to the Discussion (p. 9, l. 40-46). We now state:
+“Our findings entail several recommendations for policymakers. Voluntary emission offsetting is not associated with positive corporate environmental performance or high costs for companies. Therefore, it is not a reliable alternative to regulatory measures, such as compliance carbon pricing. Policymakers should, therefore, focus on strengthening regulatory mechanisms to ensure substantial corporate contributions to emission reductions. Moreover, companies in emission-intensive sectors typically offset only small portions of their total Scope 1, 2, and 3 emissions and allocate minimal funds to purchasing carbon credits. This highlights the importance of policies such as the European Union’s Green Claims Directive (Green Claims Directive, 2023), which aims to prevent companies from making excessive environmental claims.”
+References
+
+Ascui, F., Brander, M., Cojoianu, T., & Li, Q. (2021). Moral hazard and the market-based method: Does using REAs affect corporate emissions performance? Academy of Management Proceedings, 2021(1), 15686. https://doi.org/10.5465/AMBPP.2021.15686abstract
+
+Baker, T. (1996). On the Genealogy of Moral Hazard. Texas Law Review, 75(2), 237–292.
+
+Belkhir, L., Bernard, S., & Abdelgadir, S. (2017). Does GRI reporting impact environmental sustainability? A cross-industry analysis of CO2 emissions performance between GRI-reporting and non-reporting companies. Management of Environmental Quality: An International Journal, 28(2), 138–155. https://doi.org/10.1108/MEQ-10-2015-0191
+
+Bjørn, A., Lloyd, S. M., Brander, M., & Matthews, H. D. (2022). Renewable energy certificates threaten the integrity of corporate science-based targets. Nature Climate Change, 12(6), 539–546. https://doi.org/10.1038/s41558-022-01379-5
+
+Bjørn, A., Lloyd, S., & Matthews, D. (2021). From the Paris Agreement to corporate climate commitments: Evaluation of seven methods for setting ‘science-based’ emission targets. Environmental Research Letters, 16(5), 054019. https://doi.org/10.1088/1748-9326/abe57b
+
+Bolay, A.-F., Bjørn, A., Weber, O., & Margni, M. (2022). Prospective sectoral GHG benchmarks based on corporate climate mitigation targets. Journal of Cleaner Production, 376, 134220. https://doi.org/10.1016/j.jclepro.2022.134220
+
+Brander, M., Gillenwater, M., & Ascui, F. (2018). Creative accounting: A critical perspective on the market-based method for reporting purchased electricity (scope 2) emissions. Energy Policy, 112, 29–33. https://doi.org/10.1016/j.enpol.2017.09.051
+
+Byers, E., Krey, V., Kriegler, E., Riahi, K., Schaeffer, R., Kikstra, J., Lamboll, R., Nicholls, Z., Sandstad, M., Smith, C., van der Wijst, K., Al -Khourdajie, A., Lecocq, F., Portugal-Pereira, J., Saheb, Y., Stromman, A., Winkler, H., Auer, C., Brutschin, E., ... van Vuuren, D. (2022). AR6 Scenarios Database (Version 1.1) [Dataset]. Intergovernmental Panel on Climate Change. https://doi.org/10.5281/zenodo.7197970
+
+Dahlmann, F., Branicki, L., & Brammer, S. (2019). Managing Carbon Aspirations: The Influence of Corporate Climate Change Targets on Environmental Performance. Journal of Business Ethics, 158(1), 1–24. https://doi.org/10.1007/s10551-017-3731-z
+
+Downar, B., Ernstberger, J., Reichelstein, S., Schwenen, S., & Zaklan, A. (2021). The impact of carbon disclosure mandates on emissions and financial operating performance. Review of Accounting Studies, 26(3), 1137–1175. https://doi.org/10.1007/s11142-021-09611-x
+
+Gillenwater, M., Lu, X., & Fischlein, M. (2014). Additionality of wind energy investments in the U.S. voluntary green power market. Renewable Energy, 63, 452–457. https://doi.org/10.1016/j.renene.2013.10.003
+
+Haque, F., & Ntim, C. G. (2020). Executive Compensation, Sustainable Compensation Policy, Carbon Performance and Market Value. British Journal of Management, 31(3), 525–546. https://doi.org/10.1111/1467-8551.12395
+
+Hsueh, L. (2019). Voluntary climate action and credible regulatory threat: Evidence from the carbon disclosure project. Journal of Regulatory Economics, 56(2), 188–225. https://doi.org/10.1007/s11149-019-09390-z
+Ioannou, I., Li, S. X., & Serafeim, G. (2016). The Effect of Target Difficulty on Target Completion: The Case of Reducing Carbon Emissions. The Accounting Review, 91(5), 1467–1492. https://doi.org/10.2308/accr-51307
+Klaassen, L., & Stoll, C. (2021). Harmonizing corporate carbon footprints. Nature Communications, 12(1), Article 1. https://doi.org/10.1038/s41467-021-26349-x
+Proposal for a DIRECTIVE OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL on Substantiation and Communication of Explicit Environmental Claims (Green Claims Directive), No. COM(2023) 166, European Comission (2023).
+Qian, W., & Schaltegger, S. (2017). Revisiting carbon disclosure and performance: Legitimacy and management views. The British Accounting Review, 49(4), 365–379. https://doi.org/10.1016/j.bar.2017.05.005
+SBTi. (2023). SBTi Corporate Net-Zero Standard v1.1 (Version 1.1).
+Simon Dietz, Dan Gardiner, Nikolaus Hastreiter, Valentin Jahn, & Jolien Noels. (2021). Carbon Performance assessment of oil & gas producers: Note on methodology. Transition Pathway Initiative. https://transitionpathwayinitiative.org/publications/uploads/2021-carbon-performance-assessment-of-oil-gas-producers-note-on-methodology
+The Greenhouse Gas Protocol—A Corporate Accounting and Reporting Standard. (2015).
+UNFCCC. (2023). Regional Climate Week Asia-Pacific—The role of carbon credits. https://unfccc.int/sites/default/files/resource/Session%201d_The%20role%20of%20carbon%20credits.pdf

0b813482168c9364c8a5829646c9bed96c4d8b16dc8516b90be2106af51fc4ec/metadata.json

@@ -0,0 +1,166 @@
+{
+    "title": "Genomic signatures of pre-resistance in Mycobacterium tuberculosis",
+    "pre_title": "Genomic Signatures of Pre-Resistance in Mycobacterium tuberculosis",
+    "journal": "Nature Communications",
+    "published": "15 December 2021",
+    "supplementary_0": [
+        {
+            "label": "Supplementary Information",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-021-27616-7/MediaObjects/41467_2021_27616_MOESM1_ESM.pdf"
+        },
+        {
+            "label": "Peer Review File",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-021-27616-7/MediaObjects/41467_2021_27616_MOESM2_ESM.pdf"
+        },
+        {
+            "label": "Description of Additional Supplementary Files",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-021-27616-7/MediaObjects/41467_2021_27616_MOESM3_ESM.pdf"
+        },
+        {
+            "label": "Supplementary Data 1",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-021-27616-7/MediaObjects/41467_2021_27616_MOESM4_ESM.csv"
+        },
+        {
+            "label": "Supplementary Data 2",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-021-27616-7/MediaObjects/41467_2021_27616_MOESM5_ESM.csv"
+        },
+        {
+            "label": "Supplementary Data 3",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-021-27616-7/MediaObjects/41467_2021_27616_MOESM6_ESM.csv"
+        },
+        {
+            "label": "Reporting Summary",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-021-27616-7/MediaObjects/41467_2021_27616_MOESM7_ESM.pdf"
+        }
+    ],
+    "supplementary_1": NaN,
+    "supplementary_2": NaN,
+    "source_data": [
+        "/articles/s41467-021-27616-7#MOESM5",
+        "/articles/s41467-021-27616-7#MOESM6"
+    ],
+    "code": [
+        "https://github.com/arturotorreso/mtb_pre-resistance.git"
+    ],
+    "subject": [
+        "Antimicrobial resistance",
+        "Bacterial genetics",
+        "Genome-wide association studies",
+        "Phylogenetics",
+        "Tuberculosis"
+    ],
+    "license": "http://creativecommons.org/licenses/by/4.0/",
+    "preprint_pdf": "https://www.researchsquare.com/article/rs-364747/v1.pdf?c=1646868071000",
+    "research_square_link": "https://www.researchsquare.com//article/rs-364747/v1",
+    "nature_pdf": "https://www.nature.com/articles/s41467-021-27616-7.pdf",
+    "preprint_posted": "25 May, 2021",
+    "research_square_content": [
+        {
+            "section_name": "Abstract",
+            "section_text": "Recent advances in bacterial whole-genome sequencing have resulted in the identification of a comprehensive catalogue of genomic signatures of antibiotic resistance in Mycobacterium tuberculosis. With a view to pre-empting the emergence of drug-resistance, we hypothesized that pre-existing balanced polymorphisms in drug susceptible genotypes (\u201cpre-resistance mutations\u201d) could increase the risk of acquiring antimicrobial resistance in the future.\r\nIn order to identify a pathogen genomic signature of future drug resistance we undertook whole-genome sequencing on 3135 culture positive isolates from different patients sampled over a 17-year period in Lima, Peru.\r\nReconstructing ancestral whole genomes on time-calibrated phylogenetic trees we identified no single drug resistance in Peru predating 1940. Moving forward in evolutionary time through the phylogenetic tree from 1940, we apply a novel genome-wide survival analysis to determine the hazard of drug resistance acquisition at the level of lineage, mono-resistance state, and single-nucleotide polymorphism. We demonstrate that lineage 2 has a significantly higher incidence of drug resistance acquisition than lineage 4 (HR 3.36, 95% CI 2.10 - 5.38,p-value =4.25\u00d710-7) and estimate that the hazard of evolving rifampicin following isoniazid resistance acquisition is 14 times that of genomes with a susceptible background (HR 14.45,95% CI 8.46 - 15.50, p-value<10\u221215). Our findings are validated in a separate publicly available dataset from Samara, Russia. After controlling for population structure, we also show that a deletion in a gene coding for the cell surface protein lppP predisposes to the acquisition of drug resistance in susceptible genotypes (HR 6.71, 95% CI4.82-11.22, p-value =1.17\u00d710\u22129).\r\nPrediction of future drug resistance in susceptible pathogens together with targeted expanded therapy has the potential to prevent drug resistance emergence in Mycobacterium tuberculosis and other pathogens. Prospective cohort studies of participants with and with-out these polymorphisms should be undertaken with a view to implementing personalized pathogen genomic therapy. This approach could be employed to preempt and prevent the emergence of drug resistance and other important traits in other organisms.General MicrobiologyTuberculosisphylogeneticsdrug resistanceAMR",
+            "section_image": []
+        },
+        {
+            "section_name": "Additional Declarations",
+            "section_text": "There is NO Competing Interest.",
+            "section_image": []
+        },
+        {
+            "section_name": "Supplementary Files",
+            "section_text": "suppMaterialpreresistancearticle.pdf",
+            "section_image": []
+        }
+    ],
+    "nature_content": [
+        {
+            "section_name": "Abstract",
+            "section_text": "Recent advances in bacterial whole-genome sequencing have resulted in a comprehensive catalog of antibiotic resistance genomic signatures in Mycobacterium tuberculosis. With a view to pre-empt the emergence of resistance, we hypothesized that pre-existing polymorphisms in susceptible genotypes (pre-resistance mutations) could increase the risk of becoming resistant in the future. We sequenced whole genomes from 3135 isolates sampled over a 17-year period. After reconstructing ancestral genomes on time-calibrated phylogenetic trees, we developed and applied a genome-wide survival analysis to determine the hazard of resistance acquisition. We demonstrate that M. tuberculosis lineage 2 has a higher risk of acquiring resistance than lineage 4, and estimate a higher hazard of rifampicin resistance evolution following isoniazid mono-resistance. Furthermore, we describe loci and genomic polymorphisms associated with a higher risk of resistance acquisition. Identifying markers of future antibiotic resistance could enable targeted therapy to prevent resistance emergence in M. tuberculosis and other pathogens.",
+            "section_image": []
+        },
+        {
+            "section_name": "Introduction",
+            "section_text": "Mycobacterium tuberculosis is estimated to have killed 1 billion people over the last 200 years1 and remains one of the world\u2019s most deadly pathogens2. Drug resistance in bacteria, particularly the Enterobacteriaceae and Mycobacterium tuberculosis, imposes an unsustainable burden on health programs worldwide with some strains so extensively resistant that they are untreatable with existing antibiotic therapy3. Although recent advances in bacterial whole-genome sequencing have significantly improved the identification of drug resistance4, post hoc approaches to diagnosis miss the opportunity to preempt the emergence of drug resistance and implement preventive measures prior to the acquisition and spread of antibiotic resistant disease.\n\nAn increased risk of drug resistance emergence is often attributed to inadequate implementation of control measures5, but bacterial factors have also been proposed as potential contributors to drug resistance6. Evidence of differential drug resistance acquisition at the M. tuberculosis sublineage level is conflicting. Epidemiological and in vitro studies have suggested that the Beijing family, belonging to lineage 2, is hyper-mutable7 with a propensity to develop resistance at a higher frequency than other lineages8,9,10,11, while others cite evidence to the contrary12,13. Pre-existing resistance to one antibiotic (mono-resistance) is another factor that may influence the acquisition of multidrug-resistance14. Mono-resistance to isoniazid or rifampicin has been associated with increased rates of multidrug-resistance acquisition15,16, but the relative risk of either remains unclear. Similarly, phylogenetic analyses suggest a stepwise progression towards multidrug-resistance, where mutations conferring isoniazid resistance tend to precede those linked to rifampicin resistance17,18,19,20.\n\nPhylogenetic trees have been increasingly used to study pathogen dynamics and evolutionary processes of a wide range of phenotypes of epidemiological interest, including virulence and drug resistance acquisition21,22. A necessary focus on improving the molecular diagnosis of drug resistance has led to the generation of large strain collections of drug resistant pathogens. However, unrepresentative samples of this kind enriched for drug-resistant isolates limit the ability to characterize the evolution and dynamics of drug resistance from a diverse background of ancestral susceptible strains. Inadequate sampling without comprehensive population level coverage or sufficient temporal span compounds this problem, while the monomorphic nature of the M. tuberculosis genome makes constructing time-calibrated phylogenetic trees particularly challenging. As a consequence, a single mutation rate is often applied to the data, but this assumption inappropriately forces lineages and sub-lineages to conform to the same global mutation rate, thus limiting the inferences that can be made from the data.\n\nOvercoming these issues, we present findings from samples collected over a 17-year time span with population level coverage in the hyperendemic suburbs of Lima, Peru. We apply a genome-wide survival analysis to a time-calibrated phylogeny of 3135 M. tuberculosis strains, and show the existence of pre-resistance mutations among drug susceptible genotypes that increase the risk of future drug resistance emergence in M. tuberculosis. We demonstrate significant differences in the acquisition of drug resistance between lineages, on mono-resistant backgrounds, and at the level of nucleotide polymorphisms. Our findings were then tested and replicated in an independent publicly available data set of 1027 whole genomes collected in Samara, Russia, and in a collection of 1573 isolates from multiple countries to demonstrate that they can be globally generalized.",
+            "section_image": []
+        },
+        {
+            "section_name": "Results",
+            "section_text": "A total of 3432 M. tuberculosis genomes from Lima (Peru) were analyzed, of which 3135 passed genomic quality filters. Of this, 2037 were part of a population level study carried out in 2009 where sputum samples were taken from all patients presenting tuberculosis symptoms in the Lima areas of Callao and Lima South23 (Supplementary Data File\u00a01). Comparison of drug resistance prevalence between the population level sampling using molecular genotyping and reports of epidemiological data in Peru2 are consistent: 1.5% (32/2037) of samples were rifampicin mono-resistant; 5% were isoniazid mono-resistant (105/2037), and 13% were multidrug-resistant (251/2037) (Supplementary Table\u00a01). The remaining samples were collected from cohort studies covering a 17-year period of research in the regions of Lima and Callao in order to achieve a sufficient temporal span in our sampling window (Supplementary Fig.\u00a01). Both lineage 2 and lineage 4 had a similar distribution of sampling dates (Supplementary Fig.\u00a02).\n\nThe isolates were first aligned to the reference genome H37Rv, then lineages and sublineages were assigned using clade specific SNPs24. Lineage 4 (L4, Euro-American) consisted of 2807 samples, while lineage 2 (L2, Beijing) had 327 isolates (Table\u00a01). There was a single representative of lineage 1 (Indo-Oceanic), which was used to root the phylogenetic tree. The remaining samples from the data set, which included 5 M. caprae isolates, were not used in the downstream analysis. Lineage 4 had the highest diversity, comprising 1235 isolates from lineage 4.3 (LAM), 935 from lineage 4.1.2.1 (Haarlem), 271 from lineage 4.1.1 (X-Type), and 312 from lineages denoted as Type T, which encompasses lineages 4.5, 4.7, 4.8 and 4.9. Other minor sublineages included lineage 4.2.2 (TUR) and lineage 4.4. All isolates from Lineage 2 were part of the Beijing sublineage (lineage 2.2), mainly from the sublineage 2.2.1 or Modern Beijing, and with only one representative of the sublineage 2.2.2 or Asia Ancestral25 (Table\u00a01).\n\nThe alignment of the isolates to the reference genome resulted in 64,586 SNPs, of which 18,022 were singletons (28%). Most SNPs were not widely distributed across the population, and only 8088 variants had a frequency in the dataset higher than 1%. A total of 59,789 SNPs (16,934 singletons, 28%) were identified for lineage 4, and 4821 SNPs (1370 singletons, 28%) for lineage 2. We applied the same analysis to a publicly available data set of 1027 isolates from Samara, Russia, as a validation set where, unlike the Peruvian dataset, lineage 2 constitutes the main lineage. We identified a total of 28,414 SNPs, consistent with previous publications of this data19.\n\nPatient demographic metadata was available for 2220 samples, of which 88% were smear positive, 27% were previously treated for tuberculosis, and 2.8% were HIV positive, consistent with previous population level estimates in Peru26. The median age was 28 years (IQR 21\u201341). In our cohort, 86% of lineage 2 and 89% of lineage 4 were smear positive.\n\nThe maximum likelihood phylogeny constructed using these alignments grouped the isolates by lineage similarly to previously published global data sets24 (Fig.\u00a01a, Supplementary Fig.\u00a03). To study the temporal dynamics of drug resistance acquisition at the population level, the maximum likelihood phylogeny was time-calibrated using the sampling dates of the isolates, which extended from 1999 to 2016. Dated phylogenies were built separately for lineage 2 and lineage 4 in order to avoid the confounding effect of the temporal and population structures27. Before time calibration of the phylogeny, we tested the adequacy of the temporal correlation of evolutionary change to reliably infer the model parameters. First, a root-to-tip linear regression of the number of substitutions from the root and the sampling times was fitted to confirm a positive association between time and evolutionary change. As uneven sampling may bias the root-to-tip regression28, a date-randomization test was additionally performed using the full Bayesian model implemented in BactDating29 with the original dataset and 100 randomizations where the sampling times were permuted, representing the expectations of the model parameters in the absence of temporal signal. The substitution rate estimated for the original dataset and for the 100 randomizations was compared to verify a lack of overlap between the 95% credible intervals. Both lineage 4 and lineage 2 datasets showed a clear temporal signal (Supplementary Fig.\u00a04), and thus model parameters could be confidently inferred from the data30,31. We used the relaxed clock model implemented in BactDating29, allowing the mutation rate to vary in each branch independently. We ran the MCMC until convergence of the chains was achieved, with an effective sample size (ESS) of at least 100 (Supplementary Fig.\u00a05). The estimated rate for lineage 2 was 0.45 substitutions per genome per year (0.32\u20130.57 95% CI), while lineage 4 had a clock rate of 0.299 (0.25\u20130.36 95% CI) (Fig.\u00a01b). The estimates of the molecular clock for both lineages were consistent with previous reports32. The most common recent ancestor (tMRCA) for our samples was placed at 565 CE (263;826 95% CI) for lineage 4 (Fig.\u00a01c), while lineage 2 had a tMRCA in 1325 CE (1070;1499 95% CI) (Fig.\u00a01d).\n\nColors represent different lineages and sublineages. a Maximum likelihood phylogeny. Scale in number of substitutions per genome. b Violin plots showing the posterior density distribution of the inferred substitution rate in substitutions per genome per year derived by sampling from 107 MCMC iterations. The substitution rate was estimated separately for lineage 4 (blue) and lineage 2 (red). Box plots inside the violin indicate the median value of the distribution (black horizontal line) and the interquartile range. Whiskers denote 1.5x the interquartile range, while outliers are plotted as individual points. c Time-calibrated phylogeny of lineage 4. d Time-calibrated phylogeny of lineage 2.\n\nDrug resistance was inferred for all isolates at the tips of the phylogenetic tree using well-established drug resistance associated SNPs33. In addition to molecular typing of drug resistance, all isolates included in the analysis had Drug Susceptibility Testing (DST) performed either by MODS or by the proportional method in agar. DST and molecular typing showed consistent results for 96% of the samples for rifampicin resistance and 92% for isoniazid resistance. The time of emergence of drug resistant mutations was inferred by reconstructing the ancestral sequences of the internal nodes in the phylogenetic tree. The time of emergence of a specific antibiotic resistance mutation was approximated to the inferred year of the internal node where such mutation first appeared. The phylogenetic estimates of the emergence of drug resistance conferring mutations in lineage 4 occurred around the time of the known introductions of the corresponding drug. In contrast the emergence of drug resistance in lineage 2 was observed to have arisen many years after the introduction of antituberculous drugs. This is consistent with the geographic spread of lineage 4 in Europe together with early widespread use of drugs in this region (Table\u00a02, Fig.\u00a02). For both lineage 2 and lineage 4, the earliest inferred occurrence of resistance was to isoniazid, by the Ser315Thr mutation in the gene KatG, around 1957 (1928;1978 95% CI) for lineage 2, and 1942 (1913;1960 95% CI) for lineage 4, in line with the reported wide introduction of isoniazid in 195234. The rifamycins were first isolated in 195734, and we estimate the date for the first acquisition of resistance to rifampicin due to the rpoB mutation Ser450Leu to have emerged around 1951 (1931;1971 95% CI) for lineage 4 and 1974 (1953;1988 95% CI) for lineage 2. None of the drug resistant nodes reverted to susceptible along the branches of the two phylogenetic trees.\n\nPosterior density distribution inferred using a time-calibrated phylogeny for both lineage 4 and lineage 2. Arrows represent the approximate time of antibiotic introduction.\n\nIt has been shown that secondary mutations arising after the acquisition of drug resistance may alleviate the fitness cost associated to antibiotic resistance mutations35, but little is known about their temporal dynamics. Data on M. tuberculosis drug resistance compensatory mechanisms is mainly limited to isoniazid and rifampicin resistance36,37. Non-synonymous mutations in the gene rpoC have been suggested as secondary compensatory mutations for rifampicin associated mutations in the rpoB gene37. A total of 34% (258/755) of lineage 4 isolates harboring rpoB mutations also had rpoC non-synonymous polymorphisms; for lineage 2, 38% (33/87) of isolates with rpoB mutations carried rpoC polymorphisms. No significant differences were observed between lineages in a logistic regression model (OR\u2009=\u20090.85, 95% CI 0.54\u20131.35, p-value = 0.49). Overall, 62% of rifampicin resistant isolates carried Ser450Leu rpoB mutations (525/842). Rifampicin resistant isolates harboring Ser450Leu rpoB mutations had a higher probability of carrying mutations in the rpoC gene (52%, 272/525) than isolates with other rpoB mutations (6%, 19/317) in a logistic regression model (OR\u2009=\u200916.86, 95% CI 10.55\u201328.50, p-value = 4\u2009\u00d7\u200910\u221229). Only 3% (9/291) of the isolates carried two non-synonymous mutations in the rpoC gene, while the rest had only one.\n\nTo understand the emergence of rpoC non-synonymous mutations, we scanned the phylogenetic branches of rifampicin resistant isolates from the root to the tip, using the inferred sequences of the ancestral nodes to determine the time of emergence of rpoC non-synonymous mutations. The analysis was repeated in 100 bootstrap phylogenies to infer confidence values around our estimates. In both lineage 2 and lineage 4, the emergence of rpoC non-synonymous mutations occurred immediately after or at the same time as the emergence of rifampicin resistance, and continued steadily over time (Fig.\u00a03). For lineage 2, there was not a single emergence of rpoC mutations occurring prior to the rifampicin resistance conferring mutations. In the case of lineage 4, two rpoC mutations emerged before rifampicin resistance: c.765150\u2009G\u2009>\u2009A and c.765590\u2009C\u2009>\u2009A. Both mutations appeared once independently in the entire phylogeny. The mutation c.765150\u2009G\u2009>\u2009A emerged in our dataset around the year 936 CE (727;1110 95% CI), in the MRCA of the clades X-type and Haarlem, all of which present the c.765150\u2009G\u2009>\u2009A mutation (1206/2807 isolates). For c.765590\u2009C\u2009>\u2009A, the estimated year of emergence was around 1867 CE (1813;1907 95% CI), and all the tips harboring the mutation belonged to a cluster of isolates part of the LAM11 sublineage (33/2807 isolates). Given that both mutations showed a clear phylogenetic structure and emerged independently only once in our dataset, they were considered phylogenetic mutations and were removed from the final analysis. The emergence of non-synonymous mutations in the rpoC gene was similar for isolates carrying the Ser450Leu rpoB mutation and for those isolates with other rpoB mutations (Supplementary Fig.\u00a06).\n\na, b Cumulative number of non-synonymous mutations in rpoC over time. The x-axis represents the years since the inferred time of rifampcin resistance (time 0). Dark blue line shows the cumulative number of mutations for the ML tree, while the 95% confidence interval (shaded area) is inferred by repeating the analysis in 100 bootstrap phylogenies. The analysis was performed separately for a lineage 2 and b lineage 4.\n\nTo estimate the year of M. tuberculosis introductions to Peru, we subsampled the Peruvian isolates and analyzed them alongside global representatives of both lineage 2 and lineage 4 for which both collection date and geographic origin were known (Supplementary Data File\u00a02). The phylogenies were time-calibrated using a relaxed clock model as implemented in BactDating29, and MCMC convergence was assessed using the traces of the model parameters (Supplementary Fig.\u00a07).\n\nThe phylogeographic history was inferred by reconstructing the ancestral states by maximum likelihood. The geographical origin of the isolates was treated as a discrete character, and we assumed that the time of introduction occurred at the first Peruvian node of each clade.\n\nTwo early introductions of lineage 2 from China in 1880 (1856;1905 95% CI) and 1906 (1881;1929 95% CI) accounted for 84% of the Peruvian dataset (275/327 isolates), and 77% (90/117) of the drug resistance isolates as well as 77% (20/26) of the independent resistance evolutionary events (Fig.\u00a04a). Later introductions to Peru from China occurred between 1981 (1969;1989 95% CI) and 2004 (1999;2007 95% CI), with one introduction in 1986 (1974;1994 95% CI) from South Asia representing 6 isolates, one of which is drug resistant.\n\na, b Inferred introductions of Mycobacterium tuberculosis in Peru. The top part shows a time-calibrated phylogeny, with inferred introductions to Peru highlighted in the nodes with colors representing the country from which the clade was introduced. Peruvian clades are shown in blue. The bottom part shows the estimated year of introduction. The analysis was done separately for a lineage 2 and b lineage 4. For visual representation purposes, only the year of introduction of clades with more than 10 tips are shown for lineage 4.\n\nLineage 4 was inferred to have been introduced in Peru several times over the years, mainly from Europe and Brazil (Fig.\u00a04b). Lineage 4.3 (LAM) represents the first and main introduction from Brazil around 1512 (1383;1598 95% CI), shortly after the well documented arrival of the Europeans to the continent. Subsequent smaller introductions from Europe and Brazil around 1644 (1545;1713 95% CI) and 1743 (1671;1796 95% CI) shaped the LAM and T type lineages. We inferred that most of L4.1.2.1, part of the Haarlem sublineage, likely evolved from two introductions from Europe around 1693 (1614;1749 95% CI) and 1834 (1781;1869 95% CI). The most likely introductions of the X type clade (L4.1.1) occurred from Brazil in 1692 (1597;1754 95% CI) and South Africa for L4.1.1.3 in 1706 (1615;1773 95% CI).\n\nThe risk of acquiring drug resistance was calculated as the Cox Proportional Hazard Ratio (HR) using the time between sensitive internal nodes and the first drug resistant node in the time-calibrated phylogenetic trees. The Kaplan\u2013Meier curves showed that lineage 2 had a higher probability of acquiring drug resistance than lineage 4 (Log-rank test p-value = 1.2\u2009\u00d7\u200910\u22129; Fig.\u00a05a). The estimated hazard ratio of drug resistance acquisition for lineage 2 was estimated to be 3.36 when compared to lineage 4 (HR 3.36, 95% CI 2.10\u20135.38, Likelihood ratio test p-value = 4.25\u2009\u00d7\u200910\u22127). A similar trend was observed in the Samara dataset (HR 4.82, 95% CI 3.74\u20136.21, Likelihood ratio test p-value = 6.8\u2009\u00d7\u200910\u221234; Kaplan\u2013Meier curve Log-rank test p-value = 1\u2009\u00d7\u200910\u221239; Fig.\u00a05b). The risk of drug resistance acquisition was also higher in lineage 2 when compared to all sublineages of lineage 4 in the Peruvian dataset, using LAM3 as a reference (lineage 2 HR 3.32, 95% CI 1.84\u20136.28, Likelihood ratio test p-value = 1.9\u2009\u00d7\u200910\u22124, all other p-values > 0.2; Kaplan\u2013Meier curve Log-rank test p-value = 6.9\u2009\u00d7\u200910\u22128; Fig.\u00a05c). To assess the adequacy of our data to the proportional hazard assumption, we calculated the relationship between the Schoenfeld residuals against time. In all cases, a non-significant association between the Schoenfeld residuals and time supported the use of the proportional hazards model (Supplementary Fig.\u00a08).\n\na\u2013c Top: Hazard ratio (HR). Points and error bars represent the HR estimate and the 95% CI, respectively. The p-value for the HR was calculated using the likelihood ratio test. Bottom: Kaplan\u2013Meier curve and numbers at risk. Y-axis represents the probability of remaining susceptible to any antibiotic, while the X-axis shows the time in years or the distance in branch length. Shaded areas show the 95% CI. Kaplan\u2013Meier curves were compared and p-values were derived using the log-rank test. a Depicts the HR of lineage 2 compared to lineage 4 in the Peruvian dataset (HR 3.36, 95% CI 2.10\u20135.38, Likelihood ratio test p-value = 4.25\u2009\u00d7\u200910\u22127) and the different Kaplan\u2013Meier curve for lineage 2 and lineage 4 (Log-rank test p-value = 1.2\u2009\u00d7\u200910\u22129). b Same metrics for the Samara dataset (HR 4.82, 95% CI 3.74\u20136.21, Likelihood ratio test p-value = 6.8\u2009\u00d7\u200910\u221234; Kaplan\u2013Meier curve Log-rank test p-value = 1\u2009\u00d7\u200910\u221239). c Shows HR between lineage 2 and the different sublineages of lineage 4 found in the Peruvian dataset (LAM9, LAM3, LAM11, Haarlem, X type and T type), using LAM3 as a reference (lineage 2 HR 3.32, 95% CI 1.84\u20136.28, Likelihood ratio test p-value = 1.9\u2009\u00d7\u200910\u22124, all other p-values > 0.2; Kaplan\u2013Meier curve Log-rank test p-value = 6.9\u2009\u00d7\u200910\u22128). Statistical significance of the hazard ratio differences presented next to the CI bars (*p\u2009<\u20090.05; **p\u2009<\u20090.01; ***p\u2009<\u20090.001).\n\nIn order to evaluate the robustness of our maximum likelihood phylogeny, we repeated both the dating and the survival analysis on 100 phylogenetic bootstrap replicates. In both lineage 2 and lineage 4, the Cox Proportional Hazard ratio was not significantly different between the maximum likelihood phylogeny and the bootstrap replicates (Supplementary Fig.\u00a09a, b). Additionally, the Kaplan-Meier curve of the 100 replicates was similar to that of the maximum likelihood tree (Supplementary Fig.\u00a09c, d).\n\nBoth phylogenetic trees were subsampled to include only the isolates from the 2009 population level study to prevent any biases due to inclusion of datasets enriched for drug resistance isolates (Supplementary Table\u00a01). The results were congruent with those obtained with the entire dataset. Lineage 2 was characterized by a higher risk of drug resistance acquisition when compared to lineage 4 (HR 4.84, 95% CI 2.78\u20138.45, Likelihood ratio test p-value = 2.7\u2009\u00d7\u200910\u22128; Kaplan-Meier curve Log-rank test p-value = 7.9\u2009\u00d7\u200910\u221210; Supplementary Fig.\u00a010a). Moreover, lineage 2 also had a higher hazard ratio than any sublineage of lineage 4 (lineage 2 HR 5.1, 95% CI 2.17\u201311.9, Likelihood ratio test p-value = 1.8\u2009\u00d7\u200910\u22124, all other p-values > 0.2; Kaplan-Meier curve Log-rank test p-value = 3.02\u2009\u00d7\u200910\u22127; Supplementary Fig.\u00a010b).\n\nSeveral confounders may also explain the differential rate of drug resistance acquisition between lineages (Supplementary Fig.\u00a011). We found no significant association between M. tuberculosis sublineages and HIV (Supplementary Fig.\u00a011a) and smear positivity (Supplementary Fig.\u00a011b) in a logistic regression model (n\u2009=\u20092133, all p-values >\u20090.1). It has been previously shown that prison conditions increase and amplify drug-resistance tuberculosis in Peru38,39. Our study population showed a higher proportion of prison infection in lineage 2 (5.8%, 12/206 patients) when compared to lineage 4 (1.4%, 28/2011 patients, Supplementary Fig.\u00a011c), and a higher risk of lineage 2 infection within the prison population (n\u2009=\u20092217, OR\u2009=\u20094.4, 95% CI 2.1\u20138.5, p\u2009<\u20090.001). This finding did not explain the higher rate of drug resistance acquisition in lineage 2, since all the lineage 2 samples taken from prisoners belonged to the same cluster and none of them harbored drug resistance conferring mutations. Since previous treatment with antituberculous drugs has been associated to an increased risk of acquiring drug resistance38, we tested the association between previous treatment and the different sublineages of our cohort. Previous treatment history was available for 2236 samples contained metadata regarding previous treatment of tuberculosis. We found no significant association between M. tuberculosis sublineages and previous treatment with antituberculous drugs in a logistic regression model (n\u2009=\u20092236, all p-values >\u20090.1), suggesting that the between lineage differences in drug resistance acquisition observed in the survival analysis are not confounded by a differential distribution of antibiotics between sublineages (Supplementary Fig.\u00a011d). Behavioral patterns may also affect the dynamics of drug resistance acquisition and transmission. Patient sex (Supplementary Fig.\u00a011e) was not significantly associated with any of the M. tuberculosis sublineages (n\u2009=\u20092205, p-value = 0.3). The mean age for patients with lineage 4 infection was 33.3 (22\u201341 IQR), while the mean age for patients with lineage 2 was 30.4 (21\u201336 IQR). Although the distribution of patient age was similar between the two lineages (Supplementary Fig.\u00a011f), lineage 4 showed an incident rate of age 1.09 higher than lineage 2 in a quasi-Poisson model (lineage 4 estimate = 0.09, standard error = 0.02, p\u2009=\u20090.001). On the other hand, age was not significantly associated with a higher risk of drug resistance acquisition in a logistic model (OR 0.999, 95%CI 0.994\u20131.003, p\u2009=\u20090.69).\n\nTo determine the effect of mono-resistance on the acquisition of further multidrug-resistance, the hazard ratio of acquiring rifampicin resistance was calculated for isoniazid mono-resistant ancestral genotypes versus susceptible ancestral strains. Genotypes with an isoniazid mono-resistant background had 15 times the hazard of developing rifampicin resistance tuberculosis relative to wild type susceptible strains (HR 15.12, 95% CI 10.54\u201321.69, Likelihood ratio test p-value <\u200910\u221215; Kaplan\u2013Meier curve Log-rank test p-value = 2, 7\u2009\u00d7\u200910\u221263; Fig.\u00a06a). A larger hazard ratio was obtained in the Samara dataset (HR 37.28, 95% CI 18.81\u201373.88, Likelihood ratio test p-value = 3.4\u2009\u00d7\u200910\u221225; Kaplan\u2013Meier curve p-value = 4.6\u2009\u00d7\u200910\u221263; Fig.\u00a06b), although the low prevalence of mono-resistance clades in the Samara set may bias this estimate.\n\na, b Top: Hazard ratio (HR). Points and error bars represent the HR estimate and the 95% CI, respectively. The p-value for the HR was calculated using the likelihood ratio test. Bottom: Kaplan\u2013Meier curve and numbers at risk. Y-axis represents the probability of remaining susceptible to rifampicin, while the X-axis shows the time in years or the distance in branch length. Shaded areas show the 95% confidence interval. P-values for the Kaplan\u2013Meier curves were calculated using the log-rank test. a Depicts the risk of acquiring rifampicin resistance from an already isoniazid mono-resistant background compared to a drug susceptible one (HR 15.12, 95% CI 10.54\u201321.69, Likelihood ratio test p-value = 1.3\u2009\u00d7\u200910\u221240) and the Kaplan\u2013Meier curves for the different backgrounds (Log-rank test p-value = 2.7\u2009\u00d7\u200910\u221263). b Same metrics for the Samara dataset (HR 37.28, 95% CI 18.81\u201373.88, Likelihood ratio test p-value = 3.4\u2009\u00d7\u200910\u221225; Kaplan\u2013Meier curve p-value = 4.6\u2009\u00d7\u200910\u221263). Statistical significance of the hazard ratio differences presented next to the CI bars (*p\u2009<\u20090.05; **p\u2009<\u20090.01; ***p\u2009<\u20090.001).\n\nMultidrug-resistance was preceded by rifampicin mono-resistance only one time in the phylogenetic tree. Due to the infrequent occurrence of rifampicin mono-resistance prior to multidrug-resistance emergence, the risk of developing multidrug-resistance from a rifampicin mono-resistance background could not be reliably estimated.\n\nGenome-wide survival analysis was performed using a Cox Proportional Hazard regression model to identify genetic variants in phylogenetic nodes inferred to be drug susceptible but associated with a higher risk of progression towards drug resistance. Resistant nodes were defined as those inferred to be resistant to any antibiotic in order to identify common pathways of increased risk of acquiring resistance regardless of the specific antibiotic. The association analysis was performed in lineage 4 and lineage 2 separately, and we further corrected for population structure using a kinship matrix, which reduced the genomic inflation factor (\u03bb) to 1.16 (Supplementary Fig.\u00a012).\n\nSix variants in drug susceptible ancestral genotypes were associated with a higher risk of acquiring drug resistance in lineage 4 after population and multiple testing correction, three of which were in previously annotated genes (Fig.\u00a07a). The variant with the lowest p-value corresponded to a 9\u2009bp deletion at location 2,604,157 in the locus lppP, which encodes a lipoprotein and has been predicted to be required for growth in macrophages40. This deletion had a frequency of 1.7% in the population, and it evolved 12 times independently along the phylogenetic tree. Genotypes with this variant had a hazard ratio 7.36 times greater than those with an intact lppP (Fig.\u00a07b, HR 7.36 95% CI 3.85\u201314.04, p-value = 7.46\u2009\u00d7\u200910\u221210). We replicated our findings in a global data set of 1573 L4 isolates (Supplementary Data File\u00a03), which was relatively enriched for drug resistance (55% of samples were resistant to any drug). The lppP deletion had a frequency of 9% and inferred susceptible genotypes with the deletion had a hazard ratio 3.6 times greater than those without it (HR 3.6, 95% CI 1.9\u20136.9, p-value = 8.7\u2009\u00d7\u200910\u22125). Two synonymous polymorphisms at esx genes were found to be associated with a higher risk of acquiring drug resistance in inferred drug susceptible genotypes. The esx gene family encodes protein secretion systems described to be critical for growth, pathogenesis, and mycobacterial\u2013host interactions41. The two polymorphisms were detected in the gene esxL at position 1,341,044 with a hazard ratio of 3.2 (HR 3.2 95% CI 1.91\u20135.37, p-value = 1.01\u2009\u00d7\u200910\u22126), and at position 2,626,011 in the gene esxO (HR 11.12, 95% CI 5.50\u201322.5 p-value = 1.52\u2009\u00d7\u200910\u22125) with a frequency in the population of 17 and 5%, respectively. In the L4 global dataset, the esxL SNP had a frequency of 19.5% while the esxO polymorphism was present in 10% of the isolates. Inferred susceptible genotypes in the global data set carrying the mutation in esxO had a risk of acquiring drug resistance 3.1 times higher than those with the reference genotype (HR 3.1, 95% CI 1.3\u20137.3, p-value = 0.009), while those carrying the mutation in esxL had a risk 1.4 higher, although differences where not statistically significant (HR 1.4, 95% CI 0.7\u20133.0, p-value = 0.3). Visual inspection of the short-read alignments around the described genes was undertaken to confirm high quality alignments over these regions (Supplementary Fig.\u00a013).\n\na Manhattan plot for GWAS on increased risk of drug resistance acquisition in lineage 4. The red line represents the Bonferroni corrected p-value threshold of 3.37\u00d710\u22125. Labels show the genes where the significant hits are located. Colors indicate the hazard ratio, with a scale of blue representing hazards ratio lower than 1 and a scale of reds for hazard ratios higher than 1. b Kaplan\u2013Meier curve and numbers at risk of a 9\u2009bp deletion in the gene lppP comparing the probability of remaining susceptible between those nodes without the deletion (blue) and those with it (red). Shaded areas represent the 95% CI. The p-value for the Kaplan\u2013Meier curves was calculated using the log-rank test.\n\nFor the gene-based GWAS, non-synonymous variants were aggregated for each locus and a binary matrix was created reflecting whether internal nodes and tips contained at least one non-synonymous polymorphism for each gene. After population and Bonferroni multiple testing correction, a total of 35 variants had a p-value lower than the significance threshold of 4.38\u2009\u00d7\u200910\u22125 (Table\u00a03). Functional annotations curated in the Mycobrowser42 showed that most genes were related to metabolism and cellular respiration (frdB, cyp135A1, Rv3113, hemE, pyrG, galK, icd1, Rv1096, fum, galE2, Rv1751, Rv3712, asnB) as well as cell wall processes (kdpA, lppP, cfp2, lytB1, mmpL1, Rv1417, lgt, Rv0226c, caeA, murF, pstB).\n\nNo significant associations were identified for lineage 2 after correcting for population structure, possibly due to the lower diversity of lineage 2 and the strong lineage effect on the phenotype. The analysis could not be replicated in the Samara dataset as the Samara dataset was significantly smaller and hence lacked sufficient statistical power.",
