Mardiyyah/TAPT_data_V2_split · Datasets at Hugging Face

PMCID stringclasses 24 values	Title stringclasses 24 values	Sentences stringlengths 2 40.7k
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	We used Drug2Cell (D2C) to score the expression levels of 2,395 drug target signatures from the CHEMBL database in single cells from HEOCA, and used scDECAF to select drug target signatures that exhibited global covariation in two or more cell types to identify multicellular drug signatures (Extended Data Fig. 10a and Supplementary Table 6).
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	Comparison between ASC-, FSC- and PSC-derived intestine and lung organoid models showed substantial differences in drug target pathway activities (Extended Data Fig. 10b,c).
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	Druggable targets in categories including alimentary tract metabolism, systemic hormones, anti-infectives, antiparasitics, and antineoplastic and immunomodulating agents were implicated in signatures that varied between cell types and stem cell sources.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	Comparison between lung and intestine organoid models across all cell types suggested that many druggable targets are distinct between the two tissues and cell types common to both intestine and lung (for example, stem and goblet cells) have unique features (Extended Data Fig. 10d–f).
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	In summary, our analyses demonstrate that HEOCA is a technically and biologically diverse cohort that can be leveraged to evaluate organoid models, identify pathways impacted by perturbations, and, more broadly, explore the ontogeny of human biology.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	Single-cell transcriptome sequencing technologies have advanced organoid research by offering a powerful set of experimental and computational tools to investigate cell types present in these complex 3D models.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	Despite immense progress, it remains a challenge to understand and quantify organoid fidelity and to place variation between organoid datasets into a larger context.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	To begin to address these challenges, we have built an integrated cell atlas of organoids that model endoderm-derived tissues, incorporating organoid datasets that have been generated from multiple different types of stem cells and protocols.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	We have established a framework for integration and harmonized cell-type annotation, which makes interpreting cell heterogeneity between organoid datasets tractable.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	Harmonization of cell-type annotation and nomenclature is challenging, and we envision that comprehensive integrated reference atlases across the human lifespan will enhance the robustness of cell annotation in organoid datasets.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	With regard to atlas building, single-cell transcriptome data from diverse experimental designs can introduce strong technical noise because of batch effects, protocol variation, genomic method and other technical biases, making data integration challenging.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	To overcome this, we evaluated existing integration methods and identified a suitable model based on bioconservation and integration metrics.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	This integration method, scPoli, is structured to incorporate additional data, enabling rapid comparison of datasets through the sc2heoca package or ArchMap website (https://www.archmap.bio).
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	We find that there is notable variation in organoid cell composition, prevalence of off-target cells and overall cell state similarity.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	This variation, and comparison with available reference atlases, revealed that current organoid technologies cover a large diversity of human cell types and states, and particularly that organoids can model both early stages of fetal development as well as stages of adulthood.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	This result helps to clarify the use of human organoid technologies to explore development, model disease and test therapeutics.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	Through cross-organ, multiorganoid integration, it was possible to identify off-target cells, a particular problem in PSC-derived organoids because of incomplete specification, as well as to distinguish cell states that markedly differed from states present in the atlas.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	This ability to distinguish nonpresent states is helpful to assess protocols, as well as to identify features of disease models that are absent in normal, healthy organoids.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	Indeed, the integrated HEOCA presents an opportunity to place an endodermal organoid dataset into a relationship with datasets generated from a technically and biologically diverse cohort.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	There is still a major challenge with organoid fidelity quantification, particularly with rare cell types or transient ontogenetic states, because there is not yet a complete and integrated atlas of human cell-type diversity during development and adulthood from primary tissues.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	Comprehensive integrated reference atlases across the human lifespan, in health and disease, together with diverse organoid models in normal and perturbed conditions, will help to clarify the full potential of the human genome.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	Altogether, the HEOCA will serve as a valuable resource for the organoid research community and a foundation to expand the ability to model human biology.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	The experiments conducted in this study did not require approval from a specific ethics board.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	To integrate the atlas, all available datasets were included, with no sample exclusion.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	For integration method comparisons and sample variance effect analyses, random samples were selected from the full dataset.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	Reproducibility codes for the analyses are available online via GitHub, as detailed in the ‘Code availability’ section.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	All statistical methods are described in the corresponding sections of the paper.