update_datasets_splits

.gitattributes

@@ -1,62 +1,3 @@
-*.7z filter=lfs diff=lfs merge=lfs -text
-*.arrow filter=lfs diff=lfs merge=lfs -text
-*.bin filter=lfs diff=lfs merge=lfs -text
-*.bz2 filter=lfs diff=lfs merge=lfs -text
-*.ckpt filter=lfs diff=lfs merge=lfs -text
-*.ftz filter=lfs diff=lfs merge=lfs -text
-*.gz filter=lfs diff=lfs merge=lfs -text
-*.h5 filter=lfs diff=lfs merge=lfs -text
-*.joblib filter=lfs diff=lfs merge=lfs -text
-*.lfs.* filter=lfs diff=lfs merge=lfs -text
-*.lz4 filter=lfs diff=lfs merge=lfs -text
-*.mds filter=lfs diff=lfs merge=lfs -text
-*.mlmodel filter=lfs diff=lfs merge=lfs -text
-*.model filter=lfs diff=lfs merge=lfs -text
-*.msgpack filter=lfs diff=lfs merge=lfs -text
-*.npy filter=lfs diff=lfs merge=lfs -text
-*.npz filter=lfs diff=lfs merge=lfs -text
-*.onnx filter=lfs diff=lfs merge=lfs -text
-*.ot filter=lfs diff=lfs merge=lfs -text
-*.parquet filter=lfs diff=lfs merge=lfs -text
-*.pb filter=lfs diff=lfs merge=lfs -text
-*.pickle filter=lfs diff=lfs merge=lfs -text
-*.pkl filter=lfs diff=lfs merge=lfs -text
-*.pt filter=lfs diff=lfs merge=lfs -text
-*.pth filter=lfs diff=lfs merge=lfs -text
-*.rar filter=lfs diff=lfs merge=lfs -text
-*.safetensors filter=lfs diff=lfs merge=lfs -text
-saved_model/**/* filter=lfs diff=lfs merge=lfs -text
-*.tar.* filter=lfs diff=lfs merge=lfs -text
-*.tar filter=lfs diff=lfs merge=lfs -text
-*.tflite filter=lfs diff=lfs merge=lfs -text
-*.tgz filter=lfs diff=lfs merge=lfs -text
-*.wasm filter=lfs diff=lfs merge=lfs -text
-*.xz filter=lfs diff=lfs merge=lfs -text
-*.zip filter=lfs diff=lfs merge=lfs -text
-*.zst filter=lfs diff=lfs merge=lfs -text
-*tfevents* filter=lfs diff=lfs merge=lfs -text
-# Audio files - uncompressed
-*.pcm filter=lfs diff=lfs merge=lfs -text
-*.sam filter=lfs diff=lfs merge=lfs -text
-*.raw filter=lfs diff=lfs merge=lfs -text
-# Audio files - compressed
-*.aac filter=lfs diff=lfs merge=lfs -text
-*.flac filter=lfs diff=lfs merge=lfs -text
-*.mp3 filter=lfs diff=lfs merge=lfs -text
-*.ogg filter=lfs diff=lfs merge=lfs -text
-*.wav filter=lfs diff=lfs merge=lfs -text
-# Image files - uncompressed
-*.bmp filter=lfs diff=lfs merge=lfs -text
-*.gif filter=lfs diff=lfs merge=lfs -text
-*.png filter=lfs diff=lfs merge=lfs -text
-*.tiff filter=lfs diff=lfs merge=lfs -text
-# Image files - compressed
-*.jpg filter=lfs diff=lfs merge=lfs -text
-*.jpeg filter=lfs diff=lfs merge=lfs -text
-*.webp filter=lfs diff=lfs merge=lfs -text
-# Video files - compressed
-*.mp4 filter=lfs diff=lfs merge=lfs -text
-*.webm filter=lfs diff=lfs merge=lfs -text
-*.h5ad.gz filter=lfs diff=lfs merge=lfs -text
-*.h5ad filter=lfs diff=lfs merge=lfs -text
-norman19_cpa_hvg_normalized_curated.h5ad.gz filter=lfs diff=lfs merge=lfs -text
+version https://git-lfs.github.com/spec/v1
+oid sha256:694981c975d0061427a5983a2a1ce262963a3f2b16a2bc09afcfe25459734bd2
+size 2634

README.md

@@ -1,22 +1,3 @@
----
-license: cc-by-4.0
----
-
-The dataset contains data used in work:
-"Perturbench: Benchmarking machine learning models for cellular perturbation analysis."