+            "section_image": [
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-021-27616-7/MediaObjects/41467_2021_27616_Fig1_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-021-27616-7/MediaObjects/41467_2021_27616_Fig2_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-021-27616-7/MediaObjects/41467_2021_27616_Fig3_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-021-27616-7/MediaObjects/41467_2021_27616_Fig4_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-021-27616-7/MediaObjects/41467_2021_27616_Fig5_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-021-27616-7/MediaObjects/41467_2021_27616_Fig6_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-021-27616-7/MediaObjects/41467_2021_27616_Fig7_HTML.png"
+            ]
+        },
+        {
+            "section_name": "Discussion",
+            "section_text": "This study represents the largest population level genomic analysis of Mycobacterium tuberculosis to date. Our 17-year sampling time frame provided a unique opportunity to study drug resistance acquisition dynamics and evolution. To our knowledge this is the first description and evaluation of pathogen pre-resistance (pre-existing polymorphisms that predispose to the acquisition of future drug resistance).\n\nUsing an ancestral state genome-wide survival analysis to move in time through the phylogenetic tree, we show that M. tuberculosis is predisposed to acquire drug resistance mutations at the lineage level, after mono-resistance, and at the level of nucleotide polymorphisms. Identifying pathogen genetic factors that predispose strains to evolve drug resistance could help prevent the acquisition and spread of resistance as well as treatment failure by expanding treatments to those strains most likely to become resistant in the future.\n\nPrevious studies of acquired drug resistance at the sublineage level in M. tuberculosis have led to contradictory outcomes, with small sample sizes in fluctuation assays7,12 or using amalgamated sub-population level samples. Here we demonstrate that lineage 2 acquired resistance to antibiotics more rapidly than lineage 4. There were no significant differences observed in drug resistance acquisition between the sub-lineages of the most diverse lineage 4. Even though lineage 2 showed an increased risk in drug-resistance acquisition, lineage 4 evolved resistance earlier than lineage 2 for almost all drugs analyzed, with the exception of streptomycin. This may be explained by the Euro-American distribution of lineage 4 and the earlier widespread implementation of antibiotics in these regions. Our analysis also suggests that the acquisition dynamic of compensatory mutations was similar for both lineage 2 and lineage 4. After rifampicin resistance associated mutations evolve in a clade, non-synonymous mutations in rpoC start to occur and steadily accumulate over time. Thus, pre-resistance mutations emerge independently of compensatory mutations.\n\nThe identification and control of mono-resistant strains is a key component of tuberculosis public health infection prevention and control efforts. Mono-resistance is associated with worse clinical outcomes43 and an increased probability of progressing to multidrug-resistance15. At the population level, our study quantifies this risk and shows that isoniazid mono-resistant strains have at least 15 times the hazard of developing multidrug-resistance relative to wild type susceptible strains. Despite the use of new therapeutics, multi-drug resistant tuberculosis continues to require polypharmacy with increased toxicity and longer treatment duration2. Globally, molecular rapid drug resistance surveillance is focused primarily on rifampicin, with the widely implemented GeneXpert MTB/RIF PCR based assay unable to detect isoniazid mono-resistance. Although Drug Susceptibility Testing (DST) is the current gold standard for identification of drug resistance isolates and can detect isoniazid mono-resistance, known diagnostic delays associated with it may limit its use in reducing mono-resistance amplification44,45. Inadequate diagnosis of isoniazid mono-resistance will inevitably lead to inappropriate treatment and could fuel rapid evolution of multidrug-resistance, thus posing a significant threat to tuberculosis control.\n\nWe also identified loci associated with higher risk of future drug resistance acquisition. To remain polymorphic, these variants must be under balancing selection and only be positively selected once exposed to drug therapy. Rather than causing resistance directly, these variants could promote resistance acquisition by compensating for the fitness costs of resistance in vivo35 or by increasing drug tolerance46.\n\nAt the gene level, most non-synonymous mutations associated with pre-resistance genotypes were located in genes related to cell wall processes and metabolism. Functional studies and prospective clinical trials are warranted to confirm their association with future drug resistance acquisition.\n\nThe variant with the lowest p-value corresponded to a 9\u2009bp deletion in the gene lppP present at a frequency of 1.7% in the population and arose 12 times independently. Deletions in lipoproteins have been well characterized in the past47, and lppP has been predicted to be required for growth in macrophages40. Lipoproteins can act as antigenic proteins47, and thus deletions in the genes encoding them may alter the interaction between the bacilli and the macrophages, potentially conferring a selective advantage in the presence of drug and increasing the probability of acquiring drug resistance. This variant was also present in a global dataset for L4 at a frequency of 9%, which could be explained by the higher prevalence of drug resistance isolates in most publicly available data sets.\n\nTwo additional synonymous variants were identified in the genes esxL and esxO, which encode the ESAT-6-like proteins esxL and esxO. These genes are part of a family of genes that encode immunogenic secreted proteins that play a role in mycobacterial growth, pathogenesis, and host-pathogen interactions41. Moreover, esxO has been associated with pathogenesis by inducing autophagy in infected macrophages48. Synonymous homoplastic variants in esx genes have been previously identified23, but their phenotypic effects are still unclear.\n\nThis study benefited from an unbiased population level coverage of both drug resistant and drug susceptible strains that enabled us to reliably correct for the founder effect and control for the influence of pre-existing population diversity. The large sampling size and time frame\u2014a consequence of 17-years of continued research in the same location\u2014allowed us to generate time-calibrated phylogenies without imposing a global mutation rate. This was a pre-requisite for the downstream analyses and our GWAS survival analysis approach. We were also able to replicate our sublineage and mono-resistance dependent hazards of acquired resistance in the smaller Samara dataset. However, the time scale and size of this publicly available data was insufficient to allow us to confirm the effect of the lppP deletion in a second independent dataset.\n\nAlthough our phylogenetic analysis reveals trends of drug resistance acquisition over evolutionary time, prospective cohort studies are required to determine the effect of these mutations at the individual patient and household level. Non-bacterial factors are unlikely to influence our findings, since they would have to have been disproportionately and consistently associated to a specific lineage over long periods of time. Nevertheless, we explored the influence of confounding variables on our dataset. There was no difference observed in the proportion of patients receiving previous antituberculous treatment between the two lineages. This makes our findings unlikely to be influenced by differential exposure to drugs among lineages. There was no difference in sputum smear grade between lineages, suggesting that our findings are not a consequence of increased pathogenicity of lineage 2 in comparison to lineage 4. Moreover, other factors that may affect the rates of drug resistance acquisition such as HIV status, sex, or imprisonment, did not show differences between lineages. Even though the age of patients with lineage 4 infection was significantly higher than that of patients with lineage 2, the difference was small. Additionally, patient age was not associated to a higher incidence of drug resistance, and therefore it is unlikely that age differences explain the higher risk of acquiring antibiotic resistance of lineage 2. Differential healthcare systems influence the acquisition and transmission of drug resistance tuberculosis, and thus importation events to Peru of resistant strains from specific lineages could have affected the dynamics of drug resistance acquisition. We showed that lineage 2 in Peru is characterized by two importations around 1900 CE, which is consistent with major immigration events of laborers from China to Peru at the end of the 19th century to work in the railroads, guano mines, and cotton and sugarcane plantations49,50. Conversely, the majority of lineage 4 clades were imported from Europe and Brazil between the 16th and 18th centuries, compatible with European colonial expansion51. Therefore, significant immigration to Peru occurred well before the advent of antibiotics, which limits the influence of imported drug resistant strains. Moreover, the majority of introductions occurring in recent times were of drug susceptible clades. Nevertheless, it is possible that some resistance events may have arisen as a result of importation of resistant strains from countries with different drug selection pressures.\n\nIn summary, this population wide 17 year-long epidemiological study of M. tuberculosis genetics provides the first description and evaluation of pre-resistant polymorphisms in susceptible genotypes that predispose to the acquisition of future drug resistance. Prediction of future drug resistance in susceptible pathogens together with targeted expanded therapy has the potential to prevent drug resistance emergence in M. tuberculosis and other pathogens. Prospective cohort studies of participants with and without these polymorphisms should be undertaken to inform clinical trials of personalized pathogen genomic therapy. This ancestral state genome-wide survival analysis could also be employed to predict and prevent the emergence of resistance or indeed any important trait of interest in other organisms.",
+            "section_image": []
+        },
+        {
+            "section_name": "Methods",
+            "section_text": "Ethical approval for sample collection and processing was obtained from the Institutional Review Board of Universidad Peruana Cayetano Heredia and the Peruvian Ministry of Health for all individual studies from which this data was derived38.\n\nSamples were selected from previous projects taking place across the region of Lima. The first project consisted of a population level study carried out between 2008 and 2010 as part of the population level implementation of Microscopic Observation Drug Susceptibility (MODS) testing52,53. A total of 2139 unselected patients of tuberculosis were collected (Supplementary Data File\u00a01). Of these patients, 284 were analyzed in previous studies (PRJEB5280)23, while 1855 were processed as part of this project (PRJEB39837).\n\nA second set of 213 MDR-TB samples was obtained from a 3-year long household follow-up study conducted between 2010\u2013201338, of which 185 randomly selected samples underwent whole-genome sequencing (PRJEB5280)23.\n\nAdditionally, 42 unpublished whole-genome sequences of samples collected from 1999 to 2007 in different regions of Lima were added to the study (PRJEB47846), as well as 575 samples that were collected between 2003 and 2013 as part of the CRyPTIC Consortium (PRJEB32234)54, and 489 samples from the TANDEM Consortium (PRJEB23245)55 taken between 2014 and 2016.\n\nAll samples without collection date, as well as reference clinical samples, were excluded from the final list. Drug Susceptibility Testing (DST) was performed either by MODS53 or by the proportions method on agar56.\n\nTo replicate our findings, all the analyses were repeated in a publicly available independent dataset of 1027 isolates from Samara, Russia (PRJEB2138), as the sampling was also representative of the population19. We also included a global data set of 1573 publicly available lineage 4 samples relatively enriched for drug resistance (Supplementary Data File\u00a03).\n\nQuality analysis of the raw reads was performed using FastQC57. A de-novo assembly of the short reads was done with SPAdes genome assembler v3.14.058 across kmers of size 21, 33, 45, 55, 65, 75, 81, 101, 111, and 121. The resulting assembly contigs were mapped to the well annotated H37Rv reference genome (Gene bank: AL123456) using minimap259 with the asm20 option. Single nucleotide polymorphisms (SNPs) and small insertions and deletions (indels) were identified with BCFtools mpileup and BCFtools call v1.960 using the multiallelic calling algorithm, keeping the information about every single site in the genome in a VCF file. Lastly, indels were left-aligned and normalized using BCFtools norm. A consensus sequence was created from the VCF file. In order to determine the quality of the variants, the raw reads were mapped against the resulting consensus sequence using the mem algorithm implemented in BWA v0.7.1761, after which the alignments were sorted using SAMtools v1.962 and filtered for possible PCR and optical duplicates using Picard v2.19.063. Local realignment around indels was performed using the GATK v3.8-1-0 \u2018IndelRealigner\u2019 module64. The mean coverage for each sample was calculated as the number of mapped bases (excluding soft-clipped bases) divided by the genome size. Samples with a mean coverage lower than 15x were excluded from subsequent analysis. SNPs and indels were detected as described in the previous step.\n\nVariants that did not meet the quality criteria were filtered using a combination of BCFtools filter and custom scripts in Python v3.7.3 with the following cutoffs: minimum Phred-scaled quality score (QUAL) of 20; minimum mapping quality (MQ) of 20; minimum genotype quality (QG) of 20; minimum read position bias (RPB), mapping quality bias (MQB), and strand bias (SP) of 0.001; minimum depth (DP) of 10 and a maximum of 5 times the mean coverage; minimum of reads supporting the alternate allele (AD) of 75% of the total depth in that position, with no less than two reads in the forward (ADF) and the reverse (ADR) strands. Additionally, SNPs within 2\u2009bp of an indel and indels within 3\u2009bp of another indel were removed, as both situations can be indicative of mapping artifacts. Positions that did not meet the quality criteria were annotated using the IUPAC ambiguity codes65. Samples with more than 25 high quality heterozygous calls were removed to avoid the inclusion of putative mixed infections. Variants that overlapped 100\u2009bp intervals around known hypervariable regions, such repetitive elements and transposases66, were removed from the analysis as this may affect the reliability of the alignment. Similarly, recombinant regions in genes coding for proline-glutamate (PE) or proline-proline-glutamate (PPE)67, and SNPs implicated in drug resistance33 were excluded in order to minimize homoplasies that could disrupt the tree topology. The resulting sequences were concatenated to generate a multiple sequence alignment. Sites with a proportion of ambiguous bases higher than 10% were excluded from the analysis. Last, samples with a proportion of ambiguous sites in the alignment higher than 5% were excluded.\n\nThe functional consequence of variants was assessed using the Variant Effect predictor (VEP) v104.368\n\nAll the phylogenetic analyses were performed based on the alignment containing both lineage 4 and lineage 2 samples, as well as separately for each lineage.\n\nA maximum likelihood phylogeny was inferred using RAxML-NG69 with the GTR model, 20 starting trees (10 random and 10 parsimony), 100 bootstrap replicates, and a minimum branch length of 10\u22129. A Lineage 2 sample randomly selected from our dataset was selected as an outgroup for the Lineage 4 phylogeny. Likewise, a random Lineage 4 isolate was used as a root for the Lineage 2 phylogeny. The tree containing both lineage 4 and lineage 2 samples was rooted using a lineage 1 isolate.\n\nIn order to construct a time-calibrated phylogeny, we tested whether there was a detectable amount of evolutionary change between samples collected at different times30,31. This was done in lineage 4 and lineage 2 separately in order to avoid population structure confounding in the temporal signal27. Two different tests of the temporal signal were applied: the root-to-tip regression method and the date-randomization test28. For the former, BactDating29 was used to perform a linear regression between the collection dates and their root-to-tip genetic distance in the maximum likelihood tree. Additionally, we carried out a date-randomization test, where evolutionary rates estimated by BactDating29 were compared between the observed data set and 100 data sets obtained by permutation of sampling dates70.\n\nBactDating29 was used to time-calibrate the tree using the mixed model for 107 iterations to achieve both convergence of the MCMC chains and an effective sample size of at least 100.\n\nThe phylogenetic global context of the Peruvian isolates was investigated by subsampling the phylogenies and repeating the analysis alongside publicly available isolates representative of the global diversity of M. tuberculosis (Supplementary Data File\u00a02). To subsample the phylogenies, first a random sample was selected from each phylogenetic cluster with a branch length lower than 1 SNPs per genome. The phylogeny was then divided into clusters of samples 50 SNPs apart, and a maximum of 20 samples for lineage 2 and 5 samples for lineage 4 were randomly selected for each cluster, unless it consisted of only 1 sample, in which case it was ignored. The phylogeny with the Peruvian subsamples and the global representatives was inferred separately for lineage 2 and for lineage 4 as described above.\n\nThe subsequent phylogenetic analysis was performed using the R package ape71. Marginal reconstruction of the ancestral sequences was carried out by maximum likelihood as implemented in Phangorn72, including gaps and ambiguity codes to reflect prior probabilities of character states65.\n\nTime-to-event analysis was performed on the tree using the R package Survival73. The time was measured for all pairs of nodes as the distance between the older and the younger node in the time-calibrated phylogeny. An observation was defined as censored if both nodes were drug sensitive. On the other hand, an event occurred if the older node was drug sensitive and the younger node was drug resistant. Only the first acquisition of resistance was considered. Observations taking place before 1940 were discarded. The Kaplan-Meier survival curve and the Cox proportional hazard ratio were calculated. The Kaplan-Meier curves for different groups were compared using the log-rank test, where the null hypothesis is that there is no difference in survival between the different groups. Differences in the hazard ratio were tested using the likelihood ratio test. The entire pipeline was repeated for 100 phylogenetic bootstrap replicates.\n\nMissing base calls at the tips were imputed by maximum likelihood using the re-rooting method74 and the IUPAC ambiguous codes to reflect tip state prior probabilities. In short, for each tip the phylogeny was re-rooted at that tip and the marginal probabilities for the missing bases were calculated for that node using the R package phytools75. Association analysis was performed at the gene and SNP level using the variant sites alignment for the tips of the phylogenetic tree, as well as the reconstructed sequences of the internal nodes. The phenotype was defined as leading to resistance in the phylogenetic tree, and thus only drug susceptible nodes that immediately preceded the first resistant node of each branch were considered. For the gene level analysis, non-synonymous variants were aggregated, excluding lineage specific SNPs. Loci with a frequency of non-synonymous variants in the dataset lower than 1% were not considered in the analysis. At the SNP-level, variants with a frequency in the population (tips of the tree) lower than 1% were excluded. Furthermore, only those variants that were polymorphic at the node level were used. Genome-wide association was performed using the Cox proportional hazard model and the time between nodes. In order to correct for population structure, a genetic distance matrix was calculated using SNPs with a frequency in the population higher than 5%, and the eigenvectors were used as covariates in the Cox regression model. The genomic inflation factor (\u03bb) was calculated as the ratio of the median of the empirically observed \u03c72 to the median of the expected \u03c72. The p-values were corrected for multiple testing using a Bonferroni correction. Functional annotation of the genomic variants was assessed using Mycobrowser42. Alignments were visually inspected for a random selection of samples using the Integrative genomics viewer (IGV)76 and the R software package Gviz77.\n\nFurther information on research design is available in the\u00a0Nature Research Reporting Summary linked to this article.",
+            "section_image": []
+        },
+        {
+            "section_name": "Data availability",
+            "section_text": "All raw sequencing data are available with accession numbers listed in the Methods section. Samples sequenced as part of this study have been submitted to the European Nucleotide Archive under accessions PRJEB39837 and PRJEB47846.\n\nPublicly available datasets used in this study include PRJEB5280, PRJEB32234, PRJEB23245, and PRJEB2138.\n\nAll other publicly available datasets are listed in Supplementary Data File\u00a02 and Supplementary Data File\u00a03.",
+            "section_image": []
+        },
+        {
+            "section_name": "Code availability",
+            "section_text": "All custom code used in this article can be accessed at https://github.com/arturotorreso/mtb_pre-resistance.git.",
+            "section_image": []
+        },
+        {
+            "section_name": "References",
+            "section_text": "Paulson, T. Epidemiology: a mortal foe. Nature 502, S2\u2013S3 (2013).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWorld Health Organization. Global tuberculosis report 2020 (2020).\n\nRaviglione, M. XDR-TB: Entering the post-antibiotic era? Int. J. Tuberculosis Lung Dis. 10, 1185\u20131187 (2006).\n\n\n                    Google Scholar\u00a0\n                \n\nWalker, T. M. et al. Whole-genome sequencing for prediction of Mycobacterium tuberculosis drug susceptibility and resistance: a retrospective cohort study. Lancet Infect. Dis. 15, 1193\u20131202 (2015).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGandhi, N. R. et al. Multidrug-resistant and extensively drug-resistant tuberculosis: a threat to global control of tuberculosis. Lancet 375, 1830\u20131843 (2010).\n\nArticle\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGygli, S. M., Borrell, S., Trauner, A. & Gagneux, S. Antimicrobial resistance in Mycobacterium tuberculosis: mechanistic and evolutionary perspectives. FEMS Microbiol. Rev. 41, 354\u2013373 (2017).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nFord, C. B. et al. Mycobacterium tuberculosis mutation rate estimates from different lineages predict substantial differences in the emergence of drug-resistant tuberculosis. Nat. Genet. 45, 784\u2013790 (2013).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nCox, H. S. et al. The Beijing genotype and drug resistant tuberculosis in the Aral Sea region of Central Asia. Respiratory Res. 6, 134 (2005).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nDrobniewski, F. Drug-resistant tuberculosis, clinical virulence, and the dominance of the Beijing strain family in Russia. JAMA 293, 2726 (2005).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nTaype, C. et al. Genetic diversity, population structure and drug resistance of Mycobacterium tuberculosis in Peru. Infect. Genet. Evolution 12, 577\u2013585 (2012).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nPang, Y. et al. Spoligotyping and drug resistance analysis of Mycobacterium tuberculosis strains from national survey in China. PLoS ONE 7, e32976 (2012).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWerngren, J. & Hoffner, S. E. Drug-susceptible Mycobacterium tuberculosis Beijing genotype does not develop mutation-conferred resistance to rifampin at an elevated rate. J. Clin. Microbiol. 41, 1520\u20131524 (2003).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGlynn, J. R., Whiteley, J., Bifani, P. J., Kremer, K. & van Soolingen, D. Worldwide occurrence of Beijing/W strains of Mycobacterium tuberculosis: a systematic review. Emerg. Infect. Dis. 8, 843\u2013849 (2002).\n\nArticle\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBorrell, S. et al. Epistasis between antibiotic resistance mutations drives the evolution of extensively drug-resistant tuberculosis. Evolution Med. Public Health 2013, 65\u201374 (2013).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSaunders, N. J. et al. Deep resequencing of serial sputum isolates of Mycobacterium tuberculosis during therapeutic failure due to poor compliance reveals stepwise mutation of key resistance genes on an otherwise stable genetic background. J. Infect. 62, 212\u2013217 (2011).\n\nArticle\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMenzies, D. et al. Effect of duration and intermittency of Rifampin on tuberculosis treatment outcomes: a systematic review and meta-analysis. PLoS Med. 6, e1000146 (2009).\n\nArticle\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nEldholm, V. et al. Four decades of transmission of a multidrug-resistant Mycobacterium tuberculosis outbreak strain. Nat. Commun. 6, 7119 (2015).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nCohen, K. A. et al. Evolution of extensively drug-resistant tuberculosis over four decades: whole genome sequencing and dating analysis of Mycobacterium tuberculosis Isolates from KwaZulu-Natal. PLOS Med. 12, e1001880 (2015).\n\nArticle\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nCasali, N. et al. Evolution and transmission of drug-resistant tuberculosis in a Russian population. Nat. Genet. 46, 279\u2013286 (2014).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nManson, A. L. et al. Genomic analysis of globally diverse Mycobacterium tuberculosis strains provides insights into the emergence and spread of multidrug resistance. Nat. Genet. 49, 395\u2013402 (2017).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGrenfell, B. T. Unifying the epidemiological and evolutionary dynamics of pathogens. Science 303, 327\u2013332 (2004).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGalvani, A. P. Epidemiology meets evolutionary ecology. Trends Ecol. Evolution 18, 132\u2013139 (2003).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGrandjean, L. et al. Convergent evolution and topologically disruptive polymorphisms among multidrug-resistant tuberculosis in Peru. PLoS ONE 12, e0189838 (2017).\n\nArticle\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nColl, F. et al. A robust SNP barcode for typing Mycobacterium tuberculosis complex strains. Nat. Commun. 5, 4812 (2014).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nAjawatanawong, P. et al. A novel ancestral Beijing sublineage of Mycobacterium tuberculosis suggests the transition site to modern Beijing sublineages. Sci. Rep. 9, 1\u201312 (2019).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWorld Health Organization. Global tuberculosis report 2009 (2009).\n\nMurray, G. G. R. et al. The effect of genetic structure on molecular dating and tests for temporal signal. Methods Ecol. Evol. 7, 80\u201389 (2016).\n\nArticle\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nRieux, A. & Balloux, F. Inferences from tip-calibrated phylogenies: a review and a practical guide. Mol. Ecol. 25, 1911\u20131924 (2016).\n\nArticle\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nDidelot, X., Croucher, N. J., Bentley, S. D., Harris, S. R. & Wilson, D. J. Bayesian inference of ancestral dates on bacterial phylogenetic trees. Nucleic Acids Res. 46, e134\u2013e134 (2018).\n\nArticle\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nDrummond, A. J., Pybus, O. G., Rambaut, A., Forsberg, R. & Rodrigo, A. G. Measurably evolving populations. Trends Ecol. Evol. 18, 481\u2013488 (2003).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBiek, R., Pybus, O. G., Lloyd-Smith, J. O. & Didelot, X. Measurably evolving pathogens in the genomic era. Trends Ecol. Evol. 30, 306\u2013313 (2015).\n\nArticle\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMenardo, F., Duch\u00eane, S., Brites, D. & Gagneux, S. The molecular clock of Mycobacterium tuberculosis. PLOS Pathog. 15, e1008067 (2019).\n\nArticle\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nColl, F. et al. Rapid determination of anti-tuberculosis drug resistance from whole-genome sequences. Genome Med. 7, 51 (2015).\n\nArticle\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKeshavjee, S. & Farmer, P. E. Tuberculosis, drug resistance, and the history of modern medicine. N. Engl. J. Med. 367, 931\u2013936 (2012).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGagneux, S. et al. Variable host-pathogen compatibility in Mycobacterium tuberculosis. Proc. Natl Acad. Sci. USA 103, 2869\u20132873 (2006).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSherman, D. R. et al. Compensatory ahpC gene expression in isoniazid-resistant Mycobacterium tuberculosis. Science 272, 1641\u20131643 (1996).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nComas, I. et al. Whole-genome sequencing of rifampicin-resistant Mycobacterium tuberculosis strains identifies compensatory mutations in RNA polymerase genes. Nat. Genet.\u200944, 106\u2013110 (2011).\n\nArticle\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGrandjean, L. et al. Transmission of multidrug-resistant and drug-susceptible tuberculosis within households: a prospective cohort study. PLOS Med. 12, e1001843 (2015).\n\nArticle\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWarren, J. L. et al. Investigating spillover of multidrug-resistant tuberculosis from a prison: a spatial and molecular epidemiological analysis. BMC Med. 16, 122 (2018).\n\nArticle\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nRengarajan, J., Bloom, B. R. & Rubin, E. J. From the cover: genome-wide requirements for Mycobacterium tuberculosis adaptation and survival in macrophages. Proc. Natl Acad. Sci. 102, 8327\u20138332 (2005).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nTufariello, J. M. et al. Separable roles for Mycobacterium tuberculosis ESX-3 effectors in iron acquisition and virulence. Proc. Natl Acad. Sci. USA 113, E348\u2013E357 (2016).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKapopoulou, A., Lew, J. M. & Cole, S. T. The MycoBrowser portal: a comprehensive and manually annotated resource for mycobacterial genomes. Tuberculosis 91, 8\u201313 (2011).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nVan der Heijden, Y. F. et al. Isoniazid-monoresistant tuberculosis is associated with poor treatment outcomes in Durban, South Africa. Int. J. Tuberculosis Lung Dis. 21, 670\u2013676 (2017).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nAhmad, S. & Mokaddas, E. Current status and future trends in the diagnosis and treatment of drug-susceptible and multidrug-resistant tuberculosis. J. Infect. Public Health 7, 75\u201391 (2014).\n\nArticle\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nNguyen, T. N. A., Anton-Le Berre, V., Ba\u00f1uls, A.-L. & Nguyen, T. V. A. Molecular diagnosis of drug-resistant tuberculosis; a literature review. Front. Microbiol. 10, 794 (2019).\n\nArticle\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHicks, N. D. et al. Clinically prevalent mutations in Mycobacterium tuberculosis alter propionate metabolism and mediate multidrug tolerance. Nat. Microbiol. 3, 1032\u20131042 (2018).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nTsolaki, A. G. et al. Functional and evolutionary genomics of Mycobacterium tuberculosis: Insights from genomic deletions in 100 strains. Proc. Natl Acad. Sci. USA 101, 4865\u20134870 (2004).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMohanty, S. et al. Mycobacterium tuberculosis EsxO (Rv2346c) promotes bacillary survival by inducing oxidative stress mediated genomic instability in macrophages. Tuberculosis 96, 44\u201357 (2016).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGonzales, M. J. Chinese plantation workers and social conflict in Peru in the late nineteenth century. J. Latin Am. Studies 21, 385\u2013424 (1989).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nRitacco, V. et al. Mycobacterium tuberculosis strains of the Beijing genotype are rarely observed in tuberculosis patients in South America. Mem. do Inst. Oswaldo Cruz 103, 489\u2013492 (2008).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBrynildsrud, O. B. et al. Global expansion of Mycobacterium tuberculosis lineage 4 shaped by colonial migration and local adaptation. Sci. Adv. 4, eaat5869 (2018).\n\nArticle\u00a0\n    ADS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMoore, D. A. et al. Microscopic-observation drug-susceptibility assay for the diagnosis of TB. N. Engl. J. Med. 355, 1539\u20131550 (2006).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nCaviedes, L. et al. Rapid, efficient detection and drug susceptibility testing of Mycobacterium tuberculosis in sputum by microscopic observation of broth cultures. J. Clin. Microbiol. 38, 1203\u20131208 (2000).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nThe CRyPTIC Consortium; 100000 Genomes Project. Prediction of susceptibility to first-line tuberculosis drugs by DNA sequencing. N. Engl. J. Med. 379, 1403\u20131415 (2018).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nUgarte-Gil, C. et al. Diabetes mellitus among pulmonary tuberculosis patients from 4 tuberculosis-endemic countries: the tandem study. Clin. Infect. Dis. 70, 780\u2013788 (2020).\n\nPubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nClinical and Laboratory Standards Institute (CLSI). Susceptibility Testing of Mycobacteria, Nocardiae, and Other Aerobic Actinomycetes; Approved Standard\u2014Second Edition tech. rep. (2011).\n\nAndrews, S. & Babraham Bioinformatics. FastQC: a quality control tool for high throughput sequence data. Manual (2010).\n\nBankevich, A. et al. SPAdes: a new genome assembly algorithm and its applications to single-cell sequencing. J. Comput. Biol. 19, 455\u2013477 (2012).\n\nArticle\u00a0\n    MathSciNet\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLi, H. Minimap2: Pairwise alignment for nucleotide sequences. Bioinformatics 34, 3094\u20133100 (2018).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nDanecek, P. et al. The variant call format and VCFtools. Bioinformatics 27, 2156\u20132158 (2011).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLi, H. & Durbin, R. Fast and accurate long-read alignment with Burrows\u2013Wheeler transform. Bioinformatics 26, 589\u2013595 (2010).\n\nArticle\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLi, H. et al. The sequence alignment/map format and SAMtools. Bioinformatics 25, 2078\u20132079 (2009).\n\nArticle\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nThomer, A. K., Twidale, M. B., Guo, J. & Yoder, M. J. Picard Tools in Conference on Human Factors in Computing Systems - Proceedings (2016).\n\nMcKenna, A. et al. The genome analysis toolkit: a MapReduce framework for analyzing next-generation DNA sequencing data. Genome Res. 20, 1297\u20131303 (2010).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nCornish-Bowden, A. Nomenclature for incompletely specified bases in nucleic acid sequences: recommendations 1984. Nucleic Acids Res. 13, 3021\u20133030 (1985).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nCole, S. T. et al. Deciphering the biology of Mycobacterium tuberculosis from the complete genome sequence. Nature 393, 537\u2013544 (1998).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nPhelan, J. E. et al. Recombination in pe/ppe genes contributes to genetic variation in Mycobacterium tuberculosis lineages. BMC Genomics 17, 151 (2016).\n\nArticle\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMcLaren, W. et al. The Ensembl variant effect predictor. Genome Biol. 17, 1\u201314 (2016).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKozlov, A. M., Darriba, D., Flouri, T., Morel, B. & Stamatakis, A. RAxML-NG: a fast, scalable and user-friendly tool for maximum likelihood phylogenetic inference. Bioinformatics 35, 4453\u20134455 (2019).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nDuch\u00eane, S., Duch\u00eane, D., Holmes, E. C. & Ho, S. Y. The performance of the date-randomization test in phylogenetic analyses of time-structured virus data. Mol. Biol. Evolution 32, 1895\u20131906 (2015).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nParadis, E., Claude, J. & Strimmer, K. APE: analyses of phylogenetics and evolution in R language. Bioinformatics 20, 289\u2013290 (2004).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSchliep, K. P. phangorn: phylogenetic analysis in R. Bioinformatics 27, 592\u2013593 (2011).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nTerry M. Therneau & Patricia M. Grambsch. Modeling Survival Data: Extending the Cox Model (Springer, 2000).\n\nYang, Z., Kumar, S. & Nei, M. A new method of inference of ancestral nucleotide and amino acid sequences. Genetics 141, 1641\u20131650 (1995).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nRevell, L. J. phytools: an R package for phylogenetic comparative biology (and other things). Methods Ecol. Evol. 3, 217\u2013223 (2012).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nRobinson, J. T. et al. Integrative genomics viewer. Nat. Biotechnol. 29, 24\u201326 (2011).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHahne, F. & Ivanek, R. in Statistical Genomics: Methods and Protocols 335\u2013351 (2016).\n\nDownload references",
+            "section_image": []
+        },
+        {
+            "section_name": "Acknowledgements",
+            "section_text": "We would like to thank the participants of the study. LG was supported by the Wellcome Trust (201470/Z/16/Z), the National Institute of Allergy and Infectious Diseases of the National Institutes of Health under award number 1R01AI146338 and by the GOSH/ICH Biomedical Research Centre. OMK was supported by the Imperial Biomedical Research Centre (NIHR Imperial BRC, grant P45058). XD was supported by the NIHR Health Protection Research Unit in Genomics and Enabling Data. We thank the CRyPTIC project and the Tandem project for making whole-genome data available in the public domain. All authors acknowledge UCL Computer Science Technical Support Group (TSG) and the UCL Department of Computer Science High Performance Computing Cluster.",
+            "section_image": []
+        },
+        {
+            "section_name": "Author information",
+            "section_text": "Imperial College London, Department of Infectious Diseases, London, UK\n\nArturo Torres Ortiz\u00a0&\u00a0Louis Grandjean\n\nUniversidad Peruana Cayetano Heredia, Lima, Per\u00fa\n\nJorge Coronel\n\nUnidad T\u00e9cnica de Tuberculosis MDR, Ministerio de Salud, Lima, Per\u00fa\n\nJulia Rios Vidal\u00a0&\u00a0Cesar Bonilla\n\nUniversidad Privada San Juan Bautista, Lima, Per\u00fa\n\nCesar Bonilla\n\nLondon School of Hygiene and Tropical Medicine, London, UK\n\nDavid A. J. Moore\n\nJohns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA\n\nRobert H. Gilman\n\nUCL Genetics Institute, London, UK\n\nFrancois Balloux\n\nRespiratory Medicine, National Heart and Lung Institute, Imperial College London, London, UK\n\nOnn Min Kon\n\nUniversity of Warwick, School of Life Sciences and Department of Statistics, Warwick, UK\n\nXavier Didelot\n\nUCL Department of Infection, Institute of Child Health, London, UK\n\nLouis Grandjean\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nA.T.O., L.G., X.D., O.M.K., F.B., J.R.V., R.G., C.B. and D.M. conceived and designed the study. J.C. performed and supervised tuberculosis cultures, DNA extraction, laboratory, and sequencing work. L.G., X.D., F.B. and O.M.K. jointly supervised the research. A.T.O., L.G., F.B. and X.D. performed and advised on computational analyses. A.T.O. and L.G. wrote the manuscript with input from all co-authors. All authors read and approved the final manuscript.\n\nCorrespondence to\n                Louis Grandjean.",
+            "section_image": []
+        },
+        {
+            "section_name": "Ethics declarations",
+            "section_text": "The authors declare no competing interests.",
+            "section_image": []
+        },
+        {
+            "section_name": "Peer review information",
+            "section_text": "Nature Communications thanks the anonymous reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available.",
+            "section_image": []
+        },
+        {
+            "section_name": "Additional information",
+            "section_text": "Publisher\u2019s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.",
+            "section_image": []
+        },
+        {
+            "section_name": "Rights and permissions",
+            "section_text": "Open Access  This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article\u2019s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article\u2019s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.\n\nReprints and permissions",
+            "section_image": []
+        },
+        {
+            "section_name": "About this article",
+            "section_text": "Torres Ortiz, A., Coronel, J., Vidal, J.R. et al. Genomic signatures of pre-resistance in Mycobacterium tuberculosis.\n                    Nat Commun 12, 7312 (2021). https://doi.org/10.1038/s41467-021-27616-7\n\nDownload citation\n\nReceived: 18 May 2021\n\nAccepted: 29 November 2021\n\nPublished: 15 December 2021\n\nVersion of record: 15 December 2021\n\nDOI: https://doi.org/10.1038/s41467-021-27616-7\n\nAnyone you share the following link with will be able to read this content:\n\nSorry, a shareable link is not currently available for this article.\n\n\n\n\n                            Provided by the Springer Nature SharedIt content-sharing initiative\n                        ",
+            "section_image": [
+                "https://data:image/svg+xml;base64,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"
+            ]
+        },
+        {
+            "section_name": "This article is cited by",
+            "section_text": "Nature Reviews Microbiology (2025)\n\nScientific Reports (2025)\n\nScientific Reports (2025)\n\nScientific Reports (2025)\n\nScientific Reports (2024)",
+            "section_image": []
+        }
+    ]
+}

0b99a6927a5c75064b7bc56db653ee830f9e7bf163e4237af0445d8c8895bd96/metadata.json

@@ -0,0 +1,139 @@
+{
+    "title": "Fourier-based three-dimensional multistage transformer for aberration correction in multicellular specimens",
+    "pre_title": "Fourier-Based 3D Multistage Transformer for Aberration Correction in Multicellular Specimens",
+    "journal": "Nature Methods",
+    "published": "01 October 2025",
+    "supplementary_0": [
+        {
+            "label": "Supplementary Information",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41592-025-02844-7/MediaObjects/41592_2025_2844_MOESM1_ESM.pdf"
+        },
+        {
+            "label": "Reporting Summary",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41592-025-02844-7/MediaObjects/41592_2025_2844_MOESM2_ESM.pdf"
+        },
+        {
+            "label": "Peer Review file",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41592-025-02844-7/MediaObjects/41592_2025_2844_MOESM3_ESM.pdf"
+        }
+    ],
+    "supplementary_1": NaN,
+    "supplementary_2": NaN,
+    "source_data": [
+        "https://github.com/cell-observatory/aovift",
+        "https://github.com/cell-observatory/beads_simulator"
+    ],
+    "code": [
+        "https://github.com/cell-observatory/aovift",
+        "https://github.com/cell-observatory/aovift/pkgs/container/aovift",
+        "https://github.com/abcucberkeley/PetaKit5D"
+    ],
+    "subject": [
+        "Computational models",
+        "Machine learning"
+    ],
+    "license": "http://creativecommons.org/licenses/by/4.0/",
+    "preprint_pdf": "https://www.researchsquare.com/article/rs-6273247/v1.pdf?c=1759403703000",
+    "research_square_link": "https://www.researchsquare.com//article/rs-6273247/v1",