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	Human intestinal tissue samples were obtained and experimental procedures performed within the framework of the nonprofit foundation HTCR (Munich, Germany) including informed patient consent.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	Ileal organoids were derived and maintained according to previously published culture conditions.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	For the cytokine treatments, organoids were dissociated into five- to ten-cell fragments using TrypLE (Invitrogen) and reseeded in Matrigel.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	After 6 days, organoids were treated for 6 days by supplementing the culture medium with 50 ng ml TNF and 200 ng ml RANKL (Acro Biosystems) for TNF treatment or with 400 ng ml IL-4 and 40 ng ml IL-13 (Acro Biosystems) for IL-13 and IL-4 treatment.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	To model host cell responses to viral infection, organoids were treated for 1 day with 1 ng ml IFNα, 1 ng ml IFNγ (Acro Biosystems) and 5 ng ml IFNβ (PeproTech).
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	To model acute pathogenic inflammation, organoids were treated for 1 day with 10 ng ml TNF, 10 ng ml IL-6, 500 ng ml IL-17A, 1 ng ml IFNγ (Acro Biosystems), 50 ng ml IL-18, 100 ng ml OSM (BioLegend) and 100 ng ml SCF (MedChemExpress).
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	After the indicated treatment durations, organoids were dissociated for scRNA-seq using the Neural Tissue Dissociation Kit (P) (Miltenyi Biotec) as described previously.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	First, culture medium was removed and organoids were incubated in Cell Recovery Solution (Corning) for 40 min at 4 °C.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	Next, organoids were transferred to 1% bovine serum albumin (BSA)-coated tubes using HBSS–1% BSA buffer while pipetting thoroughly to fragmentize the organoids.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	Organoid fragments were centrifuged at 500g, 5 min, 4 °C.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	Each cell pellet was resuspended in prewarmed buffer X mixed with 25 μl of enzyme P. Cells were incubated for 15 min at 37 °C combined with mechanical dissociation by pipetting every 5 min.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	Next, 5 μl of enzyme A in 10 μl of buffer Y was added to the digest and incubated for a further 10 min combined with pipetting every 5 min.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	Cells were subsequently washed twice with HBSS–1% BSA buffer and filtered through a 40-μm filter coated with 1% BSA.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	Single cells were counted using a Countess 3 FL Automated Cell Counter (Invitrogen) and kept on ice.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	Dilutions of ~1,000 cells per μl in 50–60 μl of HBSS–1% BSA buffer were prepared and immediately processed using the 10x Chromium Next GEM Single Cell 3′ Reagent Kit (v.3.1) according to the manufacturer’s instructions.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	Libraries were sequenced on Illumina’s NovaSeq6000.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	The scRNA-seq data used in this study were obtained from the original papers (Supplementary Table 1).
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	If the raw fastq files were available, they were downloaded.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	The seq2science (v.1.2.2) method was used to download the raw fastq files from the Gene Expression Omnibus database (https://www.ncbi.nlm.nih.gov/geo/) or BioStudies database (https://www.ebi.ac.uk/biostudies/).
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	The reads were aligned to the GRCh38 genome and Ensembl 98 gene annotation using STARsolo (STAR v.2.7.10b).
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	In cases in which the raw FASTQ files were not available, the raw counts were downloaded instead.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	The downloaded counts and the counts obtained from the realigned reads were merged for subsequent analysis.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	To integrate the data, we combined the count data from all the samples into a unified dataset.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	For subsequent analysis, we retained only the genes classified as protein-coding genes and long noncoding RNA genes.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	The low-quality cells in each sample were filed.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	The raw counts were then normalized to a total count of 10,000 and log-transformed.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	Given these normalized counts, the top 3,000 highly variable genes were identified using the default settings in Scanpy.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	These highly variable genes were selected for further downstream analysis.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	Cell-type annotation was performed using the snapseed method (https://github.com/devsystemslab/snapseed).
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	For each sample, the raw counts were normalized to a total count of 10,000 and then log-transformed.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	From these normalized counts, the top 3,000 highly variable genes were identified using the default settings in Scanpy.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	These highly variable genes were selected as the subset for further downstream analysis.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	Principal component analysis (PCA) was performed on the normalized data, and the top 30 principal components were chosen for calculating the k nearest neighbors (kNN).