-
-The data comes from the following publications:
-
--  Norman, T. M., Horlbeck, M. A., Replogle, J. M., Ge, A. Y., Xu, A., Jost, M., Gilbert, L. A., and Weissman, J. S. (2019). *Exploring genetic interaction manifolds constructed from rich single-cell phenotypes.* Science, 365(6455):786–793.
--  Srivatsan, S. R., McFaline-Figueroa, J. L., Ramani, V., Saunders, L., Cao, J., Packer, J., Pliner, H. A., Jackson, D. L., Daza, R. M., Christiansen, L., Zhang, F., Steemers, F., Shendure, J., and Trapnell, C. (2020). *Massively multiplex chemical transcriptomics at single-cell resolution.* Science, 367(6473):45–51.
--  Frangieh, C. J., Melms, J. C., Thakore, P. I., Geiger-Schuller, K. R., Ho, P., Luoma, A. M., Cleary, B., Jerby-Arnon, L., Malu, S., Cuoco, M. S., Zhao, M., Ager, C. R., Rogava, M., Hovey,
-L., Rotem, A., Bernatchez, C., Wucherpfennig, K. W., Johnson, B. E., Rozenblatt-Rosen, O.,
-Schadendorf, D., Regev, A., and Izar, B. (2021). *Multimodal pooled Perturb-CITE-seq screens in
-patient models define mechanisms of cancer immune evasion*. Nat. Genet., 53(3):332–341.
-- Jiang, L., Dalgarno, C., Papalexi, E., Mascio, I., Wessels, H.-H., Yun, H., Iremadze, N., Lithwick Yanai, G., Lipson, D., and Satija, R. (2024a). *Systematic reconstruction of molecular pathway signatures using scalable single-cell perturbation screens.* bioRxiv, page 2024.01.29.576933.
-- McFaline-Figueroa, J. L., Srivatsan, S., Hill, A. J., Gasperini, M., Jackson, D. L., Saunders,
-L., Domcke, S., Regalado, S. G., Lazarchuck, P., Alvarez, S., et al.
-(2024). *Multiplex single cell chemical genomics reveals the kinase dependence of the response to targeted therapy*. Cell Genomics, 4(2)
-- Szałata, A., Benz, A., Cannoodt, R., Cortes, M., Fong, J., Kuppasani, S., Lieberman, R., Liu, T., Mas-Rosario, J. A., Meinl, R., Nourisa, J., Tumiel, J., Tunjic, T. M., Wang, M., Weber, N., Zhao, H., Anchang, B., Theis, F. J., Luecken, M. D., Burkhardt, D. B. (2024). *A benchmark for prediction of transcriptomic responses to chemical perturbations across cell types.* NeurIPS, (38).
-
-Datasets with fixed splits have their splits included as `.csv` files with two colums: the first corresponds to the cell ID (which is the `.obs_names` of the respective h5ad file) and second to the split value (`train`, `val`, `test`). Some datasets contain multiple splits in which case the split files are in a `tar.gz`. 