+    "nature_pdf": "https://www.nature.com/articles/s41592-025-02844-7.pdf",
+    "preprint_posted": "01 Apr, 2025",
+    "research_square_content": [
+        {
+            "section_name": "Abstract",
+            "section_text": "High-resolution tissue imaging is often compromised by sample-induced optical aberrations that degrade resolution and contrast. While wavefront sensor-based adaptive optics (AO) can measure these aberrations, such hardware solutions are typically complex, expensive to implement, and slow when serially mapping spatially varying aberrations across large fields of view. Here, we introduce AOViFT (Adaptive Optical Vision Fourier Transformer)---a machine learning-based aberration sensing framework built around a 3D multistage Vision Transformer that operates on Fourier domain embeddings. AOViFT infers aberrations and restores diffraction-limited performance in puncta-labeled specimens with substantially reduced computational cost, training time, and memory footprint compared to conventional architectures or real-space networks. We validated AOViFT on live gene-edited zebrafish embryos, demonstrating its ability to correct spatially varying aberrations using either a deformable mirror or post-acquisition deconvolution. By eliminating the need for the guide star and wavefront sensing hardware and simplifying the experimental workflow, AOViFT lowers technical barriers for high-resolution volumetric microscopy across diverse biological samples.Biological sciences/Computational biology and bioinformatics/Machine learningBiological sciences/Computational biology and bioinformatics/Computational models",
+            "section_image": []
+        },
+        {
+            "section_name": "Additional Declarations",
+            "section_text": "There is NO Competing Interest.",
+            "section_image": []
+        }
+    ],
+    "nature_content": [
+        {
+            "section_name": "Abstract",
+            "section_text": "High-resolution tissue imaging is often compromised by sample-induced optical aberrations that degrade resolution and contrast. Although wavefront sensor-based adaptive optics (AO) can measure these aberrations, such hardware solutions are typically complex, expensive to implement and slow when serially mapping spatially varying aberrations across large fields of view. Here we introduce AOViFT (adaptive optical vision Fourier transformer)\u2014a machine learning-based aberration sensing framework built around a three-dimensional multistage vision transformer that operates on Fourier domain embeddings. AOViFT infers aberrations and restores diffraction-limited performance in puncta-labeled specimens with substantially reduced computational cost, training time and memory footprint compared to conventional architectures or real-space networks. We validated AOViFT on live gene-edited zebrafish embryos, demonstrating its ability to correct spatially varying aberrations using either a deformable mirror or postacquisition deconvolution. By eliminating the need for the guide star and wavefront sensing hardware and simplifying the experimental workflow, AOViFT lowers technical barriers for high-resolution volumetric microscopy across diverse biological samples.",
+            "section_image": []
+        },
+        {
+            "section_name": "Main",
+            "section_text": "As we peer deeper into living organisms to reveal their inner workings, our view is increasingly compromised by sample-induced optical aberrations. Numerous AO methods exist to compensate for these by using a wavefront shaping device that responds to a measurement of sample-induced aberration1. These methods differ in their complexity, generality, robustness and practicality. In our laboratory, dependable success was had using a Shack\u2013Hartmann (SH) sensor to measure the aberrations imparted on a guide star (GS) created by two-photon excitation (TPE) fluorescence within the specimen2, and we have used this approach extensively in adaptive optical (AO) lattice light sheet (LLS) microscopy (AO-LLSM) to study four-dimensional (4D) subcellular dynamic processes within the native environment of whole multicellular organisms3.\n\nSeveral recent approaches dispense with the cost and complexity of hardware-based wavefront measurement in favor of directly inferring aberrations from the microscope images themselves through machine learning (ML)4,5,6,7,8 (Supplementary Table 1). Based on our experience with a variety of specimens, any ML-AO approach suitable for AO-LLSM must meet the following specifications:\n\nSpeed: to maximize the range of spatiotemporal events that can be visualized, the time for the ML model to infer the aberrations across any volume should be less than the time needed to image it\u2014typically a few seconds in LLSM for a volume that encompasses a handful of cells.\n\nRobustness: the model must accurately predict the vast majority of aberrations encountered in practice\u2014for AO-LLSM in zebrafish embryos, typically up to 5\u03bb peak-to-valley (P\u2013V), where \u03bb is the free-space wavelength, in any combination of the first 15 Zernike modes (\\({Z}_{0}^{\\,0}\\) through \\({Z}_{4}^{\\,\\pm 4}\\) Supplementary Fig. 1).\n\nAccuracy: the method should be able to recover close to the theoretical three-dimensional (3D) resolution limits of the microscope, regardless of the distribution of spatial frequencies within the specimen.\n\nNoninvasiveness: the method should provide accurate correction without unduly depleting the fluorescence photon budget within the specimen or perturbing its native physiology.\n\nAs none of the aforementioned ML-AO methods meet all these specifications, we endeavored to create one better suited to the needs of AO-LLSM. Our baseline model architecture, selected from an ablation study (Supplementary Note A, Supplementary Figs. 2\u20137), contains two transformer stages with patches of 32 and 16 pixels, respectively (Fig. 1c).\n\na, AOViFT correction. An aberrated 3D volume is preprocessed and cast into a Fourier embedding, which is passed to a 3D vision transformer model to predict the detection wavefront. A DM compensates for this aberration, enabling acquisition of a corrected volume (D, depth; H, height; W, width). b, The Fourier embedding, \\({\\mathcal{E}}\\). The Fourier transform of the 3D volume is embedded into a lower space (\\({\\mathcal{E}}\\in {{\\mathbb{R}}}^{\\ell \\times d\\times d}\\)), consisting of three amplitude planes (\u03b11, \u03b12, \u03b13) and three phase planes (\u03c61, \u03c62, \u03c63), each of size d \u00d7 d where d is the Fourier embedding size. c, AOViFT model. The Fourier embedding is input to a dual-stage 3D vision transformer model. At each stage, the \u2113 Fourier planes are tiled into k patches (Patchify), applying a radially encoded positional embedding to each patch. These patches are passed through n Transformer layers. At the end of each stage, a residual connection is added, and the patches are merged back to the shape matching the stage input (Merge patches). After all stages, the resulting patches are pooled (GlobalAvgPool) and connected with a dense layer to output the z Zernike coefficients.\n\nPriors can greatly improve the performance of any ML approach. For our method, we depend on the prior that each isoplanatic subvolume (that is, having the same aberration) within the larger volume of interest contains one or more fluorescent puncta of true subdiffractive size. Here we introduce these by using genome-edited specimens expressing fluorescent protein-fused versions of AP2\u2014an adapter protein that targets clathrin-coated pits (CCPs) ubiquitously present at CCPs located on the plasma membrane of all cells (Methods). While this entails a one-time upfront cost for each specimen type, it noninvasively produces a robust signal for AO correction that does not preclude simultaneously imaging another subcellular target that occupies the same fluorescence channel, provided they are computationally separable9.",
+            "section_image": [
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41592-025-02844-7/MediaObjects/41592_2025_2844_Fig1_HTML.png"
+            ]
+        },
+        {
+            "section_name": "Results",
+            "section_text": "We created five variants of AOViFT by varying the numbers of layers and heads in each stage (Supplementary Table 2) to explore the tradeoffs between model size (number of parameters and memory footprint), speed (floating-point operations (FLOPs) required, training time and latency) and prediction accuracy (Supplementary Fig. 3). To compare these to existing state-of-the-art architectures, we developed 3D versions of ViT and ConvNeXt for AO inference in three and four different size variants, respectively (Supplementary Note B). We trained all models with the same set of 2\u2009\u00d7\u2009106 synthetic image volumes chosen to capture the full diversity of aberrations and imaging conditions likely to be encountered in AO-LLSM (Methods) and tested AOViFT on a separate set of 105 image volumes created to find the limits of its accuracy when presented with an even larger range of aberration magnitudes, signal-to-noise ratio (SNR) and number of fluorescent puncta (Supplementary Note C). We also tested the performance of all models and variants on 104 image volumes from a test set that contained only a single punctum in each (Fig. 2, Supplementary Figs. 8\u20139 and Supplementary Table 3).\n\na\u2013d, Comparison of model variants ConvNeXt-T/S/B/L (blue), ViT/16-S/B (orange) and AOViFT-T/S/B/L/H (gray). a, Total number of trainable parameters. b, Maximum predictions per second, using a batch size of 1,024 on a single A100 GPU. Higher values are better. c, Training time on eight H100 GPUs. d, Median \u03bb RMS residuals over 10,000 test samples after one correction, with aberrations ranging between 0.2\u03bb and 0.4\u03bb, simulated with 50,000 to 200,000 integrated photons. e,f, Median \u03bb RMS residuals using our Small variant of AOViFT model for a single bead over a wide range of SNR. g,h, Median \u03bb RMS residuals using our Small variant of AOViFT model for several beads (up to 150 beads), simulated at photon levels from 50,000 to 200,000 per bead. Lower values are better for all performance indicators listed here, except for b. CDF, cumulative distribution function; KDE, kernel density estimation.\n\nAlthough all models but the smallest variant of ConvNeXt were able to reduce the median residual error in a single iteration of aberration prediction to less than the diffraction limit (Fig. 2d), AOViFT excelled in its parsimonious use of compute resources: training time using a node with eight NVIDIA H100 GPUs (Fig. 2c), training FLOPs (Supplementary Fig. 8c) and memory footprint (Supplementary Fig. 8f). This reflects the benefits of our multistage architecture: faster convergence by learning features across different scales, accurate prediction even at comparatively modest model size (Fig. 2a), highest inference rate among the models tested (Fig. 2b) and fastest single-shot inference time (\u2018latency\u2019; Supplementary Fig. 8h). Given its small size and low latency, we chose the Small variant of AOViFT as our primary model for evaluation.\n\nDiffraction-limited performance is defined conventionally by wavefront distortions below \u22480.075\u03bb root mean square (RMS) or \u03bb/4 peak-to-valley, corresponding to a Strehl ratio of 0.8 under the Rayleigh quarter-wave criterion10,11. In silico evaluation using the 104 single punctum test images shows that AOViFT recovers diffraction-limited performance in a single iteration in nearly all trials where the initial aberration is <0.30\u03bb\u2009RMS and the integrated signal is >5\u2009\u00d7\u2009104 photons (Fig. 2e). The corrective range increases to 0.40, 0.50, 0.55 and 0.6\u03bb\u2009RMS for two to five iterations, respectively, although ~5\u2009\u00d7\u2009104 photons remains the floor of required signal (Fig. 2f and Supplementary Fig. 10). This is comparable to the signal needed for SH wavefront sensing2,3 and at least three times lower than that needed for PhaseRetrieval. In comparison, PhaseNet4 and PhaseRetrieval12 extend the diffraction-limited range only slightly (initial aberration <0.15\u03bb\u2009RMS) after a single iteration on the same test data (Supplementary Fig. 11a,b) and, in contrast to AOViFT, do not appreciably increase this range after several iterations (Supplementary Fig. 12a\u2013f). PhaseRetrieval does advantageously reduce residuals after a single iteration over a much broader range of initial aberration than AOViFT, and this trend continues with further iteration, albeit never back to the diffraction limit (Supplementary Fig. 12a\u2013c). However, this advantage is lost if the fiducial bead is not centered in the field of view (FOV), and the predictive power of PhaseNet is lost completely under the same circumstances (Supplementary Fig. 11d,e) because the widefield 3D image of the bead is then clipped. Furthermore, PhaseRetrieval and PhaseNet assume a priori the existence of only a single bead. AOViFT is trained on one of five puncta falling anywhere within the FOV but, thanks to the normalization step to eliminate phase fringes from several puncta (Supplementary Fig. 24dd), produces inferences comparably accurate to a single punctum for up to 150 puncta, provided their mean nearest neighbor distance is >400\u2009nm (Fig. 2g,h and Supplementary Fig. 13). Indeed, AOViFT relies on the combined signal of several native but dim subdiffractive biological assemblies such as CCPs to achieve accurate inferences.\n\nWe performed all experiments using the AO-LLSM microscope schematized in Supplementary Fig. 14. For initial characterization of the ability of AOViFT to correct a wide range of possible aberrations, we performed 66 separate experiments wherein we:\n\nintroduced aberration by applying to the deformable mirror (DM) one of the 66 possible combinations of one or two Zernike modes (from the first 15, excluding piston, tip/tilt and defocus), with each mode set to 0.2\u03bb\u2009RMS amplitude\n\nused AO-LLSM with the MBSq-35 LLS excitation profile of Supplementary Table 6 to image a field of 100\u2009nm diameter fluorescent beads with this aberration;\n\nused AOViFT to predict the aberration\n\napplied the corrective pattern to the DM\n\nrepeated (1)\u2013(5) for five iterations\n\nIn 45 cases, we recovered diffraction-limited performance in two iterations (Fig. 3) and, in five iterations for 11 more cases (Supplementary Fig. 15). In the remaining ten cases, aberrations were reduced by at least 50% after five iterations.\n\na\u2013l, Four examples where the initial aberration was applied artificially by the DM. a,b,c, O-Astig and H-Coma \\(({Z}_{n = 2}^{\\,m = \\,{-}\\!2}+{Z}_{n = 3}^{\\,m = 1})\\); d,e,f, O-Quadrafoil and P-Spherical \\(({Z}_{n = 4}^{\\,m = \\,{-}\\!4}+{Z}_{n = 4}^{\\,m = 0})\\); g,h,i, V-Astig and V-Trefoil \\(({Z}_{n = 2}^{\\,m = 2}+{Z}_{n = 3}^{\\,m = \\,{-}\\!3})\\); j,k,l, V-Coma and O-Astig2 \\(({Z}_{n = 3}^{\\,m = \\,{-}1}+{Z}_{n = 4}^{\\,m =\\, {-}\\!2})\\). Iteration 0 shows XY maximum intensity projection (MIP) of four beads with initial aberration imaged using LLS, providing initial conditions for AOViFT predictions (a,d,g,j). Iteration\u20091 shows the resulting field of beads after applying AOViFT prediction to the DM (b,e,h,k). Iteration\u20092 shows the results after applying the AOViFT prediction measured from Iteration\u20091 (c,f,i,l). Insets: the AOViFT predicted wavefront over the NA\u2009=\u20091.0 pupil with a dashed line at NA\u2009=\u20090.85. m, Heatmap of the residual aberrations (measured by PhaseRetrieval on isolated bead) after applying AOViFT predictions, starting with a single Zernike mode up to Mode 14 (\\({Z}_{n = 4}^{\\,m = 4}\\)) across up to five iterations.\n\nWe next tested the ability of AOViFT to correct aberrations during live cell imaging under biologically relevant conditions of limited signal, dense puncta and specimen motion. To this end, we applied aberrations to the DM and imaged cultured SUM159 human breast-cancer-derived cells gene edited to produce endogenous levels of the clathrin adapter protein AP2 tagged with eGFP. This yielded numerous membrane-bound CCPs at various stages of maturation that were suitable for aberration measurement. In one example (Fig. 4a), we applied a 2.9\u03bb peak-to-valley (P\u2013V) aberration to the DM consisting of a mix of horizontal coma and oblique trefoil (\\({Z}_{3}^{1}\\) and \\({Z}_{3}^{3}\\)), and recovered near diffraction-limited performance after two iterations (Supplementary Table 4). Peak signal at the CCPs increased twofold to threefold postcorrection, and the spatial frequency content as seen in orthoslices through the 3D fast Fourier transform (FFT) (insets at bottom) increased in every iteration. In another case (Fig. 4b), we reduced a 3.1\u03bb P\u2013V aberration composed of a combination of horizontal coma and primary spherical (\\({Z}_{3}^{1}\\) and \\({Z}_{4}^{\\,0}\\)) to 0.069\u03bb\u2009RMS after two iterations, increasing CCP signal by threefold to fourfold (Supplementary Table 4). Four more examples of correction on cells and fiducial beads after applying single modes of 1\u03bb P\u2013V aberration are given in Supplementary Fig. 16 and five more examples of two-mode correction are shown in Supplementary Fig. 17.\n\na, 3D volume SUM159-AP2 cells represented as xyMIPs and yzMIPs covering a 15.7\u2009\u00d7\u200955.6\u2009\u00d7\u200925.6\u2009\u03bcm3 FOV after applying a 2.9\u03bb P\u2013V aberration to the DM. This aberration combines horizontal coma \\({Z}_{3}^{1}\\) and oblique trefoil \\({Z}_{3}^{3}\\). b, xyMIPs and yzMIPs of a similar FOV with 3.1\u03bb P\u2013V aberration composed of horizontal coma (\\({Z}_{3}^{1}\\)) and primary spherical (\\({Z}_{4}^{\\,0}\\)). In both cases, near diffraction-limited performance was recovered after two iterations. Insets: FFTs and corresponding wavefronts for each iteration.\n\nAs a transparent vertebrate, zebrafish are a popular model organism for imaging studies. However, the spatially heterogenous refractive index within multicellular organisms and the discontinuity of refractive index at their surfaces with respect to the imaging medium result in aberrations that vary throughout their interiors (Fig. 5a). We corrected a ~ 2\u03bb P\u2013V aberration in one such region (Fig. 5b, top) with AOViFT (Fig. 5b, bottom) near the notochord of a transgenic zebrafish embryo 72\u2009h postfertilization expressing AP2-mNeonGreen in CCPs at the membranes of all cells (ap2s1:ap2s1-mNeonGreenbk800; Methods) and recovered spatial frequencies across the corrected volume (FFTs at right) comparable to SH correction over the same region (Fig. 5b, middle).\n\na, xyMIP of a 72-hpf gene-edited zebrafish embryo expressing endogenous AP2-mNeonGreen, exhibiting native and spatially varying aberrations near the notochord. b, Enlarged view of the dashed blue box in a. The xyMIPs and yzMIPs, along with the corresponding FFTs of a 12.5\u2009\u00d7\u200912.5\u2009\u00d7\u200912.8\u2009\u03bcm3 FOV, show ~2\u03bb P\u2013V of sample-induced aberration without AO (top row), corrected by SH (middle row) and corrected by AOViFT (bottom row). The contrast for each volume was scaled to its 1st and 99.99th percentile intensity values. c, xyMIPs and yzMIPs of a different gene-edited zebrafish embryo expressing exogenous AP2-mNeonGreen and injected mRNA for mChilada-Cox8a (to visualize mitochondria). The AP2 signal was used to infer the underlying aberration, and the same correction was applied to both channels. Top: ~1.5\u03bb P\u2013V aberration; middle: AOViFT correction after two iterations. The third and fourth rows present the results of OMW deconvolution without and with AOViFT corrected volumes, respectively.\n\nIn a second embryo expressing AP2-mNeonGreen in CCPs and mChilada-Cox8a in mitochondria (Fig. 5c), we used the mNeonGreen signal to correct a ~1.5\u03bb P\u2013V aberration (top row) in one region, which provided an aberration-corrected view of both CCPs and mitochondria (second row). Deconvolution of the aberrated images using an assumed ideal point spread function (PSF) amplified only high frequency artifacts (third row), but provided a more accurate representation of sample structure (bottom row) for the aberration-corrected ones by compensating for known attenuation of high spatial frequencies in the ideal optical transfer function (OTF).\n\nWith GS-illuminated SH sensors, aberration measurement is not accurate unless it is confined to a single isoplanatic region. However, these are often much smaller than the volume of interest, and their boundaries are not generally known a priori. Consequently, microscopists are often forced to map aberrations by serial SH measurement over many small, tiled subregions whose dimensions are a matter of educated guesswork. On the other hand, with AOViFT we generated a complete map of 204 aberrations (Fig. 6a) at 6.3-\u03bcm intervals over 37\u2009\u00d7\u2009211\u2009\u00d7\u200912.8\u2009\u03bcm3 in a live zebrafish embryo 48\u2009hpf (Fig. 6b,d) in ~1.5\u2009min on a single node of four A100 graphics processing units (GPUs). Unfortunately, it is not possible to apply a corrective pattern to a single pupil conjugate DM and thereby correct this spatially varying aberration across the entire FOV. One option would be to apply each aberration in turn and image the tiles one by one, or together in groups of similar aberration. Although slow, this would recover the full information of which the microscope is capable. However, a much faster and simpler alternative is to deconvolve each raw image tile with its own unique aberrated PSF (Fig. 6c,e). This does not recover full diffraction-limited performance, but it does suppress aberration-induced artifacts and provides a more faithful representation of the underlying sample structure (Fig. 6f\u2013h).\n\na, Isoplanatic patch map determined by AOViFT for 204 tiles (6.3\u2009\u00d7\u20096.3\u2009\u00d7\u200912.8\u2009\u03bcm3 each), spanning 37\u2009\u00d7\u2009211\u2009\u00d7\u200912.8\u2009\u03bcm3 FOV in a live, gene-edited zebrafish embryo expressing endogenous AP2-mNeonGreen. The yellow box marks areas with insufficient spatial features to accurately determine aberrations; an ideal PSF was used for OMW deconvolution in these regions. b, xyMIP of the AP2 signal without AO. c, xyMIP of each tile after deconvolution with spatially varying PSFs predicted by AOViFT. d,e, Raw (d) and deconvolved (e) xyMIPs of the mitochondria channel for the same region. f, Enlarged view of the wavefronts within the black dashed box in a. g,h, Zoomed-in views of AP2 (g) and mitochondria (h) structures from b\u2013e, comparing No AO to OMW deconvolution using either an ideal PSF or spatially varying tile-specific aberrated PSFs predicted by AOViFT.",
+            "section_image": [
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41592-025-02844-7/MediaObjects/41592_2025_2844_Fig2_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41592-025-02844-7/MediaObjects/41592_2025_2844_Fig3_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41592-025-02844-7/MediaObjects/41592_2025_2844_Fig4_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41592-025-02844-7/MediaObjects/41592_2025_2844_Fig5_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41592-025-02844-7/MediaObjects/41592_2025_2844_Fig6_HTML.png"
+            ]
+        },
+        {
+            "section_name": "Discussion",
+            "section_text": "AOViFT provides accurate mapping of spatially varying sample-induced aberrations in specimens having subdiffractive puncta. Although AOViFT can be slower than using SH for a single region of interest, it gains a substantial net speed advantage when mapping several regions of interest across a large FOV due to its parallelizable inference framework (Supplementary Table S5). Moreover, its throughput can be further accelerated by distributed GPU processing across several nodes and by compiling the model with TensorRT (https://docs.nvidia.com/deeplearning/tensorrt/pdf/TensorRT-Developer-Guide.pdf) for optimized inference. Unlike AOViFT, SH measurement with a TPE GS has the key advantage of being agnostic to the fluorescence distribution within each isoplanatic region, but requires additional hardware (TPE laser, galvos, SH sensor) and the TPE power level must be carefully monitored to minimize photodamage. In addition, since the isoplanatic regions are not known a priori, the initial measurement grid for SH sensing must be very dense to accurately map aberrations and their rate of change across the FOV, requiring additional time at an additional cost to the photon budget. Conversely, AOViFT determines the aberration map from a single large 3D image volume, and can therefore iteratively adjust tile sizes or positions in silico as needed until the map converges to an accurate solution.\n\nAlthough trained for a specific LLS type (Supplementary Table 6), AOViFT retained predictive capability when tested in silico with other light sheets as well (Supplementary Fig. 20). Although training specifically for such light sheets might increase the predictive range even further, a more fruitful path might be to augment the synthetic training data with light sheets axially offset from the detection focal plane to replace the closed-loop hardware-based mitigation of such offsets needed now3. Future models might leverage ubiquitous subcellular markers, such as plasma membranes or organelles, rather than genetically expressed diffraction-limited puncta, provided these markers contain sufficient high spatial frequency content for accurate inference of aberrations. Finally, to enhance generalizability of AOViFT and reduce overfitting to narrowly defined imaging scenarios, future models should incorporate a more diverse range of light sheets, specimen types and labeling strategies.\n\nDevelopment of AOViFT highlighted the challenges of constructing a 3D transformer-based architecture for AO correction. Each iteration of model design, training and testing required specialized simulated data pipelines, large GPU resources and extensive hyperparameter tuning\u2014leading to lengthy model development cycles. A key bottleneck is the absence of universally applicable, large pretrained models for volumetric imaging data\u2014a limitation that extends beyond AO applications.\n\nUnlike the natural image domain, where ViT benefited from extensive training on standardized two-dimensional (2D) datasets, a comparable \u2018foundation model\u2019 for 3D microscopy is pending the collection of similar datasets. This gap severely limits how far and how quickly new methods like AOViFT can be generalized. Although our work highlights the feasibility of building a solution for a given task (for example, AO corrections under specific imaging conditions), adapting to new scenarios such as new sample types, microscope geometries or aberration ranges typically requires substantial retraining and additional data curation.\n\nThese limitations highlight the need for pretrained foundation models in volumetric microscopy. We consider AOViFT\u2014a 3D vision transformer model\u2014as a stepping stone towards the more ambitious goal of creating a 4D model pretrained on massive volumetric microscopy datasets. Such a model could be fine-tuned for tasks across spatial (from molecules to organisms) and temporal (from stochastic molecular kinetics to embryonic development) scales13. Realizing this vision would require petabytes of high-quality curated 4D datasets and significant computational resources. However, successful implementation would dramatically shorten development timelines, improve generalization and reduce the overhead of custom training for varied experimental setups or microscope configurations.",
+            "section_image": []
+        },
+        {
+            "section_name": "Methods",
+            "section_text": "Imaging was performed using an AO-LLS microscope similar to one described previously3 (Supplementary Fig. 14 and Supplementary Table 7). Briefly, 488-nm and 560nm lasers (500\u2009mW 2RU-VFL-P-500-488-B1R and 1,000\u2009mW 2RU-VFL-P-1000-560-B1R, MPB Communications Inc.) were modulated using an acousto-optical tunable filter (Quanta-Tech, AA OptoElectronic, AOTFnC-400.650-CPCh-TN) and shaped into a stripe by a Powell lens (Laserline Optics Canada, LOCP-8.9R20-2.0) and a pair of 50- and 250-mm cylindrical lenses (25-mm diameter; Thorlabs, ACY254-050 and LJ1267RM-A). The stripe illuminated a reflective, phase-only, gray-scale spatial light modulator (Meadowlark Optics, AVR Optics, P1920-0635-HDMI; 1,920\u2009\u00d7\u20091,152 pixels) located at a sample conjugate plane. An eight-bit phase pattern written to the spatial light modulator generated the desired light-sheet pattern in the sample, and an annular mask (Thorlabs Imaging) at a pupil conjugate plane blocked unwanted diffraction orders before the light passed through the excitation objective (Thorlabs, TL20X-MPL). A pair of pupil conjugate galvanometer mirrors (Cambridge Technology, Novanta Photonics, 6SD11226 and 6SD11587) scanned the light sheet at the sample plane. The sample was positioned at the common foci of the excitation and detection objectives by a three-axis XYZ stage (Smaract; MLS-3252-S, SLS-5252-S and SLS-5252-S). Fluorescence emission from the sample was collected by a detection objective (Zeiss, \u00d720 1.0 numerical aperture (NA), 421452-9800-000), reflected off a pupil conjugate DM (ALPAO, DM69) that applied aberration corrections, and then recorded on two sample conjugate cameras (Hamamatsu ORCA Fusion).\n\nSH measurements (Supplementary Fig. 18) were performed on the same microscope by localizing the intensity maxima (on a Hamamatsu ORCA Fusion) formed by the emitted light after passage through a pupil conjugate lenslet array (Edmund Optics, 64-479). The positional shifts of these maxima relative to those seen with no specimen present encode the pupil wavefront phase2, which can then be reconstructed.\n\nAOViFT inference is performed routinely on the microscope acquisition PC (Intel Xeon, W5-3425, Windows 11, 512\u2009GB RAM, NVIDIA A6000 with 48\u2009GB VRAM). Inferences are made in an Ubuntu Docker container based on the TensorFlow NGC Container (24.02-tf2-py3) running in parallel with the microscope control software. Data communication between AOViFT and the microscope control software is handled through the computer\u2019s file system. Image files and command-line parameters are passed to the model, and an output text file reports the resultant DM actuator values (Supplementary Fig. 19). When a volume is large enough to require tiling and dozens of volumes need to be processed, model inferences are parallelized and run using a SLURM compute cluster consisting of four nodes, each node containing four NVIDIA A100 80GB.\n\nThe 25-mm coverslips (Thorlabs, CG15XH) used for imaging beads, cells and zebrafish embryos were first cleaned by sonication in 70% ethanol followed by Milli-Q water, each for at least 30\u2009min. They were then stored in Milli-Q water until use. Gene-edited SUM159-AP2-eGFP cells14 were grown in Dulbecco\u2019s modified Eagle\u2019s medium (DMEM)/F12 with GlutaMAX (Gibco, 10565018) supplemented with 5% fetal bovine serum (FBS; Avantor Seradigm, 89510-186), 10\u2009mM HEPES (Gibco 15630080), 1\u2009\u03bcg\u2009ml\u22121 hydrocortisone (Sigma, H0888), 5\u2009\u03bcg\u2009ml\u22121 insulin (Sigma, I9278). Fluorescent beads (0.2-\u03bcm diameter, Invitrogen FluoSpheres Carboxylate-Modified Microspheres, 505/515 nm, cat. no. F8811 or 0.2-\u03bcm diameter Tetraspeck, Thermo Fisher Scientific Invitrogen, T7280) alone or with cells at 30\u201350% confluency were deposited onto plasma-treated and poly-d-lysine (Sigma-Aldrich, P0899)-treated 25-mm coverslips. Cells were cultured under standard conditions (37\u2009\u00b0C, 5% CO2, 100% humidity) with twice weekly passaging. The SUM159-AP2-eGFP cells were imaged in Leibovitz\u2019s L-15 medium without phenol red (Gibco,21083027) with 5% FBS (American Type Culture Collection, SCRR-30-2020), 100\u2009\u03bcM Trolox (Tocris, 6002) and 100\u2009\u03bcg\u2009ml\u22121 Primocin (InvivoGen, ant-pm-1) at 37\u2009\u00b0C. Aberrations of approximately 1\u03bb P\u2013V were induced using a DM in ten configurations of Zernike modes (\\({Z}_{2}^{2}\\), \\({Z}_{3}^{-3}\\), \\({Z}_{3}^{-1}\\), \\({Z}_{4}^{\\,0}\\) and their pairwise combinations). Widefield PSFs were collected from 0.2-\u03bcm fluorescent beads to confirm the aberrations applied and residual aberrations after correction (Supplementary Table 7).\n\nGenome-edited ap2s1-expressing zebrafish (genome editing of ap2s1, ap2s1:ap2s1-mNeonGreenbk800; Supplementary Note D) were injected with cox8-mChilada mRNA for two color experiments. The N-terminal 34 amino acids of Cox8a were cloned into a pMTB backbone with a linker and mChilada coding sequence on the C terminus (unpublished, gift from N. Shaner). The plasmid was linearized, and mRNA was synthesized using a SP6 mMessage mMachine transcription kit (Thermo Fisher). RNA was purified using an RNeasy kit (Qiagen) and embryos were injected with 2\u2009nl of 10\u2009ng\u2009\u03bcl\u22121 Cox8a-mChilada, 100\u2009mM KCl, 0.1% phenol red, 0.1\u2009mM EDTA and 1\u2009mM Tris, pH\u20097.5. Zebrafish embryos were first nanoinjected with 3\u2009nl of a solution containing 0.86\u2009ng\u2009\u03bcl\u22121 \u03b1-bungarotoxin protein, 1.43\u2009\u00d7\u2009PBS and 0.14% phenol red. The injected embryos were mounted for imaging using a custom, volcano-shaped agarose mount. Each mount was constructed by solidifying a few drops of 1.2% (w/w) high-melting agarose (Invitrogen UltraPure Agarose, 16500\u2013100, in 1\u00d7 Danieau buffer) between a 25-mm glass coverslip and a 3D-printed mold (Formlabs Form 3+, printed in clear v.4 resin). This created ridges that formed a narrow groove. A hair-loop was used to orient the embryo within the agarose groove, positioning the left lateral side upward. Subsequently, 10\u201320\u2009\u03bcl of 0.5% (w/w) low-melt agarose (Invitrogen UltraPure LMP Agarose, 16520\u2013100, in 1\u00d7 Danieau buffer) preheated to 40\u2009\u00b0C, containing 0.2\u2009\u03bcm Tetraspeck microspheres, was added on top of the embryo. This layer solidified around the embryo to secure it while providing fiducial beads for sample finding. Once the low-melt agarose solidified, the volcano-shaped mount was held by a custom sample holder for imaging. The embryo was oriented so that its anterior\u2013posterior axis lay parallel to the sample x\u2009axis, with the anterior end facing the excitation objective and the posterior end facing the detection objective. The microscope objectives and the sample was immersed in a bath of ~50\u2009ml bath Danieau buffer and were fully submerged, ensuring the embryo remained in buffered medium. Measurements for AOViFT and SH were done serially on the same FOV to compare the aberration corrections of both methods (Supplementary Table 7).\n\nTo compensate for sample-induced aberrations postacquisition, we performed a tile-based spatially varying deconvolution on each 3D volume. Each volume was first subdivided into several 3D tiles approximating isoplanatic patches. A AOViFT predicted PSF (for compensation) or an ideal PSF (for no compensation) was assigned to each tile, and aberrations were corrected using OTF masked Wiener (OMW) deconvolution15. To minimize boundary artifacts during deconvolution, the tile size was extended by half the PSF width at each boundary (32 pixels); after deconvolution, these overlaps were removed and the deconvolved core regions were stitched together to form the final corrected volume. All computations were done in MATLAB v.2024a (Mathworks).\n\nTo train a model for predicting optical aberrations from images of subdiffractive objects in biological samples, we generated synthetic datasets encompassing a range of relevant variables (for example, aberration modes and amplitudes, number and density of puncta, SNR). This synthetic dataset generation procedure is as follows.\n\nFor a single subdiffractive punctum, the electric field in the rear pupil of the detection objective is given by:\n\nwhere A(kx, ky) is the pupil amplitude with coordinates kx, ky, and \u03d5(kx, ky) is the pupil phase. Under aberration-free conditions, \u03d5(kx, ky) is a constant. We can empirically determine A(kx, ky) by acquiring a widefield image of an isolated subdiffractive object (100-nm fluorescent bead), performing phase retrieval12,16 and applying the opposite of the retrieved phase using a pupil conjugate DM so that \u03d5(kx, ky) becomes a constant.\n\nThe electric field for the image of a single aberrated punctum is:\n\nwhere the \u03d5abb(kx, ky) is described as a weighted sum of Zernike modes of unique amplitudes:\n\nEmpirically, zebrafish induced aberrations for the microscopes used here are well described by combinations of 11 of the first 15 Zernike modes17 (Supplementary Fig. 1), for which n\u2009\u2264\u20094, excluding piston (\\({Z}_{0}^{\\,0}\\)), tip (\\({Z}_{1}^{-1}\\)), tilt (\\({Z}_{1}^{1}\\)) and defocus (\\({Z}_{2}^{\\,0}\\)) (as these represent phase offsets or sample translation). The distributions and amplitudes of the remainder are used to build the training set as discussed below.\n\nThe aberrated 3D detection PSF of a subdiffractive punctum is approximated by:\n\nwhere \\({k}_{z}=\\sqrt{{(\\frac{2\\pi \\eta }{\\lambda })}^{2}-{k}_{x}^{2}-{k}_{y}^{2}}\\), \u03b7 is the refractive index of the imaging medium and \u03bb is the free-space wavelength of the fluorescence emission.\n\nFor light sheet microscopy, the aberrated 3D overall PSF is:\n\nwhere PSFexc(z) is given by the cross-section of the swept light sheet used for imaging. Examples of these PSFs are shown in Supplementary Fig. 20 I\u2013V, with MBSq-35 in Supplementary Table 6 used for training and imaging (see ref. 18 for additional information on these light sheets).\n\nEach synthetic training volume sample V is 64\u2009\u00d7\u200964\u2009\u00d7\u200964 voxels in size spanning 8\u2009\u00d7\u20098\u2009\u00d7\u200912.8\u2009\u03bcm3 (with 125\u2009\u00d7\u2009125\u2009\u00d7\u2009200\u2009nm3 voxels) and containing between J =\u20091 to J\u2009=\u20095 puncta chosen from a uniform distribution and located randomly at points (xj, yj, zj) within the volume. Each punctum is modeled as a Gaussian of full width at half maximum wj chosen randomly from the set [100, 200, 300, 400] nm, allowing for slightly larger than the diffraction-limit features. The image of each punctum is generated by its convolution with the aberrated PSF:\n\nThe integrated photons No per punctum were selected from a uniform distribution of 1 to 200,000 photons. The total intensity distribution is:\n\nwhere\n\nAs the signal from each aberrated punctum can exceed the boundary of V, total signal SV within V is:\n\nAfter accounting for partial signal contributions (SV) the photons per voxel were converted to camera counts by applying the quantum efficiency QE, Poisson shot noise \u03b7 and camera read noise \u03f5 to arrive at the final synthetic training set example:\n\nTo ensure diversity in the training set to cover potential aberrations, each training example was chosen from the amplitudes of the 11 included aberration modes shown in color in Supplementary Fig. 1 with equal probability from one of four different distributions:\n\nSingle mode (Supplementary Fig. 21b) One mode is randomly chosen, with amplitude \u03b1 chosen randomly from 0\u2009\u2264\u2009\u03b1\u2009\u2264\u20090.5\u2009\u03bb RMS.\n\nBimodal (Supplementary Fig. 21c) An initial target for the total amplitude \u03b1t is chosen randomly from 0\u2009\u2264\u2009\u03b1t\u2009\u2264\u20090.5\u2009\u03bb RMS. A second partitioning factor \u03f5 is chosen randomly from 0\u2009\u2264\u2009\u03f5\u2009\u2264\u20091. The amplitudes of the two modes are then \u03b11\u2009=\u2009\u03f5\u03b1t and \u03b12\u2009=\u2009(1\u2009\u2212\u2009\u03f5)\u03b1t.\n\nPowerlaw (Supplementary Fig. 21d) An initial target for the total amplitude \u03b1t is chosen randomly from 0\u2009\u2264\u2009\u03b1t\u2009\u2264\u20090.5\u2009\u03bb RMS. The initial partitioning factors \u03f5n for the modes are chosen randomly from a Lomax (that is, Pareto II) distribution19:\n\nwhere each xn is chosen randomly from 0\u2009\u2264\u2009xn\u2009\u2264\u20091. They are then renormalized:\n\nand the final amplitudes of the modes are \\({\\alpha }_{n}={\\epsilon }_{n}^{{\\prime} }{\\alpha }_{t}\\).\n\nDirichlet (Supplementary Fig. 21e) An initial target for the total amplitude \u03b1t is chosen randomly from 0\u2009\u2264\u2009\u03b1t\u2009\u2264\u20090.5\u2009\u03bb RMS. The initial partitioning factors \u03f5n for the modes are chosen randomly from 0\u2009\u2264\u2009\u03f5n\u2009\u2264\u20091. They are then renormalized:\n\nand the final amplitudes of the modes are \\({\\alpha }_{n}={\\epsilon }_{n}^{{\\prime} }{\\alpha }_{t}\\).\n\nTogether, the training examples from these four distributions create a diverse set of overall aberration amplitudes and number of significant modes in the training data, with all 11 modes contributing equally across the dataset (Supplementary Fig. 21a).\n\nFor the model training, a dataset of 2\u2009million synthetic 3D volumes was created, with aberration magnitude uniform sampled from 0.0 to 0.5 \u03bb RMS (at wavelength \u03bb\u2009=\u2009510\u2009nm), uniform distribution of the number of objects between 1 and 5, and photons ranging between 1 and 200,000 integrated photons per object.\n\nTo evaluate our models, we created a test dataset with 100,000 3D volumes. The parameter distribution was the same as training, but extended the aberration magnitude up to 1.0 \u03bb RMS, and up to 500,000 integrated photons. To test the operational limit of our models, this test dataset included up to 150 objects in any given volume.\n\nMost ML vision models operate on real-space representations of the data, which lack clearly defined limits on image size or feature descriptors of their content. Instead, we used Fourier domain embeddings (Supplementary Fig. 24). These are bound by the microscope\u2019s OTF. Aberrations within an isoplanatic patch globally effect all photons within that patch, producing a unique, learnable \u2018fingerprint\u2019 pattern in the FFT amplitude and phase (Supplementary Note A.1 and Supplementary Figs. 5\u20137).\n\nTo create Fourier embeddings (Fig. 1b) for our model, we preprocess the input 3D image stack W of CCPs within an isoplanatic region to suppress noise and edge artifacts (Fig. 1a),\n\nThe preprocessing module (\u03d2) begins with a set of filters to extract sharp-edged objects that reveal the aberration signatures: a Gaussian high-pass filter to remove inhomogeneous background and a low-pass filter through a Fourier frequency filter, with cutoff set at the detection NA limit (\u03c3\u2009=\u20093 voxels). A Tukey window (Tukey cosine fraction\u2009=\u20090.5, in \\(\\hat{x}\\hat{y}\\) only) is applied to remove FFT edge artifacts from the volume borders. No windowing is applied along the axial direction, \\(\\hat{z}\\), because embeddings are constructed near kz\u2009=\u20090 where aberration information is maximized.\n\nOnce preprocessed, a ratio of the resultant 3D FFT amplitude, to the 3D FFT amplitude of the ideal PSF (undergoing identical preprocessing steps) is used to generate the amplitude embedding, \u03b1(kz) at each kz plane:\n\nwhere \\({\\mathcal{F}}\\) denotes the 3D Fourier transform. The most useful information content is located at kz\u2009=\u20090, the principal plane located at the midpoint of the \\(\\hat{{k}_{z}}\\)-axis. Three 2D planes from \u03b11, \u03b12 and \u03b13 along \\(\\hat{{k}_{z}}\\)-axis as are necessary to extract axial information for inputs to the model as follows:\n\nwhere \u03b11 is the principal plane along the kx-axis and ky-axis, \u03b12 is the mean of five consecutive 2D planes starting from the principal plane and \u03b13 is the mean of five consecutive 2D planes starting from the kz\u2009=\u20095 plane (Supplementary Figs. 7 and 24a,c).\n\nFor the phase embedding, \u03c6, we first remove interference from several puncta in the FOV that may obscure the aberration signature in the phase image. The interference patterns are removed using: peak local maxima (PLM; https://scikit-image.org/docs/stable/auto_examples/segmentation/plot_peak_local_max.html) for peak detection in real space using normalized cross-correlation (NCC; https://scikit-image.org/docs/stable/auto_examples/registration/plot_masked_register_translation.html) with a kernel cropped from the highest peak in V. The neighboring voxels around the detected puncta peaks are masked off, creating a volume, \\({\\mathcal{S}}\\). The OTF with interference removed, \\({\\tau }^{{\\prime} }\\), can now be obtained as well as a real space reconstructed volume, \\({V}^{{\\prime} }\\), through inverse FFT,\n\nThe phase \u03c6(kz) at each kz plane is then given by the unwrapped phase of \u03c4 at that plane (Supplementary Fig. 24b,d). We calculate the three phase embeddings in the same manner as our amplitude embedding such that:\n\nCombining the six planes together, we define the input to the model as a Fourier embedding,\n\nA notable advantage of this approach is that, although the signal from each individual CCP is weak, those in the same isoplanatic region contain near-identical spatial frequency distributions that add together to yield Fourier embeddings of high SNR suitable for accurate inference of the underlying aberration (Supplementary Fig. 6).