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	Using the kNN, a UMAP was generated to visualize the data in a lower-dimensional space.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	To cluster the data, the Leiden clustering method with a resolution of 2 was applied.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	This clustering approach helped to identify distinct groups of cells based on their gene expression patterns.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	Previously defined marker genes associated with specific cell types were used to guide the annotation process.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	To annotate cell types in each cluster, the snapseed method was used.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	This method calculates the area under the receiver operating characteristic (ROC) curve (AUC) and fold change values for each marker gene in relation to the cluster.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	If multiple markers were available for a particular cell type, the maximum AUC and fold change values were selected.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	The average AUC and fold change values were used to represent the specific cell type, and the most specific cell type was annotated for each cluster based on these criteria.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	For gene expression level, we merged all counts in each organoid sample by genes using the adpbulk method and applied a natural logarithm transformation to one plus the counts.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	We then selected the top 500 highly variable genes and calculated PCA based on their expression.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	From principal components 1 and 2 (PC1 and PC2), we selected the top 200 and bottom 200 loading genes for GO enrichment analysis using the GSEApy method.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	For the scPoli embedding level, we calculated the mean scPoli embedding for each organoid sample using the adpbulk method, followed by PCA based on the mean embedding.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	A linear model was then used to calculate the covariance between principal components and sample counts, stem cell source, scRNA-seq method, tissue type and publication.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	To benchmark and compare different integration methods, we selected ten random samples from the dataset for validation, repeating this process ten times.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	Twelve integration methods, including PCA, Seurat (v.3, v.4 and v.5), scVI, scANVI, scPoli, bbknn, harmony, combat, CSS (pearson) and CSS (spearman) were applied to the data to assess their performance in integrating the samples.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	The scIB method, a benchmarking tool, was used to evaluate and compare the results obtained from these integration methods.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	In the scPoli model, we configured the following parameters for effective training and integration: embedding_dim was set to 3; hidden_layer_sizes were determined as the square root of the total number of cells.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	During the training phase, we used the following settings: early_stopping_metric was set to val_prototype_loss; mode was set to min; threshold was set to 0; patience was set to 20; reduce_lr was enabled, with lr_patience set to 13 and lr_factor set to 0.1; n_epochs were set to 5; pretraining_epochs were set to 4; eta was set to 10; alpha_epoch_anneal was set to 100.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	To benchmark and compare how different sample variances affect integration, we selected ten random samples from the dataset as a control and another ten random samples with the same sample variance, such as samples from the same organoid tissue type.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	Each group of selected samples was integrated using the scPoli method with the same settings as in the HEOCA integration.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	The scIB method was then used to benchmark the different integrations.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	The difference in scIB output between the control and each sample variance pair was calculated to represent the effect of sample variance on integration.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	For benchmarking the effect of cell number variance, we selected the same ten samples as the control and performed a random subset of each sample to the median or mean number of cells in all the HEOCA samples.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	The subsequent comparison followed the same procedure as the other sample variance benchmarks.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	After integration, we recluster all cells in the atlas based on the scPoli integrated embedding using the Leiden method with a resolution of 10 (HEOCA and HIOCA) and 10 (HLOCA), respectively.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	Annotations were then assigned to each cluster using the dominant cell type per cluster.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	Some clusters of cells were adjusted according to the marker genes expression.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	We randomly subset 100,000 cells from the atlas.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	For each cell type, we used the Wilcoxon rank-sum test to identify DEGs, selecting the top ten genes as marker genes for each cell type.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	We combined the selected marker genes and performed hierarchical clustering on the resulting gene set.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	The human fetal endoderm tissue atlas was downloaded.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	The normal endoderm tissues including the esophagus, lung, liver, intestine, stomach and pancreas were subsetted.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	The top 3,000 highly variable genes were subsetted for data integration.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	The cells in each tissue were integrated using the scPoli method, with the cell_type serving as the cell-type key for integration and with the same parameters used in the HEOCA integration.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	The scPoli model was saved for the downstream comparison.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	The Tabula Sapiens multiple-organ adult single-cell transcriptomic atlas of humans was downloaded (https://tabula-sapiens-portal.ds.czbiohub.org/).
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	The endoderm tissues including the liver, lung, pancreas, small intestine, large intestine, prostate and stomach were subsetted.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	The endothelial, epithelial and stromal compartments of cells were subsetted.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	The top 3,000 highly variable genes were subsetted for data integration.
PMC12081310	An integrated transcriptomic cell atlas of human endoderm-derived organoids.	The cells in each tissue were integrated using the scPoli method, with the cell_ontology_class serving as the cell-type key for integration and with the same parameters used in the HEOCA atlas integration.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