+version https://git-lfs.github.com/spec/v1
+oid sha256:c6df38cbbfc289ea8013bbb87406dc23d09916fab785c8040f82a3973d3a83b8
+size 2544

croissant.json

@@ -1,219 +0,0 @@
-{
-  "@context": {
-    "@vocab": "https://schema.org/",
-    "cr": "http://mlcommons.org/schema/",
-    "sc": "https://schema.org/"
-  },
-  "@type": "sc:Dataset",
-  "name": "PerturBench",
-  "description": "A dataset containing single-cell RNA-seq data with genetic and chemical perturbations.",
-  "url": "https://huggingface.co/datasets/altoslabs/perturbench",
-  "license": "https://creativecommons.org/licenses/by-nc/4.0/deed.en",
-  "citation": "Yan Wu, Esther Wershof, Sebastian M Schmon, Marcel Nassar, B\u0142a\u017cej Osi\u0144ski, Ridvan Eksi, Zichao Yan, Rory Stark, Kun Zhang, and Thore Graepel (2025). PerturBench: Benchmarking Machine Learning Models for Cellular Perturbation Analysis. Proceedings of the 39th Conference on Neural Information Processing Systems (NeurIPS 2025).",
-  "datePublished": "2025-05-15",
-  "version": "1.0",
-  "conformsTo": "http://mlcommons.org/croissant/1.0",
-  "distribution": [
-    {
-      "@type": "sc:FileObject",
-      "@id": "srivatsan20-h5ad",
-      "name": "srivatsan20_h5ad_file",
-      "description": "Gzipped HDF5 file for the Srivatsan20 dataset.",
-      "contentUrl": "https://huggingface.co/datasets/altoslabs/perturbench/resolve/main/srivatsan20_highest_dose_processed.h5ad.gz",
-      "encodingFormat": "application/gzip",
-      "sha256": "ddd46251ea3942c0e3799f64cada9718b455ffdeabb1a8fe8b3333ebec4946f0"
-    },
-    {
-      "@type": "sc:FileObject",
-      "@id": "norman19-h5ad",
-      "name": "norman19_h5ad_file",
-      "description": "Gzipped HDF5 file for the Norman19 dataset.",
-      "contentUrl": "https://huggingface.co/datasets/altoslabs/perturbench/resolve/main/norman19_processed.h5ad.gz",
-      "encodingFormat": "application/gzip",
-      "sha256": "8f880e3d85bb68a73a6d044cae13f62a86b8132b0302c6e1bc4b4e5564e6530f"
-    },
-    {
-      "@type": "sc:FileObject",
-      "@id": "norman19-cpa-h5ad",
-      "name": "norman19_cpa_h5ad_file",
-      "description": "CPA publication version of the norman19 dataset, subset to highly variable genes only.",
-      "contentUrl": "https://huggingface.co/datasets/altoslabs/perturbench/resolve/main/norman19_cpa_hvg_normalized_curated.h5ad.gz",
-      "encodingFormat": "application/gzip",
-      "sha256": "f04b53e1e2390d4e8b59029fc6e482f160a77b73c6965d6b487975954b76ad49"
-    },
-    {
-      "@type": "sc:FileObject",
-      "@id": "norman19-cpa-splits",
-      "name": "norman19_cpa_splits_file",
-      "description": "CPA publication splits for their version of the norman19 dataset",
-      "contentUrl": "https://huggingface.co/datasets/altoslabs/perturbench/resolve/main/norman19_cpa_hvg_normalized_splits.tar.gz",
-      "encodingFormat": "application/gzip",
-      "sha256": "d4369b3ca4864e6348ca8a43f77feae20b36fede7f3d9dfa8e6e31abd1156962"
-    },
-    {
-      "@type": "sc:FileObject",
-      "@id": "frangieh21-h5ad",
-      "name": "frangieh21_h5ad_file",
-      "description": "The gzipped HDF5 file containing the processed perturbation data.",
-      "contentUrl": "https://huggingface.co/datasets/altoslabs/perturbench/resolve/main/frangieh21_processed.h5ad.gz",
-      "encodingFormat": "application/gzip",
-      "sha256": "d8e93131e33c02b9fce9d1a1ba8e4924ae6cc4907341b45138ef82cb4a162c79"
-    },
-    {
-      "@type": "sc:FileObject",
-      "@id": "frangieh21-csv",
-      "name": "frangieh21_csv_file",
-      "description": "CSV file containing the data splits for the dataset.",
-      "contentUrl": "https://huggingface.co/datasets/altoslabs/perturbench/resolve/main/frangieh21_split.csv",
-      "encodingFormat": "text/csv",
-      "sha256": "ac2e8361a44239eb07aa813ca9025985b399aecce4851ef879a534bb1a72b711"
-    },
-    {
-      "@type": "sc:FileObject",
-      "@id": "mcfalinefigueroa23-h5ad",
-      "name": "mcfalinefigueroa23_h5ad_file",
-      "description": "Gzipped HDF5 file for the McFalineFigueroa23 dataset.",
-      "contentUrl": "https://huggingface.co/datasets/altoslabs/perturbench/resolve/main/mcfaline23_gxe_processed.h5ad.gz",