\n\nBelow, we outline the key components of AOViFT, which uses a 3D multistage vision transformer architecture. This model efficiently captures Fourier domain features at several spatial scales, enabling robust aberration prediction.\n\nRecent advances in attention-based transformers have demonstrated scalability, generalizability and multi-modality for a range of computer vision applications20,21,22,23,24.\n\nMultiscale (or hierarchical) vision transformers, such as Swin25 and MViT26, are designed with specialized modules (for example, shifted-window partitioning25 and hybrid window attention27) to excel at a variety of detection tasks for 2D natural images using supervised training on ImageNet28. Although these variants are more efficient than their ViT counterparts in terms of FLOPs and number of parameters, they often incorporate specialized modules as noted above. Hiera29 showed that these designs can be streamlined without performance loss by leveraging large-scale self-supervised pretraining.\n\nCurrent multiscale architectures use a feature pyramid network scheme30\u2014downsampling the spatial resolution of the image for each stage while expanding the embedding size for deeper layers. Instead, in our work, we use \u03a9 stages and do not downsample during any of the stages, but rather select different patch sizes for each stage (Fig. 1). This allows the embedding dimension within each stage to be fixed to the number of voxels in the patch of that stage, rather than expanding with increasing depth as in some hierarchical models.\n\nThe input to the model is the Fourier embedding, a 3D tensor \\({\\mathcal{E}}\\in {{\\mathbb{R}}}^{\\ell \\times d\\times d}\\), where \u2113\u2009=\u20096 is the number of 2D planes each with a height and width of d. For each model stage, i, patchifying begins by dividing the input tensor \\({\\mathcal{E}}\\) into nonoverlapping 2D tiles (each pi\u2009\u00d7\u2009pi) that are each flattened into a one-dimensional patch for a total of ki patches in a plane. After patchifying, the input tensor is transformed into \\({x}_{p}\\in {{\\mathbb{R}}}^{\\ell \\times {k}_{i}\\times {p}_{i}^{2}}\\) (Fig. 1b).\n\nThe initial ViT model uses a set of consecutive transformer layers with a fixed patch size for all transformers, where each transformer layer can capture local and global dependencies between patches through self-attention20. The computation needed for the self-attention layers scales quadratically with reference to the number of patches (that is, sequence length). Although using a smaller patch size could be useful to capture visual patterns at a finer resolution, using a large patch size is computationally cheaper.\n\nOur baseline model uses a two-stage design with patch sizes of 32 and 16 pixels, respectively (Fig. 1c). Supplementary Note A shows an ablation study using several stages with patch sizes ranging between 8 and 32 pixels.\n\nRather than adopting the Cartesian positional encoding of ViT20, we use a polar coordinate system (r, \u03b8) to encode the position of each patch. This choice is motivated by the radial symmetries of the Zernike polynomials and the efficiencies gained in NeRF31, coordinate-based MLPs32 and RoFormer33. For a given plane in \\({\\mathcal{E}}\\) (Eq. 27), the radial positional encoding vector (RPE) is calculated for every patch,\n\nwhere (r, \u03b8) are the polar coordinates for the center of each patch, and m\u2009=\u200916. All patches and their positional encoding are then mapped into a sequence of learnable linear projections \\(\\zeta \\in {{\\mathbb{R}}}^{\\ell \\times {k}_{i}\\times {p}_{i}^{2}}\\) that we use as our input to the transformer layers in the model.\n\nEach stage has n transformer layers, where each layer has h multihead attention (MHA) layers that map the interdependencies between patches, followed by a multilayer perceptron block (MLP) that learns the relationship between pixels within a patch. The stage\u2019s embedding size, \\({\\epsilon }_{i}={p}_{i}^{2}\\), is set to match the number of voxels in a patch for that stage. The MLP block is four times wider than the embedding size (Supplementary Fig. 2c). Layer normalization (LN)34 is applied before each step, and a skip/residual connection35 is added after each step:\n\nIn addition to the skip connections in each transformer layer, we also add a skip connection between the input and output of each stage. We use a dropout rate of 0.1 for each dense layer36 and stochastic depth rate of 0.1 (ref. 37). The patches from the final stage are pooled using a global average along the last dimension and passed to a fully connected layer to output z Zernike coefficients.\n\nWe use self-attention38 as our default attention module for all transformer layers in our model. Complementary to our approach, recent studies have looked into alternative attention methods to reduce the quadratic scaling of self-attention23,39,40. Our architecture is compatible with these attention mechanisms, which would further improve our model\u2019s efficiency.\n\nSupplementary Note A shows an ablation study of our synthetic data simulator (Supplementary Note A.2), our multistage design (Supplementary Note A.3), our training dataset size (Supplementary Note A.4 and Supplementary Fig. 2) and details of our training hyperparameters (Supplementary Note A.5, Supplementary Table 2 and Supplementary Table 8). We also introduce a new way of measuring prediction confidence of our model using digital rotations in Supplementary Note A.6 (Supplementary Fig. 23).\n\nWe present a detailed cost analysis benchmark comparing our architecture with other widely used models such as ConvNeXt41 and ViT20 in Supplementary Note B. To further diagnose our model\u2019s performance, we carried out a series of experiments to understand our model\u2019s sensitivity to SNR (Supplementary Note C.1 and Supplementary Figs. 24\u201325), generalizability to other light sheets (Supplementary Note C.2), number of objects in the FOV (Supplementary Note C.3) and object size (Supplementary Note C.4 and Supplementary Fig. 26).\n\nAll experiments with zebrafish were done in accordance with protocols approved by the University of California, Berkeley\u2019s Animal Care and Use Committee and following standard protocols (animal use protocol number AUP-2019-09-12560-1). All zebrafish used in this study were embryos younger than 72\u2009h postfertilization. Sex determination was not a factor in our experiments. All husbandry and experiments with zebrafish were done in accordance with protocols approved by the University of California, Berkeley\u2019s Animal Care and Use Committee and following standard protocols (animal use protocol numbers AUP-2019-09-12560-1 (Upadhyayula laboratory), AUP-2020-10-13737-1 (Swinburne laboratory) and AUP-2021-05-14347-1 (Zebrafish Facility Core Protocol)).\n\nFurther information on research design is available in the Nature Portfolio Reporting Summary linked to this article.",
+            "section_image": []
+        },
+        {
+            "section_name": "Data availability",
+            "section_text": "Data for demos is available on our Github repository at https://github.com/cell-observatory/aovift. The full datasets for training and testing are too large to be hosted on public repositories, they can be shared upon reasonable request. Our synthetic data generator is also available at https://github.com/cell-observatory/beads_simulator to enable users to simulate their own datasets for training and evaluation.",
+            "section_image": []
+        },
+        {
+            "section_name": "Code availability",
+            "section_text": "Source code for training and evaluation (and all pretrained models) are available at https://github.com/cell-observatory/aovift. Docker image is available at https://github.com/cell-observatory/aovift/pkgs/container/aovift. Deconvolution was performed using PetaKit5D (https://github.com/abcucberkeley/PetaKit5D).",
+            "section_image": []
+        },
+        {
+            "section_name": "References",
+            "section_text": "Zhang, Q. et al. Adaptive optics for optical microscopy. Biomed. Opt. Express 14, 1732\u20131756 (2023).\n\nArticle\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nWang, K. et al. Rapid adaptive optical recovery of optimal resolution over large volumes. Nat. Methods 11, 625\u2013628 (2014).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLiu, T.-L. et al. Observing the cell in its native state: imaging subcellular dynamics in multicellular organisms. Science 360, 1392 (2018).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSaha, D. et al. Practical sensorless aberration estimation for 3D microscopy with deep learning. Optics Express 28, 29044\u201329053 (2020).\n\nArticle\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nRai, M. R., Li, C., Ghashghaei, H. T. & Greenbaum, A. Deep learning-based adaptive optics for light sheet fluorescence microscopy. Biomed. Opt. Express 14, 2905\u20132919 (2023).\n\nArticle\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nZhang, P. et al. Deep learning-driven adaptive optics for single-molecule localization microscopy. Nat. Methods 20, 1748\u20131758 (2023).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHu, Q. et al. Universal adaptive optics for microscopy through embedded neural network control. Light Sci. Appl. 12, 270 (2023).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKang, I., Zhang, Q., Yu, S. X. & Ji, N. Coordinate-based neural representations for computational adaptive optics in widefield microscopy. Nat. Mach. Intell. 6, 714\u2013725 (2024).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nAshesh, A. et al. Microsplit: Semantic unmixing of fluorescent microscopy data. Preprint at bioRxiv https://doi.org/10.1101/2025.02.10.637323 (2025).\n\nMahajan, V. N. Strehl ratio for primary aberrations: some analytical results for circular and annular pupils. J. Optic. Soc. Am. 72, 1258\u20131266 (1982).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBentley, J. & Olson, C. Field Guide to Lens Design (SPIE Press, 2012)\n\nHanser, B. M., Gustafsson, M. G. L., Agard, D. A. & Sedat, J. W. Phase-retrieved pupil functions in wide-field fluorescence microscopy. J. Microsc. 216, 32\u201348 (2004).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBetzig, E. A cell observatory to reveal the subcellular foundations of life. Nat. Methods 22, 646\u2013649 (2025).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nAguet, F. et al. Membrane dynamics of dividing cells imaged by lattice light-sheet microscopy. Mol. Biol. Cell 27, 3418\u20133435 (2016).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nRuan, X. et al. Image processing tools for petabyte-scale light sheet microscopy data. Nat. Methods 21, 2342\u20132352 (2024).\n\nArticle\u00a0\n    CAS\u00a0\n    PubMed\u00a0\n    PubMed Central\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nCampbell, H. I., Zhang, S., Greenaway, A. H. & Restaino, S. Generalized phase diversity for wave-front sensing. Opt. Lett. 29, 2707\u20132709 (2004).\n\nArticle\u00a0\n    PubMed\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLakshminarayanan, V. & Fleck, A. Zernike polynomials: a guide. J. Mod. Opt. 58, 545\u2013561 (2011).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLiu, G. et al. Characterization, comparison, and optimization of lattice light sheets. Sci. Adv. 9, 6623 (2023).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLomax, K. S. Business failures: another example of the analysis of failure data. J. Am. Stat. Assoc. 49, 847\u2013852 (1954).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nDosovitskiy, A. et al. An image is worth 16 \u00d7 16 words: transformers for image recognition at scale. In Proc. International Conference on Learning Representations (ICLR, 2021).\n\nCarion, N. et al. End-to-end object detection with transformers. In European Conference on Computer Vision (eds Vedaldi, A. et al.) 213\u2013229 (Springer, 2020).\n\nArnab, A. et al. Vivit: a video vision transformer. In Proc. IEEE/CVF International Conference on Computer Vision 6836\u20136846 (IEEE, 2021).\n\nCheng, B., Misra, I., Schwing, A.G., Kirillov, A. & Girdhar, R. Masked-attention mask transformer for universal image segmentation. In Proc. IEEE/CVF Conference on Computer Vision and Pattern Recognition 1290\u20131299 (IEEE, 2022).\n\nHe, K. et al. Masked autoencoders are scalable vision learners. In Proc. IEEE/CVF Conference on Computer Vision and Pattern Recognition 16000\u201316009 (IEEE, 2022).\n\nLiu, Z. et al. Swin transformer: hierarchical vision transformer using shifted windows. In Proc. IEEE/CVF International Conference on Computer Vision 10012\u201310022 (IEEE, 2021).\n\nFan, H. et al. Multiscale vision transformers. In Proc. IEEE/CVF International Conference on Computer Vision 6824\u20136835 (IEEE, 2021).\n\nLi, Y. et al. Mvitv2: improved multiscale vision transformers for classification and detection. In Proc. IEEE/CVF Conference on Computer Vision and Pattern Recognition 4804\u20134814 (IEEE, 2022).\n\nDeng, J. et al. Imagenet: a large-scale hierarchical image database. In Proc. 2009 IEEE Conference on Computer Vision and Pattern Recognition 248\u2013255 (IEEE, 2009).\n\nRyali, C. et al. Hiera: a hierarchical vision transformer without the bells-and-whistles. In Proc. International Conference on Machine Learning 29441\u201329454 (PMLR, 2023).\n\nLin, T.-Y. et al. Feature pyramid networks for object detection. In Proc. IEEE Conference on Computer Vision and Pattern Recognition 2117\u20132125 (IEEE, 2017).\n\nMildenhall, B. et al. Nerf: representing scenes as neural radiance fields for view synthesis. Commun. ACM 65, 99\u2013106 (2021).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nTancik, M. et al. in Advances in Neural Information Processing Systems Vol. 33 (eds Larochelle, H. et al.) 7537\u20137547 (Curran Associates, 2020).\n\nSu, J. et al. Roformer: enhanced transformer with rotary position embedding. Neurocomputing 568, 127063 (2024).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBa, L.J., Kiros, J.R. & Hinton, G.E. Layer normalization. Preprint at https://arxiv.org/abs/1607.06450 (2016).\n\nHe, K., Zhang, X., Ren, S. & Sun, J. Deep residual learning for image recognition. In. Proc. IEEE Conference on Computer Vision and Pattern Recognition 770\u2013778 (IEEE, 2016).\n\nSrivastava, N., Hinton, G., Krizhevsky, A., Sutskever, I. & Salakhutdinov, R. Dropout: a simple way to prevent neural networks from overfitting. J. Mach. Learn. Res. 15, 1929\u20131958 (2014).\n\n\n                    Google Scholar\u00a0\n                \n\nHuang, G., Sun, Y., Liu, Z., Sedra, D. & Weinberger, K.Q. Deep networks with stochastic depth. In Proc. Computer Vision\u2013ECCV 2016: 14th European Conference Part IV 14. 646\u2013661 (Springer, 2016).\n\nVaswani, A. et al. in Advances in Neural Information Processing Systems Vol. 30 (eds Guyon, I. et al.) 2440\u20132448 (Curran Associates, 2017).\n\nDao, T., Fu, D., Ermon, S., Rudra, A. & R\u00e9, C. in Advances in Neural Information Processing Systems (NeurIPS 2022) Vol. 35 (eds Koyejo, S. et al.) 16344\u201316359 (Curran Associates, 2022).\n\nYang, J., Li, C., Dai, X. & Gao, J. in Advances in Neural Information Processing Systems (NeurIPS 2022) Vol. 35 (eds Koyejo, S. et al.) 4203\u20134217 (Curran Associates, 2022).\n\nLiu, Z. et al. A convnet for the 2020s. In Proc IEEE/CVF Conference on Computer Vision and Pattern Recognition 11976\u201311986 (IEEE, 2022).\n\nDownload references",
+            "section_image": []
+        },
+        {
+            "section_name": "Acknowledgements",
+            "section_text": "We thank X. Ruan, M. Mueller, P. Zwart and H. York for helpful discussions and comments. SUM159 cells used in this study were a gift from the Kirchhausen laboratory. We thank N. Shaner for providing the mChilada fluorescent protein plasmid to I.A.S., which was used to generate reagents for this study. We gratefully acknowledge the support of this work by the Laboratory Directed Research and Development (LDRD) Program of Lawrence Berkeley National Laboratory under US Department of Energy contract no. DE-AC02-05CH11231. We thank J. White for managing our computing cluster. T.A., G.L., F.G., J.L.H. and S.U. are partially supported by the Philomathia Foundation (awarded to E.B. and S.U.). T.A. and G.L. are partially supported by the Chan Zuckerberg Initiative (awarded to S.U.). T.A. and S.U. are partially supported by Lawrence Berkeley National Laboratory\u2019s LDRD program 7647437 and 7721359 (awarded to S.U.). T.A., D.E.M. and E.B. are supported by HHMI (awarded to E.B.). C. Shirazinnejad. and D.G.D. are partially supported by NIH Grant R35GM118149 (awarded to D.G.D.). C. Simmons, I.S.A. and I.A.S. are partially supported by NIH Grant 1R01DC021710 (awarded to I.A.S.). F.G. is partially funded by the Feodor Lynen Research Fellowship, Humboldt Foundation. S.U. is funded by the Chan Zuckerberg Initiative Imaging Scientist program 2019-198142 and 2021-244163. E.B. is an HHMI Investigator. S.U. is a Chan Zuckerberg Biohub\u2013San Francisco Investigator.",
+            "section_image": []
+        },
+        {
+            "section_name": "Author information",
+            "section_text": "Howard Hughes Medical Institute, Ashburn, VA, USA\n\nThayer Alshaabi,\u00a0Daniel E. Milkie\u00a0&\u00a0Eric Betzig\n\nDepartment of Molecular and Cell Biology, University of California Berkeley, Berkeley, CA, USA\n\nThayer Alshaabi,\u00a0Gaoxiang Liu,\u00a0Cyna Shirazinejad,\u00a0Jason L. Hong,\u00a0Kemal Achour,\u00a0Frederik G\u00f6rlitz,\u00a0Ana Milunovic-Jevtic,\u00a0Cat Simmons,\u00a0Ibrahim S. Abuzahriyeh,\u00a0Erin Hong,\u00a0Samara Erin Williams,\u00a0Nathanael Harrison,\u00a0Evan Huang,\u00a0Eun Seok Bae,\u00a0Alison N. Killilea,\u00a0Ian A. Swinburne,\u00a0David G. Drubin,\u00a0Srigokul Upadhyayula\u00a0&\u00a0Eric Betzig\n\nDepartment of Molecular Biophysics and Integrated Bioimaging, Lawrence Berkeley National Laboratory, Berkeley, CA, USA\n\nSrigokul Upadhyayula\n\nChan Zuckerberg Biohub, San Francisco, CA, USA\n\nSrigokul Upadhyayula\n\nDepartment of Physics, University of California Berkeley, Berkeley, CA, USA\n\nEric Betzig\n\nHelen Wills Neuroscience Institute, Berkeley, CA, USA\n\nEric Betzig\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nT.A. designed models, developed training pipelines and evaluation benchmarks. D.E.M. designed the preprocessing algorithms and developed the microscope software for the imaging experiments. T.A. and D.E.M. designed Fourier embedding, and developed the synthetic data generator for training and validation. C. Shirazinnejad, C. Simmons, I.S.A., S.E.W., N.H., E. Hong, E. Huang, E.S.B., A.N.K., D.G.D. and I.A.S. generated the zebrafish reagents. A.N.K. prepared the cultured SUM159 cells. C. Shirazinnejad, J.L.H., K.A. and A.M.-J. prepared samples. G.L. and J.L.H. performed the imaging experiments with zebrafish. J.L.H. and K.A. performed the imaging experiments with cells. G.L., K.A. and F.G. performed the imaging experiments with beads. T.A., D.E.M. and S.U. performed analysis and prepared figures. T.A. wrote the paper with input from all co-authors. D.E.M., E.B. and S.U. edited the paper. T.A., E.B. and S.U. supervised the project.\n\nCorrespondence to\n                Thayer Alshaabi, Srigokul Upadhyayula or Eric Betzig.",
+            "section_image": []
+        },
+        {
+            "section_name": "Ethics declarations",
+            "section_text": "The authors declare no competing interests.",
+            "section_image": []
+        },
+        {
+            "section_name": "Peer review",
+            "section_text": "Nature Methods thanks Gordon Love, Chang Qiao and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available. Primary Handling Editor: Rita Strack, in collaboration with the Nature Methods team.",
+            "section_image": []
+        },
+        {
+            "section_name": "Additional information",
+            "section_text": "Publisher\u2019s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.",
+            "section_image": []
+        },
+        {
+            "section_name": "Rights and permissions",
+            "section_text": "Open Access  This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article\u2019s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article\u2019s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.\n\nReprints and permissions",
+            "section_image": []
+        },
+        {
+            "section_name": "About this article",
+            "section_text": "Alshaabi, T., Milkie, D.E., Liu, G. et al. Fourier-based three-dimensional multistage transformer for aberration correction in multicellular specimens.\n                    Nat Methods 22, 2171\u20132179 (2025). https://doi.org/10.1038/s41592-025-02844-7\n\nDownload citation\n\nReceived: 20 March 2025\n\nAccepted: 22 August 2025\n\nPublished: 01 October 2025\n\nVersion of record: 01 October 2025\n\nIssue date: October 2025\n\nDOI: https://doi.org/10.1038/s41592-025-02844-7\n\nAnyone you share the following link with will be able to read this content:\n\nSorry, a shareable link is not currently available for this article.\n\n\n\n\n                            Provided by the Springer Nature SharedIt content-sharing initiative\n                        ",
+            "section_image": [
+                "https://data:image/svg+xml;base64,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"
+            ]
+        }
+    ]
+}

0c8e96cefef7de3e586d6f95c286e441bad381dad34d27b920d415c13c90e794/preprint/preprint.md

@@ -0,0 +1,1305 @@
+Genetic associations with human longevity are enriched for oncogenic genes
+
+Junyoung Park
+jpark01@stanford.edu
+
+Stanford University    https://orcid.org/0000-0002-9098-2858
+Andrés Peña-Tauber
+Stanford University    https://orcid.org/0000-0001-9501-4920
+Lia Talozzi
+Stanford University
+Michael Greicius
+Stanford University School of Medicine
+Yann Le Guen
+Stanford University    https://orcid.org/0000-0001-6649-8364
+
+Article
+
+Keywords:
+
+Posted Date: August 28th, 2024
+
+DOI: https://doi.org/10.21203/rs.3.rs-4837717/v1
+
+License: This work is licensed under a Creative Commons Attribution 4.0 International License.
+Read Full License
+
+Additional Declarations: There is NO Competing Interest.
+
+Version of Record: A version of this preprint was published at Nature Communications on February 28th, 2025. See the published version at https://doi.org/10.1038/s41467-025-57315-6.
+Genetic associations with human longevity are enriched for oncogenic genes
+
+Junyoung Park, PhD1, Andrés Peña-Tauber, BA1, Lia Talozzi, PhD1, Michael D. Greicius, MD1*, Yann Le Guen, PhD2*
+
+Affiliations
+1Department of Neurology and Neurological Sciences, Stanford University, Stanford, CA, 94305, USA.
+2Quantitative Sciences Unit, Department of Medicine, Stanford University, Stanford, CA, 94304, USA.
+
+Corresponding authors:
+Junyoung Park
+Stanford Neuroscience Health Center
+290 Jane Stanford Way,
+Stanford, CA 94305-5090
+jpark01@stanford.edu
+(650) 666-2696
+
+* These authors contributed equally
+Abstract
+
+Human lifespan is shaped by both genetic and environmental exposures and their interaction. To enable precision health, it is essential to understand how genetic variants contribute to earlier death or prolonged survival. In this study, we tested the association of common genetic variants and the burden of rare non-synonymous variants in a survival analysis, using age-at-death (N = 35,551, median [min, max] = 72.4 [40.9, 85.2]), and last-known-age (N = 358,282, median [min, max] = 71.9 [52.6, 88.7]), in European ancestry participants of the UK Biobank. The associations we identified seemed predominantly driven by cancer, likely due to the age range of the cohort. Common variant analysis highlighted three longevity-associated loci: \( APOE \), \( ZSCAN23 \), and \( MUC5B \). We identified six genes whose burden of loss-of-function variants is significantly associated with reduced lifespan: \( TET2 \), \( ATM \), \( BRCA2 \), \( CKMT1B \), \( BRCA1 \) and \( ASXL1 \). Additionally, in eight genes, the burden of pathogenic missense variants was associated with reduced lifespan: \( DNMT3A \), \( SF3B1 \), \( CHL1 \), \( TET2 \), \( PTEN \), \( SOX21 \), \( TP53 \) and \( SRSF2 \). Most of these genes have previously been linked to oncogenic-related pathways and some are linked to and are known to harbor somatic variants that predispose to clonal hematopoiesis. A direction-agnostic (SKAT-O) approach additionally identified significant associations with \( Clorf52 \), \( TERT \), \( IDH2 \), and \( RLIM \), highlighting a link between telomerase function and longevity as well as identifying additional oncogenic genes.
+
+Our results emphasize the importance of understanding genetic factors driving the most prevalent causes of mortality at a population level, highlighting the potential of early genetic testing to identify germline and somatic variants increasing one’s susceptibility to cancer and/or early death.
+Introduction
+Longevity is a complex trait influenced by both genetic and environmental factors and their interactions [1]. According to previous studies, genetics accounts for as much as 40% of the heritability of longevity [2-4]. Identifying the genetic variants that contribute to earlier death or prolonged survival can highlight key biological pathways linked to lifespan and inform genetic testing for general health and screening and enabling precision health. Previous genome-wide association studies (GWAS) have identified over 20 associated loci including APOE [5, 6], CHRNA3/5 [7], HLA-DQA1 and LPA [8]. Recently, a burden analysis of protein-truncating variants from whole-exome sequencing (WES) data identified four additional genes (BRCA2, BRCA1, ATM, and TET2) linked to reduced lifespan [9]. However, most previous research on lifespan genetics has predominantly used proxy data, such as parents’ age at death, due to a lack of proband lifespan data. While proxy-based GWAS have been necessary for large cohorts of primarily middle-aged individuals with limited mortality data, this approach restricts the accuracy and scope of findings, as it may fail to comprehensively capture the genetic influences that directly impact an individual’s lifespan [10]. On the other hand, some studies have employed logistic regression models on cases of extreme longevity and younger controls [11-13]. This approach may offer new insights by focusing on exceptionally long-lived individuals, yet they can be limited and costly. Moreover, replication of borderline significant variants remains an issue due to varying case definitions across studies, with some defining cases as individuals who survive to ages beyond 90 or 100 years or using the 90th or 99th survival percentiles as the age cutoff.
+
+In this study, we carried out a genetic analysis of direct mortality data in the UK Biobank, the genetic database with the largest number of reported deaths (35,551 subjects) and aged individuals (344,237 subjects over 60 years old). To assess the association of genetic variants with longevity in a survival analysis, we performed GWAS of common variants imputed from microarray data as well as burden/sequence kernel association test-optimized (SKAT-O) association of rare non-synonymous variants from WES data.
+
+Results
+
+Genome-wide association analyses in imputed array data
+Our GWAS assessed 6,127,227 common variants (minor allele frequency (MAF) \( \geq 1\% \)) using Martingale residuals on 393,833 individuals including 35,551 deceased subjects (mean age at death: 71.2 years) and 358,282 living subjects (mean current age: 70.7) from UKB (Supplementary Table 1) [14]. Two loci reached genome-wide significance (GWS) (\( p < 5.0 \times 10^{-8} \)) on chromosomes 19 and 6 (Figure 1A). On chromosome 19, rs429358 was the lead variant at the APOE locus (\( \beta = 0.013, p = 6.4 \times 10^{-47}, \text{MAF} = 15.6\% \)). We tested whether the presence of APOE-\( \varepsilon 4 \) was enriched in certain primary causes of death. Among the top four causes of death, each representing over 5% of total deaths (Figure 1B), only those due to ‘Diseases of the circulatory system’ (Chi-square \( p = 1.6 \times 10^{-16} \)) and ‘Diseases of the nervous system’ (\( p = 1.1 \times 10^{-71} \)) showed a significant enrichment in the proportion of \( \varepsilon 4 \) carriers compared to the prevalence of \( \varepsilon 4 \) carriers among all subjects (Figure 1C). In the chromosome 6 locus, overlapping ZSCAN23, two variants were GWS: rs6902687, located 2.2 kb upstream of the transcription start site (TSS), and rs13190937 situated in the 5' untranslated region (UTR) (rs6902687_C: \( \beta = 0.004, p = 1.5 \times 10^{-8}, \text{MAF} = 36.6\%; \) rs13190937_A: \( = 0.004, p = p = 1.5 \times 10^{-8}, \text{MAF} = 36.6\%, \) Figure 1D). To explore a potential regulatory function for variants at the ZSCAN23 locus, we investigated whether the lead SNPs were expression quantitative trait loci (eQTLs) in the Genotype-Tissue Expression Project (GTEx) v8 database. rs13190937 was significantly associated with increased ZSCAN23 expression in pancreatic tissue and the
+longevity GWAS signal colocalized with the ZSCAN23 expression quantitative trait loci (eQTL) (posterior probability of colocalization (PP4) = 0.94; Figure 1E). Phenome-wide association study analysis (PheWAS) using PheWeb [15] shows that the main associations of rs13190937 are with celiac disease and intestinal malabsorption (\( p = 1.8 \times 10^{-57} \)) (Supplementary Figure I).
+
+In sex-stratified GWAS (180,970 males and 212,863 females), the \( APOE \) locus was again linked to longevity in both males and females (Supplementary Table 1 and Supplementary Figure 2A and B). In males, we observed an additional GWS association for rs35705950_T located between \( MUC5AC \) and \( MUC5B \) on chromosome 11 (\( \beta = 0.01, p = 2.1 \times 10^{-8}, \text{MAF} = 11.2\% \)) (Supplementary Figure 2C), while no additional association was found in females. This variant was notably linked to increased \( MUC5B \) expression in lung tissue with the longevity GWAS signal aligning with a \( MUC5B \) eQTL (PP4 = 0.99; Supplementary Figure 2D). We also confirmed through PheWAS that rs35705950 is associated with a diagnosis of pulmonary fibrosis (\( p = 4.4 \times 10^{-13} \)) and “Other interstitial pulmonary diseases with fibrosis” listed as primary cause of death (\( p = 1.7 \times 10^{-5} \)), and Illness of the father “Lung cancer” (\( p = 2.1 \times 10^{-4} \)), but not with the mother’s (\( p = 0.07 \)) (Supplementary Figure 2E).
+
+Gene-based rare variant association analyses in whole-exome data
+
+Among 26,230,624 variants with MAF < 1%, 1,830,070 variants (17,174 genes) were annotated as loss-of-function (LoF) or missense variants. Of these, 628,362 were predicted LoF variants (17,071 genes with a median of 28 variants per gene), 985,950 were missense variants predicted as damaging by AlphaMissense (15,891 genes with a median of 47 variants per gene), and 349,791 were missense variants predicted as damaging by rare exome variant ensemble learner (REVEL) (12,219 genes with a median of 12 variants per gene). Of variants classified by each, 23.7% of AlphaMissense and 66.9% of REVEL variants were also pathogenic by the other classifier.
+
+We identified six genes whose burden of LoF variants is significantly associated with reduced lifespan: \( TET2 \) (\( p = 3.8 \times 10^{-30} \)), \( ATM \) (\( p = 6.0 \times 10^{-10} \)), \( BRC42 \) (\( p = 1.3 \times 10^{-34} \)), \( CKMT1B \) (\( p = 4.5 \times 10^{-7} \)), \( BRCA1 \) (\( p = 4.9 \times 10^{-12} \)) and \( ASXL1 \) (\( p = 2.3 \times 10^{-44} \)) (Figure 2A and Table 1). All of these but \( CKMT1B \) also showed gene-wide significance in a direction-agnostic (SKAT-O) approach (Supplementary Figure 3A). Additionally, in eight genes, the burden of missense variants predicted as pathogenic by AlphaMissense was associated with reduced lifespan: \( DNMT3A \) (\( p = 1.4 \times 10^{-9} \)), \( SF3B1 \) (\( p = 6.7 \times 10^{-12} \)), \( CHL1 \) (\( p = 5.0 \times 10^{-7} \)), \( TET2 \) (\( p = 4.2 \times 10^{-7} \)), \( PTEN \) (\( p = 1.0 \times 10^{-8} \)), \( SOX21 \) (\( p = 2.9 \times 10^{-8} \)), \( TP53 \) (\( p = 3.1 \times 10^{-15} \)) and \( SRSF2 \) (\( p = 9.8 \times 10^{-89} \)) (Figure 2B). Lastly, three genes showed gene-wide significance for burden of missense variants predicted by REVEL: \( DNMT3A \) (\( p = 6.6 \times 10^{-9} \)), \( PTEN \) (\( p = 6.6 \times 10^{-6} \)), and \( TP53 \) (\( p = 6.2 \times 10^{-9} \)) (Supplementary Figure 4 and Supplementary Table 2). SKAT-O identified additional associations with pathogenic missense variants predicted by AlphaMissense in \( Clor52 \) (\( p = 2.1 \times 10^{-7} \)), \( IDH2 \) (\( p = 5.3 \times 10^{-39} \)) and \( RLIM \) (\( p = 3.7 \times 10^{-9} \)) (Supplementary Figure 3B), and by REVEL in \( TERT \) (\( p = 8.1 \times 10^{-10} \)) (Supplementary Figure 3C and Supplementary Table 2).
+
+For sex-specific gene-based analysis, an additional six genes not identified in the whole-cohort analysis showed gene-wide significance in males by either burden or SKAT-O: \( CDKN1A \) and \( PTPRK \) (LoF); \( COA7 \) and \( TG \) (AlphaMissense); \( NMNAT2 \) and \( PITRM1 \) (REVEL) (Supplementary Figure 5A, 6A and Supplementary Table 3). In females, we identified three additional genes associated with reduced lifespan: \( PORCN \) (AlphaMissense); \( UGT1A8 \) and \( OLIG1 \) (REVEL) (Supplementary Figure 5B, 6B and Supplementary Table 4).
+
+Gene-burden survival analysis
+For the 13 gene-wide significant genes in the burden analyses, we assessed the association of variant carrier status with lifespan using Cox proportional hazards regression. Carriers of LoF variants in six genes were associated with decreased survival compared to non-carriers: CKMT1B (HR=3.9, \( p = 2.1 \times 10^{-6} \)), ASXL1 (HR=2.2, \( p = 3.8 \times 10^{-26} \)) (Figure 3A), TET2 (HR=1.7, \( p = 2.7 \times 10^{-18} \)), ATM (HR=1.7, \( p = 2.5 \times 10^{-10} \)), BRCA2 (HR=2.4, \( p = 1.1 \times 10^{-40} \)), and BRCA1 (HR =2.2, \( p = 1.0 \times 10^{-12} \)) (Supplementary Figure 7A). Similarly, carriers of AlphaMissense-predicted pathogenic variants exhibited significantly earlier mortality compared to non-carriers on the following genes: DNMT3A (HR=1.4, \( p = 2.0 \times 10^{-7} \)), SF3B1 (HR=2.1, \( p = 2.0 \times 10^{-8} \)), CHL1 (HR=1.3, \( p = 3.4 \times 10^{-7} \)), PTEN (HR=4.2, \( p = 3.4 \times 10^{-10} \)), SOX21 (HR=1.9, \( p = 5.2 \times 10^{-8} \)), TP53 (HR=3.7, \( p = 1.7 \times 10^{-13} \)), SRSF2 (HR=5.0, \( p = 2.4 \times 10^{-52} \)) (Figure 3B) and TET2 (HR=1.5, \( p = 3.8 \times 10^{-6} \)) (Supplementary Figure 7B). Carriers of pathogenic variants predicted by REVEL showed similar trends: DNMT3A (HR=1.5, \( p = 1.9 \times 10^{-6} \)), PTEN (HR=5.3, \( p = 1.3 \times 10^{-10} \)), and TP53 (HR=2.4, \( p = 6.6 \times 10^{-8} \)) (Supplementary Figure 7C).
+
+To explore the contribution of individual rare variants to mortality in each gene-wide significant gene in the burden and SKAT-O tests, we conducted Cox proportional hazards regression for each variant with a minor allele count (MAC) of three or more (Table 2). In total, 587 variants including LoF, AlphaMissense and REVEL variants were examined. After applying a Bonferroni correction for multiple testing, setting the significance threshold at \( 8.5 \times 10^{-5} \) (0.05/587), we identified significant associations with reduced lifespan for four LoF variants: rs370735654 in TET2 (MAC=17, HR=7.0, \( p = 5.6 \times 10^{-9} \)), rs87779834 in ATM (MAC=113, HR=2.5, \( p = 2.7 \times 10^{-5} \)), rs80359520 in BRCA2 (MAC=10, HR=6.1, \( p = 1.8 \times 10^{-6} \)), rs750318549 in ASXL1 (MAC=201, HR=2.5, \( p = 2.3 \times 10^{-15} \)). Additionally, significant associations with AlphaMissense variants were noted in six genes, impacting lifespan: rs769009649 in Clorf52 (MAC=62, HR=3.2, \( p = 7.4 \times 10^{-7} \)), rs147001633 in DNMT3A (MAC=269, HR=1.8, \( p = 3.7 \times 10^{-5} \)), rs377023736 in SF3B1 (MAC=12, HR=6.0, \( p = 8.5 \times 10^{-8} \)), rs116421102 in CHL1 (MAC=1,842, HR=1.3, \( p = 2.5 \times 10^{-5} \)), rs121913502 in IDH2 (MAC=45, HR=5.7, \( p = 1.0 \times 10^{-20} \)), rs11540652 in TP53 (MAC=5, HR=10.0, \( p = 6.6 \times 10^{-8} \)) and rs751713049 in SRSF2 (MAC=51, HR=5.8, \( p = 1.9 \times 10^{-26} \)). For missense variants predicted by REVEL, rs1043358053 in TERT (MAC=5, HR=11.9, \( p = 7.4 \times 10^{-5} \)) and rs11540652 in TP53 were significantly linked to reduced lifespan (Supplementary Table 5).
+
+Phenome-wide association studies
+For the nine novel longevity genes identified in the burden test (CKMT1B, ASXL1, DNMT3A, SF3B1, CHL1, PTEN, SOX21, TP53 and SRSF2), we examined the burden of LoF or pathogenic missense variants through PheWASs across 1,670 UKB phenotypes including disease occurrences derived from electronic health record, self-reported family history, and physical measures (Supplementary Figure 8). The burden of LoF variants in ASXL1 and AlphaMissense variants in DNMT3A, SF3B1, PTEN, TP53 and SRSF2 were strongly linked to an increased risk of leukemia: acute myeloid leukemia (ASXL1: Odds Ratio (OR)=1.05; \( p = 8.6 \times 10^{-170} \); DNMT3A: OR=1.03 , \( p = 2.1 \times 10^{-150} \); SRSF2: OR=1.3 , \( p = 1.2 \times 10^{-195} \); TP53: OR=1.05, \( p = 4.7 \times 10^{-35} \)), monocytic leukemia (DNMT3A: OR=1.01, \( p = 2.5 \times 10^{-9} \)), chronic lymphoid leukemia (SF3B1: OR=1.07, \( p = 4.1 \times 10^{-68} \)) and acute lymphoid leukemia (PTEN: OR=1.01, \( p = 2.1 \times 10^{-14} \)). Additionally, the burden of LoF in CKMT1B was associated with hypopharynx cancer (OR=1.03, \( p = 3.9 \times 10^{-26} \)), vertiginous syndromes (OR=1.03, \( p = 3.0 \times 10^{-17} \)) and salivary glands cancer (OR=1.03; \( p = 3.2 \times 10^{-12} \)). SOX21 burden was associated with increased acne (OR=1.01, \( p = 6.9 \times 10^{-7} \)) and spinocerebellar disease (OR=1.01, \( p = 2.3 \times 10^{-6} \)).
+Somatic mutation and clonal hematopoiesis of indeterminate potential
+
+We computed the variant allelic fraction (VAF) per carrier for each variant included in the analysis. Generally, germline variants have a mean VAF close to 50%, while somatic variants’ mean VAF will be lower [16]. Thus, when an association is linked to clonal hematopoiesis of indeterminate potential (CHIP), we expect the distribution of VAF to be left-shifted compared to a normal distribution centered at VAF = 50%. Considering LoF variants, *TET2* (mean VAF across variants [95% bootstrap confidence interval for the mean VAF] = 0.33 [0.31,0.34]) and *ASXL1* (mean VAF =0.32 [0.31,0.33]) burden test associations are supported by variants with a left-shifted VAF distribution (Supplementary Table 6, Supplementary Figure 9A). Similarly, considering pathogenic Alpha Missense variants, in *DNMT3A* (mean VAF= 0.24 [0.23-0.24]), *TET2* (mean VAF=0.36 [0.34,0.38]), *TP53* (mean VAF=0.28 [0.24,0.34]), *SRSF2* (mean VAF=0.30 [0.28,0.31]), *SF3B1* (mean VAF=0.31 [0.26,0.37]) and CHL1 (mean VAF=0.37[0.28-0.45]) are also left-shifted and the observed associations may be linked to CHIP (Supplementary Table 5, Supplementary Figure 9B).
+
+Methods
+
+Study participants
+
+The UKB is a large population-based longitudinal cohort study with recruitment from 2006 to 2010 in the United Kingdom [17]. In total, 502,664 participants aged 40-69 years were recruited and underwent extensive phenotyping including health and demographic questionnaires, clinic measurements, and blood draw at one of 22 assessment centers, of whom 468,541 subjects have been genotyped by both SNP array and WES.
+
+We restricted our analysis to 393,833 individuals who self-reported their ethnic background as ‘white British’ and were categorized as European ancestry based on genetic ethnic grouping (Field: 21000). Among them, 35,551 subjects were reported deceased, and their ages at death were recorded from the UK Death Registry (Field: 40007). For the other 358,282 subjects without death records, we assumed they were still alive by the latest censoring date (November 30, 2023). We determined their last known ages by subtracting their year and month of birth (Field: 33) from the censoring date.
+
+SNP array genotyping and QC
+
+A total of 488,000 UKB participants were genotyped using one of two closely related Affymetrix microarrays (UKB Axiom Array or UK BiLEVE Axiom Array) for ~ 820,000 variants. Quality control (QC), phasing, and imputation were performed as described previously [18]. Briefly, the genotyped dataset was phased and imputed into UK10K, 1000 Genomes Project phase 3, and Haplotype Reference Consortium reference panels, resulting in approximately 97 million variants. Additionally, we removed SNPs with imputation quality score < 0.3, genotype missing rate > 0.05, minor allele frequency (MAF) < 1%, and Hardy-Weinberg equilibrium \( p < 1.0 \times 10^{-6} \).
+
+Genome-wide association studies
+
+We performed linear regression models using PLINK v2.0 [19] to test the association of common variants (MAF \( \geq 1\% \)) with longevity for the entire cohort, as well as stratified by sex. For all three analyses, we used Martingale residuals calculated using the Cox proportional hazards model as the outcome variable. The procedure for calculating Martingale residuals was as follows. First, a Cox proportional hazards model [20] was fitted without genotype,
+
+\[
+H(t|z) = H_0(t)e^{z\beta'}
+\]
+where \( H_0(t) \) is the baseline hazard function at time point \( t \) given the last-known age and dead/alive status, \( Z = [Z_1, ..., Z_k] \) is a covariate matrix, and \( \beta = [\beta_1, ..., \beta_k] \) is a coefficient matrix for \( Z \). Here, we included sex and the first five principal components (PC) as covariates, but for sex-specific analyses, sex was excluded. Then, Martingale residuals were calculated as:
+
+\[
+\overline{M}_i = \delta_i - \overline{H}_0(t)e^{z\hat{\beta}'}
+\]
+
+where \( \delta_i \) is the dead/alive status (0=alive, 1=dead) of the ith subject and \( \hat{\beta} \) is the estimated coefficient matrix. We adapted the coxph function from the survival (v.3.6.4) R package [21] to compute the Martingale residuals. Genome-wide significance threshold was set at the standard GWAS level (\( p = 5.0 \times 10^{-8} \)). We used LocusZoom [22] to generate regional plots and Python v.3.7 to create Manhattan plots.
+
+Gene expression and colocalization analysis
+To evaluate the effect of the significant loci identified in our GWAS, we examined expression quantitative trait loci (eQTLs) across 49 tissues having at least 73 samples from the Genotype-Tissue Expression Project (GTEx) version 8 [23]. Bayesian colocalization analysis was employed using the COLOC package (v.5.2.3) [24] in R and the posterior probability of colocalization (PP4) was calculated between GWAS findings and eQTL associations within a 1 megabase (Mb) window. Additionally, colocalization was visualized using the locuscompareR package [25].
+
+Whole-exome sequencing and QC
+Whole exome sequencing (WES) data was available for 469,835 UKB participants. The dataset was generated by the Regeneron Genetics Center [26]. Details about the production and QC for the WES data are described previously [26]. We restricted the WES analysis to rare variants (MAF < 1%).
+
+Rare variant annotation
+Rare variants in WES data were annotated using Variant Effect Predictor (v. 112) provided by Ensembl [27]. We defined LoF variants as those with predicted consequences: splice acceptor, splice donor, stop gained, frameshift, start loss, stop loss, transcript ablation, feature elongation, or feature truncation. Missense variants were annotated using AlphaMissense [28] and REVEL [29] plugins and included if they had an AlphaMissense score \( \geq 0.7 \) or REVEL score \( \geq 0.75 \). All annotation was conducted based on GRCh38 genome coordinates.
+
+Gene-based rare variant association studies
+For testing groups of rare variants, genotype matrices were first transformed into a binary variable describing whether samples carry a variant of a given class as follows:
+
+\[
+G_i = \begin{cases}
+1, & \text{if } \sum_{j=1}^k g_{ij} > 0 \\
+0, & \text{if } \sum_{j=1}^k g_{ij} = 0
+\end{cases}
+\]
+
+Where \( g_{ij} \) is the minor allele count observed for subject \( i \) at variant \( j \) in the gene and \( k \) is the number of variants in the gene. We carried out two gene-based tests: Burden test and sequence kernel association test-optimized (SKAT-O) [30]. The burden test is a mean-based test that assumes the same direction of effects for all variants within a gene. On the other hand, SKAT-O employs a weighted average of the burden test and SKAT [31], the latter a variance-based test that does not lose power when variants have opposing directions of effect.