We used Drug2Cell (D2C) to score the expression levels of 2,395 drug target signatures from the CHEMBL database in single cells from HEOCA, and used scDECAF to select drug target signatures that exhibited global covariation in two or more cell types to identify multicellular drug signatures (Extended Data Fig. 10a and Supplementary Table 6).

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

Comparison between ASC-, FSC- and PSC-derived intestine and lung organoid models showed substantial differences in drug target pathway activities (Extended Data Fig. 10b,c).

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

Druggable targets in categories including alimentary tract metabolism, systemic hormones, anti-infectives, antiparasitics, and antineoplastic and immunomodulating agents were implicated in signatures that varied between cell types and stem cell sources.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

Comparison between lung and intestine organoid models across all cell types suggested that many druggable targets are distinct between the two tissues and cell types common to both intestine and lung (for example, stem and goblet cells) have unique features (Extended Data Fig. 10d–f).

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

In summary, our analyses demonstrate that HEOCA is a technically and biologically diverse cohort that can be leveraged to evaluate organoid models, identify pathways impacted by perturbations, and, more broadly, explore the ontogeny of human biology.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

Single-cell transcriptome sequencing technologies have advanced organoid research by offering a powerful set of experimental and computational tools to investigate cell types present in these complex 3D models.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

Despite immense progress, it remains a challenge to understand and quantify organoid fidelity and to place variation between organoid datasets into a larger context.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

To begin to address these challenges, we have built an integrated cell atlas of organoids that model endoderm-derived tissues, incorporating organoid datasets that have been generated from multiple different types of stem cells and protocols.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

We have established a framework for integration and harmonized cell-type annotation, which makes interpreting cell heterogeneity between organoid datasets tractable.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

Harmonization of cell-type annotation and nomenclature is challenging, and we envision that comprehensive integrated reference atlases across the human lifespan will enhance the robustness of cell annotation in organoid datasets.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

With regard to atlas building, single-cell transcriptome data from diverse experimental designs can introduce strong technical noise because of batch effects, protocol variation, genomic method and other technical biases, making data integration challenging.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

To overcome this, we evaluated existing integration methods and identified a suitable model based on bioconservation and integration metrics.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

This integration method, scPoli, is structured to incorporate additional data, enabling rapid comparison of datasets through the sc2heoca package or ArchMap website (https://www.archmap.bio).

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

We find that there is notable variation in organoid cell composition, prevalence of off-target cells and overall cell state similarity.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

This variation, and comparison with available reference atlases, revealed that current organoid technologies cover a large diversity of human cell types and states, and particularly that organoids can model both early stages of fetal development as well as stages of adulthood.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

This result helps to clarify the use of human organoid technologies to explore development, model disease and test therapeutics.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

Through cross-organ, multiorganoid integration, it was possible to identify off-target cells, a particular problem in PSC-derived organoids because of incomplete specification, as well as to distinguish cell states that markedly differed from states present in the atlas.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

This ability to distinguish nonpresent states is helpful to assess protocols, as well as to identify features of disease models that are absent in normal, healthy organoids.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

Indeed, the integrated HEOCA presents an opportunity to place an endodermal organoid dataset into a relationship with datasets generated from a technically and biologically diverse cohort.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

There is still a major challenge with organoid fidelity quantification, particularly with rare cell types or transient ontogenetic states, because there is not yet a complete and integrated atlas of human cell-type diversity during development and adulthood from primary tissues.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

Comprehensive integrated reference atlases across the human lifespan, in health and disease, together with diverse organoid models in normal and perturbed conditions, will help to clarify the full potential of the human genome.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

Altogether, the HEOCA will serve as a valuable resource for the organoid research community and a foundation to expand the ability to model human biology.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

The experiments conducted in this study did not require approval from a specific ethics board.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

To integrate the atlas, all available datasets were included, with no sample exclusion.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

For integration method comparisons and sample variance effect analyses, random samples were selected from the full dataset.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

Reproducibility codes for the analyses are available online via GitHub, as detailed in the ‘Code availability’ section.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

All statistical methods are described in the corresponding sections of the paper.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

Human intestinal tissue samples were obtained and experimental procedures performed within the framework of the nonprofit foundation HTCR (Munich, Germany) including informed patient consent.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