-      "encodingFormat": "application/gzip",
-      "sha256": "f6639e1ee8f24d2f2d0eea0011446fad370cfba96009ffc257b85fa604aa36f8"
-    },
-    {
-      "@type": "sc:FileObject",
-      "@id": "mcfalinefigueroa23-splits",
-      "name": "mcfalinefigueroa23_splits_file",
-      "description": "Gzipped tar archive containing the data splits for the McFalineFigueroa23 dataset. Each split corresponds to a different data scale (small, medium, full).",
-      "contentUrl": "https://huggingface.co/datasets/altoslabs/perturbench/resolve/main/mcfaline23_gxe_splits.tar.gz",
-      "encodingFormat": "application/gzip",
-      "sha256": "a383a9e8e38e63756e5dd9e1a4df7844a97669ef0d553c70de8f1441e40f0aa7"
-    },
-    {
-      "@type": "sc:FileObject",
-      "@id": "jiang24-h5ad",
-      "name": "jiang24_h5ad_file",
-      "description": "Gzipped HDF5 file for the Jiang24 dataset.",
-      "contentUrl": "https://huggingface.co/datasets/altoslabs/perturbench/resolve/main/jiang24_processed.h5ad.gz",
-      "encodingFormat": "application/gzip",
-      "sha256": "e92dec6610e261f31eb15d50aff0c12e19e99d992452e1f14d560477e741ccfe"
-    },
-    {
-      "@type": "sc:FileObject",
-      "@id": "jiang24-csv",
-      "name": "jiang24_csv_file",
-      "description": "CSV file containing the data split for the Jiang24 dataset.",
-      "contentUrl": "https://huggingface.co/datasets/altoslabs/perturbench/resolve/main/jiang24_split.csv",
-      "encodingFormat": "text/csv",
-      "sha256": "f1822ab1252b363727f30bbe8889384c2b306369889364390ecccb80e75f87d7"
-    },
-    {
-      "@type": "sc:FileObject",
-      "@id": "op3-h5ad",
-      "name": "op3_h5ad_file",
-      "description": "Gzipped HDF5 file for the OP3 dataset.",
-      "contentUrl": "https://huggingface.co/datasets/altoslabs/perturbench/resolve/main/op3_processed.h5ad.gz",
-      "encodingFormat": "application/gzip",
-      "sha256": "a95d7b3717128f89d34abdc5031e38bd04ed616296b9344f720f9e1e9d265b99"
-    },
-    {
-      "@type": "sc:FileObject",
-      "@id": "op3-csv",
-      "name": "op3_csv_file",
-      "description": "CSV file containing the data split for the OP3 dataset.",
-      "contentUrl": "https://huggingface.co/datasets/altoslabs/perturbench/resolve/main/op3_split.csv",
-      "encodingFormat": "text/csv",
-      "sha256": "433e23e944ba7865185520a41510323186f755f7bc5a5a413d11704e2db60e90"
-    },
-    {
-      "@type": "sc:FileSet",
-      "@id": "srivatsan20",
-      "name": "Srivatsan20",
-      "description": "This dataset is a modified version of the Srivatsan20 dataset published by Srivatsan et al. via GEO:GSE139944. It contains 188 chemical perturbations subset to the highest dose only applied across 3 cell types. The full data preprocessing notebook can be found at: https://github.com/altoslabs/perturbench/blob/main/notebooks/neurips2025/data_curation/curate_Srivatsan20.ipynb.",
-      "encodingFormat": "application/gzip",
-      "cr:includes": [
-        "srivatsan20-h5ad"
-      ]
-    },
-    {
-      "@type": "sc:FileSet",
-      "@id": "norman19",
-      "name": "Norman19",
-      "description": "This dataset is a modified version of the Norman19 dataset published by Norman et al. via GEO:GSE133344. It contains 287 genetic perturbations (131 duals) applied to k562 cells. The full data preprocessing notebook can be found at: https://github.com/altoslabs/perturbench/blob/main/notebooks/neurips2025/data_curation/curate_Norman19.ipynb.",
-      "encodingFormat": "application/gzip",
-      "cr:includes": [
-        "norman19-h5ad",
-        "norman19-cpa-h5ad",
-        "norman19-cpa-splits"
-      ]
-    },
-    {
-      "@type": "sc:FileSet",
-      "@id": "frangieh21",
-      "name": "Frangieh21",
-      "description": "This dataset is a modified version of the Frangieh21 dataset published by Frangieh et al. via https://singlecell.broadinstitute.org/single_cell/study/SCP1064/multi-modal-pooled-perturb-cite-seq-screens-in-patient-models-define-novel-mechanisms-of-cancer-immune-evasion. It contains 248 genetic perturbations applied to 3 melanoma cell models. The full data preprocessing notebook can be found at: https://github.com/altoslabs/perturbench/blob/main/notebooks/neurips2025/data_curation/curate_Frangieh21.ipynb.",