+Association tests were performed for each gene and rare variant class, including separately LoF variants, missense variants with an AlphaMissense score \( \geq 0.7 \), and missense variants with a REVEL score \( \geq 0.75 \), using Martingale residuals as the phenotype as in the common variant analyses. We excluded genes with fewer than 10 variant carriers to ensure the reliability of our analyses. A gene-wide significance threshold was established at \( p = 7.4 \times 10^{-7} \) based on the Bonferroni method accounting for the number of genes, variant classes, and statistical methods. Gene-based analyses were carried out using the SKAT package (v.2.2.5) in R.
+
+To characterize the impacts of gene burden in significant genes, we compared lifespan survival depending on gene burden using Kaplan–Meier survival curves, log-rank tests, and Cox proportional hazard regression analyses. Additionally, we performed Cox proportional hazards regression to assess the effect of each rare variant in a gene. The survival (v.3.6.4) package in R was utilized for the survival analysis.
+
+Phenome-wide association studies
+For gene-wide significant genes, we conducted phenome-wide association studies (PheWAS) of variant carrier status across 1,670 phenotypes in the UKB derived from binary, categorical, and continuous traits. Phenotypes included the International Classification of Disease 10 (ICD-10) codes, family history (e.g. father’s illness, father’s age at death), blood count (e.g. white blood cell count), blood biochemistry (e.g. Glucose levels), infectious diseases (e.g. pp 52 antigen for Human Cytomegalovirus), physical measures (e.g. BMI), cognitive test (e.g. pairs matching) and brain measurements (e.g. subcortical volume of hippocampus). For ICD-10 codes, we excluded phenotypes from the following ICD-10 chapters: 'Injuries, poisonings, and certain other consequences of external causes' (Chapter XIX), 'External causes of morbidity and mortality' (Chapter XX), 'Factors influencing health status and contacts with health services' (Chapter XXI), and 'Codes for special purposes' (Chapter XXII). The ICD-10 codes were then converted into Phecodes (v.1.2) [32] which combine correlated ICD codes into a distinct code and improve alignment with diseases commonly used in clinical practice.
+
+For binary traits, we removed phenotypes with fewer than 100 cases, and for continuous traits, those with fewer than 100 participants were excluded. Depending on the phenotype, we employed various regression models including binary logistic regression, ordinal logistic regression, multinomial logistic regression, and linear regression. All analyses included age, sex, and first five PCs as covariates. Phenome-wide significance threshold was set at \( p = 2.9 \times 10^{-5} \) based on the number of phenotypes.
+
+Variant allelic fraction
+To investigate whether some gene-level associations are enriched for somatic variants, we computed the variant allele frequency (VAF) for each heterozygous sample, reporting the mean VAF and VAF distribution per gene per variant class. VAF is defined as the number of reads with an alternate allele divided by the read depth at a given variant position. We also calculated the confidence interval for the mean VAF per gene using 10,000 bootstrap samples to ensure robust statistical analysis.
+
+Discussion
+In this study, we report several known and novel findings related to genetic risks associated with longevity analyzing 393,833 European participants from the UKB. In the common variant GWAS, three independent loci associated with increased mortality risk were identified. In the gene-based analysis of rare non-synonymous variants, 17 genes had their burden/SKAT-O test associated with longevity.
+Consistent with previous reports, rs429358, determining the \( APOE\text{-}\varepsilon4 \) allele dosage, was associated with decreased lifespan across both sexes. \( APOE\text{-}\varepsilon4 \) is well known for its associations with Alzheimer’s Disease [33] and cardiovascular disease [34]. In our dataset, the proportion of \( \varepsilon4 \) carriers was significantly higher for deaths caused by ‘Disease of the circulatory system’ and ‘Diseases of the nervous system’ compared to the general prevalence of \( \varepsilon4 \) carriers, which could explain the effect of \( \varepsilon4 \) on longevity. Examining the subcategories of these ICD-10 chapters, 'Disease of the circulatory system' includes cardiovascular disease (I51.6), while 'Diseases of the nervous system' covers Alzheimer's disease (G30). We also identified a GWS association at the ZSCAN23 locus, which had not been previously reported. Our colocalization analysis revealed that the longevity-associated signal colocalizes with a ZSCAN23 eQTL in pancreatic tissue with increased expression observed in minor allele carriers. Although the role of ZSCAN23 remains unclear, recent studies have linked its expression to pancreatic tumors, supporting our colocalization findings [35]. For sex-specific GWAS, a GWS association specific to males was found between \( MUC5AC \) and \( MUC5B \), which highly colocalizes with a \( MUC5B \) eQTL in lung tissue and many studies have linked this variant to pulmonary disease like idiopathic pulmonary fibrosis [36, 37] and COVID-19 [38, 39]. Previously reported SNP associations with longevity were concordant in our dataset but none of these passed the GWAS suggestive threshold (\( p=1.0\times10^{-5} \)) except for those at the \( APOE \) locus. This phenomenon likely resulted from previous studies relying on proxy data such as parental age at death, which may capture a different set of genetic factors than direct proband mortality data.
+
+In our gene-based rare variant analysis, 17 genes achieved gene-wide significance (\( p <7.4\times10^{-7} \)) in either the burden or SKAT-O test. Four of these, \( TET2 \), \( ATM \), \( BRCA2 \), and \( BRCA1 \), were reported in a previous rare-variant analysis of longevity in UKB [9]. We identified 13 novel genes associated with longevity—\( CKMT1B \), \( ASXL1 \), \( DNMT3A \), \( SF3B1 \), \( PTEN \), \( SOX21 \), \( TP53 \), \( SRSF2 \), \( Clorf52 \), \( CHL1 \), \( IDH2 \), \( RLIM \), and \( TERT \)—when assessing variants causing genetic LoF or missense variants classified as pathogenic by REVEL or AlphaMissense. Of note, LoF and missense variant analyses identified mostly separate genes with only one overlap (\( TET2 \)). This supports the use of both categories in rare variant analyses and may indicate that missense variants as classified by AlphaMissense capture a wider range of variation missed when only assessing LoF variants, which are generally interpreted as resulting in haploinsufficiency. Importantly, missense variants may lead to increased or decreased protein function. In our analyses, \( IDH2 \) was not gene-wide significant with the burden test (\( p=1.9\times10^{-4} \), Table 1) but was highly significant with SKAT-O (\( p=5.3\times10^{-39} \), Table 1). Since SKAT-O does not lose power when variants have differing directions of effect, this suggests that different mutations in \( IDH2 \) can lead to either increased or decreased longevity. These results underline the gain in information achieved when studying rare missense variants as well as LoF using appropriate statistical techniques.
+
+Strikingly, most of the genes we identified carrying longevity-associated rare variants have been previously linked to cancer. \( TET2 \), \( ASXL1 \), \( DNMT3A \), and \( SF3B1 \) are all known to harbor causal leukemia variants [40-43], and somatic variants in \( SRSF2 \) have been described in myelodysplastic syndrome [44]. \( ATM \), \( BRCA2 \), and \( BRCA1 \) mutations have been well characterized in breast, ovarian, and other cancers [45-47]. \( RLIM \) appears to be a regulator of estrogen-dependent transcription, an important pathway in breast cancer [48], and has been recently described as a potential tumor suppressor [49]. \( PTEN \) and \( TP53 \) are well studied due to their critical role in genomic stability and are the two most mutated genes in human cancer [50]. \( IDH2 \) is also frequently mutated in many kinds of cancer [51]. The antisense long noncoding RNA \( SOX21\text{-}ASI \), but not \( SOX21 \), has been linked to oral, cervical, and breast cancer [52-54]. A recent study found potential for \( CKMT1B \) expression as a prognostic
+biomarker in glioma [55]; similarly, alterations in CHL1 expression have been associated with development and metastasis of many types of cancer [56]. Finally, variation in both the coding and promoter sequences of TERT has been associated with a variety of cancer types [57, 58].
+Our PheWAS results also suggest that most of these genes are associated with cancer, specifically blood-based tumors such as myeloid leukemia. Combined with the common ZSCAN23 locus we identified, associated with pancreatic tumors, this points to cancer being the major genetic factor currently affecting lifespan in UKB. This is consistent with a previous study of healthspan that found cancer to be the first emerging disease in over half of disease cases in UKB [59]. These results likely reflect the characteristics of the cohort, comprised of predominantly middle-aged individuals, with age-at-death ranging from 40.9 to 85.2 years and last-known ages between 52.6 and 88.7 years.
+For sex-specific rare variant analyses, we identified six novel genes (CDKN1A, PTPRK, COA7, TG, NMNAT2 and PITRM1) in males and three genes (PORCN, UGT1A8 and OLIG1) in females. Some of these genes have been found to associate with sex-specific diseases. In one study, advanced prostate cancer patients had a higher frequency of a variant on the 3'UTR of CDKN1A [60] and the gene has received attention as a potential therapeutic target for prostate cancer [61]. PORCN is located on the X chromosome and mutations on it can cause Goltz-Gorlin Syndrome [62], but it has also been found to regulate a signaling pathway that controls cancer cell growth [63]. UGT1A8 expression is altered in endometrial cancer [64] and amino acid substitutions in it may modulate estradiol metabolism leading to increased risk of breast and endometrial cancer [65].
+Since UKB collected DNA from peripheral blood mononuclear cell samples, we explored whether the variants were potentially of somatic origin, picked up by WES genotyping due to CHIP. The VAF distribution of variants included in our analysis emphasizes that several associations are likely linked to CHIP, and notably include the well-established CHIP-related genes TET2, ASXL1, DNMT3A, SF3B1, TP53 and SRSF2. While WES heterozygote genotypes for these variants will not include all variants with some degree of CHIP within these genes (as evidenced by many more individuals having non-zero alternate allele count at these locations, data not shown), it does capture CHIP-related somatic variants sufficiently to establish robust associations with longevity. In UKB the mean duration between the primary visit (blood draw date) and death is currently 9.2 years (± 3.8) and suggests that WES screening for CHIP variants may be used as a precision health tool to contribute to earlier cancer detection by assessing individuals with higher susceptibility risks. In addition to known cancer variants, such as breast cancer-related BRCA1/BRCA2, our study highlights novel associations that should be considered in cancer susceptibility screenings.
+By combining large-scale GWAS with rare variant analysis, this study enhances our understanding of the genetic basis of human longevity. Our results emphasize the importance of understanding the genetic factors driving the most prevalent causes of mortality on a population level, highlighting the potential for early genetic testing to identify germline and somatic variants that place some individuals at risk of early death. Understanding the biological pathways through which these genes influence cancer and aging, as well as the environmental factors interacting with these pathways, will be essential for developing therapeutic targets aimed at extending healthy lifespan. Our study's implications thus extend beyond genetics, as they touch on the broader aspects of health care, public health policy, and preventive strategies against age-related diseases.
+In conclusion, this study enhances our understanding of the genetic basis of human longevity by combining large-scale GWAS with detailed rare variant analysis. The novel loci identified warrant further exploration to understand their biological roles and interactions with
+environmental factors, which will be crucial for unraveling the complex nature of aging and developing strategies to mitigate its adverse effects.
+Figures and Tables
+
+Figure 1. Common variant GWAS of longevity. (A) Manhattan plot. (B) The proportion of cause of death for the top 4 categories, each accounting for more than 5% of total deaths) (C) Association of causes of death with \( APOE-e4 \) genotype. (D) Locuszoom and (E) colocalization plots at the ZSCAN23 locus, colocalized with ZSCAN23 eQTL in pancreatic tissue in GTEx. PP4: posterior probability of colocalization.
+
+(A)
+
+![Manhattan plot showing genetic variants and their association with longevity](page_232_370_1012_246.png)
+
+(B)
+
+![Bar chart showing proportions of diseases by category](page_232_629_388_246.png)
+
+(C)
+
+![Bar chart showing proportions of diseases by category and APOE-e4 genotype](page_629_629_1012_246.png)
+
+(D)
+
+![Locuszoom plot showing genetic variants and their association with longevity](page_232_888_388_246.png)
+
+(E)
+
+![Colocalization plot showing PP4 and rs13190937 association](page_629_888_1012_246.png)
+Figure 2. Rare variant burden association with longevity, considering loss-of-functions (A) and Alpha Missense pathogenic variants (B). Novel genes are highlighted in red.
+
+(A)
+
+![Loss of Function plot showing chromosomes 1-22 and X, with points labeled TET2, BRCA2, ATM, CKMT1B, and ASXL1 above the threshold line.](page_347_186_1057_312.png)
+
+(B)
+
+![Alpha Missense plot showing chromosomes 1-22 and X, with points labeled DNMT3A, SF3B1, CHL1, TET2, PTEN, SOX21, TP53, and SRSF2 above the threshold line.](page_347_591_1057_312.png)
+Figure 3. Survival curves comparing carriers and non-carriers of variants on genes with a significant burden of loss-of-function (A) and AlphaMissense pathogenic (B) variants.
+
+(A) Loss of Function
+
+![Survival curves for CKMT1B and ASXL1 showing HR and p-value for carriers vs. non-carriers](page_184_156_579_246.png)
+
+(B) Alpha Missense
+
+![Survival curves for DNMT3A, SF3B1, CHL1, PTEN, SOX21, TP53, and SRSF2 showing HR and p-value for carriers vs. non-carriers](page_184_406_1082_482.png)
+Supplementary Figure 1. Phenome-wide association of rs13190937 on ZSCAN23. This analysis is based on PheWeb (https://pheweb.org/UKB-Neale/).
+
+![Phenome-wide association plot showing rs13190937 on ZSCAN23](page_246_355_1012_180.png)
+Supplementary Figure 2. Sex-stratified common variant GWAS of longevity. Manhattan plot in males (A), and females (B). Locuszoom (C) and colocalization (D) plots at the MUC5B locus in males, colocalized with MUC5B eQTL in lung tissue in GTEx. PP4: posterior probability of colocalization. (E) Phenom-wide association of rs35705950. This analysis is based on PheWeb (https://pheweb.org/UKB-Neale/).
+
+(A)
+
+![Manhattan plot for males showing chromosomes 1-22 and X, with highlighted regions for MUC5B and APOE](page_186_340_670_246.png)
+
+(B)
+
+![Manhattan plot for females showing chromosomes 1-22 and X, with highlighted regions for MUC5B and APOE](page_186_601_670_246.png)
+
+(C)
+
+![Locuszoom plot for rs35705950 in males](page_866_340_670_246.png)
+
+(D)
+
+![Colocalization plots for rs35705950 in males](page_866_601_670_246.png)
+
+(E)
+
+![Phenom-wide association plot for rs35705950](page_186_866_1076_120.png)
+Supplementary Figure 3. Rare variant SKAT-O association with longevity considering 3 categories: Loss-of-function (A), Alpha Missense (B), and REVEL (C). Novel genes are highlighted in red.
+
+(A)
+
+![Loss of Function plot showing chromosomes 1-22 and X, with highlighted genes TET2, BRCA2, ATM, BRC41, ASXL1](page_256_287_1047_377.png)
+
+(B)
+
+![Alpha Missense plot showing chromosomes 1-22 and X, with highlighted genes SRSF2, IDH2, SOX21, TP53, RLIM, PTEN, SOX21, DNMT3A, SF3B1, C1orf52](page_256_682_1047_377.png)
+
+(C)
+
+![Missense (REVEL) plot showing chromosomes 1-22 and X, with highlighted genes TERT, PTEN, TP53, DNMT3A](page_256_1077_1047_377.png)
+Supplementary Figure 4. Rare variant burden association with longevity considering REVEL pathogenic missense variants. Novel genes are highlighted in red.
+
+![Rare variant burden association with longevity considering REVEL pathogenic missense variants. Novel genes are highlighted in red.](page_328_312_1096_377.png)
+Supplementary Figure 5. Sex-stratified rare variant burden association with longevity considering 3 categories for each sex: Loss-of-function (A), Alpha Missense (B), and REVEL (C). Novel genes are highlighted in red.
+
+![Sex-stratified rare variant burden association with longevity considering 3 categories for each sex: Loss-of-function (A), Alpha Missense (B), and REVEL (C). Novel genes are highlighted in red.](page_184_276_1082_670.png)
+Supplementary Figure 6. Sex-stratified rare variants SKAT-O association with longevity considering 3 categories for each sex: Loss-of-function (A), Alpha Missense (B), and REVEL (C). Novel genes are highlighted in red.
+
+![Sex-stratified rare variants SKAT-O association with longevity considering 3 categories for each sex: Loss-of-function (A), Alpha Missense (B), and REVEL (C). Novel genes are highlighted in red.](page_184_273_1082_693.png)
+Supplementary Figure 7. Survival curves comparing carriers and non-carriers of variants considered on genes with a significant burden of loss-of-function (TET2, ATM, BRCA2 and BRCA1) (A), AlphaMissense pathogenic (B) variants (TET2), missense variants predicted by REVEL (DNMT3A, PTEN and TP53) (C)
+
+(A) Loss of Function
+
+![Survival curves for TET2, ATM, BRCA2, BRCA1 loss-of-function variants](page_184_186_1082_232.png)
+
+(B) Alpha Missense
+
+(C) Missense (REVEL)
+
+![Survival curves for TET2, DNMT3A, PTEN, TP53 missense variants](page_184_420_1082_232.png)
+Supplementary Figure 8. Phenome-wide association of the burden of rare variants at the nine novel genes identified in our burden test. Variants considered correspond to loss-of-function and Alpha missense defined variants. P-values less than \(1.0 \times 10^{-50}\) are capped at 50.
+
+![Phenome-wide association plots for CKMT1B, ASXL1, DNMT3A, SF3B1, CHL1, PTEN, SOX21, TP53, SRSF2](page_184_370_1082_1016.png)
+Supplementary Figure 9. Variant allelic fraction distribution per gene for variants considered in each category: Loss-of-function (A) and Alpha Missense (B).
+
+(A) Loss of Function
+
+(B) Alpha Missense
+
+![Variant allelic fraction distribution per gene for variants considered in each category: Loss-of-function (A) and Alpha Missense (B).](page_184_232_1076_579.png)
+Table 1. Significant genes for rare variants association with burden and SKAT-O tests (\( p<7.4 \times 10^{-7} \)). Gene names in bold font represent novel associations.
+
+<table>
+  <tr>
+    <th>Variant Class</th>
+    <th>Chr</th>
+    <th>Gene</th>
+    <th># of variants</th>
+    <th># of carriers</th>
+    <th>Burden \( p\)-value</th>
+    <th>SKAT-O \( p\)-value</th>
+  </tr>
+  <tr>
+    <td rowspan="6">LoF</td>
+    <td>4</td>
+    <td><i>TET2</i></td>
+    <td>243</td>
+    <td>563</td>
+    <td>3.8 \times 10^{-30}</td>
+    <td>3.4 \times 10^{-54}</td>
+  </tr>
+  <tr>
+    <td>11</td>
+    <td><i>ATM</i></td>
+    <td>247</td>
+    <td>1,170</td>
+    <td>6.0 \times 10^{-10}</td>
+    <td>4.8 \times 10^{-11}</td>
+  </tr>
+  <tr>
+    <td>13</td>
+    <td><i>BRCA2</i></td>
+    <td>245</td>
+    <td>1,271</td>
+    <td>1.3 \times 10^{-34}</td>
+    <td>1.2 \times 10^{-42}</td>
+  </tr>
+  <tr>
+    <td>15</td>
+    <td><b><i>CKMT1B</i></b></td>
+    <td>15</td>
+    <td>40</td>
+    <td>4.5 \times 10^{-7}</td>
+    <td>1.8 \times 10^{-6}</td>
+  </tr>
+  <tr>
+    <td>17</td>
+    <td><i>BRCA1</i></td>
+    <td>120</td>
+    <td>456</td>
+    <td>4.9 \times 10^{-12}</td>
+    <td>1.3 \times 10^{-11}</td>
+  </tr>
+  <tr>
+    <td>20</td>
+    <td><i>ASXL1</i></td>
+    <td>72</td>
+    <td>533</td>
+    <td>2.3 \times 10^{-44}</td>
+    <td>2.9 \times 10^{-46}</td>
+  </tr>
+  <tr>
+    <td rowspan="10">Alpha Missense</td>
+    <td>1</td>
+    <td><i>C1orf52</i></td>
+    <td>23</td>
+    <td>175</td>
+    <td>4.5 \times 10^{-5}</td>
+    <td>2.1 \times 10^{-7}</td>
+  </tr>
+  <tr>
+    <td>2</td>
+    <td><i>DNMT3A</i></td>
+    <td>167</td>
+    <td>1,229</td>
+    <td>1.4 \times 10^{-9}</td>
+    <td>1.8 \times 10^{-10}</td>
+  </tr>
+  <tr>
+    <td>2</td>
+    <td><i>SF3B1</i></td>
+    <td>64</td>
+    <td>195</td>
+    <td>6.7 \times 10^{-12}</td>
+    <td>2.5 \times 10^{-16}</td>
+  </tr>
+  <tr>
+    <td>3</td>
+    <td><i>CHL1</i></td>
+    <td>33</td>
+    <td>3,666</td>
+    <td>5.0 \times 10^{-7}</td>
+    <td>1.1 \times 10^{-6}</td>
+  </tr>
+  <tr>
+    <td>4</td>
+    <td><i>TET2</i></td>
+    <td>159</td>
+    <td>826</td>
+    <td>4.2 \times 10^{-7}</td>
+    <td>7.4 \times 10^{-7}</td>
+  </tr>
+  <tr>
+    <td>10</td>
+    <td><i>PTEN</i></td>
+    <td>50</td>
+    <td>71</td>
+    <td>1.0 \times 10^{-8}</td>
+    <td>3.9 \times 10^{-11}</td>
+  </tr>
+  <tr>
+    <td>13</td>
+    <td><i>SOX21</i></td>
+    <td>52</td>
+    <td>463</td>
+    <td>2.9 \times 10^{-8}</td>
+    <td>6.7 \times 10^{-9}</td>
+  </tr>
+  <tr>
+    <td>15</td>
+    <td><i>IDH2</i></td>
+    <td>89</td>
+    <td>349</td>
+    <td>1.9 \times 10^{-4}</td>
+    <td>5.3 \times 10^{-39}</td>
+  </tr>
+  <tr>
+    <td>17</td>
+    <td><i>TP53</i></td>
+    <td>35</td>
+    <td>90</td>
+    <td>3.1 \times 10^{-15}</td>
+    <td>2.5 \times 10^{-16}</td>
+  </tr>
+  <tr>
+    <td>17</td>
+    <td><i>SRSF2</i></td>
+    <td>14</td>
+    <td>141</td>
+    <td>9.8 \times 10^{-89}</td>
+    <td>8.3 \times 10^{-108}</td>
+  </tr>
+  <tr>
+    <td>X</td>
+    <td><i>RLIM</i></td>
+    <td>25</td>
+    <td>51</td>
+    <td>8.8 \times 10^{-7}</td>
+    <td>3.7 \times 10^{-9}</td>
+  </tr>
+</table>
+
+LoF: Loss of Function; Chr: chromosome
+Table 2. Lead variant association per gene among significant genes in the burden and SKAT-O tests. Only variants with at least 3 minor alleles are reported.
+
+<table>
+  <tr>
+    <th>Variant Class</th>
+    <th>Chr</th>
+    <th>Gene</th>
+    <th>Variant</th>
+    <th>MA</th>
+    <th>MAC</th>
+    <th>AM</th>
+    <th>HR</th>
+    <th>p-value</th>
+    <th>Reported</th>
+  </tr>
+  <tr>
+    <td rowspan="6">LoF</td>
+    <td>4</td>
+    <td><i>TET2</i></td>
+    <td>rs370735654</td>
+    <td>T</td>
+    <td>17</td>
+    <td>-</td>
+    <td>7.0</td>
+    <td>5.6 \times 10^{-9}</td>
+    <td>-</td>
+  </tr>
+  <tr>
+    <td>11</td>
+    <td><i>ATM</i></td>
+    <td>rs587779834</td>
+    <td>A</td>
+    <td>113</td>
+    <td>-</td>
+    <td>2.5</td>
+    <td>2.7 \times 10^{-5}</td>
+    <td>-</td>
+  </tr>
+  <tr>
+    <td>13</td>
+    <td><i>BRCA2</i></td>
+    <td>rs80359520</td>
+    <td>C</td>
+    <td>10</td>
+    <td>-</td>
+    <td>6.1</td>
+    <td>1.8 \times 10^{-6}</td>
+    <td>Breast Cancer [66]</td>
+  </tr>
+  <tr>
+    <td>15</td>
+    <td><i>CKMT1B</i></td>
+    <td>rs1355844751</td>
+    <td>T</td>
+    <td>8</td>
+    <td>-</td>
+    <td>4.9</td>
+    <td>5.9 \times 10^{-3}</td>
+    <td>-</td>
+  </tr>
+  <tr>
+    <td>17</td>
+    <td><i>BRCA1</i></td>
+    <td>rs80357508</td>
+    <td>C</td>
+    <td>44</td>
+    <td>-</td>
+    <td>2.8</td>
+    <td>3.5 \times 10^{-3}</td>
+    <td>-</td>
+  </tr>
+  <tr>
+    <td>20</td>
+    <td><i>ASXL1</i></td>
+    <td>rs750318549</td>
+    <td>AG</td>
+    <td>201</td>
+    <td>-</td>
+    <td>2.5</td>
+    <td>2.3 \times 10^{-15}</td>
+    <td>-</td>
+  </tr>
+  <tr>
+    <td rowspan="6">Alpha Missense</td>
+    <td>1</td>
+    <td><i>C1orf52</i></td>
+    <td>rs769009649</td>
+    <td>A</td>
+    <td>62</td>
+    <td>0.876</td>
+    <td>3.2</td>
+    <td>7.4 \times 10^{-7}</td>
+    <td>-</td>
+  </tr>
+  <tr>
+    <td>2</td>
+    <td><i>DNMT3A</i></td>
+    <td>rs147001633</td>
+    <td>T</td>
+    <td>269</td>
+    <td>0.995</td>
+    <td>1.8</td>
+    <td>3.7 \times 10^{-5}</td>
+    <td>Leukemia [67, 68]</td>
+  </tr>
+  <tr>
+    <td>2</td>
+    <td><i>SF3B1</i></td>
+    <td>rs377023736</td>
+    <td>A</td>
+    <td>12</td>
+    <td>0.999</td>
+    <td>6.0</td>
+    <td>8.5 \times 10^{-8}</td>
+    <td>-</td>
+  </tr>
+  <tr>
+    <td>3</td>
+    <td><i>CHL1</i></td>
+    <td>rs116421102</td>
+    <td>C</td>
+    <td>1,842</td>
+    <td>0.745</td>
+    <td>1.3</td>
+    <td>2.5 \times 10^{-5}</td>
+    <td>-</td>
+  </tr>
+  <tr>
+    <td>4</td>
+    <td><i>TET2</i></td>
+    <td>rs76428136</td>
+    <td>G</td>
+    <td>5</td>
+    <td>0.913</td>
+    <td>8.1</td>
+    <td>2.9 \times 10^{-4}</td>
+    <td>-</td>
+  </tr>
+  <tr>
+    <td>10</td>
+    <td><i>PTEN</i></td>
+    <td>rs587782350</td>
+    <td>T</td>
+    <td>3</td>
+    <td>0.941</td>
+    <td>20.4</td>
+    <td>2.6 \times 10^{-3}</td>
+    <td>-</td>
+  </tr>
+  <tr>
+    <td>13</td>
+    <td><i>SOX21</i></td>
+    <td>rs1172148601</td>
+    <td>A</td>
+    <td>67</td>
+    <td>0.856</td>
+    <td>2.3</td>
+    <td>1.7 \times 10^{-3}</td>
+    <td>-</td>
+  </tr>
+  <tr>
+    <td>15</td>
+    <td><i>IDH2</i></td>
+    <td>rs121913502</td>
+    <td>T</td>
+    <td>45</td>
+    <td>0.987</td>
+    <td>5.7</td>
+    <td>1.0 \times 10^{-20}</td>
+    <td>Leukemia [68, 69]</td>
+  </tr>
+  <tr>
+    <td>17</td>
+    <td><i>TP53</i></td>
+    <td>rs11540652</td>
+    <td>T</td>
+    <td>5</td>
+    <td>0.996</td>
+    <td>10.0</td>
+    <td>6.6 \times 10^{-5}</td>
+    <td>Gastric Cancer [70], Ovarian Cancer [71]</td>
+  </tr>
+  <tr>
+    <td>17</td>
+    <td><i>SRSF2</i></td>
+    <td>rs751713049</td>
+    <td>T</td>
+    <td>51</td>
+    <td>0.982</td>
+    <td>5.8</td>
+    <td>1.9 \times 10^{-26}</td>
+    <td>-</td>
+  </tr>
+  <tr>
+    <td>X</td>
+    <td><i>RLIM</i></td>
+    <td>X:74592998:C:A</td>
+    <td>A</td>
+    <td>3</td>
+    <td>0.970</td>
+    <td>3.9</td>
+    <td>1.3 \times 10^{-3}</td>
+    <td>-</td>
+  </tr>
+</table>
+
+Chr: chromosome; MAC: minor allele count; AM: AlphaMissense score; HR: hazard ratio
+Supplementary Table 1. Demographics of European ancestry in the analyses.
+
+<table>
+  <tr>
+    <th rowspan="2"></th>
+    <th colspan="3">All</th>
+    <th colspan="3">Male</th>
+    <th colspan="3">Female</th>
+  </tr>
+  <tr>
+    <th>Total</th>
+    <th>Living</th>
+    <th>Deceased</th>
+    <th>Total</th>
+    <th>Living</th>
+    <th>Deceased</th>
+    <th>Total</th>
+    <th>Living</th>
+    <th>Deceased</th>
+  </tr>
+  <tr>
+    <td>N</td>
+    <td>393,833</td>
+    <td>358,282</td>
+    <td>35,551</td>
+    <td>180,970</td>
+    <td>159,911</td>
+    <td>21,059</td>
+    <td>212,863</td>
+    <td>198,371</td>
+    <td>14,492</td>
+  </tr>
+  <tr>
+    <td>Last known age</td>
+    <td>70.8 ± 7.9</td>
+    <td>70.7 ± 8.0</td>
+    <td>71.2 ± 7.5</td>
+    <td>70.9 ± 8.0</td>
+    <td>70.8 ± 8.1</td>
+    <td>71.3 ± 7.4</td>
+    <td>70.7 ± 7.9</td>
+    <td>70.7 ± 7.9</td>
+    <td>71.1 ± 7.6</td>
+  </tr>
+  <tr>
+    <td><i>APOE ε4</i> carrier</td>
+    <td>113,437 (28.8%)</td>
+    <td>102,360 (28.6%)</td>
+    <td>11,077 (31.2%)</td>
+    <td>52,190 (28.8%)</td>
+    <td>45,635 (28.5%)</td>
+    <td>6,555 (31.1%)</td>
+    <td>61,247 (28.8%)</td>
+    <td>56,725 (28.6%)</td>
+    <td>4,522 (31.2%)</td>
+  </tr>
+</table>
+Supplementary Table 2. Significant genes for burden and SKAT-O association of rare variants, considering missense variants with REVEL > 75. Gene names in bold font represent novel associations in REVEL.
+
+<table>
+  <tr>
+    <th>Variant Class</th>
+    <th>Chr</th>
+    <th>Gene</th>
+    <th># of variants</th>
+    <th># of carriers</th>
+    <th>Burden <i>p</i>-value</th>
+    <th>SKAT-O <i>p</i>-value</th>
+  </tr>
+  <tr>
+    <td rowspan="4">REVEL (&gt;75)</td>
+    <td>2</td>
+    <td><b>DNMT3A</b></td>
+    <td>116</td>
+    <td>831</td>
+    <td>6.6 \times 10^{-9}</td>
+    <td>3.5 \times 10^{-8}</td>
+  </tr>
+  <tr>
+    <td>5</td>
+    <td><b>TERT</b></td>
+    <td>31</td>
+    <td>60</td>
+    <td>2.6 \times 10^{-3}</td>
+    <td>8.1 \times 10^{-10}</td>
+  </tr>
+  <tr>
+    <td>10</td>
+    <td><b>PTEN</b></td>
+    <td>42</td>
+    <td>45</td>
+    <td>6.6 \times 10^{-8}</td>
+    <td>3.6 \times 10^{-10}</td>
+  </tr>
+  <tr>
+    <td>17</td>
+    <td><b>TP53</b></td>
+    <td>47</td>
+    <td>173</td>
+    <td>6.2 \times 10^{-9}</td>
+    <td>6.5 \times 10^{-9}</td>
+  </tr>
+</table>
+Supplementary Table 3. Significant genes for burden and SKAT-O association of rare variants in males. Genes in bold font represent novel associations in males.
+
+<table>
+  <tr>
+    <th>Variant Class</th>
+    <th>Chr</th>
+    <th>Gene</th>
+    <th># of variants</th>
+    <th># of carriers</th>
+    <th>Burden <i>p</i>-value</th>
+    <th>SKAT-O <i>p</i>-value</th>
+  </tr>
+  <tr>
+    <td rowspan="5">LoF</td>
+    <td>4</td>
+    <td><i>TET2</i></td>
+    <td>162</td>
+    <td>280</td>
+    <td>4.1 \times 10^{-20}</td>
+    <td>1.5 \times 10^{-32}</td>
+  </tr>
+  <tr>
+    <td>6</td>
+    <td><b>CDKN1A</b></td>
+    <td>8</td>
+    <td>33</td>
+    <td>1.1 \times 10^{-4}</td>
+    <td>6.3 \times 10^{-8}</td>
+  </tr>
+  <tr>
+    <td>6</td>
+    <td><b>PTPRK</b></td>
+    <td>26</td>
+    <td>40</td>
+    <td>5.9 \times 10^{-3}</td>
+    <td>4.5 \times 10^{-7}</td>
+  </tr>
+  <tr>
+    <td>11</td>
+    <td><i>ATM</i></td>
+    <td>318</td>
+    <td>520</td>
+    <td>2.6 \times 10^{-10}</td>
+    <td>2.0 \times 10^{-10}</td>
+  </tr>
+  <tr>
+    <td>13</td>
+    <td><i>BRCA2</i></td>
+    <td>172</td>
+    <td>596</td>
+    <td>2.3 \times 10^{-16}</td>
+    <td>1.5 \times 10^{-19}</td>
+  </tr>
+  <tr>
+    <td>20</td>
+    <td><i>ASXL1</i></td>
+    <td>59</td>
+    <td>347</td>
+    <td>5.4 \times 10^{-36}</td>
+    <td>6.4 \times 10^{-37}</td>
+  </tr>
+  <tr>
+    <td rowspan="6">Alpha Missense (>70)</td>
+    <td>1</td>
+    <td><i>C1orf52</i></td>
+    <td>19</td>
+    <td>76</td>
+    <td>2.2 \times 10^{-5}</td>
+    <td>2.7 \times 10^{-10}</td>
+  </tr>
+  <tr>
+    <td>1</td>
+    <td><b>COA7</b></td>
+    <td>8</td>
+    <td>11</td>
+    <td>1.0 \times 10^{-4}</td>
+    <td>3.1 \times 10^{-8}</td>
+  </tr>
+  <tr>
+    <td>2</td>
+    <td><i>SF3B1</i></td>
+    <td>43</td>
+    <td>122</td>
+    <td>1.6 \times 10^{-11}</td>
+    <td>2.8 \times 10^{-14}</td>
+  </tr>
+  <tr>
+    <td>4</td>
+    <td><i>TET2</i></td>
+    <td>107</td>
+    <td>405</td>
+    <td>1.1 \times 10^{-8}</td>
+    <td>1.6 \times 10^{-8}</td>
+  </tr>
+  <tr>
+    <td>8</td>
+    <td><b>TG</b></td>
+    <td>113</td>
+    <td>657</td>
+    <td>2.4 \times 10^{-7}</td>
+    <td>1.2 \times 10^{-6}</td>
+  </tr>
+  <tr>
+    <td>15</td>
+    <td><i>IDH2</i></td>
+    <td>58</td>
+    <td>171</td>
+    <td>2.2 \times 10^{-3}</td>
+    <td>5.5 \times 10^{-29}</td>
+  </tr>
+  <tr>
+    <td>17</td>
+    <td><i>TP53</i></td>
+    <td>24</td>
+    <td>48</td>
+    <td>1.5 \times 10^{-9}</td>
+    <td>1.1 \times 10^{-9}</td>
+  </tr>
+  <tr>
+    <td>17</td>
+    <td><i>SRSF2</i></td>
+    <td>10</td>
+    <td>104</td>
+    <td>5.4 \times 10^{-63}</td>
+    <td>1.7 \times 10^{-70}</td>
+  </tr>
+  <tr>
+    <td rowspan="2">REVEL (>75)</td>
+    <td>1</td>
+    <td><i>NMNAT2</i></td>
+    <td>21</td>
+    <td>34</td>
+    <td>1.2 \times 10^{-4}</td>
+    <td>1.9 \times 10^{-8}</td>
+  </tr>
+  <tr>
+    <td>10</td>
+    <td><b>PITRM1</b></td>
+    <td>6</td>
+    <td>10</td>
+    <td>1.5 \times 10^{-6}</td>
+    <td>3.6 \times 10^{-8}</td>
+  </tr>
+</table>
+Supplementary Table 4. Significant genes for burden and SKAT-O association of rare variants in females. Genes in bold font represent novel associations in females.
+
+<table>
+  <tr>
+    <th>Variant Class</th>
+    <th>Chr</th>
+    <th>Gene</th>
+    <th># of variants</th>
+    <th># of carriers</th>
+    <th>Burden \( p \)-value</th>
+    <th>SKAT-O \( p \)-value</th>
+  </tr>
+  <tr>
+    <td rowspan="4">LoF</td>
+    <td>4</td>
+    <td><i>TET2</i></td>
+    <td>151</td>
+    <td>283</td>
+    <td>7.8 \times 10^{-11}</td>
+    <td>5.1 \times 10^{-24}</td>
+  </tr>
+  <tr>
+    <td>13</td>
+    <td><i>BRCA2</i></td>
+    <td>182</td>
+    <td>675</td>
+    <td>1.2 \times 10^{-20}</td>
+    <td>4.8 \times 10^{-27}</td>
+  </tr>
+  <tr>
+    <td>17</td>
+    <td><i>BRCA1</i></td>
+    <td>80</td>
+    <td>217</td>
+    <td>3.6 \times 10^{-12}</td>
+    <td>2.4 \times 10^{-11}</td>
+  </tr>
+  <tr>
+    <td>20</td>
+    <td><i>ASXL1</i></td>
+    <td>46</td>
+    <td>186</td>
+    <td>1.1 \times 10^{-6}</td>
+    <td>2.0 \times 10^{-7}</td>
+  </tr>
+  <tr>
+    <td rowspan="6">Alpha Missense (>70)</td>
+    <td>2</td>
+    <td><i>DNMT3A</i></td>
+    <td>135</td>
+    <td>671</td>
+    <td>7.6 \times 10^{-9}</td>
+    <td>2.3 \times 10^{-10}</td>
+  </tr>
+  <tr>
+    <td>5</td>
+    <td><i>TERT</i></td>
+    <td>22</td>
+    <td>27</td>
+    <td>1.7 \times 10^{-3}</td>
+    <td>2.2 \times 10^{-7}</td>
+  </tr>
+  <tr>
+    <td>13</td>
+    <td><i>SOX21</i></td>
+    <td>34</td>
+    <td>251</td>
+    <td>2.0 \times 10^{-7}</td>
+    <td>1.6 \times 10^{-7}</td>
+  </tr>
+  <tr>
+    <td>17</td>
+    <td><i>SRSF2</i></td>
+    <td>10</td>
+    <td>37</td>
+    <td>3.3 \times 10^{-16}</td>
+    <td>8.4 \times 10^{-27}</td>
+  </tr>
+  <tr>
+    <td>17</td>
+    <td><i>TP53</i></td>
+    <td>23</td>
+    <td>52</td>
+    <td>1.1 \times 10^{-7}</td>
+    <td>1.8 \times 10^{-8}</td>
+  </tr>
+  <tr>
+    <td>X</td>
+    <td><b>PORCN</b></td>
+    <td>11</td>
+    <td>32</td>
+    <td>5.6 \times 10^{-4}</td>
+    <td>3.7 \times 10^{-7}</td>
+  </tr>
+  <tr>
+    <td rowspan="2">REVEL (>75)</td>
+    <td>2</td>
+    <td><i>UGT1A8</i></td>
+    <td>2</td>
+    <td>18</td>
+    <td>2.8 \times 10^{-7}</td>
+    <td>2.8 \times 10^{-7}</td>
+  </tr>
+  <tr>
+    <td>21</td>
+    <td><i>OLIG1</i></td>
+    <td>7</td>
+    <td>18</td>
+    <td>1.3 \times 10^{-5}</td>
+    <td>7.0 \times 10^{-7}</td>
+  </tr>
+</table>
+Supplementary Table 5. Lead variant association per gene among significant genes in the burden and SKAT-O tests. Only significant variant associations with at least 3 minor allele counts per gene are reported in this table.
+
+<table>
+  <tr>
+    <th>Variant Class</th>
+    <th>Chr</th>
+    <th>Gene</th>
+    <th>Variant</th>
+    <th>MA</th>
+    <th>MAC</th>
+    <th>AM</th>
+    <th>HR</th>
+    <th>p-value</th>
+    <th>Reported</th>
+  </tr>
+  <tr>
+    <td rowspan="4">REVEL (&gt;75)</td>
+    <td>2</td>
+    <td><i>DNMT3A</i></td>
+    <td>rs367909007</td>
+    <td>G</td>
+    <td>14</td>
+    <td>0.983</td>
+    <td>4.3</td>
+    <td>1.2 \times 10^{-3}</td>
+    <td>-</td>
+  </tr>
+  <tr>
+    <td>5</td>
+    <td><i>TERT</i></td>
+    <td>rs1043358053</td>
+    <td>C</td>
+    <td>5</td>
+    <td>0.926</td>
+    <td>11.9</td>
+    <td>7.4 \times 10^{-7}</td>
+    <td>-</td>
+  </tr>
+  <tr>
+    <td>10</td>
+    <td><i>PTEN</i></td>
+    <td>rs587782350</td>
+    <td>C</td>
+    <td>3</td>
+    <td>0.941</td>
+    <td>20.4</td>
+    <td>2.6 \times 10^{-3}</td>
+    <td>-</td>
+  </tr>
+  <tr>
+    <td>17</td>
+    <td><i>TP53</i></td>
+    <td>rs11540652</td>
+    <td>T</td>
+    <td>5</td>
+    <td>0.996</td>
+    <td>10.0</td>
+    <td>6.6 \times 10^{-5}</td>
+    <td>Gastric Cancer [70], Ovarian Cancer [71]</td>
+  </tr>
+</table>
+Supplementary Table 6. Mean variant allelic fraction per gene across participants included in the corresponding gene-level Burden/SKAT–O analysis.