Ileal organoids were derived and maintained according to previously published culture conditions.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

For the cytokine treatments, organoids were dissociated into five- to ten-cell fragments using TrypLE (Invitrogen) and reseeded in Matrigel.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

After 6 days, organoids were treated for 6 days by supplementing the culture medium with 50 ng ml TNF and 200 ng ml RANKL (Acro Biosystems) for TNF treatment or with 400 ng ml IL-4 and 40 ng ml IL-13 (Acro Biosystems) for IL-13 and IL-4 treatment.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

To model host cell responses to viral infection, organoids were treated for 1 day with 1 ng ml IFNα, 1 ng ml IFNγ (Acro Biosystems) and 5 ng ml IFNβ (PeproTech).

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

To model acute pathogenic inflammation, organoids were treated for 1 day with 10 ng ml TNF, 10 ng ml IL-6, 500 ng ml IL-17A, 1 ng ml IFNγ (Acro Biosystems), 50 ng ml IL-18, 100 ng ml OSM (BioLegend) and 100 ng ml SCF (MedChemExpress).

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

After the indicated treatment durations, organoids were dissociated for scRNA-seq using the Neural Tissue Dissociation Kit (P) (Miltenyi Biotec) as described previously.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

First, culture medium was removed and organoids were incubated in Cell Recovery Solution (Corning) for 40 min at 4 °C.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

Next, organoids were transferred to 1% bovine serum albumin (BSA)-coated tubes using HBSS–1% BSA buffer while pipetting thoroughly to fragmentize the organoids.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

Organoid fragments were centrifuged at 500g, 5 min, 4 °C.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

Each cell pellet was resuspended in prewarmed buffer X mixed with 25 μl of enzyme P. Cells were incubated for 15 min at 37 °C combined with mechanical dissociation by pipetting every 5 min.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

Next, 5 μl of enzyme A in 10 μl of buffer Y was added to the digest and incubated for a further 10 min combined with pipetting every 5 min.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

Cells were subsequently washed twice with HBSS–1% BSA buffer and filtered through a 40-μm filter coated with 1% BSA.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

Single cells were counted using a Countess 3 FL Automated Cell Counter (Invitrogen) and kept on ice.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

Dilutions of ~1,000 cells per μl in 50–60 μl of HBSS–1% BSA buffer were prepared and immediately processed using the 10x Chromium Next GEM Single Cell 3′ Reagent Kit (v.3.1) according to the manufacturer’s instructions.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

Libraries were sequenced on Illumina’s NovaSeq6000.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

The scRNA-seq data used in this study were obtained from the original papers (Supplementary Table 1).

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

If the raw fastq files were available, they were downloaded.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

The seq2science (v.1.2.2) method was used to download the raw fastq files from the Gene Expression Omnibus database (https://www.ncbi.nlm.nih.gov/geo/) or BioStudies database (https://www.ebi.ac.uk/biostudies/).

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

The reads were aligned to the GRCh38 genome and Ensembl 98 gene annotation using STARsolo (STAR v.2.7.10b).

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

In cases in which the raw FASTQ files were not available, the raw counts were downloaded instead.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

The downloaded counts and the counts obtained from the realigned reads were merged for subsequent analysis.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

To integrate the data, we combined the count data from all the samples into a unified dataset.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

For subsequent analysis, we retained only the genes classified as protein-coding genes and long noncoding RNA genes.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

The low-quality cells in each sample were filed.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

The raw counts were then normalized to a total count of 10,000 and log-transformed.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

Given these normalized counts, the top 3,000 highly variable genes were identified using the default settings in Scanpy.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

These highly variable genes were selected for further downstream analysis.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

Cell-type annotation was performed using the snapseed method (https://github.com/devsystemslab/snapseed).

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

For each sample, the raw counts were normalized to a total count of 10,000 and then log-transformed.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

From these normalized counts, the top 3,000 highly variable genes were identified using the default settings in Scanpy.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

These highly variable genes were selected as the subset for further downstream analysis.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

Principal component analysis (PCA) was performed on the normalized data, and the top 30 principal components were chosen for calculating the k nearest neighbors (kNN).