-      "encodingFormat": "application/gzip",
-      "cr:includes": [
-        "frangieh21-h5ad",
-        "frangieh21-csv"
-      ]
-    },
-    {
-      "@type": "sc:FileSet",
-      "@id": "mcfalinefigueroa23",
-      "name": "McFalineFigueroa23",
-      "description": "This dataset is a modified version of the McFalineFigueroa23 dataset published by McFaline-Figueroa et al. via GEO:GSE225775. It contains ~200 perturbations applied across 6 cell lines and 5 cytokine treatments (30 unique biological states). The data preprocessing occured in two steps and both files can be found at: https://github.com/altoslabs/perturbench/blob/main/notebooks/neurips2025/data_curation/ with the curate_McFalineFigueroa_2023 prefix.",
-      "encodingFormat": "application/gzip",
-      "cr:includes": [
-        "mcfalinefigueroa23-h5ad",
-        "mcfalinefigueroa23-splits"
-      ]
-    },
-    {
-      "@type": "sc:FileSet",
-      "@id": "jiang24",
-      "name": "Jiang24",
-      "description": "This dataset is a modified version of the Jiang24 dataset published by Jiang et al. via https://zenodo.org/records/14518762. It contains 525 genetic perturbations applied across 3 cell lines and 5 chemical treatments (15 unique biological states). The data preprocessing occured in two steps and both files can be found at: https://github.com/altoslabs/perturbench/blob/main/notebooks/neurips2025/data_curation/ with the curate_Jiang_2024 prefix.",
-      "encodingFormat": "application/gzip",
-      "cr:includes": [
-        "jiang24-h5ad",
-        "jiang24-csv"
-      ]
-    },
-    {
-      "@type": "sc:FileSet",
-      "@id": "op3",
-      "name": "OP3",
-      "description": "This dataset is a modified version of the OpenProblems perturbation prediction challenge dataset that a Kaggle competition part of the NeurIPS 2023 competition track by Burkhardt et al. via https://openproblems.bio/benchmarks/perturbation_prediction?version=v1.0.0. It contains 144 chemical perturbations applied to PBMCs with at least 4 mature cell types. The data preprocessing can be found at: https://github.com/altoslabs/perturbench/blob/main/notebooks/neurips2025/data_curation/curate_op3.ipynb.",
-      "encodingFormat": "application/gzip",
-      "cr:includes": [
-        "op3-h5ad",
-        "op3-csv"
-      ]
-    }
-  ],
-  "recordSet": [
-    {
-      "@type": "cr:RecordSet",
-      "name": "treatment_classification",
-      "description": "Records for the perturbation response prediction task using scRNA-seq datasets stored in h5ad files. Records are sourced from six manifest files found in the distribution list, with one file per dataset, and stored in the `.obs` slot of each h5ad file. The perturbation identity is the 'condition' column, with the `control` value reserved for the DMSO or non-targeting CRISPR controls. The cell type is stored in the `cell_type` column. Additional cytokine or chemical treatments are stored in the `treatment` column. Splits are defined dynamically by the PerturBench library.",
-      "field": [
-        {
-          "@type": "cr:Field",
-          "name": "condition",
-          "description": "Chemical or genetic perturbation applied to cells. This is the key conditioning label for the perturbation response prediction task.",
-          "dataType": "sc:Text"
-        },
-        {
-          "@type": "cr:Field",
-          "name": "cell_type",
-          "description": "Cell line or cell type of the sample.",
-          "dataType": "sc:Text"
-        },
-        {
-          "@type": "cr:Field",
-          "name": "treatment",
-          "description": "Additional cytokine or chemical treatments applied to cells, used together with `cell_type` to define the biological state of a sample.",
-          "dataType": "sc:Text"
-        }
-      ]
-    }
-  ]
-}

norman19_cpa_hvg_normalized_curated.h5ad.gz → norman19_cpa_hvg_normalized_curated.h5ad

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f4d5a9aff863df16c167012bf7af6e17dc4b30c0353c88507279f91f52bdd59a
-size 295522019
+oid sha256:428a4ee1ddb7476bb7a8c378e821763ec6f71e560434685553066bc66bda37d8
+size 838510096