+
+<table>
+  <tr>
+    <th>Variant Class</th>
+    <th>Chr</th>
+    <th>Gene</th>
+    <th># of subjects</th>
+    <th># of variants</th>
+    <th>Mean VAF (SD)</th>
+  </tr>
+  <tr>
+    <td rowspan="6">LoF</td>
+    <td>4</td>
+    <td><i>TET2</i></td>
+    <td>266</td>
+    <td>133</td>
+    <td>0.33 (0.14)</td>
+  </tr>
+  <tr>
+    <td>11</td>
+    <td><i>ATM</i></td>
+    <td>734</td>
+    <td>128</td>
+    <td>0.46 (0.10)</td>
+  </tr>
+  <tr>
+    <td>13</td>
+    <td><i>BRCA2</i></td>
+    <td>1,061</td>
+    <td>162</td>
+    <td>0.44 (0.12)</td>
+  </tr>
+  <tr>
+    <td>15</td>
+    <td><i>CKMT1B</i></td>
+    <td>29</td>
+    <td>8</td>
+    <td>0.56 (0.17)</td>
+  </tr>
+  <tr>
+    <td>17</td>
+    <td><i>BRCA1</i></td>
+    <td>302</td>
+    <td>74</td>
+    <td>0.46 (0.09)</td>
+  </tr>
+  <tr>
+    <td>20</td>
+    <td><i>ASXL1</i></td>
+    <td>502</td>
+    <td>30</td>
+    <td>0.32 (0.11)</td>
+  </tr>
+  <tr>
+    <td rowspan="8">Alpha Missense</td>
+    <td>1</td>
+    <td><i>C1orf52</i></td>
+    <td>87</td>
+    <td>11</td>
+    <td>0.50 (0.07)</td>
+  </tr>
+  <tr>
+    <td>2</td>
+    <td><i>DNMT3A</i></td>
+    <td>593</td>
+    <td>33</td>
+    <td>0.24 (0.10)</td>
+  </tr>
+  <tr>
+    <td>2</td>
+    <td><i>SF3B1</i></td>
+    <td>23</td>
+    <td>5</td>
+    <td>0.31 (0.16)</td>
+  </tr>
+  <tr>
+    <td>3</td>
+    <td><i>CHL1</i></td>
+    <td>12</td>
+    <td>3</td>
+    <td>0.37 (0.18)</td>
+  </tr>
+  <tr>
+    <td>4</td>
+    <td><i>TET2</i></td>
+    <td>123</td>
+    <td>41</td>
+    <td>0.36 (0.15)</td>
+  </tr>
+  <tr>
+    <td>10</td>
+    <td><i>PTEN</i></td>
+    <td>5</td>
+    <td>3</td>
+    <td>0.39 (0.04)</td>
+  </tr>
+  <tr>
+    <td>13</td>
+    <td><i>SOX21</i></td>
+    <td>22</td>
+    <td>6</td>
+    <td>0.49 (0.06)</td>
+  </tr>
+  <tr>
+    <td>15</td>
+    <td><i>IDH2</i></td>
+    <td>46</td>
+    <td>14</td>
+    <td>0.45 (0.11)</td>
+  </tr>
+  <tr>
+    <td>17</td>
+    <td><i>TP53</i></td>
+    <td>19</td>
+    <td>7</td>
+    <td>0.28 (0.14)</td>
+  </tr>
+  <tr>
+    <td>17</td>
+    <td><i>SRSF2</i></td>
+    <td>164</td>
+    <td>6</td>
+    <td>0.30 (0.11)</td>
+  </tr>
+  <tr>
+    <td>X</td>
+    <td><i>RLIM</i></td>
+    <td>4</td>
+    <td>2</td>
+    <td>0.44 (0.41)</td>
+  </tr>
+</table>
+Acknowledgements
+This research has been conducted using the UK Biobank Resource under application number 45420. We thank all the participants and researchers of UK Biobank for making these data open and accessible to the research community.
+
+Authors’ contributions
+J.P conducted all the analyses, prepared all figures and wrote the manuscript. A.P.T contributed to the manuscript writing. L.T provided critical comment on the manuscript. M.D.G and Y.L.G planned, organized and supervised the entire study and revised the manuscript. All authors have approved the submitted version.
+
+Funding
+This research was supported by the Dean’s Postdoctoral Fellowship at the School of Medicine, Stanford University. Additionally, this research was partially supported by the Biostatistics Shared Resource (B-SR) of the NCI-sponsored Stanford Cancer Institute: P30CA124435 and by the following NIH funding source of Stanford’s Center for Clinical and Translational Education and Research award, under the Biostatistics, Epidemiology and Research Design (BERD) Program: 1UM1TR004921-01.
+
+Data availability
+GWAS summary statistics for this study are available in the GWAS Catalog. Data supporting the findings of this study are available from the UK Biobank (UKB). Access to these data is available from the authors with UKB permission
+
+Code availability
+The codes used for analyses in the present study are available at the following link:
+https://github.com/Junkkkk/Lifespan-studies
+
+Consent for publication
+Not applicable.
+
+Competing interests
+The authors declare that they have no competing interests.
+References
+1. Passarino, G., F. De Rango, and A. Montesanto, Human longevity: Genetics or Lifestyle? It takes two to tango. Immunity & Ageing, 2016. 13: p. 1-6.
+2. Ruby, J.G., et al., Estimates of the Heritability of Human Longevity Are Substantially Inflated due to Assortative Mating. Genetics, 2018. 210(3): p. 1109-1124.
+3. v, B.H.J., et al., Genetic influence on human lifespan and longevity. Hum Genet, 2006. 119(3): p. 312-21.
+4. van den Berg, N., et al., Longevity defined as top 10% survivors and beyond is transmitted as a quantitative genetic trait. Nature Communications, 2019. 10(1): p. 35.
+5. Ryu, S., et al., Genetic landscape of APOE in human longevity revealed by high-throughput sequencing. Mech Ageing Dev, 2016. 155: p. 7-9.
+6. Sebastiani, P., et al., APOE Alleles and Extreme Human Longevity. J Gerontol A Biol Sci Med Sci, 2019. 74(1): p. 44-51.
+7. Joshi, P.K., et al., Variants near CHRNA3/5 and APOE have age- and sex-related effects on human lifespan. Nature Communications, 2016. 7(1): p. 11174.
+8. Joshi, P.K., et al., Genome-wide meta-analysis associates HLA-DQA1/DRB1 and LPA and lifestyle factors with human longevity. Nature Communications, 2017. 8(1): p. 910.
+9. Liu, J.Z., et al., The burden of rare protein-truncating genetic variants on human lifespan. Nature Aging, 2022. 2(4): p. 289-294.
+10. Liu, J.Z., Y. Erlich, and J.K. Pickrell, Case–control association mapping by proxy using family history of disease. Nature Genetics, 2017. 49(3): p. 325-331.
+11. Bae, H., et al., A genome-wide association study of 2304 extreme longevity cases identifies novel longevity variants. International Journal of Molecular Sciences, 2022. 24(1): p. 116.
+12. Sebastiani, P., et al., Four Genome-Wide Association Studies Identify New Extreme Longevity Variants. The Journals of Gerontology: Series A, 2017. 72(11): p. 1453-1464.
+13. Deelen, J., et al., A meta-analysis of genome-wide association studies identifies multiple longevity genes. Nature Communications, 2019. 10(1): p. 3669.
+14. Therneau, T.M., PM. Grambsch, and T.R. Fleming, Martingale-based residuals for survival models. Biometrika, 1990. 77(1): p. 147-160.
+15. Gagliano Taliun, S.A., et al., Exploring and visualizing large-scale genetic associations by using PheWeb. Nature Genetics, 2020. 52(6): p. 550-552.
+16. Baer, C., et al., "Somatic" and "pathogenic" - is the classification strategy applicable in times of large-scale sequencing? Haematologica, 2019. 104(8): p. 1515-1520.
+17. Sudlow, C., et al., UK biobank: an open access resource for identifying the causes of a wide range of complex diseases of middle and old age. PLoS Med, 2015. 12(3): p. e1001779.
+18. Bycroft, C., et al., The UK Biobank resource with deep phenotyping and genomic data. Nature, 2018. 562(7726): p. 203-209.
+19. Purcell, S., et al., PLINK: a tool set for whole-genome association and population-based linkage analyses. The American journal of human genetics, 2007. 81(3): p. 559-575.
+20. Cox, D.R., Regression models and life-tables. Journal of the Royal Statistical Society: Series B (Methodological), 1972. **34**(2): p. 187-202.
+21. Therneau, T., A package for survival analysis in S. R package version, 2015. **2**(7): p. 2014.
+22. Pruim, R.J., et al., LocusZoom: regional visualization of genome-wide association scan results. Bioinformatics, 2010. **26**(18): p. 2336-2337.
+23. Mohammadi Pejman 5 6 Park YoSon 11 Parsana Princy 12 Segre Ayellet V. 1 Strober Benjamin J. 9 Zappala Zachary 7 8, G.C.L.a.A.F.B.A.A.C.S.E.D.J.R.H.Y.J.B., et al., Genetic effects on gene expression across human tissues. Nature, 2017. **550**(7675): p. 204-213.
+24. Giambartolomei, C., et al., Bayesian test for colocalisation between pairs of genetic association studies using summary statistics. PLoS genetics, 2014. **10**(5): p. e1004383.
+25. Liu, B., et al., Abundant associations with gene expression complicate GWAS follow-up. Nature Genetics, 2019. **51**(5): p. 768-769.
+26. Szustakowski, J.D., et al., Advancing human genetics research and drug discovery through exome sequencing of the UK Biobank. Nature genetics, 2021. **53**(7): p. 942-948.
+27. McLaren, W., et al., The Ensembl Variant Effect Predictor. Genome Biology, 2016. **17**(1): p. 122.
+28. Cheng, J., et al., Accurate proteome-wide missense variant effect prediction with AlphaMissense. Science, 2023. **381**(6664): p. eadg7492.
+29. Ioannidis, N.M., et al., REVEL: An Ensemble Method for Predicting the Pathogenicity of Rare Missense Variants. Am J Hum Genet, 2016. **99**(4): p. 877-885.
+30. Lee, S., et al., Optimal unified approach for rare-variant association testing with application to small-sample case-control whole-exome sequencing studies. Am J Hum Genet, 2012. **91**(2): p. 224-37.
+31. Wu, M.C., et al., Rare-variant association testing for sequencing data with the sequence kernel association test. Am J Hum Genet, 2011. **89**(1): p. 82-93.
+32. Wu, P., et al., Mapping ICD-10 and ICD-10-CM codes to phecodes: workflow development and initial evaluation. JMIR medical informatics, 2019. **7**(4): p. e14325.
+33. Bellenguez, C., et al., New insights into the genetic etiology of Alzheimer’s disease and related dementias. Nature Genetics, 2022. **54**(4): p. 412-436.
+34. Mahley, R.W., Apolipoprotein E: from cardiovascular disease to neurodegenerative disorders. Journal of Molecular Medicine, 2016. **94**(7): p. 739-746.
+35. Du, Q., et al., Epigenetic silencing ZSCAN23 promotes pancreatic cancer growth by activating Wnt signaling. Cancer Biology & Therapy, 2024. **25**(1): p. 2302924.
+36. Lee, M.-G. and Y.H. Lee, A meta-analysis examining the association between the MUC5B rs35705950 T/G polymorphism and susceptibility to idiopathic pulmonary fibrosis. Inflammation Research, 2015. **64**(6): p. 463-470.
+37. Nakano, Y., et al., MUC5B promoter variant rs35705950 affects MUC5B expression in the distal airways in idiopathic pulmonary fibrosis. American journal of respiratory and critical care medicine, 2016. **193**(4): p. 464-466.
+38. van Moorsel, C.H., et al., The MUC5B promoter polymorphism associates with severe
+COVID-19 in the European population. Frontiers in medicine, 2021. 8: p. 668024.
+
+39. Verma, A., et al., A MUC5B gene polymorphism, rs35705950-T, confers protective effects against COVID-19 hospitalization but not severe disease or mortality. American journal of respiratory and critical care medicine, 2022. 206(10): p. 1220-1229.
+
+40. Quivoron, C., et al., TET2 Inactivation Results in Pleiotropic Hematopoietic Abnormalities in Mouse and Is a Recurrent Event during Human Lymphomagenesis. Cancer Cell, 2011. 20(1): p. 25-38.
+
+41. Abdel-Wahab, O., et al., ASXL1 Mutations Promote Myeloid Transformation through Loss of PRC2-Mediated Gene Repression. Cancer Cell, 2012. 22(2): p. 180-193.
+
+42. Yang, L., R. Rau, and M.A. Goodell, DNMT3A in haematological malignancies. Nature Reviews Cancer, 2015. 15(3): p. 152-165.
+
+43. Wang, L., et al., <i>SF3B1</i> and Other Novel Cancer Genes in Chronic Lymphocytic Leukemia. New England Journal of Medicine, 2011. 365(26): p. 2497-2506.
+
+44. Kim, E., et al., SRSF2 Mutations Contribute to Myelodysplasia by Mutant-Specific Effects on Exon Recognition. Cancer Cell, 2015. 27(5): p. 617-630.
+
+45. Cremona, C.A. and A. Behrens, ATM signalling and cancer. Oncogene, 2014. 33(26): p. 3351-3360.
+
+46. Wooster, R., et al., Identification of the breast cancer susceptibility gene BRCA2. Nature, 1995. 378(6559): p. 789-792.
+
+47. Ford, D., et al., Risks of cancer in BRCA1-mutation carriers. The Lancet, 1994. 343(8899): p. 692-695.
+
+48. Johnsen, S.A., et al., Regulation of Estrogen-Dependent Transcription by the LIM Cofactors CLIM and RLIM in Breast Cancer. Cancer Research, 2008. 69(1): p. 128-136.
+
+49. Gao, R., et al., E3 Ubiquitin Ligase RLIM Negatively Regulates c-Myc Transcriptional Activity and Restains Cell Proliferation. PLOS ONE, 2016. 11(9): p. e0164086.
+
+50. Yin, Y. and W.H. Shen, PTEN: a new guardian of the genome. Oncogene, 2008. 27(41): p. 5443-5453.
+
+51. Yang, H., et al., IDH1 and IDH2 Mutations in Tumorigenesis: Mechanistic Insights and Clinical Perspectives. Clinical Cancer Research, 2012. 18(20): p. 5562-5571.
+
+52. Yang, C.-M., et al., Aberrant DNA hypermethylation-silenced SOX21-AS1 gene expression and its clinical importance in oral cancer. Clinical Epigenetics, 2016. 8(1): p. 129.
+
+53. Zhang, X., et al., Long noncoding RNA SOX21-AS1 promotes cervical cancer progression by competitively sponging miR-7/VDAC1. Journal of Cellular Physiology, 2019. 234(10): p. 17494-17504.
+
+54. Sheng, X.-Y., et al., Long-Chain Non-Coding SOX21-AS1 Promotes Proliferation and Migration of Breast Cancer Cells Through the PI3K/AKT Signaling Pathway. Cancer Management and Research, 2020. 12(null): p. 11005-11014.
+
+55. Shi, H., et al., CKMT1B is a potential prognostic biomarker and associated with immune infiltration in Lower-grade glioma. PLOS ONE, 2021. 16(1): p. e0245524.
+
+56. Senchenko, V.N., et al., Differential expression of CHL1 gene during development of major
+human cancers. PloS one, 2011. **6**(3): p. e15612.
+
+57. Baird, D.M., *Variation at the TERT locus and predisposition for cancer*. Expert Reviews in Molecular Medicine, 2010. **12**: p. e16.
+
+58. Vinagre, J., et al., *Frequency of TERT promoter mutations in human cancers*. Nature Communications, 2013. **4**(1): p. 2185.
+
+59. Zenin, A., et al., *Identification of 12 genetic loci associated with human healthspan*. Communications Biology, 2019. **2**(1): p. 41.
+
+60. Kibel, A.S., et al., *CDKN1A and CDKN1B Polymorphisms and Risk of Advanced Prostate Carcinoma1*. Cancer Research, 2003. **63**(9): p. 2033-2036.
+
+61. Li, S., et al., *miR-3619-5p inhibits prostate cancer cell growth by activating CDKN1A expression*. Oncology Reports, 2017. **37**(1): p. 241-248.
+
+62. Lombardi, M.P., et al., *Mutation update for the PORCN gene*. Human Mutation, 2011. **32**(7): p. 723-728.
+
+63. Covey, T.M., et al., *PORCN moonlights in a Wnt-independent pathway that regulates cancer cell proliferation*. Plos one, 2012. **7**(4): p. e34532.
+
+64. Zhao, F., et al., *The function of uterine UDP-glucuronosyltransferase 1A8 (UGT1A8) and UDP-glucuronosyltransferase 2B7 (UGT2B7) is involved in endometrial cancer based on estrogen metabolism regulation*. Hormones, 2020. **19**: p. 403-412.
+
+65. Thibaudeau, J., et al., *Characterization of common UGT1A8, UGT1A9, and UGT2B7 variants with different capacities to inactivate mutagenic 4-hydroxylated metabolites of estradiol and estrone*. Cancer research, 2006. **66**(1): p. 125-133.
+
+66. Yang, X.R., et al., *Prevalence and spectrum of germline rare variants in BRCA1/2 and PALB2 among breast cancer cases in Sarawak, Malaysia*. Breast Cancer Res Treat, 2017. **165**(3): p. 687-697.
+
+67. Clay-Gilmour, A., et al., *Pathogenic and likely pathogenic germline variation in patients with myeloid malignancies and their unrelated HLA-matched hematopoietic stem cell donors*. 2024.
+
+68. Petiti, J., et al., *Comprehensive Molecular Profiling of NPM1-Mutated Acute Myeloid Leukemia Using RNAseq Approach*. International Journal of Molecular Sciences, 2024. **25**(7): p. 3631.
+
+69. Bayram, D.M., F.M. Lafta, and B.F. Matti, *Impact of IDH Mutations on DNA Methylation of Acute Myeloid Leukemia Related Genes: A Review Article*. Journal of the Faculty of Medicine Baghdad, 2024. **66**(1): p. 116-125.
+
+70. Hnatyszyn, A., et al., *Mutations in Helicobacter pylori infected patients with chronic gastritis, intestinal type of gastric cancer and familial gastric cancer*. Hereditary Cancer in Clinical Practice, 2024. **22**(1): p. 9.
+
+71. Richau, C.S., et al., *BRCA1, BRCA2, and TP53 germline and somatic variants and clinicopathological characteristics of Brazilian patients with epithelial ovarian cancer*. Cancer Medicine, 2024.

0cada618bb488f1258412c269438e0dbe8f9b287946d4660b5442737b0d7019d/metadata.json

@@ -0,0 +1,144 @@
+{
+    "title": "Sondheimer oscillations as a probe of non-ohmic flow in WP2 crystals",
+    "pre_title": "Sondheimer oscillations as a probe of non-ohmic flow in type-II Weyl semimetal WP2",
+    "journal": "Nature Communications",
+    "published": "10 August 2021",
+    "supplementary_0": [
+        {
+            "label": "Supplementary Information",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-021-25037-0/MediaObjects/41467_2021_25037_MOESM1_ESM.pdf"
+        },
+        {
+            "label": "Peer Review File",
+            "link": "https://static-content.springer.com/esm/art%3A10.1038%2Fs41467-021-25037-0/MediaObjects/41467_2021_25037_MOESM2_ESM.pdf"
+        }
+    ],
+    "supplementary_1": NaN,
+    "supplementary_2": NaN,
+    "source_data": [
+        "https://doi.org/10.5281/zenodo.4675599",
+        "/articles/s41467-021-25037-0#Fig4"
+    ],
+    "code": [],
+    "subject": [
+        "Electronic devices",
+        "Electronic properties and materials"
+    ],
+    "license": "http://creativecommons.org/licenses/by/4.0/",
+    "preprint_pdf": "https://www.researchsquare.com/article/rs-131719/v1.pdf?c=1637613590000",
+    "research_square_link": "https://www.researchsquare.com//article/rs-131719/v1",
+    "nature_pdf": "https://www.nature.com/articles/s41467-021-25037-0.pdf",
+    "preprint_posted": "11 Feb, 2021",
+    "research_square_content": [
+        {
+            "section_name": "Abstract",
+            "section_text": "As conductors in electronic applications shrink, microscopic conduction processes lead to strong deviations from Ohm\u2019s law. Depending on the length scales of momentum conserving (lMC) and relaxing (lMR) electron scattering, and the device size (d), current flows may shift from ohmic to ballistic to hydrodynamic regimes and more exotic mixtures thereof. So far, an in situ, in-operando methodology to obtain these parameters self-consistently within a micro/nanodevice, and thereby identify its conduction regime, is critically lacking. In this context, we exploit Sondheimer oscillations, semi-classical magnetoresistance oscillations due to  helical electronic motion, as a method to obtain lMR in micro-devices even when lMR>>d. This gives information on the bulk lMR complementary to quantum oscillations, which are sensitive to all scattering processes. We extract lMR from the Sondheimer amplitude in the topological semi-metal WP2, at elevated temperatures up to T~50 K, in a range most relevant for hydrodynamic transport phenomena. Our data on \u03bcm-sized devices are in excellent agreement with experimental reports of the large bulk lMR  and thus confirm that WP2 can be microfabricated without degradation. Indeed, the measured scattering rates match well with those of theoretically predicted electron-phonon scattering, thus supporting the notion of strong momentum exchange between electrons and phonons in WP2 at these temperatures. These results conclusively establish Sondheimer oscillations as a quantitative probe of lMR in micro-devices in studying non-ohmic electron flow.Electronic Materials and DevicesHard Condensed-matter PhysicsNon-ohmic Electron FlowElectron ScatteringConduction RegimeMagnetoresistance OscillationsMomentum Exchange",
+            "section_image": []
+        },
+        {
+            "section_name": "Figures",
+            "section_text": "Figure 1Figure 2Figure 3Figure 4",
+            "section_image": [
+                "https://assets-eu.researchsquare.com/files/rs-131719/v1/413bbbdc84de23becb38d09a.png",
+                "https://assets-eu.researchsquare.com/files/rs-131719/v1/b57d89d62bde1dcc5ae31f03.png",
+                "https://assets-eu.researchsquare.com/files/rs-131719/v1/1915a3cf803fc45d3cea0299.png",
+                "https://assets-eu.researchsquare.com/files/rs-131719/v1/f656a078113500970bfcb444.png"
+            ]
+        },
+        {
+            "section_name": "Full Text",
+            "section_text": "Due to technical limitations, full-text HTML conversion of this manuscript could not be completed. However, the latest manuscript can be downloaded and  accessed as a PDF.",
+            "section_image": []
+        },
+        {
+            "section_name": "Additional Declarations",
+            "section_text": "There is NO Competing Interest.",
+            "section_image": []
+        },
+        {
+            "section_name": "Supplementary Files",
+            "section_text": "Supplementaryinformation.pdf",
+            "section_image": []
+        }
+    ],
+    "nature_content": [
+        {
+            "section_name": "Abstract",
+            "section_text": "As conductors in electronic applications shrink, microscopic conduction processes lead to strong deviations from Ohm\u2019s law. Depending on the length scales of momentum conserving (lMC) and relaxing (lMR) electron scattering, and the device size (d), current flows may shift from ohmic to ballistic to hydrodynamic regimes. So far, an in situ methodology to obtain these parameters within a micro/nanodevice is critically lacking. In this context, we exploit Sondheimer oscillations, semi-classical magnetoresistance oscillations due to helical electronic motion, as a method to obtain lMR even when lMR\u2009\u226b\u2009d. We extract lMR from the Sondheimer amplitude in WP2, at temperatures up to T\u2009~\u200940 K, a range most relevant for hydrodynamic transport phenomena. Our data on \u03bcm-sized devices are in excellent agreement with experimental reports of the bulk lMR and confirm that WP2 can be microfabricated without degradation. These results conclusively establish Sondheimer oscillations as a quantitative probe of lMR in micro-devices.",
+            "section_image": []
+        },
+        {
+            "section_name": "Introduction",
+            "section_text": "In macroscopic metallic wires, the flow of electric current is well described by Ohm\u2019s law, which assigns a metal a spatially uniform \u2018bulk\u2019 conductivity. The underlying assumption is that the complex and frequent scattering events of charge carriers occur on the microscopic length scale of a mean free path, which is much smaller than the size of the conductor, d, leading to diffusive behavior. In addition to the scattering processes of bulk systems, the resistance of microscopic conductors is mostly dominated by boundary scattering, thereby masking the internal scattering processes of the bulk in resistance measurements. Here, we present a method to uncover these bulk processes in micro-scale metals, which are of technological importance for fabrication of quantum electronic devices, and simultaneously critical to a fundamental understanding of microscopic current flow patterns. It is instructive to classify the bulk scattering processes into two categories: those that relax the electron momentum, such as electron\u2013phonon, Umklapp or inelastic scattering, occurring at length-scale lMR; and those that conserve the electron momentum, such as direct or phonon-mediated electron\u2013electron scattering, associated with a length-scale lMC.\n\nWithin a kinetic theory framework, these three length scales, namely d, lMR, and lMC, can be used to describe the current flow in micro-scale conductors. When momentum-conserving interactions are negligible, ohmic flow at the macro-scale (lMC\u2009\u226b\u2009d\u2009\u226b\u2009lMR) gives way to ballistic transport in clean metals where lMR, lMC\u2009\u226b\u2009d. Conversely, when momentum-conserving interactions occur over similar or smaller length scales to momentum-relaxing interactions, a third regime of \u2018hydrodynamic\u2019 transport (lMR\u2009\u226b\u2009d\u2009\u226b\u2009lMC) is observable1,2. In this regime, the static transport properties of electron fluids can be described by an effective viscosity that captures the momentum diffusion of the system2,3. These electron fluids exhibit classical fluid phenomena such as Poiseuille flow, whereby the current flow density is greatly decreased at the conductor boundary. Recently, advances in both experimental probes and theoretical descriptions have enabled direct observation of these effects using spatially resolved current density imaging, and have hinted towards the rich landscape of electron hydrodynamics in micro-scale crystals3,4,5.\n\nWhile such local-probe experiments provide means of quantifying electron\u2013electron interactions, and thus extracting lMC, direct measurement of the intrinsic momentum-relaxing processes (lMR) within micron-scale conductors remains elusive, yet is greatly needed. From a practical perspective, lMR describes the overall scattering from impurities and the lattice vibrations within the metallic microstructure, which at low temperature is an important feedback parameter of quality control in fabrication. Furthermore, given both the reduction of sample size and the improved crystal quality, seemingly exotic transport scenarios where lMR\u2009\u226b\u2009d\u2009\u226b\u2009lMC is satisfied are expected to become more prevalent in technology. An accurate description of these length scales is necessary to predict the overall resistance and thus voltage drops and heat dissipation in the nanoelectronic devices. For example, the resistive processes in a hydrodynamic conductor occur at the boundaries rather than homogeneously distributed in the bulk, which alters the spatial distribution of Joule heating and thereby has significant impacts on thermal design.\n\nReal devices will operate at some intermediate state in the d, lMR, and lMC parameter space, departing from the well-understood limiting cases of ohmic, ballistic, and hydrodynamic flow. Rich landscapes of distinct hydrodynamic transport regimes are predicted depending on the relative sizes of the relevant length scales6. Effective understanding, modeling and prediction of transport requires an experimental method to estimate these parameters reliably in every regime. In large, ohmic conductors, the bulk mean free path lMR can be simply estimated from the device resistance using a Drude model. Yet when lMC, lMR\u2009\u2273\u2009d, boundary scattering dominates the resistance, and hence estimates of the bulk scattering parameters are highly unreliable. This leaves the worrying possibility of misinterpreting the transport situation in a conductor, in that the microfabrication itself may introduce defects or changes in the bulk properties that remain undetected by macroscopic observables such as the resistance, but have profound impact on the microscopic current distribution. These effects are already noticeable in state-of-the-art transistors, owing to the low carrier density of semi-conductors7, but have similarly been reported in metallic conductors5. With the increased technological interest in quantum and classical electronics operating at cryogenic temperatures, such questions about unconventional transport regimes are also of practical relevance in next generation electronics8.\n\nIn this context, we propose to exploit a magneto-oscillatory phenomenon, Sondheimer oscillations (SO), as a self-consistent method to obtain the transport scattering length lMR in-situ, even in constricted channels when lMR\u2009\u226b\u2009d. In general, a magnetic field (B) applied perpendicular to a thin metal forces the carriers on the Fermi surface (FS) to undergo cyclotron motion. Those on extremal orbits of the FS are localized in space due to the absence of a net velocity component parallel to the magnetic field. These localized trajectories can become quantum-coherent, and their interference causes the well-known Shubnikov-de Haas (SdH) oscillations. The states away from extremal orbits also undergo cyclotron motion, yet they move with a net velocity along the magnetic field, analogous to the helical trajectories of free electrons in a magnetic field (Fig.\u00a01). These states are responsible for the Sondheimer size effect which manifests itself as a periodic-in-B oscillation of the transport coefficients, as discovered in the middle of the past century for clean elemental metals9.\n\na Illustration of the Sondheimer effect. Left: the applied magnetic field is B\u2009=\u20093\u0394B and the electron (red) makes an integer number of rotations, with no contribution to transport. Right: B\u2009\u2260\u2009n\u0394B. The electron hits the top surface at a different position than its origin on the bottom surface, leading to a contribution to the conductivity. b Resistivity as function of temperature for a WP2 microdevice. Inset: false-color SEM image of a typical device used in this study. c Sondheimer oscillations seen in the Hall resistivity of a WP2 microdevice, for different temperatures. The oscillation period of \u0394B\u2009=\u20091.6 T is highlighted.\n\nFor any given state, the magnitude of B sets the helical radius and thus determines how many revolutions the electron completes while traveling from one surface to the other in a microdevice. If an integer number of revolutions occurs, the charge carrier will have performed no net motion along the channel, and hence is semi-classically localized (Fig.\u00a01a). However, if the number of revolutions is non-integer, a net motion along or perpendicular to the channel exists, delocalizing the carriers, resulting in oscillatory longitudinal and transverse magnetotransport behavior. Large-angle bulk scattering events dephase the trajectory, hence the strong sensitivity of SO to the bulk lMR even in nanostructures. These SO are an inherent property of mesoscale confined conductors in three dimensions and have no counterpart in 2D metals like graphene.\n\nThe period of the SO is derived by considering a classical charged particle on a helical trajectory between two surfaces perpendicular to the magnetic field10. One compares the time it takes to travel the distance d between the surfaces, td\u2009=\u2009d/v\u2225, to the time to complete a single cyclotron revolution, \u03c4c\u2009=\u20092\u03c0/\u03c9c\u2009=\u20092\u03c0m*/eB (m*: effective mass, e: electron charge, \u03c9c\u2009=\u2009eB/m*: cyclotron frequency). Their ratio describes the number of revolutions of the trajectory. For certain fields the helix is commensurate with the finite structure and the number of revolutions is integer, n, such that td\u2009=\u2009n\u03c4c. This occurs periodically in field, with the period given by:\n\nThe useful identity \\({v}_{\\parallel }=\\frac{\\hslash }{2\\pi {m}^{* }}\\left(\\frac{\\partial A}{\\partial {k}_{\\parallel }}\\right)\\), derived by Harrison11, directly relates the SO period to the FS geometry, where v\u2225 and k\u2225 denote the velocity and momentum component parallel to the magnetic field and A is the FS cross-sectional area encircled by the orbit in k-space. Note the contrast to conventional QO which appear around extremal orbits, where \\(\\frac{\\partial A}{\\partial {k}_{\\parallel }}=0\\).\n\nAll conduction electrons undergo cyclotron motion, yet depending on \\(\\frac{\\partial A}{\\partial {k}_{\\parallel }}\\), they experience different commensurability fields with a structure of given size d. Hence oscillatory contributions to the total conductivity are washed out, unless a macroscopic number of states share the same \\({v}_{\\parallel }\\propto {\\left(\\frac{\\partial A}{\\partial {k}_{\\parallel }}\\right)}_{{E}_{f}}\\)10. In earlier days of Fermiology12, geometric approximations for FSs, such as elliptical endpoints, were introduced to identify those generalized geometric features that lead to extended regions of constant \\(\\frac{\\partial A}{\\partial {k}_{\\parallel }}\\). The computational methods available nowadays allow a more modern approach to the problem. FSs calculated by ab-initio methods can be numerically sliced in order to calculate their cross-section A(k\u2225). We propose to extend this routine procedure, used to find extremal orbits relevant for QO \\(\\left(\\frac{\\partial A}{\\partial {k}_{\\parallel }}=0\\right)\\), to identify SO-active regions \\(\\left(\\frac{{\\partial }^{2}A}{\\partial {k}_{\\parallel }^{2}} \\sim 0\\right)\\), based on the Fermi-surface slicing code SKEAF13 (see Methods for details on implementation).\n\nSO are caused by the real-space motion of charge carriers and hence also pose some conditions on the shape of the conductor. First, surface scattering needs to be mostly diffusive. If an electron undergoes specular scattering N times before scattering diffusively, it contributes towards the SO as if the sample had an effective thickness Nd14, leading to overtones. Naturally, SO vanish in the (unrealistic) limit of perfectly specular boundary conditions, as such ideal kinetic mirrors remove any interaction of the electron system with the finite size of the conductor. Secondly, the conductor must feature two parallel, plane surfaces perpendicular to the magnetic field to select only one spiral trajectory over the entire structure. The parallelicity requirement is simply given by a fraction of the pitch of the spiral at a certain field (maximal thickness variation \\({{\\Delta }}d \\, < \\, {v}_{\\parallel }{\\tau }_{c}=d\\frac{{{\\Delta }}B}{B}\\))10. These requirements are naturally fulfilled in planar electronic devices.\n\nIt is instructive to briefly compare SO to the more widely known QO of resistance, the SdH effect. Both are probes of the FS geometry based on cyclotron orbits, yet the microscopics are strikingly different. While QO frequencies are exclusively determined by FS properties via the Onsager relation and are thus independent of the sample shape, SO are finite-size effects. SO emerge from extended regions on the FS, unlike SdH oscillations to which only states in close vicinity of extremal orbits contribute. While SdH oscillations are quantum interference phenomena, SO are semi-classical, which is key to their use as a robust probe of exotic transport regimes. If both can be observed, powerful statements on the scattering microscopics can be made, as SdH is sensitive to all dephasing collision events and SO separates out the large-angle ones15. However, the much more stringent conditions of phase coherence in SdH severely limit their observations at higher temperatures. SO are observable up to relatively high temperatures at which the rapidly shrinking lMR(T) leads to a transition into an ohmic state, when lMR(T)\u2009<\u2009d. As such, they are ideally suited to explore the exotic transport regimes in which, for example, hydrodynamic effects occur.\n\nWe apply these theoretical considerations to experimentally investigate the scattering mechanisms in micron-sized crystalline bars of the type-II Weyl semimetal WP216 exploiting the Sondheimer effect. Bulk single crystals of WP2 are known for their long lMR, in the range of 100\u2013500\u2009\u03bcm17,18,19, comparable to the elemental metals in which SO were initially discovered20,21,22,23. These are an ideal test case for non-ohmic electron flow, as hydrodynamic transport signatures and nontrivial electron\u2013phonon dynamics have been observed in various topological semimetals17,18,24,25,26. These ulta-pure crystals are then reduced in size by nanofabrication techniques into constricted channels, to study hydrodynamic or ballistic corrections to the current flow.\n\nHere we employ Focused Ion Beam (FIB) micromachining27, which allows precise control over the channel geometry in 3D. In this technique, we accelerate Xe ions at 30\u2009kV to locally sputter the target crystal grown by chemical vapor transport (CVT)19,28 until a slab of desired dimensions in the \u03bcm-range remains. This technique leads to an amorphized surface of around 10 nm thickness, yet has been shown to leave bulk crystal structures pristine29. Naturally, reducing the size of a conductor even without altering its bulk mean free path significantly changes the device resistance at low temperatures due to finite size corrections30. Hence, measurements of the constricted device resistance alone cannot exclude the possibility of bulk degradation due to the fabrication. Thus far, one could only argue based on size-dependent resistance studies that the values smoothly extrapolate to the bulk resistivity value in the limit of infinite device size18,31. Measuring SO directly in the microfabricated devices themselves, however, quantitatively supports that the ultra-high purity of the parent crystal remains unchanged by our fabrication. We note that the fundamental question of the bulk parameters is universal in mesoscopic conducting structures irrespective of the fabrication technique, and these considerations are thus equally applicable to structures obtained by mechanically or chemically thinned samples as well as epitaxially grown crystalline films. SO should provide general insights into the material quality in the strongly confined regime, allowing to contrast different fabrication techniques.",
+            "section_image": [
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-021-25037-0/MediaObjects/41467_2021_25037_Fig1_HTML.png"
+            ]
+        },
+        {
+            "section_name": "Results",
+            "section_text": "We measure our \u03bcm-confined devices using standard lock-in techniques with applied currents between 50 and 100\u2009\u03bcA, low enough to limit self-heating, and magnetic fields up to 18 T. At high temperatures, the measured resistivity agrees well with previous reports on high quality bulk crystals, as expected given the momentum-relaxing limited mean free path of charge carriers in this regime (Fig.\u00a01b). Yet in the low temperature limit, the device resistance exceeds that of bulk crystals by more than an order of magnitude17,19,32. Conversely, the residual resistance ratios in our devices (RRR\u2009\u2248\u2009160\u2013300) are also considerably lower than in bulk crystals32. The main question we address by SO is whether this excess resistance points to fabrication-induced damage, finite size corrections, or a mixture thereof. At low temperatures around 3 K, a drop in resistance signals a superconducting transition. As WP2 in bulk form is not superconducting, this likely arises from an amorphous W-rich surface layer due to the FIB fabrication similar to observations made in NbAs33 and TaP34. In Fig.\u00a01c, we show the Hall resistivity, \u03c1xy, of one of our devices as a function of the magnetic field, for different temperatures. The Hall signal comprises oscillations with a period of \u0394B\u2009=\u20091.6 T, resolved above approximately B\u2009=\u20092 T.\n\nA hallmark signature of SO is their linear frequency dependence on the device thickness perpendicular to the field. For this reason, we fabricated crystalline devices with multiple sections of different thickness to study the dependence on the channel thickness, d, in a consistent manner. This \u2018staircase\u2019 device allows the simultaneous measurement of transport on 5 steps, as illustrated in Fig.\u00a02. SO appear in all transport coefficients, magnetoresistance (MR) and Hall effect alike, yet here we focus on the Hall effect for two practical reasons. First, the step edges induce non-uniform current flows, and hence the device would need to be considerably longer to avoid spurious voltage contributions from currents flowing perpendicular to the device in a longitudinal resistance measurement. Second, WP2 exhibits a very large MR yet a small Hall coefficient, as typical for compensated semi-metals. Therefore, the SO are more clearly distinguishable against the background in a Hall measurement, but they are also present in the longitudinal channel.\n\na False color SEM image of a staircase device, used to measure Sondheimer oscillations for different thicknesses. The crystal is colored in purple, and gold contacts in yellow. b Main: Schematic of the staircase device, illustrating all possible measurement configurations as well as the thickness of each section. Top left: SEM image of the lamella that will become the device shown in a, prior to extracting it from its parent crystal. Bottom right: SEM image of the same lamella, glued down onto a sapphire substrate, ready to define the device geometry. The lamella and glue are covered in gold (not colored) throughout the full field of view. The magnetic field is applied perpendicular to the structure, aligned along the crystallographic [011] direction.\n\nThe fabrication process of our WP2 devices follows largely the same procedure as described in ref. 27. However, for the staircase device, a few key changes were made. In the first fabrication step, the FIB is used to cut a lamella from a bulk WP2 crystal. One side is polished flat, and the other side polished into five sections, each to a different thickness (Fig.\u00a02b). It is then transferred, flat side down, into a drop of araldite epoxy on a sapphire substrate and electrically contacted by Au sputtering (Fig.\u00a02b). In a second FIB step, the staircase slab is patterned laterally into its final structure (Fig.\u00a02a). We use Xe ions for the entire FIB fabrication process in order to avoid potential issues with Ga ion implantation leading to changes in the carrier density. Indeed, experimentally, we see no indication of any charge carrier modulation.\n\nAll segments of the staircase devices show pronounced B-periodic oscillations in the Hall channel, from which the linear background is removed by taking second derivatives. (Fig.\u00a03). At the lowest fields, a weak, aperiodic structure is observed. In this regime, the cyclotron diameter does not fit into the bar, preventing the formation of the Sondheimer spirals. Note that in all devices of different thickness, this onset field of the SO is the same. This is a natural consequence of the fact that the lateral size, perpendicular to the magnetic field, by design, is the same for all steps of the staircase. Each step, however, differs in thickness d parallel to the magnetic field, and the period varies accordingly between steps (Fig.\u00a03b). At even higher fields, the onset of regular SdH oscillations hallmarks a transition into a different quantized regime. The SO frequency F\u2009=\u20091/\u0394B varies linearly with d as expected (Fig.\u00a03c, Eq. (1)).\n\na Second derivative of the Hall resistivity shown in Fig.\u00a01c at T\u2009=\u20094\u2009K. Inset: Fast Fourier Spectrum (FFT) corresponding to this data. b Second derivatives of the Hall resistivity at three different thicknesses, d\u2009=\u20094.3, 2.7 and 2.0\u2009\u03bcm, T\u2009=\u20094\u2009K. c FFTs corresponding to the data in b. d Dependence of the Sondheimer frequency on d. The error bars in the frequency, F, are derived from the width of the relevant peaks in the FFT spectra and those in d from the standard deviation of thickness measurements made with SEM. The red dashed line is calculated from the Fermi surface as determined from DFT. e Cross-sectional area, A, of the dogbone Fermi surface pocket of WP2 as a function of k parallel to the field direction of our experiments (top), and its derivative (bottom). f Location of observed Sondheimer orbits drawn in orange on the dogbone-shaped Fermi surface pocket. The magnetic field is applied along the [011]-direction, perpendicular to the current, as indicated by the red line.\n\nNext we identify the Sondheimer-active region on the FS from the ab-initio band structure, which was calculated by density functional theory (DFT) with the projected augmented wave method as implemented in the code of the Vienna ab-initio Simulation Package35. The FS of WP2 consists of two types of spin-split pockets: dogbone-shaped electron pockets and extended cylindrical hole pockets (see Fig.\u00a04 and Supplementary Fig.\u00a02 for a complete picture of the FS).\n\na FFTs of the SO at different temperatures for thicknesses of 4.3, 2.7, and 2.0\u2009\u03bcm. The data at T\u2009=\u20094\u2009K is the same as that in Fig.\u00a03c. b Temperature dependence of the Sondheimer and SdH oscillation amplitudes, for different sample thicknesses. The error bars are estimated from the variation in amplitude in the FFT spectra. The dashed lines are fits used to extrapolate to the amplitude at zero temperature, A(0) (see Methods for details). The dotted line is a Lifshitz\u2013Kosevich fit, giving an effective mass of 1.1\u2009me. Two regimes are highlighted: that of quantum coherence, where SdH oscillations exist alongside SO, and that of Sondheimer, where only SO exist. c Scattering times extracted for a 4.3\u2009\u03bcm thick section of a WP2 device using Eq. (2) and the calculated Fermi velocity, vF\u2009=\u20093.6\u2009\u00d7\u2009105\u2009m/s. An approximate quantum lifetime extracted from the SdH oscillations as well as data from refs. 18,46 are included for comparison. Errors in \u03c4MR are propagated from those in the amplitudes and the error in \u03c4q is the standard deviation of several measurement. d Calculated scattering time for all electron\u2013phonon scattering (\u03c4e\u2212ph) and e the scattering efficiency determining the momentum-relaxing scattering lifetimes (\u03c4MR) projected onto the Fermi surface at T\u2009=\u200910\u2009K.\n\nOnly one area quantitatively agrees with the observed SO periodicity: the four equivalent endpoints of the dogbone (colored orange in Fig.\u00a03f). Slicing all Fermi-surfaces using SKEAF13, their cross-sections A(k\u2225) are obtained. While in QO analysis this information is discarded once the extremal orbits are identified, it forms the basis of the SO analysis. As the dogbone is sliced from the endpoints, the area continuously grows until the two endpoint orbits merge and the area abruptly doubles. Slicing further, the area grows until the maximum orbit along the diagonal is reached. The mirror symmetry of the FS enforces then a symmetric spectrum when slicing further beyond the maximum. The quasi-linear growth at the endpoints signals an extended area of Sondheimer-active orbits. Averaging the near-constant derivative in this region, \\(\\frac{\\partial A}{\\partial {k}_{\\parallel }}\\), provides via Eq. (1) a tuning-parameter-free prediction of the thickness dependence of the SO frequency. This ab-initio prediction (red line in Fig.\u00a03c) is in excellent agreement with the observed thickness dependence.\n\nNext the temperature-dependence of the SO amplitude is used to gain direct information about the microscopic scattering processes acting on this region of the FS. In Fig.\u00a04a, b, we plot this temperature dependence and highlight two regimes: that of quantum coherence and that of purely SO. In the first regime, quantum coherence leads to SdH oscillations; however, for typical effective masses m*\u2009\u2248\u2009me, as in WP2, they are only observable at very low temperatures (T\u2009<\u20095\u2009K). Importantly, their quick demise upon increasing temperature is not driven by the temperature dependence of the scattering time, but rather by the broadening of the Fermi-Dirac distribution. This is apparent as their temperature dependence is well described by the Lifshitz\u2013Kosevich formalism based on a temperature-independent quantum lifetime, \u03c4q.\n\nThis strong temperature-suppression of QO severely limits their use to probe scattering mechanisms at elevated temperatures. SO, on the other hand, do not rely on quantum coherence and are readily observed to much higher temperatures, up to 40\u2009K in WP2, while their temperature decay allows a direct determination of the transport lifetime, \u03c4MR\u2009=\u2009lMR/vF. Hence SO make an excellent tool to study materials in the temperature range pertinent to exotic transport regimes like ballistic or hydrodynamic. They self-evidence non-diffusive transport as they only vanish when lMR\u2009~\u2009d, and hence are only absent in situations of conventional transport within a given device.",