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

Using the kNN, a UMAP was generated to visualize the data in a lower-dimensional space.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

To cluster the data, the Leiden clustering method with a resolution of 2 was applied.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

This clustering approach helped to identify distinct groups of cells based on their gene expression patterns.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

Previously defined marker genes associated with specific cell types were used to guide the annotation process.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

To annotate cell types in each cluster, the snapseed method was used.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

This method calculates the area under the receiver operating characteristic (ROC) curve (AUC) and fold change values for each marker gene in relation to the cluster.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

If multiple markers were available for a particular cell type, the maximum AUC and fold change values were selected.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

The average AUC and fold change values were used to represent the specific cell type, and the most specific cell type was annotated for each cluster based on these criteria.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

For gene expression level, we merged all counts in each organoid sample by genes using the adpbulk method and applied a natural logarithm transformation to one plus the counts.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

We then selected the top 500 highly variable genes and calculated PCA based on their expression.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

From principal components 1 and 2 (PC1 and PC2), we selected the top 200 and bottom 200 loading genes for GO enrichment analysis using the GSEApy method.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

For the scPoli embedding level, we calculated the mean scPoli embedding for each organoid sample using the adpbulk method, followed by PCA based on the mean embedding.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

A linear model was then used to calculate the covariance between principal components and sample counts, stem cell source, scRNA-seq method, tissue type and publication.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

To benchmark and compare different integration methods, we selected ten random samples from the dataset for validation, repeating this process ten times.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

Twelve integration methods, including PCA, Seurat (v.3, v.4 and v.5), scVI, scANVI, scPoli, bbknn, harmony, combat, CSS (pearson) and CSS (spearman) were applied to the data to assess their performance in integrating the samples.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

The scIB method, a benchmarking tool, was used to evaluate and compare the results obtained from these integration methods.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

In the scPoli model, we configured the following parameters for effective training and integration: embedding_dim was set to 3; hidden_layer_sizes were determined as the square root of the total number of cells.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

During the training phase, we used the following settings: early_stopping_metric was set to val_prototype_loss; mode was set to min; threshold was set to 0; patience was set to 20; reduce_lr was enabled, with lr_patience set to 13 and lr_factor set to 0.1; n_epochs were set to 5; pretraining_epochs were set to 4; eta was set to 10; alpha_epoch_anneal was set to 100.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

To benchmark and compare how different sample variances affect integration, we selected ten random samples from the dataset as a control and another ten random samples with the same sample variance, such as samples from the same organoid tissue type.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

Each group of selected samples was integrated using the scPoli method with the same settings as in the HEOCA integration.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

The scIB method was then used to benchmark the different integrations.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

The difference in scIB output between the control and each sample variance pair was calculated to represent the effect of sample variance on integration.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

For benchmarking the effect of cell number variance, we selected the same ten samples as the control and performed a random subset of each sample to the median or mean number of cells in all the HEOCA samples.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

The subsequent comparison followed the same procedure as the other sample variance benchmarks.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

After integration, we recluster all cells in the atlas based on the scPoli integrated embedding using the Leiden method with a resolution of 10 (HEOCA and HIOCA) and 10 (HLOCA), respectively.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

Annotations were then assigned to each cluster using the dominant cell type per cluster.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

Some clusters of cells were adjusted according to the marker genes expression.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

We randomly subset 100,000 cells from the atlas.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

For each cell type, we used the Wilcoxon rank-sum test to identify DEGs, selecting the top ten genes as marker genes for each cell type.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

We combined the selected marker genes and performed hierarchical clustering on the resulting gene set.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

The human fetal endoderm tissue atlas was downloaded.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

The normal endoderm tissues including the esophagus, lung, liver, intestine, stomach and pancreas were subsetted.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

The top 3,000 highly variable genes were subsetted for data integration.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

The cells in each tissue were integrated using the scPoli method, with the cell_type serving as the cell-type key for integration and with the same parameters used in the HEOCA integration.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

The scPoli model was saved for the downstream comparison.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

The Tabula Sapiens multiple-organ adult single-cell transcriptomic atlas of humans was downloaded (https://tabula-sapiens-portal.ds.czbiohub.org/).

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

The endoderm tissues including the liver, lung, pancreas, small intestine, large intestine, prostate and stomach were subsetted.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

The endothelial, epithelial and stromal compartments of cells were subsetted.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

The top 3,000 highly variable genes were subsetted for data integration.

PMC12081310

An integrated transcriptomic cell atlas of human endoderm-derived organoids.

The cells in each tissue were integrated using the scPoli method, with the cell_ontology_class serving as the cell-type key for integration and with the same parameters used in the HEOCA atlas integration.