+            "section_image": [
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-021-25037-0/MediaObjects/41467_2021_25037_Fig2_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-021-25037-0/MediaObjects/41467_2021_25037_Fig3_HTML.png",
+                "https:////media.springernature.com/lw685/springer-static/image/art%3A10.1038%2Fs41467-021-25037-0/MediaObjects/41467_2021_25037_Fig4_HTML.png"
+            ]
+        },
+        {
+            "section_name": "Discussion",
+            "section_text": "Key to observable SO is that electrons do not undergo large-angle scattering events on their path between the surfaces. We therefore have the condition that lMR\u2009>\u2009d36,37. As lMR(T) decreases with increasing temperature and the boundary scattering is assumed to be temperature-independent, the SO amplitude is suppressed as \\({e}^{-d/{l}_{{{{{{{{\\rm{MR}}}}}}}}}(T)}\\) which allows us to estimate the bulk transport mean free path within a finite-size sample, even when d\u2009\u226a\u2009lMR. It is extracted as36:\n\nwhere A(T) is the SO amplitude at temperature T. A(T\u2009=\u20090) is estimated by extrapolation, which is a robust procedure as the SO amplitude saturates at low but finite temperatures. This is analogous to the saturation of the resistivity of bulk metals at low temperatures, once bosonic scattering channels are frozen out and temperature-independent elastic defect scattering becomes dominant.\n\nIn the following discussion, we focus on the scattering time \u03c4MR to facilitate comparison of our results with literature and theory, using the average Fermi velocity on the dogbone FS determined from our band structure calculations self-consistently, vF\u2009=\u20093.6\u2009\u00d7\u2009105\u2009m/s. The \u03c4MR(T) obtained from all devices quantitatively agrees, despite their strong difference in thickness (between 1.3 and 4.6\u2009\u03bcm) and hence SO frequency, further supporting the validity of this simple analysis (see Fig.\u00a04c and Supplementary Fig.\u00a05). The lifetimes on the SO devices furthermore agree with measurements on bulk crystals18, evidencing that the increased resistivity compared to bulk can be wholly attributed to finite size corrections rather than to any fabrication-induced damage, and that FIB fabrication does not introduce significant changes to the bulk properties of WP2 that might cause misinterpretations of the scattering regime.\n\nFor our WP2 devices, a standard Dingle analysis15 of the QO yields a quantum scattering time \u03c4q\u2009~\u200910\u221213\u2009\u2212\u200910\u221212 s (Fig.\u00a04c), in agreement with published values for bulk crystals WP219. As \u03c4q is sensitive to all dephasing scattering events, but \u03c4MR only to large-angle momentum-relaxing scattering, the microscopics of the scattering processes in WP2 are brought to light. The four orders of magnitude difference between \u03c4MR and \u03c4q reflects a common observation in topological semi-metals such as Cd3As238, PtBi239, or TaAs40.\n\nLong \u03c4MR, together with a high quality, clean sample, enables the realization of the hydrodynamic regime where the momentum-conserving scattering dominates. These quantitative measurements of \u03c4q and \u03c4MR(T) can now be directly compared to theoretical models of scattering. We consider an initial electronic state with energy \u03b5nk (where n and k are the band index and wavevector respectively) scattering against a phonon with energy \u03c9q\u03bd (where \u03bd and q are the phonon polarization and wavevector respectively), into a final electronic state with energy \u03b5mk+q. The electron\u2013phonon scattering time \u03c4e\u2212ph describing such an interaction can be obtained from the electron self energy using Fermi\u2019s golden rule:\n\nwhere \u03a9BZ is the Brillouin zone volume, fnk and nq\u03bd are the Fermi\u2013Dirac and Bose\u2013Einstein distribution functions, respectively, and the electron\u2013phonon matrix element for a scattering vertex is given by\n\nHere \\(\\langle {\\psi }_{m{{{{{{{\\bf{k}}}}}}}}+{{{{{{{\\bf{q}}}}}}}}}|\\) and \\(|{\\psi }_{n{{{{{{{\\bf{k}}}}}}}}}\\rangle\\) are Bloch eigenstates and \u2202q\u03bdV is the perturbation of the self-consistent potential with respect to ion displacement associated with a phonon branch with frequency \u03c9q\u03bd. Plotting these state-resolved electron\u2013phonon lifetimes at ~10\u2009K on the FS reveals the distribution of scattering in the SO-active regions (Fig.\u00a04d). Equation (3), however, accounts, to first order, for all electron\u2013phonon interactions, irrespective of the momentum transfer or equivalently the scattering angle. To remedy this, we augment the scattering rate with an \u2018efficiency\u2019 factor41 given by the relative change of the initial and final state momentum (\\(1-\\frac{{v}_{n{{{{{{{\\bf{k}}}}}}}}}\\cdot {v}_{n{{{{{{{\\bf{k}}}}}}}}}}{| {v}_{n{{{{{{{\\bf{k}}}}}}}}}| | {v}_{n{{{{{{{\\bf{k}}}}}}}}}| }=1-\\cos \\theta\\)), where vnk is the group velocity and \u03b8 is the scattering angle:\n\nAt low temperatures, the thermally activated phonon modes have a tiny q, therefore the initial and final electronic states only differ from a small angle. It is thus important to take this momentum-relaxation efficiency factor into account in addition to \u03c4e\u2212ph, in order to estimate \u03c4MR which determines the electron mean free path in the SO-active regions. In the SO measurements, the electron orbits are located on the endpoints of the dogbone-shaped electron pockets (Fig.\u00a03f), therefore we highlight the scattering efficiency distribution on the electron FS in Fig.\u00a04e. Indeed, when the orbit is aligned along the diagonal direction, the FS cross section features very low scattering efficiency with an averaged \\(1-\\cos \\theta \\, < \\, 0.1\\). This supports our observation of frequently scattering electrons with long transport lifetimes in the SO measurement.\n\nThese results demonstrate the power of the Sondheimer size effect for the extraction of the momentum-relaxing mean free path in mesoscopic devices when d\u2009\u226a\u2009lMR via their temperature dependence. Combined with first-principles theoretical calculations, we were able to locate the states contributing to the helical motion to the elliptical endpoints of a particular FS of WP2. We note, however, that such analysis as well as the thickness dependence are only relevant for the academic purpose of robustly identifying these oscillations as SO. Once this is established, the relevant lifetimes may straightforwardly be obtained from the resistance oscillations at a single thickness. Hence, SO promise to be a powerful probe to obtain the bulk mean free path in devices with \u03bcm-scale dimensions without relying on any microscopic model assumptions. This analysis is a clear pathway to identify scattering processes in clean conductors within operating devices. It thereby provides important feedback of the materials quality after a micro-/nano-fabrication procedure and disentangles the roles of bulk and surface scattering that are inseparably intertwined in averaged transport quantities of strongly confined conductors, such as the resistance. As their origin is entirely semi-classical, they are not restricted by stringent criteria such as quantum coherence and thus span materials parameters of increased scattering rate. In particular, they survive up to significantly higher temperatures and thereby allow microscopic spectroscopy in new regimes of matter dominated by strong quasiparticle interactions, such as hydrodynamic electron transport. With this quantitative probe, it will be exciting to test recent proposals of exotic transport regimes and create devices that leverage such unconventional transport in quantum materials.",
+            "section_image": []
+        },
+        {
+            "section_name": "Methods",
+            "section_text": "High quality crystals of WP2 were grown via the CVT method using iodine as transport agent, with the following starting materials: red phosphorus (Alfa-Aesar, 99.999%) and tungsten trioxide (Alfa-Aesar, 99.998%). The starting materials were sealed in an evacuated fused silica ampoule. A two-zone-furnace with a temperature gradient of 1000\u2013900\u2009\u00b0C was used for the CVD method. After several weeks, the ampoule was removed from the furnace and quenched in water. The crystals were characterized by X-ray diffraction.\n\nIn Fig.\u00a02, we have shown the so-called staircase device which was especially designed for the measurement of SO. Here, we describe its fabrication in detail, beginning from a bulk crystal. Firstly, the crystallographic orientation of the bulk crystal is determined through XRD measurements, and it is glued in the desired orientation onto an SEM stub. It is then introduced into a Helios G4 PFIB UXe DualBeam FIB/SEM using Xe ions, hence avoiding issues associated with surface implantation as in the more common Ga-based sources. The focused beam of Xe ions is used to cut a rectangular slab we call a lamella from that crystal in three steps. The first is using a high 2.5 \u03bcA current to cut two gaps into the crystal, separated by an ~15\u2009\u03bcm thick and 150\u2009\u03bcm long section of crystal. A smaller current of 0.2\u2009\u03bcA is then used to smoothen this crystal section in order to produce moderately flat sidewalls, after which the wall is cut near the bottom and on the sides such that we have a lamella attached to the parent crystal through two beams at its top.\n\nAt this stage, a 15\u2009nA current is used to fine-cut both sides of the lamella, leaving it with parallel sidewalls and a high level of smoothness. The thickness of the lamella is now that of what will be thickest section of the staircase device. The lamella is then divided into five sections; the three middle sections are of equal length, while the sections on either end are longer for further contacts. Starting from one side, we leave the first section untouched and polish and mill the other sections to their respective target thickness using cleaning-cross-sections as the milling strategy. This is the stage shown in the top left insert of Fig.\u00a02b. After this, we cut through the beam on one side and thin down the other side.\n\nAn ex-situ micromanipulator is used to break the thinned beam and pick up the stepped lamella. On a sapphire substrate with lithographically prepatterned gold contact pads, we place a small droplet of Araldite epoxy. While the epoxy is still liquid, we place our lamella, flat side down, onto the droplet. Capillary forces then create a profile of epoxy around the lamella that extends smoothly up to each of its top surfaces, without covering any. After curing the epoxy for 1\u2009h at 150\u2009\u00b0C, we take the substrate to a sputtering machine, where it is briefly RF etched and 3 nm of Ti plus 200\u2009nm of Au are sputtered onto the lamella, glue and prepatterned gold contacts through a shadow mask. The lamella after this step is shown in the bottom right of Fig.\u00a02b.\n\nIn order to pattern the device, we again make use of the Helios G4 PFIB. The Ti/Au layer that covers what will be the active part of the device is first removed with an acceleration voltage of 5\u2009kV and an ion current of 2\u2009nA. The overall shape of the device with the position of the contacts is then cut out at 0.3\u2009nA and 30\u2009kV and the central bar of the device is gently polished in order to create smooth sidewalls. Finally, the Ti/Au layer away from the device is cut through in order to separate the contact pads and make sure that current can only flow through the device, which is then ready for measurement.\n\nIn order to check that the crystallographic orientation of the final device is as expected based on the initial XRD measurements, we perform measurements of the MR. The MR has a characteristic shape with a minimum for B\u2225c and a maximum for B\u2225b19, allowing an identification of these axes. In Supplementary Fig.\u00a01, we show an angle-dependent measurement of MR for one of our devices, which is indeed aligned 45\u2218 away from the b and c axes.\n\nThe calculations of the k-dependent cross-sectional area of the FS shown in Fig.\u00a03 and Supplementary Fig.\u00a03 were performed with a slightly modified version of SKEAF13, a commonly used software designed to numerically extract QO frequencies from calculated band structures. QO take place at extremal areas of the FS (i.e. when \\(\\frac{dA}{dk}=0\\)) and their frequency relates to the extremal area as given by the Onsager relation15: \\(F=\\frac{\\hslash }{2\\pi e}A\\). This correspondence allows us to display A(k) in units of kT rather than 1/\u00c52. As QO are more known than SO, this choice of unit facilitates their comparison.\n\nThe SKEAF algorithm, written in the Fortran 90 language, reads electronic structures calculated by DFT in the Band-XCrySDen-Structure-File format. It constructs a cubic super cell much larger than the original reciprocal unit cell and aligned with the magnetic field direction. This super cell is then divided into slices perpendicular to the magnetic field, and the software records the cross-sectional area for each slice. During regular use, SKEAF then matches the orbits over the different slices and finds the extremal ones. We, however, need the area for each slice, and hence we have added a short section of code to create a new file containing the k-values and areas (in both 1/\u00c52 and kT) for all orbits.\n\nThis file contains many copies of each FS sheet. Rather than averaging each orbit, as is done by SKEAF for the extremal orbits, we simply select one copy and plot this as in Fig.\u00a03e. This is reasonable, as the differences between the areas of the copies are consistently less than 0.1%. Finally, we take a numerical derivative of A(k) and find \\(\\frac{dA}{dk}\\), from which we can identify possible Sondheimer orbits.\n\nWe follow a step-by-step procedure in order to extract the lMR as a function of temperature. In the first step, the measured longitudinal or Hall resistivity is smoothed and differentiated twice. The required level of smoothing is adjusted for each dataset to the extent that no oscillatory component of the data is removed, while the noise is suppressed. Importantly, the same procedure is performed consistently for each temperature. After taking the second derivative, we perform an FFT of the data using a Hanning window. The relevant amplitude A(T) is found from the peak in the FFT and plotted against the temperature (see Fig.\u00a04b). In order to extract A(0), we then need to extrapolate to T\u2009=\u20090\u2009K. In order to do this, we use a fit of the form \\(A(T)={A}_{1}/\\left(1+\\exp \\left(\\frac{T-c1}{c2}\\right)\\right)\\), which provides an excellent empirical description of the data and allows us to determine A(0). As the SO amplitude saturates at low temperatures, the exact extrapolation procedure has little effect on the value of A(0) and the extrapolation is robust.\n\nFinally, we use Eq. (2) to calculate lMR(T), which we plot in Fig.\u00a04b. In Supplementary Fig.\u00a05, we show lMR(T) for several different devices, showing consistency between the values extracted for any thickness, from \u03c1xx or \u03c1xy and from measurements along different crystallographic axes.\n\nThe ab initio calculations were performed with the open source DFT code JDFTx42. We used fully relativistic Perdew\u2013Burke\u2013Ernzerhof pseudopotentials43,44,45 and included the spin-orbit coupling effect in all calculations. A kinetic energy cutoff of 28 Ha was used along with a 6\u2009\u00d7\u20096\u2009\u00d7\u20098\u0393-centered k-mesh and a Fermi-Dirac smearing of 0.01 Ha for the Brillouin zone integration. Both the lattice constants and the ion positions were relaxed until the energy differences were less than 10\u22129 Ha. To compute the electron\u2013phonon scattering time, we performed frozen phonon calculations in a 3\u2009\u00d7\u20093\u2009\u00d7\u20092 supercell, and obtained 44 maximally localized Wannier functions by projecting the plane-wave bandstructure to W d and P p orbitals, which allowed us to converge the electron scattering calculation on a much finer 66\u2009\u00d7\u200966\u2009\u00d7\u200988\\({{{{{{{\\bf{k}}}}}}}}^{\\prime}\\) and q grid for T\u2009=\u200910 K.",
+            "section_image": []
+        },
+        {
+            "section_name": "Data availability",
+            "section_text": "The data generated in this study have been deposited in the Zenodo repository, https://doi.org/10.5281/zenodo.4675599. Data presented in Fig.\u00a04d, e are available upon request.",
+            "section_image": []
+        },
+        {
+            "section_name": "References",
+            "section_text": "Zaanen, J. Electrons go with the flow in exotic material systems. Science 351, 1026\u20131027 (2016).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLevitov, L. & Falkovich, G. Electron viscosity, current vortices and negative nonlocal resistance in graphene. Nat. Phys. 12, 672\u2013676 (2016).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nVarnavides, G., Jermyn, A. S., Anikeeva, P., Felser, C. & Narang, P. Electron hydrodynamics in anisotropic materials. Nat. Commun. 11, 4710 (2020).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSulpizio, J. A. et al. Visualizing Poiseuille flow of hydrodynamic electrons. Nature 576, 75\u201379 (2019).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nVool, U. et al. Imaging phonon-mediated hydrodynamic flow in WTe2 with cryogenic quantum magnetometry. ArXiv 2009.04477 (2020).\n\nGurzhi, R. N., Kalinenko, A. N. & Kopeliovich, A. I. Hydrodynamic effects in the electrical conductivity of impure metals. Zh. Eksp. Teor. Fiz. 96, 1522\u20131532 (1989).\n\n\n                    Google Scholar\u00a0\n                \n\nBufler, F. M. et al. Monte Carlo simulation and measurement of nanoscale n-MOSFETs. IEEE Trans. Electron Devices 50, 418\u2013424 (2003).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nJazaeri, F., Beckers, A., Tajalli, A. & Sallese, J. M. A review on quantum computing: Qubits, cryogenic electronics and cryogenic MOSFET physics. ArXiv 1908.02656 (2019).\n\nSondheimer, E. H. The influence of a transverse magnetic field on the conductivity of thin metallic films. Phys. Rev. 80, 401\u2013406 (1950).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGurevich, V. L. Oscillations in the conductivity of metallic films in magnetic field. Sov. Phys. JETP 35(8), 464\u2013470 (1959).\n\n\n                    Google Scholar\u00a0\n                \n\nHarrison, W. A. Electronic structure of polyvalent metals. Phys. Rev. 118, 1190\u20131208 (1960).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nHambourger, P. D. & Marcus, J. A. Size-dependent oscillatory magnetoresistance in cadmium. Phys. Rev. B 8, 5567\u20135578 (1973).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nRourke, P. M. & Julian, S. R. Numerical extraction of de Haas-van Alphen frequencies from calculated band energies. Computer Phys. Commun. 183, 324\u2013332 (2012).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMackey, H. J. & Sybert, J. R. Harmonic content of magnetomorphic oscillations in the kinetic coefficients of electron transport due to partially specular boundary scattering. Phys. Rev. 158, 658\u2013661 (1967).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nShoenberg, D. Magnetic oscillations in metals (Cambridge University Press, 1984).\n\nAut\u00e8s, G., Gresch, D., Troyer, M., Soluyanov, A. A. & Yazyev, O. V. Robust Type-II Weyl Semimetal Phase in Transition Metal Diphosphides XP2 (X\u2009=\u2009Mo, W). Phys. Rev. Lett. 117, 066402 (2016).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nJaoui, A. et al. Departure from the Wiedemann-Franz law in WP2 driven by mismatch in T-square resistivity prefactors. npj Quantum Mater. 3, 64 (2018).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGooth, J. et al. Thermal and electrical signatures of a hydrodynamic electron fluid in tungsten diphosphide. Nat. Commun. 9, 4093 (2018).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKumar, N. et al. Extremely high magnetoresistance and conductivity in the type-II Weyl semimetals WP2 and MoP2. Nat. Commun. 8, 1642 (2017).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nGrishin, A., Lutsishin, P., Ostroukhov, Y. & Panchenko, O. Multiple Sondheimer oscillations in tungsten plates with atomically pure surfaces. Sov. J. Exp. Theor. Phys. 49, 673 (1979).\n\nADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nAlstadheim, T. & Risnes, R. Sondheimer oscillations in aluminium single crystals. Philos. Mag. 18, 885\u2013894 (1968).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKunzler, J. E. & Renton, C. A. \u201cSize effect\u201d in electrical resistivity measurements on single crystals of high-purity tin at liquid helium temperatures. Phys. Rev. 108, 1397 (1957).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBabiskin, J. & Siebenmann, P. G. New type of oscillatory magnetoresistance in metals. Phys. Rev. 107, 1249\u20131254 (1957).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nCoulter, J. et al. Uncovering electron-phonon scattering and phonon dynamics in type-I Weyl semimetals. Phys. Rev. B 100, 220301(R) (2019).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nNarang, P., Garcia, C. A. & Felser, C. The topology of electronic band structures. Nat. Mater. 20, 293\u2013300 (2021).\n\nOsterhoudt, G. B. et al. Evidence for dominant phonon-electron scattering in Weyl semimetal WP2. Phys. Rev. X 11, 01107 (2021).\n\nMoll, P. J. W. Focused ion beam microstructuring of quantum matter. Annu. Rev. Condens. Matter Phys. 9, 147\u2013162 (2018).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nShekhar, C. et al. Single crystal growth for topology and beyond. Chemical Metals Science & Solid State Chemistry. https://www.cpfs.mpg.de/ssc/crystal-growth.pdf (2018).\n\nKelley, R., Song, K., Van Leer, B., Wall, D. & Kwakman, L. Xe+ FIB milling and measurement of amorphous silicon damage. Microsc. Microanalysis 19, 862\u2013863 (2013).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nFuchs, K. The conductivity of thin metallic films according to the electron theory of metals. Math. Proc. Camb. Philos. Soc. 34, 100\u2013108 (1938).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMoll, P. J. W., Kushwaha, P., Nandi, N., Schmidt, B. & Mackenzie, A. P. Evidence for hydrodynamic electron flow in PdCoO2. Science 351, 1061\u20131064 (2016).\n\nSch\u00f6nemann, R. et al. Fermi surface of the Weyl type-II metallic candidate WP2. Phys. Rev. B 96, 121108(R) (2017).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nBachmann, M. D. et al. Inducing superconductivity in Weyl semi-metal microstructures by selective ion sputtering. Sci. Adv. 3, e1602983 288 (2017).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nVan Delft, M. R. et al. Two- and three-dimensional superconducting phases in the Weyl semimetal TaP at ambient pressure. Crystals 10, 288 (2020).\n\nArticle\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nKresse, G. & Furthm\u00fcller, J. Efficiency of ab-initio total energy calculations for metals and semiconductors using a plane-wave basis set. Computational Mater. Sci. 6, 15\u201350 (1996).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nSato, H. Phonon-limited mean free path in the Sondheimer oscillation of aluminum. J. Low. Temp. Phys. 38, 267\u2013275 (1980).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nMunarin, J. A., Marcus, J. A. & Bloomfield, P. E. Size-dependent oscillatory magnetoresistance effect in gallium. Phys. Rev. 172, 718\u2013736 (1968).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nLiang, T. et al. Ultrahigh mobility and giant magnetoresistance in Cd3As2: protection from backscattering in a Dirac semimetal. Nat. Mater. 14, 280\u2013284 (2015).\n\nGao, W. et al. Extremely large magnetoresistance in a topological semimetal candidate pyrite PtBi2. Phys. Rev. Lett. 118, 256601 (2017).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nZhang, C. L. et al. Electron scattering in tantalum monoarsenide. Phys. Rev. B 95, 085202 (2017).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nZiman, J. M. Electrons and phonons: the theory of transport phenomena in solids (Oxford University Press, 2001).\n\nSundararaman, R. et al. JDFTx: Software for joint density-functional theory. SoftwareX 6, 278\u2013284 (2017).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nPerdew, J. P., Burke, K. & Ernzerhof, M. Generalized gradient approximation made simple. Phys. Rev. Lett. 77, 3865\u20133868 (1996).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nDal Corso, A. Pseudopotentials periodic table: From H to Pu. Computational Mater. Sci. 95, 337\u2013350 (2014).\n\nArticle\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nPerdew, J. P. et al. Restoring the density-gradient expansion for exchange in solids and surfaces. Phys. Rev. Lett. 100, 136406 (2008).\n\nArticle\u00a0\n    ADS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nCoulter, J., Sundararaman, R. & Narang, P. Microscopic origins of hydrodynamic transport in the type-II Weyl semimetal WP2. Phys. Rev. B 98, 115130 (2018).\n\nArticle\u00a0\n    ADS\u00a0\n    CAS\u00a0\n    \n                    Google Scholar\u00a0\n                \n\nDownload references",
+            "section_image": []
+        },
+        {
+            "section_name": "Acknowledgements",
+            "section_text": "M.R.v.D. acknowledges funding from the Rubicon research program with project number 019.191EN.010, which is financed by the Dutch Research Council (NWO). This project was funded by the European Research Council (ERC) under the European Union\u2019s Horizon 2020 research and innovation program (grant no. 715730, MiTopMat). Y.W. is partially supported by the STC Center for Integrated Quantum Materials, NSF Grant No. DMR-1231319 for development of computational methods for topological materials. This research used resources of the National Energy Research Scientific Computing Center, a DOE Office of Science User Facility supported by the Office of Science of the U.S. Department of Energy under Contract No. DE-AC02-05CH11231 as well as resources at the Research Computing Group at Harvard University. P.N. is a Moore Inventor Fellow and gratefully acknowledges support through Grant No. GBMF8048 from the Gordon and Betty Moore Foundation. C.A.C.G. acknowledges support from the NSF Graduate Research Fellowship Program under Grant No. DGE-1745303. We acknowledge financial support from DFG through SFB 1143 (project-id 258499086) and the W\u00fcrzburg-Dresden Cluster 274 of Excellence on Complexity and Topology in Quantum Matter - ct.qmat (EXC 2147, project-id 39085490). B.G. acknowledges financial support from the Swiss National Science Foundation (grant numner CRSII5_189924). H.S and B.G thank J. Gooth for discussion and K. Moselund, S. Reidt, and A. Molinari for support, and received funding from the European Union\u2019s Horizon 2020 research and innovation program under Grant Agreement ID 829044 \u201cSCHINES\u201d.",
+            "section_image": []
+        },
+        {
+            "section_name": "Author information",
+            "section_text": "Laboratory of Quantum Materials (QMAT), Institute of Materials (IMX), \u00c9cole Polytechnique F\u00e9d\u00e9rale de Lausanne (EPFL), Lausanne, Switzerland\n\nMaarten R. van Delft,\u00a0Carsten Putzke,\u00a0Chunyu Guo,\u00a0Jonas Diaz\u00a0&\u00a0Philip J. W. Moll\n\nHarvard John A. Paulson School of Engineering and Applied Sciences, Harvard University, Cambridge, MA, USA\n\nYaxian Wang,\u00a0Georgios Varnavides,\u00a0Christina A. C. Garcia\u00a0&\u00a0Prineha Narang\n\nIBM Research Europe - Zurich, R\u00fcschlikon, Switzerland\n\nJacopo Oswald,\u00a0Heinz Schmid\u00a0&\u00a0Bernd Gotsmann\n\nMax Planck Institute for Chemical Physics of Solids, Dresden, Germany\n\nVicky S\u00fcss,\u00a0Horst Borrmann,\u00a0Yan Sun\u00a0&\u00a0Claudia Felser\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nSearch author on:PubMed\u00a0Google Scholar\n\nM.R.v.D., C.P., J.O., C.G., J.D. performed the transport experiments, as well as the microfabrication in collaboration with B.G. and H.S. The crystals were grown by V.S. and C.F., and crystallographically analyzed by H.B. Y.S. and C.F. calculated the band structures, and Y.W., G.V., C.A.C.G., P.N. performed the electron\u2013phonon scattering calculations. B.G., C.F., P.N., and P.J.W.M. conceived the experiment, and all authors contributed to writing of the manuscript.\n\nCorrespondence to\n                Maarten R. van Delft, Prineha Narang or Philip J. W. Moll.",
+            "section_image": []
+        },
+        {
+            "section_name": "Ethics declarations",
+            "section_text": "The authors declare no competing interests.",
+            "section_image": []
+        },
+        {
+            "section_name": "Additional information",
+            "section_text": "Peer review information Nature Communications thanks the anonymous reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available.\n\nPublisher\u2019s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.",
+            "section_image": []
+        },
+        {
+            "section_name": "Rights and permissions",
+            "section_text": "Open Access  This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article\u2019s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article\u2019s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.\n\nReprints and permissions",
+            "section_image": []
+        },
+        {
+            "section_name": "About this article",
+            "section_text": "van Delft, M.R., Wang, Y., Putzke, C. et al. Sondheimer oscillations as a probe of non-ohmic flow in WP2 crystals.\n                    Nat Commun 12, 4799 (2021). https://doi.org/10.1038/s41467-021-25037-0\n\nDownload citation\n\nReceived: 03 February 2021\n\nAccepted: 21 July 2021\n\nPublished: 10 August 2021\n\nVersion of record: 10 August 2021\n\nDOI: https://doi.org/10.1038/s41467-021-25037-0\n\nAnyone you share the following link with will be able to read this content:\n\nSorry, a shareable link is not currently available for this article.\n\n\n\n\n                            Provided by the Springer Nature SharedIt content-sharing initiative\n                        ",
+            "section_image": [
+                "https://data:image/svg+xml;base64,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"
+            ]
+        },
+        {
+            "section_name": "This article is cited by",
+            "section_text": "La Rivista del Nuovo Cimento (2022)",
+            "section_image": []
+        }
+    ]
+}

0d397fac72f0ff0c7a01d139e71d10c74fc08d546e5a9dce6d59c01b0601b529/preprint/preprint.md

@@ -0,0 +1,282 @@
+Direct synthesis of an iron metal-organic framework antiferromagnetic glass
+
+Guillermo Mínguez Espallargas
+guillermo.minguez@uv.es
+
+University of Valencia   https://orcid.org/0000-0001-7855-1003
+Luis León-Alcaide
+University of Valencia
+Lucía Martínez-Goyeneche
+University of Valencia
+Michele Sessolo
+Universidad de Valencia   https://orcid.org/0000-0002-9189-3005
+Bruno Vieira
+Universidade de Lisboa
+Joao Waerenborgh
+Universidade de Lisboa
+José Rodríguez-Velamazán
+Institut Laue-Langevin
+Oscar Fabelo
+Institut Laue-Langevin   https://orcid.org/0000-0001-6452-8830
+Matthew Cliffe
+University of Nottingham   https://orcid.org/0000-0002-0408-7647
+David Keen
+Rutherford Appleton Laboratory   https://orcid.org/0000-0003-0376-2767
+
+Physical Sciences - Article
+
+Keywords:
+
+Posted Date: January 16th, 2025
+
+DOI: https://doi.org/10.21203/rs.3.rs-5822781/v1
+
+License: This work is licensed under a Creative Commons Attribution 4.0 International License.
+Read Full License
+Additional Declarations: There is NO Competing Interest.
+
+Version of Record: A version of this preprint was published at Nature Communications on October 2nd, 2025. See the published version at https://doi.org/10.1038/s41467-025-63837-w.
+Direct synthesis of an iron metal-organic framework antiferromagnetic glass
+
+Luis León-Alcaide,a Lucía Martínez-Goyeneche,a Michele Sessolo,a Bruno J. C. Vieira,b João C. Waerenborgh,b J. Alberto Rodríguez-Velamazán,c Oscar Fabelo,c Matthew J. Cliffe,d David A. Keen,e and Guillermo Mínguez Espallargas*a
+
+a Instituto de Ciencia Molecular (ICMol), Universidad de Valencia, c/ Catedrático José Beltrán, 2, 46980, Paterna, Spain
+b Centro de Ciências e Tecnologias Nucleares, DECN, Instituto Superior Técnico, Universidade de Lisboa, 2695-066 Bobadela, LRS, Portugal
+c Institut Laue-Langevin, 6 rue Jules Horowitz, BP 156, 38042 Grenoble Cedex 9, France
+d School of Chemistry, University of Nottingham, Nottingham NG7 2RD, United Kingdom
+e ISIS Facility, Rutherford Appleton Laboratory, Harwell Campus, Didcot, Oxfordshire OX11 0QX, United Kingdom
+*e-mail: guillermo.minguez@uv.es
+
+Summary
+
+Glasses are ubiquitous in our everyday lives but still pose fundamental questions about the nature of order in solids. Typically formed by the rapid cooling of a liquid, these amorphous solids have broad applications, with vitreous silica (SiO₂) the most well-known example.¹ Most functional glasses are purely inorganic solids, restricting the range of functional properties feasible.² Recently, a number of new metal-organic framework (MOF) glasses containing molecular components has been discovered by heating their crystalline counterparts to a melting point followed by a rapid cooling, thereby expanding the potential applications of glass materials.³⁻⁵ However, the melt-quenched (MQ) approach is limited to MOFs that melt, which is very restrictive as most MOFs readily decompose at heating to comparatively low temperatures.⁶ The low decomposition temperatures mean that glassy MOF samples typically include impurity decomposition products detrimental to functionality, as optical, electronic and magnetic contaminants. In this work, we present a direct route to prepare a family of MOF glasses without a meltable crystalline precursor. This route produces high-purity iron (II) MOF glasses, avoiding the oxidation and partial degradation commonly associated with the conventional melt-quenching process. The absence of magnetic impurities allows us to study the magnetic properties of the MOF glass itself and show that MOF glasses are good model systems for topologically disordered amorphous antiferromagnets. We also present the functional advantages of direct-glass synthesis by creating free-standing films of glassy MOFs and integrating them in optoelectronic devices. Direct-glass synthesis is thus a powerful route to exploit the true functional potential of glassy MOFs, not only realizing new MOF glasses but also unveiling properties that can be accessed with these materials.⁷
+Introduction
+
+Metal-organic framework (MOF) glasses are a new class of molecular materials that combine the unique properties of crystalline MOFs, such as high surface area and chemical versatility,\(^8\) with the distinctive properties of a glass, such as formability.\(^9\) This makes them particularly interesting for applications that benefit from both porosity and amorphous nature,\(^{10-12}\) such as advanced gas separation,\(^{13,14}\) hybrid electrochemical systems,\(^{15,16}\) and photonic applications.\(^{17}\)
+
+However, there are currently only a handful of known MOF glasses,\(^{3,18}\) primarily derived from a subclass of MOFs known as zeolitic imidazolate frameworks (ZIFs).\(^{19}\) ZIFs are distinguished by their tetrahedral network structure, which closely resembles that of silica and zeolites.\(^{20}\) Like silica, these frameworks can melt and by rapidly cooling the melt, the MOF glass is formed.\(^{21,22}\) Conversely, unlike silica, the chemical versatility of these molecular-based solids allows the modification of both the metal centres (e.g. Zn, Co, Fe)\(^{23-25}\) and the organic ligands (e.g. imidazolates, benzimidazolates)\(^{26,27}\) causing variations in the melting temperatures. Still, the anionic imidazolate is thermally unstable in the temperature regime of melting. This instability means that melt quenching, even in inert conditions, can introduce impurities into the glass and is particularly acute for MOF glasses with paramagnetic transition metals, which typically contain impurities of metallic cobalt or Fe\(_3\)O\(_4\).\(^{28}\) Additionally, thermal processing of MOF glasses requires careful control of the time-temperature process, as melt-quenching rates, re-melting of the glass, or direct collapse of the MOF structure, can have large effects on the final properties of the material, including bubble formation, micro-cracking, surface oxidation, and extent of decomposition.\(^{29}\)
+
+We therefore developed an alternative direct route to MOF glasses, through the reaction of metal precursor and imidazolate ligand,\(^{30-32}\) that does not mirror synthetic routes for traditional inorganic glasses,\(^1\) and overcome the limitations associated with conventional methods for the preparation of molecular-based glasses. Our new route allows the preparation of nearly impurity-free MOF glasses containing paramagnetic metals, which permits the exploration of their intrinsic optical, electronic and magnetic function. MOF glasses simultaneously possess long-range structure that is completely disordered, but with a strong degree of local uniformity.\(^{33}\) Magnetic MOF glasses thus are an unusual kind of amorphous magnets, where local exchange and single ion anisotropy disorder is expected to be small, meaning the intrinsic topological disorder dominates their properties.\(^{34,35}\) This mirrors other topologically disordered materials, where the properties such as electronic or photonic bandgaps emerge from the local uniformity, as occurs in amorphous silicon or disordered metamaterials.\(^{36}\) Topologically amorphous antiferromagnets would therefore have collective magnetism. Although there are many amorphous magnets, most are ferromagnets with extensive site disorder, and the only characterized example of an amorphous antiferromagnet is ionic FeF\(_3\).\(^{37}\) Thus, achieving a molecular amorphous
+antiferromagnet would allow for a chemical versatility inaccessible with metallic systems,\(^{38}\) as demonstrated herein. Direct synthesis of MOF glasses thus not only expands the range of feasible materials, but also the range of functional properties that can be achieved in MOF glasses, as demonstrated with the realization of MOF-supported photodetectors and topologically amorphous antiferromagnets.
+
+**Direct synthesis of \(a_g\)-ZIFs**
+
+Heating at 300 °C for 6 hours a combination of ferrocene (0.16 mmol) and a mixture of imidazole:benzimidazole (0.32 mmol), sealed under vacuum, results in the formation of an orange transparent glassy monolith of formula Fe(im)\(_{2-x}\)(bim)\(_x\), denoted as **dg-MUV-29** (dg = direct-glass), which can be prepared with different amounts of imidazole and benzimidazole (\(x = 0.2\), 0.4, 0.5, and 0.6). The incorporated amount of each ligand into the framework was determined by \({}^1\)H NMR spectroscopy (Table S1). Larger amounts of benzimidazole (\(x \geq 0.8\)) yield a new crystalline phase as a byproduct, denoted **MUV-28**, of formula Fe(im)\(_1\)(bim)\(_1\), whereas the absence of benzimidazole (\(x = 0\)) produces the known crystalline **MUV-24(coi)**\(^{25}\) (Figure 1 and
+
+![Schematic representation of the structures obtained under high temperature synthetic conditions for MUV-24(coi), dg-MUV-29, and MUV-28.](page_370_670_1002_324.png)
+
+Figure 1. Schematic representation of the structures obtained under high temperature synthetic conditions (300 °C for 6 hours) for **MUV-24(coi)**, **dg-MUV-29**, and **MUV-28**. In the diagram, tetrahedral Fe\(^{II}\) centers are illustrated as orange tetrahedra. The imidazolate ligands are shown in light blue, and the benzimidazolate ligands in dark blue. The corresponding PXRD pattern for each material is presented at the bottom, together with the calculated patterns of the crystalline materials.
+Figures S7–S9). Interestingly, reaction at lower temperature (150 °C) yields mixtures of crystalline phases which have been identified as **MUV-1**,32 **MUV-28** or **IMIDFE**.30
+
+No detectable Bragg diffraction peaks are observed in the different materials of the **dg-MUV-29** family, consistent with a glassy phase. Furthermore, differential scanning calorimetry (DSC) measurements exhibit no endothermic signal related to framework melting, but instead display the distinctive glass transition corresponding to their glassy nature (Figure 2a). The glass transition temperature (\( T_g \)), defined as the peak onset temperature of the endothermic calorimetric signal of the glasses, ranges from 205 °C to 245 °C (Figure 2b) depending on the incorporated amount of benzimidazole (\( x \)), and is slightly higher than for MQ **a$_{e}$-MUV-24$^{25}$** (\( T_g = 190 \) °C), of formula Fe(im)$_2$. This effect is consistent with previous observations for the Zn analogue, where \( T_g \) also increases with bim$^-$ concentration.
+
+![Thermogravimetric analysis (TGA) and DSC curves of dg-MUV-29-bim$_{0.48}$](page_186_670_495_312.png)
+
+![DSC curves of the materials obtained via high-temperature synthesis with variable im$^-$ to bim$^-$ ratios](page_682_670_495_312.png)
+
+![Optical images of various members of dg-MUV-29](page_1178_670_246_120.png)
+
+Figure 2. a) Thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC) curves of **dg-MUV-29-bim$_{0.48}$** b) DSC of the materials obtained via high-temperature synthesis with variable im$^-$ to bim$^-$ ratios. In all cases, an initial heating cycle up to 200 °C was conducted to remove organic impurities. c) Optical images of various members of **dg-MUV-29**.
+
+The lack of long-range order in the glasses is confirmed by the absence of sharp features in their X-ray total scattering data. Pair distribution functions (PDF or \( D(r) \)) for **dg-MUV-29** were extracted from the total scattering data via appropriate corrections and subsequent Fourier transformation and are presented in Figure 3a. Comparison of this data with crystalline Fe$_3$(im)$_6$(Him)$_2$ and MQ **a$_{e}$-MUV-24** shows that the short-range correlations are similar, although a new peak (peak “C” at 2.44 Å in Figure 3a) can be detected in the **dg-MUV-29** family, corresponding to the distance between carbon atoms in the benzene ring of the benzimidazole ligand (Figure 3b). The relative intensity of this peak increases as the amount of bim increases, confirming the successful incorporation of bim rings. The PDF peaks extend up to approximately 6 Å, which matches with the distance between neighboring Fe$^{II}$ centers (5.8 – 6.1 Å in the crystalline phase **MUV-24$^{25}$**). This clearly indicates that the tetrahedral coordination of the Fe$^{II}$ centers with imidazolate linkers is maintained in **dg-MUV-29** materials, and that their chemical
+environment is similar to that of MQ \( \text{a}_g\text{-MUV-24} \). Importantly, the Bragg peaks attributed to crystalline impurities of Al and Fe$_2$N observed in the MQ \( \text{a}_g\text{-MUV-24}^{25} \) data at higher Q values are absent in dg-MUV-29 (Figure S13). These impurities prevented the study of the magnetic properties of \( \text{a}_g\text{-MUV-24} \). In contrast, the direct-glass synthesis avoids the melting process required in the melt-quenched protocol, thereby preventing typical degradation and partial oxidation observed in MQ \( \text{a}_g\text{-MUV-24}^{25} \). The phase purity is unequivocally confirmed by
+
+![X-ray PDF plots and Mössbauer spectra for different samples](page_186_370_1077_1017.png)
+
+Figure 3. a) X-ray PDF in the form of \( D(r) \) of dg-MUV-29-bim$_{0.48}$, dg-Fe-MUV-29-bim$_{0.18}$, \( \text{a}_g\text{-MUV-24} \) and IMIDFE. b) Schematic representation of the benzimidazolate bridge coordinated to two Fe$^{II}$ centers and the connectivity between tetrahedral Fe$^{II}$ ions with benzimidazolate linkers. (c-e) Mössbauer spectra of dg-MUV-29-bim$_{0.48}$ (c) \( \text{a}_g\text{-MUV-24} \) (d) and IMIDFE (e) samples, displayed with lines overlaid on the experimental data. These lines represent the sum of doublets and sextets, or distributions of QS (see Table S2), and are slightly shifted for clarity.
+Mössbauer spectroscopy, which is extremely sensitive to the geometry and oxidation state of iron centres. The characterization of the local environment of the metal ion in amorphous phases is very challenging due to the limited number of available techniques. Specific techniques such as \(^{67}\)Zn NMR has greatly advanced the understanding of Zn\(^{II}\) centers within a random network, detecting short-range disorder in Zn-based a-g-ZIFs, which has been a significant achievement in the field.\(^{33}\) The incorporation of Fe\(^{II}\) centers permits the possibility of using Mössbauer spectroscopy to study for the first time the local environment of a glassy phase. The Mössbauer spectra of dg-MUV-29 measured at 80 K reveals two broad peaks with the same relative area but different widths (Figure 3c). This is typical of amorphous materials where the Fe\(^{II}\) centres have the same coordination number but are located on sites with slightly different geometries and environments, leading to slightly different isomer shift (IS) and quadrupole splitting (QS).\(^{39}\) The spectrum can be fitted with a distribution of QS, and the average IS is consistent with tetrahedral coordination of all Fe\(^{II}\) cations as in MQ a-g-MUV-24.
+
+Chemical diversity
+
+The versatility of the direct-glass synthesis also provides the flexibility to incorporate linkers with diverse functionalities into the glassy phase. Until now, incorporating bulky benzimidazolate ligands with iron derivatives has been impossible because access to the *cag* topology is required,
+
+![Optical images and structural diagrams of dg-MUV-29-X monolithic glass, X-ray powder diffractograms, and DSC first upscan plots](page_153_670_1142_377.png)
+
+Figure 4. a) Optical images of different dg-MUV-29-X monolithic glass. b) X-ray powder diffractograms of dg-MUV-29-X. c) DSC first upscan of dg-MUV-29-X family. To calculate the \(T_g\), the material was preheated to 200 °C to remove impurities. In all cases, the amount of bim incorporated is \(x = 0.5\).
+and that phase has remained elusive. Using the direct-glass synthesis, various analogues have been successfully prepared, denoted dg-MUV-29-X, where X is a different substituent in the position 5 of the benzimidazole ligand (X = NH₂, CH₃, Br, and Cl).
+
+For the synthesis of dg-MUV-29-X, ferrocene (28 mg, 0.15 mmol) was combined with imidazole (Him) (15.3 mg, 0.23 mmol) and a 5-X-benzimidazole (0.07 mmol) and sealed under vacuum in a 4 mm diameter layering tube. The mixtures were heated at 300 °C for 6 hours, producing an orange glass in each case corresponding to dg-MUV-29-NH₂, dg-MUV-29-CH₃, dg-MUV-29-Br, and dg-MUV-29-Cl, similar in aspect to dg-MUV-29 (Figure 4a). The incorporated amount of each ligand into the framework was determined by ¹H NMR spectroscopy (Table S4 and Section S3.3) and closely matches the synthetic amount. The amount of incorporated Xbim can be varied in a controlled manner by adjusting the ratio of the reactants. The incorporation of the different benzimidazolate ligands into the glassy material can be clearly followed for dg-MUV-29-Cl and dg-MUV-29-Br using EDX-SEM, showing that Cl and Br are uniformly distributed throughout the monolithic framework (Figure S39-S40). PXRD measurements of the resultant monolithic phases revealed the absence of any Bragg diffraction for all analogues (Figure 4b).
+
+DSC measurements of the different analogues of the dg-MUV-29-X family reveals the absence of endothermic peaks corresponding to framework melting but instead show the characteristic \( T_g \) of the glassy nature of the material (Figure 4c). The glass transition is nearly identical across the samples, and reversible, allowing for multiple cycles without any alteration, as also observed in dg-MUV-29 (Figure S37).
+
+**Topologically disordered amorphous antiferromagnets**
+
+Based on the combined results of all the characterization techniques, we can conclude that the direct-glass synthesis method enables the production of a high-purity Fe⁴⁺ phase which cannot be achieved through conventional melt-quenching methods. This level of purity allows for precise measurement of magnetic correlations of the first phase-pure MOF glass containing paramagnetic ions, providing insight into the nature of the magnetic phase of dg-MUV-29. Bulk magnetic measurements confirmed antiferromagnetic interactions typical of Fe⁴⁺ imidazolate MOFs (Figure 5a), as indicated by Curie-Weiss fitting (\( T_{CW} = -24.7(4) \) K, implying \( J = 16.5(3) \) K⁴⁰ with a moment slightly larger than the spin-only value (\( C = 3.41(1) \) emu mol⁻¹ K, equivalent to \( g = 2.13 \)).⁴¹ Unlike crystalline iron imidazolates, no clear ordering transition at higher temperature was observed in static susceptibility measurements, confirming the phase purity. AC susceptibility measurements (Figure 5b) revealed a clear cusp at 5.2(2) K, consistent with the presence of cooperative magnetic interactions. The magnetization as a function of the field was linear, indicating moderately strong antiferromagnetic interactions with no ferromagnetic components (Figure S21).
+Figure 5. a) Thermal dependence of magnetic susceptibility in the temperature range 2–300 K of dg-MUV-29. The data have been fitted (red dotted line) following the Curie-Weiss model. b) Temperature dependence of the in-phase dynamic a.c. susceptibility of dg-MUV-29 measured at different frequencies. c) Mossbauer measurements at different temperatures of dg-MUV-29. The lines over the experimental points are the estimated distributions of quadrupole splittings. d) \( I(Q) \) data, offset by 5000; dots represent the experimental data of dg-MUV-29, and the line represents the Monte Carlo (MC) simulation. The MC simulation is matched to temperature, with 1.8 K data plotted against 13 K. An empirical scale, offset, and linear background have been applied to all MC data, fixed to the lowest temperature, and distances are scaled by the experimentally determined metal-metal distance \( r_0 = 6.20 \) Å. e) Real space spin correlation function derived from MC simulations showing antiferromagnetic alternation that decays with increasing distance, \( r \).
+
+The Mössbauer spectrum of dg-MUV-29 at 4 K shows magnetic hyperfine splitting (Figure 5c), revealing strong magnetic correlations consistent with the transition observed in the magnetization data. The magnetic hyperfine interactions were found to be complex, requiring detailed modeling of the hyperfine interaction parameters, including the quadrupole interaction (\( \Delta = -2.42 \) mm s\(^{-1}\)), asymmetry parameter (\( \eta = 0.70 \)), and magnetic hyperfine field (\( B_{hf} = 15 \) T). These parameters suggest significant orbital magnetic momentum and confirm the high-spin Fe\(^{II}\) state (Figure 5c). These findings support the hypothesis that dg-MUV-29 is a continuous random network (CRN) antiferromagnet. These materials exhibit a completely disordered long-range structure while maintaining a strong degree of local uniformity. This implies that exchange
+disorder, i.e. variations in the strength and sign of magnetic interactions, are small, and hence the inherent topological amorphousness of the network is driving the observed magnetic properties. Although there are many amorphous ferromagnets, amorphous antiferromagnets are extremely rare. To further investigate this hypothesis, we conducted semi-classical Monte Carlo simulations using the Wooten-Winer-Weaire CRN model of amorphous silicon,\(^{42}\) simulating a 512-spin periodic approximant of the tetrahedral continuous random network (Section S2.9). The susceptibility results were in quantitative agreement with experimental data, particularly at high temperatures. Below \( T = 15 \) K, the simulated susceptibility showed a cusp and deviation consistent with a long-range ordering transition, mirroring the behavior previously seen in analogous crystalline systems.
+
+To investigate the local spin correlations, we conducted powder neutron diffraction measurements using diffractometer D20 at the ILL on a deuterated sample across a broad temperature range (1.8 to 100 K) (Figure 5d). Subtraction of the highest temperature data (100 K) revealed short-range magnetic correlations with a distinct broad peak at 0.79 \(\text{\AA}^{-1}\). This peak is lower than the first sharp diffraction peak from structural diffraction, and occurs at similar \( Q \) to the crystalline antiferromagnet analogues,\(^{32}\) suggestive of antiferromagnetic spin-spin correlations with a correlation length similar to the structure itself. Comparison with Monte Carlo simulations, using the metal-metal distance determined from PDF measurements, \( r_0 = 6.20 \) \(\text{\AA}\), showed qualitative agreement, particularly near the ordering transition at \( T = 13 \) K. However, deviations at low temperatures suggest complexities beyond the minimal model, with spin and quadrupole correlations indicating a speromagnetic state (Figure 5e and Figure S27), characterized by a disordered arrangement of magnetic moments (spins) on a local scale, while still maintaining no net magnetization across the entire system.\(^{43}\)
+
+The combined results from neutron scattering and magnetometry strongly indicate that this material behaves as an amorphous antiferromagnet with a continuous random network structure and predominantly nearest-neighbor Heisenberg interactions. Unlike previous amorphous antiferromagnets,\(^{37}\) this material exhibits minimal exchange disorder because each Fe ion has a similar local environment with four nearest neighbors, and Fe is the sole framework-forming cation.\(^{34,35}\) This contrasts with amorphous metals, which typically exhibit significant exchange and site disorder.\(^{38}\)
+
+**Scalability and processability**
+
+The ability to bypass the melting process and eliminate the need for a crystalline intermediate enhances the processability and allows for an easy scaling of the synthesis. This enables the production of large quantities of glass in film form simply by increasing the size of the layering tube used as a reactor. To demonstrate the scalability and processability of **dg-MUV-29** we have
+used it as a substrate for a centimeter-scale thin-film photodetector based on methylammonium lead iodide (MAPbI$_3$), replacing the conventional glass typically used. MAPbI$_3$ is an archetypal material within the broader metal halide perovskite family, a class of semiconductor with promising application in optoelectronic devices such as photodetectors, among others.\textsuperscript{44} MOFs have been used in combination with perovskites as additives (blended with the active material), host materials, or functional layers in order to improve the device performances,\textsuperscript{17,45–47} but their use as substrates in optoelectronic devices remains elusive, as producing large-area homogenous and monolithic MOFs is challenging. As most perovskite optoelectronic devices are fabricated on
+
+![Scheme of the perovskite photodetector and photo of the device laminated. Current vs. applied field under illumination and in the dark. Photocurrent response of the device at an applied field of 0.4 V/μm when exposed to consecutive 40 s light pulses.](page_246_682_1012_393.png)
+
+Figure 6. a) Scheme of the perovskite photodetector and b) a photo of the device laminated. c) current vs. applied field under illumination (solid line) and in the dark (dotted line). d) Photocurrent response of the device at an applied field of 0.4 V/μm when exposed to consecutive 40 s light pulses.
+
+soda-lime glass, dg-MUV-29 is used here as the substrate for a simple proof-of-concept lateral perovskite photodetector. After isolating a centimetre-size monolithic dg-MUV-29 (see Figure S29), we used vacuum deposition\textsuperscript{48,49} to process the perovskite on top of the glass by co-sublimation of PbI$_2$ and MAI, followed by evaporation of gold contacts through a shadow mask (Figure 6a-b). The current of the resulting MOF/perovskite photodetector was measured as a function of the applied field under illumination (white light, 100 mW·cm$^{-2}$) and in the dark, as shown in Figure 6c. The device showed the expected photoconductive behavior, with the current increasing more than three order of magnitude upon illumination, with an on/off ratio > 1000 at 0.4 V/μm. The on-current was found to be rather stable for several measurements (Figure 6d), which indicate an overall good quality of the MAPbI$_3$ ilm grown on the dg-MUV-29 substrate, as well as the mechanical stability of the MOF itself.
+Conclusions
+
+The solvent-free methodology presented here facilitates the direct synthesis of a novel family of MOF-derived glasses, designated dg-MUV-29, eliminating the necessity for a pre-existing crystalline phase. This innovative approach yields high-purity Fe"II" glasses while effectively mitigating the oxidation and degradation issues commonly associated with traditional melt-quenching techniques. The versatility of this synthetic technique allows for the incorporation of a diverse array of ligands featuring varying functional groups, leading to the development of the expanded dg-MUV-29-X family. Such adaptability broadens the scope of potential applications and avenues for future research involving these materials. In this sense, we have shown that the resulting free-standing films of glassy MOFs can be used as substrate for device integration. Moreover, this methodology establishes a robust platform for probing the magnetic properties of molecular-based random networks, thereby enhancing our understanding of the magnetism inherent in disordered systems. In summary, this direct-glass synthesis method proves to be a highly effective strategy for advancing the development of novel glassy ZIFs, thus paving the way for new research initiatives and applications. The implications of glassy ZIFs in the study of disordered systems underscore their potential to significantly impact the field of materials science.
+
+References
+
+(1) Greaves, G. N.; Sen, S. Inorganic Glasses, Glass-Forming Liquids and Amorphizing Solids. Adv. Phys. 2007, 56, 1–166.
+
+(2) Wang, W. H. Bulk Metallic Glasses with Functional Physical Properties. Adv. Mater. 2009, 21, 4524–4544.
+
+(3) Ma, N.; Horike, S. Metal-Organic Network-Forming Glasses. Chem. Rev. 2022, 122, 4163–4203.
+
+(4) Wang, M.; Zhao, H.; Du, B.; Lu, X.; Ding, S.; Hu, X. Functions and Applications of Emerging Metal–Organic-Framework Liquids and Glasses. Chem. Comm. 2023, 59, 7126–7140.
+
+(5) Bennett, T. D.; Horike, S.; Mauro, J. C.; Smedskjaer, M. M.; Wondraczek, L. Looking into the Future of Hybrid Glasses. Nat. Chem. 2024, 16, 1755–1766.
+
+(6) Healy, C.; Patil, K. M.; Wilson, B. H.; Hermanspahn, L.; Harvey-Reid, N. C.; Howard, B. I.; Kleinjan, C.; Kolien, J.; Payet, F.; Telfer, S. G.; Kruger, P. E.; Bennett, T. D. The Thermal Stability of Metal-Organic Frameworks. Coord. Chem. Rev. 2020, 419, 213388.
+
+(7) Mauro, J. C.; Tandia, A.; Vargheese, K. D.; Mauro, Y. Z.; Smedskjaer, M. M. Accelerating the Design of Functional Glasses through Modeling. Chem. Mater. 2016, 28, 4267–4277.
+(8) Maurin, G.; Serre, C.; Cooper, A.; Férey, G. The New Age of MOFs and of Their Porous-Related Solids. Chem. Soc. Rev. 2017, 46, 3104–3107.
+
+(9) Qiao, A.; Bennett, T. D.; Tao, H.; Krajnc, A.; Mali, G.; Doherty, C. M.; Thornton, A. W.; Mauro, J. C.; Greaves, G. N.; Yue, Y. A Metal-Organic Framework with Ultrahigh Glass-Forming Ability. Sci. Adv. 2018, 4, eaao6827.
+
+(10) Fonseca, J.; Gong, T.; Jiao, L.; Jiang, H. L. Metal–Organic Frameworks (MOFs) beyond Crystallinity: Amorphous MOFs, MOF Liquids and MOF Glasses. J. Mater. Chem. A. 2021, 9, 10562–10611.
+
+(11) Frentzel-Beyme, L.; Kolodzeiski, P.; Weiß, J. B.; Schneemann, A.; Henke, S. Quantification of Gas-Accessible Microporosity in Metal-Organic Framework Glasses. Nat. Comm. 2022, 13, 7750
+
+(12) Zhou, C.; Longley, L.; Krajnc, A.; Smales, G. J.; Qiao, A.; Erucar, I.; Doherty, C. M.; Thornton, A. W.; Hill, A. J.; Ashling, C. W.; Qazvini, O. T.; Lee, S. J.; Chater, P. A.; Terrill, N. J.; Smith, A. J.; Yue, Y.; Mali, G.; Keen, D. A.; Telfer, S. G.; Bennett, T. D. Metal-Organic Framework Glasses with Permanent Accessible Porosity. Nat. Comm. 2018, 9, 5042.
+
+(13) Wang, Y.; Jin, H.; Ma, Q.; Mo, K.; Mao, H.; Feldhoff, A.; Cao, X.; Li, Y.; Pan, F.; Jiang, Z. A MOF Glass Membrane for Gas Separation. Angew. Chem. Int. Ed. 2020, 59, 4365–4369.
+
+(14) Yang, Z.; Belmabkhout, Y.; McHugh, L. N.; Ao, D.; Sun, Y.; Li, S.; Qiao, Z.; Bennett, T. D.; Guiver, M. D.; Zhong, C. ZIF-62 Glass Foam Self-Supported Membranes to Address CH4/N2 Separations. Nat. Mater. 2023, 22, 888–894.
+
+(15) Lin, R.; Li, X.; Krajnc, A.; Li, Z.; Li, M.; Wang, W.; Zhuang, L.; Smart, S.; Zhu, Z.; Appadoo, D.; Harmer, J. R.; Wang, Z.; Buzanich, A. G.; Beyer, S.; Wang, L.; Mali, G.; Bennett, T. D.; Chen, V.; Hou, J. Mechanochemically Synthesised Flexible Electrodes Based on Bimetallic Metal–Organic Framework Glasses for the Oxygen Evolution Reaction. Angew. Chem. Int. Ed. 2022, 61, e202112880.
+
+(16) Gao, C.; Jiang, Z.; Qi, S.; Wang, P.; Rosgaard Jensen, L.; Johansen, M.; Kolle Christensen, C.; Zhang, Y.; Bomholdt Ravnsbaek, D.; Yue, Y.; Gao, C. W.; Yue, Y. Z.; Jiang, Z. J.; Qi, S. B.; Wang, P. X.; Zhang, Y. F.; Jensen, L. R.; Johansen, M.; Christensen, C. K.; Ravnsbaek, D. B. Metal-Organic Framework Glass Anode with an Exceptional Cycling-Induced Capacity Enhancement for Lithium-Ion Batteries. Adv. Mater. 2022, 34, 2110048.
+
+(17) Hou, J.; Chen, P.; Shukla, A.; Krajnc, A.; Wang, T.; Li, X.; Doasa, R.; Tizei, L. H. G.; Chan, B.; Johnstone, D. N.; Lin, R.; Schülli, T. U.; Martens, I.; Appadoo, D.; Ari, M. S.; Wang, Z.; Wei, T.; Lo, S. C.; Lu, M.; Li, S.; Namdas, E. B.; Mali, G.; Cheetham, A. K.; Collins, S. M.;
+Chen, V.; Wang, L.; Bennett, T. D. Liquid-Phase Sintering of Lead Halide Perovskites and Metal-Organic Framework Glasses. Science. 2021, 374, 621–625.
+
+(18) Yu, Z.; Tang, L.; Ma, N.; Horike, S.; Chen, W. Recent Progress of Amorphous and Glassy Coordination Polymers. Coord. Chem. Rev. 2022, 469, 214646.
+
+(19) Bennett, T. D.; Tan, J. C.; Yue, Y.; Baxter, E.; Ducati, C.; Terrill, N. J.; Yeung, H. H. M.; Zhou, Z.; Chen, W.; Henke, S.; Cheetham, A. K.; Greaves, G. N. Hybrid Glasses from Strong and Fragile Metal-Organic Framework Liquids. Nat Commun. 2015, 6, 8079.
+
+(20) Zheng, Z.; Rong, Z.; Nguyen, H. L.; Yaghi, O. M. Structural Chemistry of Zeolitic Imidazolate Frameworks. Inorg. Chem. 2023, 62, 20861–20873.
+
+(21) Tao, H.; Bennett, T. D.; Yue, Y.; Tao, H. Z.; Yue, Y. Z.; Bennett, T. D. Melt-Quenched Hybrid Glasses from Metal–Organic Frameworks. Adv. Mater. 2017, 29, 1601705.
+
+(22) Gaillac, R.; Pullumbi, P.; Beyer, K. A.; Chapman, K.; Keen, D. A.; Bennett, T. D.; Coudert, F. X. Liquid Metal-Organic Frameworks. Nat. Mater. 2017, 16, 1149–1155.
+
+(23) Bennett, T. D.; Yue, Y.; Li, P.; Qiao, A.; Tao, H.; Greaves, N. G.; Richards, T.; Lampronti, G. I.; Redfern, S. A. T.; Blanc, F.; Farha, O. K.; Hupp, J. T.; Cheetham, A. K.; Keen, D. A. Melt-Quenched Glasses of Metal-Organic Frameworks. J. Am. Chem. Soc. 2016, 138, 3484–3492.
+
+(24) Frentzel-Beyme, L.; Kloß, M.; Kolodzeiski, P.; Pallach, R.; Henke, S. Meltable Mixed-Linker Zeolitic Imidazolate Frameworks and Their Microporous Glasses: From Melting Point Engineering to Selective Hydrocarbon Sorption. J. Am. Chem. Soc. 2019, 141, 12362–12371.
+
+(25) León-Alcaide, L.; Christensen, R. S.; Keen, D. A.; Jordá, J. L.; Brotons-Alcázar, I.; Forment-Aliaga, A.; Mínguez Espallargas, G. Meltable, Glass-Forming, Iron Zeolitic Imidazolate Frameworks. J. Am. Chem. Soc. 2023, 145, 11258–11264.
+
+(26) Hou, J.; Ríos Gómez, M. L.; Krajnc, A.; McCaul, A.; Li, S.; Bumstead, A. M.; Sapnik, A. F.; Deng, Z.; Lin, R.; Chater, P. A.; Keeble, D. S.; Keen, D. A.; Appadoo, D.; Chan, B.; Chen, V.; Mali, G.; Bennett, T. D. Halogenated Metal-Organic Framework Glasses and Liquids. J. Am. Chem. Soc. 2020, 142, 3880–3890.
+
+(27) Song, J.; Frentzel-Beyme, L.; Pallach, R.; Kolodzeiski, P.; Koutsianos, A.; Xue, W. L.; Schmid, R.; Henke, S. Modulating Liquid-Liquid Transitions and Glass Formation in Zeolitic Imidazolate Frameworks by Decoration with Electron-Withdrawing Cyano Groups. J. Am. Chem. Soc. 2023, 145, 9273–9284.
+(28) Frentzel-Beyme, L.; Kloß, M.; Pallach, R.; Salamon, S.; Moldenhauer, H.; Landers, J.; Wende, H.; Debus, J.; Henke, S. Porous Purple Glass – a Cobalt Imidazolate Glass with Accessible Porosity from a Meltable Cobalt Imidazolate Framework. J. Mater. Chem. A. **2019**, 7, 985–990.
+
+(29) Smirnova, O.; Hwang, S.; Sajzew, R.; Ge, L.; Reupert, A.; Nozari, V.; Savani, S.; Chmelik, C.; Reithofer, M. R.; Wondraczek, L.; Kärger, J.; Knebel, A. Precise Control over Gas-Transporting Channels in Zeolitic Imidazolate Framework Glasses. Nat. Mater. **2023**, 23, 262–270.
+
+(30) Rettig, S. J.; Storr, A.; Summers, D. A.; Thompson, R. C.; Trotter, J. Transition Metal Azolates from Metalloenes. 2. Synthesis, X-Ray Structure, and Magnetic Properties of a Three-Dimensional Polymetallic Iron(II) Imidazolate Complex, a Low-Temperature Weak Ferromagnet. J. Am. Chem. Soc. **1997**, *119*, 8675–8680.
+
+(31) López-Cabrelles, J.; Romero, J.; Abellán, G.; Giménez-Marqués, M.; Palomino, M.; Valencia, S.; Rey, F.; Mínguez Espallargas, G. Solvent-Free Synthesis of ZIFs: A Route toward the Elusive Fe(II) Analogue of ZIF-8. J. Am. Chem. Soc. **2019**, *141*, 7173–7180.
+
+(32) López-Cabrelles, J.; Mañas-Valero, S.; Vitórica-Yrezábal, I. J.; Bereciartua, P. J.; Rodríguez-Velamazán, J. A.; Waerenborgh, J. C.; Vieira, B. J. C.; Davidovikj, D.; Steeneken, P. G.; van der Zant, H. S. J.; Espallargas, G. M.; Coronado, E. Isoreticular Two-Dimensional Magnetic Coordination Polymers Prepared through Pre-Synthetic Ligand Functionalization. Nat. Chem. **2018**, *10*, 1001-1007.
+
+(33) Madsen, R. S. K.; Qiao, A.; Sen, J.; Hung, I.; Chen, K.; Gan, Z.; Sen, S.; Yue, Y. Ultrahigh-Field 67Zn NMR Reveals Short-Range Disorder in Zeolitic Imidazolate Framework Glasses. Science. **2020**, *367*, 1473–1476.
+
+(34) Cliffe, M.; Bartók, A.; Kerber, R.; Grey, C.; Csanyi, G.; Goodwin, A. Structural Simplicity as a Restraint on the Structure of Amorphous Silicon. Phys. Rev. B. **2017**, *95*, 224108.
+
+(35) Xie, R.; Long, G. G.; Weigand, S. J.; Moss, S. C.; Carvalho, T.; Roorda, S.; Hejna, M.; Torquato, S.; Steinhardt, P. J. Hyperuniformity in Amorphous Silicon Based on the Measurement of the Infinite-Wavelength Limit of the Structure Factor. Proc. Natl. Acad. Sci. **2013**, *110*, 13250–13254.
+
+(36) Froufe-Pérez, L. S.; Engel, M.; Damasceno, P. F.; Muller, N.; Haberko, J.; Glotzer, S. C.; Scheffold, F. Role of Short-Range Order and Hyperuniformity in the Formation of Band Gaps in Disordered Photonic Materials. Phys. Rev. Lett. **2016**, *117*, 053902.
+(37) Grenche, J. M.; Le Bail, A.; Leblanc, M.; Mosset, A.; Varret, F.; Galy, J.; Ferey, G. Structural Aspects of Amorphous Iron(III) Fluorides. J. Phys. C. **1988**, *21*, 1351.
+
+(38) Coey, J. M. D. Amorphous Magnetic Order. *J. Appl. Phys.* **1978**, *49*, 1646–1652.
+
+(39) Grenche, J. M.; Varret, F.; Leblanc, M.; Ferey, G. Mössbauer Study of an Amorphous Iron III Flouride: FeF3, XHF (0 \( \leq x \leq 1 \)). *Solid State Comm.* **1987**, *63*, 435–438.
+
+(40) López-Cabrelles, J.; Mañas-Valero, S.; Vitórica-Yrezábal, I. J.; Šiškins, M.; Lee, M.; Steenekens, P. G.; Van Der Zant, H. S. J.; Minguez Espallargas, G.; Coronado, E. Chemical Design and Magnetic Ordering in Thin Layers of 2D Metal-Organic Frameworks (MOFs). *J. Am. Chem. Soc.* **2021**, *143*, 18502–18510.
+
+(41) Burbridge, C. D.; Goodgame, D. M. L. Electronic and Mössbauer Spectra of Some Iron(II) Halide Hydrates. *J. Chem. Soc. A*, **1968**, *0*, 1410-1413.
+
+(42) Wooten, F.; Winer, K.; Weaire, D. Computer Generation of Structural Models of Amorphous Si and Ge. *Phys. Rev. Lett.* **1985**, *54*, 1392.
+
+(43) Ferey, G.; Varret, F.; Coey, J. M. D. Amorphous FeF3: A Non-Crystalline Magnet with Antiferromagnetic Interactions. *J. Phys. C.* **1979**, *12*, L531.
+
+(44) Wang, H.; Kim, D. H. Perovskite-Based Photodetectors: Materials and Devices. *Chem. Soc. Rev.* **2017**, *46*, 5204–5236.
+
+(45) Nevruzoglu, V.; Demir, S.; Karaca, G.; Tomakin, M.; Bilgin, N.; Yilmaz, F. Improving the Stability of Solar Cells Using Metal–Organic Frameworks. *J. Mater. Chem. A*. **2016**, *4*, 7930–7935.
+
+(46) Li, X.; Huang, W.; Krajnc, A.; Yang, Y.; Shukla, A.; Lee, J.; Ghasemi, M.; Martens, I.; Chan, B.; Appadoo, D.; Chen, P.; Wen, X.; Steele, J. A.; Hackbarth, H. G.; Sun, Q.; Mali, G.; Lin, R.; Bedford, N. M.; Chen, V.; Cheetham, A. K.; Tizei, L. H. G.; Collins, S. M.; Wang, L.; Hou, J. Interfacial Alloying between Lead Halide Perovskite Crystals and Hybrid Glasses. *Nat. Comm.* **2023**, *14*, 7612.
+
+(47) Tsai, H.; Shrestha, S.; Vilá, R. A.; Huang, W.; Liu, C.; Hou, C. H.; Huang, H. H.; Wen, X.; Li, M.; Wiederrecht, G.; Cui, Y.; Cotlet, M.; Zhang, X.; Ma, X.; Nie, W. Bright and Stable Light-Emitting Diodes Made with Perovskite Nanocrystals Stabilized in Metal–Organic Frameworks. *Nat. Photonics*. **2021**, *15*, 843–849.
+
+(48) Liu, M.; Johnston, M. B.; Snaith, H. J. Efficient Planar Heterojunction Perovskite Solar Cells by Vapour Deposition. *Nature*. **2013**, *501*, 395–398.
+Methods
+
+Synthesis. Ferrocene (0.15 mmol) was combined with a mixture of imidazole (Him) (range: 0.18 - 0.30 mmol) and benzimidazole (Hbim) (range: 0 - 0.12 mmol) or 5-X-benzimidazole (0.07 mmol), maintaining a total ligand amount of 0.30 mmol. The mixture was then sealed under vacuum in a layering tube with a diameter of 4 mm. The mixture was heated at 300 °C for 6 hours, yielding an orange glass. After cooling to room temperature, the layering tube was opened, and the unreacted precursors were extracted with acetonitrile.
+
+X-ray Powder Diffraction. The samples were lightly ground in an agate mortar and pestle and used to fill 0.7 mm borosilicate capillaries that were mounted and aligned on an Brucker D8 Discover powder diffractometer, using Cu Kα radiation (\( \lambda = 1.54056 \) Å). The detector was an EIGER2 R 500K, multi-mode 2D.
+
+Differential Scanning Calorimetry (DSC). Differential scanning calorimetry (DSC) measurements were conducted using a TRIOS DSC 250 instrument. Activated samples (10-15 mg) were placed in an aluminum crucible (30 μL) with a pierced lid. The measurements were performed under a continuous N₂ flow of 50 ml min⁻¹. The sample was initially heated to 40 °C, followed by a 15-minute isotherm to stabilize the environment. Subsequently, the sample was heated at a rate of 10 °C min⁻¹. Upon reaching the final temperature, a 10-minute isotherm was performed to ensure a complete phase change. The sample was then cooled back at 10 °C min⁻¹. It is important to note that DSC results are highly dependent on the experimental conditions. Therefore, when changing the equipment, parameters such as N₂ flow, sample amount, or size of the hole in the crucible lid may need to be adjusted to accurately observe the phase changes.
+
+Thermogravimetric analysis (TGA). Thermogravimetric analysis (TGA) were conducted using a TGA 550 instrument (TA Instruments). The samples were loaded in a platinum pan, under air or N₂ atmosphere. Initially, the sample was heated to 40 °C and held isothermally for 15 minutes to stabilize. Subsequently, the sample was heated at a rate of 10 °C min⁻¹.
+
+Mössbauer Spectroscopy. Mössbauer spectra were collected in transmission mode using a conventional constant-acceleration spectrometer from Wissel (Starnberg, Germany) and a 50 mCi \( ^{57}\mathrm{Co} \) source in a Rh matrix. The velocity scale was calibrated using α-Fe foil. Isomer shifts, IS, are given relative to this standard at room temperature. The absorbers were obtained by gently packing the samples into Perspex holders. Absorber thicknesses were calculated on the basis of the corresponding electronic mass-absorption coefficients for the 14.4 keV radiation.\(^{50}\) The low
+temperature measurements at 4 K and above were performed with the sample in liquid He and in He exchange gas, respectively, in a Janis (Westerville, OH, USA) bath cryostat, model SVT-400. The spectra were fitted to Lorentzian lines using a non-linear least-squares method.\(^{51}\) Line widths and relative areas of both peaks in each doublet were kept equal during the refinement. Distributions of quadrupole splittings were fitted according to the histogram method.\(^{52}\)
+
+**X-ray Total Scattering.** X-ray total scattering data were collected at room temperature using a PANalytical Empyrean laboratory diffractometer, equipped with an Ag-K\(_\alpha\) source and focusing mirrors. The samples were placed in 1 mm diameter quartz glass capillaries. Each sample underwent multiple scans, with a cumulative data collection time exceeding 24 hours per sample. Additionally, measurements were conducted on an empty capillary and the diffractometer background. The resulting X-ray total scattering patterns were processed using the GudrunX\(^{53}\) program to produce a normalized Pair Distribution Function (PDF), optimized to ensure, for example, that the low-r portion of \(g(r)\) oscillates around –1. The total scattering structure factor, S(Q), was Fourier transformed for \(0.6 < Q < 18.5\) Å\(^{-1}\) to obtain the PDF.
+
+**Neutron Total Scattering.** X-ray total scattering data were collected at temperatures of 5 K (base), 10 K, 15 K (just above the transition), 30 K, 60 K, and 120 K. Each temperature required a collection time of at least 3 hours per sample. The sample was measured using the ILL D20 instrument at a wavelength of 2.41 Å. The sample was predominantly deuterated, achieved by using deuterated imidazole (Dim) and deuterated ferrocene, with a minor hydrogen presence from benzimidazole (Hbim).
+
+**Magnetic Susceptibility (SQUID).** Magnetic susceptibility measurements were conducted on powdered samples (5–20 mg) using a SQUID magnetometer (Quantum Design MPMS-XL-5 and MPMS-XL-7). Variable-temperature (2–300 K) direct current (d.c.) magnetic susceptibility measurements were performed in applied fields of 1.0 kOe, along with variable field magnetization measurements up to ±5 T at 2.0 K. Additionally, variable-temperature (2–10 K) alternating current (a.c.) magnetic susceptibility measurements were conducted in a ±4.0 G oscillating field at frequencies ranging from 1 to 996 Hz, under a zero d.c. field.
+
+**Photodetector fabrication:** CH\(_3\)NH\(_3\)I (MAI) and PbI\(_2\) were purchased from Luminescence Technology Corp. All materials were used as received. For MAPbI\(_3\) deposition, MAI and PbI\(_2\) precursors were simultaneously evaporated following a procedure recently published by some of us: In summary, the evaporation chamber employed has only two quartz crystal microbalances (QCMs), one close to the PbI\(_2\) source designated to monitor exclusively the PbI\(_2\) precursor (PbI\(_2\)-QCM), with no cross reading, and a second one fixed at the height of the substrates to monitor simultaneously the total amount of PbI\(_2\) and MAI mass reaching the substrates (MAPbI\(_3\)-QCM). Initially, only PbI\(_2\) is heated, and the temperature is fine tuned to lead to a stable sublimation rate
+of precisely 0.50 Å s\(^{-1}\) on both PbI\(_2\)-QCM and MAPbI\(_3\)-QCM. Then, MAI is heated, and the temperature adjusted so that the sublimation detected on the MAPbI\(_3\)-QCM increases from the previous 0.50 Å s\(^{-1}\) to 0.66 Å s\(^{-1}\), while the rate on the PbI\(_2\)-QCM is kept stable at 0.50 Å s\(^{-1}\). During the evaporation the pressure of the chamber was maintained at 5·10\(^{-6}\) mbar. Gold contact were evaporated in another vacuum chamber through a shadow mask, defining lateral electrodes with channel length of 500 μm and width of 500 mm. The devices were encapsulated with Al\(_2\)O\(_3\) (30 nm), deposited by atomic layer deposition in a Arradiance reactor at 40 °C.\(^{49}\)
+
+Data availability
+
+All data generated and analyzed during this study are included in this Article and its Supplementary Information and are also available from the authors upon reasonable request. Atomic coordinates and structure factors for the reported crystal structure has been deposited in the Cambridge Crystallographic Data Centre under the accession code CCDC 2416331 (MUV-28). Copies of the data can be obtained free of charge from www.ccdc.cam.ac.uk/structures/.
+
+References
+
+(50) Long, G. J.; Cranshaw, T. E.; Longworth, G. The Ideal Mössbauer Effect Absorber Thicknesses. Mossb. Effect. Ref. Data J. **1983**, 42–49.
+
+(51) Waerenborgh, J. C.; Salamakha, P.; Sologub, O.; Gonçalves, A. P.; Cardoso, C.; Sério, S.; Godinho, M.; Almeida, M. Influence of Thermal Treatment and Crystal Growth on the Final Composition and Magnetic Properties of the YFE(x)Al(12-x) (4 ≤ x ≤ 4.2) Intermetallics. *Chem. Mater* **2000**, *12*, 1743–1749.
+
+(52) Hesse, J.; Rubartsch, A. Model Independent Evaluation of Overlapped Mossbauer Spectra. *J. Phys. E.* **1974**, *7*, 526.
+
+(53) Soper, A. GudrunN and GudrunX : Programs for Correcting Raw Neutron and X-Ray Diffraction Data to Differential Scattering Cross Section. *Rutherford Appleton Lab. Tech. Rep.* **2011**.
+
+Acknowledgements
+
+This work has been supported by the European Union (ERC-2016-CoG 724681-S-CAGE), grants PID2023-152920NB-I00, and María de Maeztu Centre of Excellence Programmes CEX2019-000919-M, funded by MCIN/AEI/10.13039/501100011033 and cofinanced by FEDER, and the Generalitat Valenciana (CIPROM/2022/48, and IDIFEDER2021/075). L.L.-A. thanks MICIN for a pre-doctoral fellowship (PRE2019-089295). This study forms part of the Advanced Materials program and was supported by MCIN with funding from European Union NextGenerationEU (PRTR-C17.11) and by Generalitat Valenciana (projects MFA/2022/31). We also thank the
+University of Valencia for research facilities (SCSIE). M.J.C. acknowledges UKRI funding (EP/X042782/1). J.C.W. and B.J.C.V. acknowledge FCT (Portugal) through contracts UID/Multi/04349/2019 and PTDC/QUI-QIN/32240/2017. M. S acknowledges financial support from the Comunitat Valenciana (CISEJI/2022/43). We acknowledge the Institut Laue-Langevin for access to beamline D20.
+
+Author contributions
+
+L.L.A. synthesized and characterized all the materials. L.M.G. and M.S. prepared and characterized the photodetector. B.J.C.V. and J.C.W. conducted and analyzed the Mössbauer measurements. J.A.R.V. and O.F. assisted with the performance and analysis of the neutron measurements. M.C. analyzed the magnetic behavior of the materials and performed the theoretical calculations. L.L.A. carried out the total scattering experiments under the supervision of D.A.K. G.M.E. conceived the idea and designed the experiments. L.L.A. and G.M.E. prepared the manuscript. All authors contributed to the preparation of the manuscript.
+
+Competing interest
+
+The authors declare no competing interests.
+
+Additional information
+
+Materials & Correspondence
+
+Correspondence and requests for materials should be addressed to G.M.E.
+Supplementary Files
+
+This is a list of supplementary files associated with this preprint. Click to download.
+
+• MUV28.cif
+• MinguezMUV29SI.pdf