FateDriver

Processed single-cell data and prior gene sets for FateDriver This dataset contains processed single-cell RNA-sequencing data and prior gene resources used in the study “FateDriver: A Dual-View Graph Neural Network for Predicting Driver Genes in Cellular Reprogramming.” All single-cell datasets included in this deposit were originally obtained from the NCBI Gene Expression Omnibus (GEO).This deposit provides processed, analysis-ready objects rather than the original raw sequencing data. The FateDriver source code and analysis scripts are available at https://github.com/Jieun-Sung/FateDriver. File contents Each dataset archive contains an anndata.h5ad file and a pseudotime.csv file. The anndata.h5ad file is a processed AnnData object containing the expression data and associated cell and gene annotations used for analysis. The pseudotime.csv file contains per-cell pseudotime values with two columns: cell_id and pseudotime. Folder structure zenodo_upload/ ├── GSE181451_Baseline_MEF.zip (~834 MB; 9,119 cells) │ └── GSE181451_Baseline_MEF/ │ ├── anndata.h5ad │ └── pseudotime.csv ├── GSE181451_Primed_MEF.zip (~462 MB; 5,381 cells) │ └── GSE181451_Primed_MEF/ │ ├── anndata.h5ad │ └── pseudotime.csv ├── GSE242423.zip (~968 MB; 10,292 cells) │ └── GSE242423/ │ ├── anndata.h5ad │ └── pseudotime.csv ├── GSE297234_GM00731.zip (~945 MB; 8,454 cells) │ └── GSE297234_GM00731/ │ ├── anndata.h5ad │ └── pseudotime.csv ├── GSE297234_GM23815.zip (~1.7 GB; 14,972 cells) │ └── GSE297234_GM23815/ │ ├── anndata.h5ad │ └── pseudotime.csv ├── GSE98567.zip (~32 MB; 394 cells) │ └── GSE98567/ │ ├── anndata.h5ad │ └── pseudotime.csv ├── GSE237173.zip (~447 MB; 6 lineage pairs) │ └── GSE237173/ │ ├── CBC_Cont-aVEC_Dox_(root_CBC_Lineage1).h5ad │ ├── CBC_Cont-aVEC_Dox_(root_CBC_Lineage1)_pseudotime.csv │ ├── EC_Cont-aVEC_Dox(root_EC_Lineage1).h5ad │ ├── EC_Cont-aVEC_Dox_(root_EC_Lineage1)_pseudotime.csv │ ├── EE_Cont-aVEC_Dox(root_EE_Lineage1).h5ad │ ├── EE_Cont-aVEC_Dox_(root_EE_Lineage1)_pseudotime.csv │ ├── PC_Cont-aVEC_Dox(root_PC_Lineage1).h5ad │ ├── PC_Cont-aVEC_Dox_(root_PC_Lineage1)_pseudotime.csv │ ├── TA_Cont-aVEC_Dox(root_TA_Lineage1).h5ad │ ├── TA_Cont-aVEC_Dox_(root_TA_Lineage1)_pseudotime.csv │ ├── TC_Cont-aVEC_Dox(root_TC_Lineage1).h5ad │ └── TC_Cont-aVEC_Dox_(root_TC_Lineage1)_pseudotime.csv └── gene_sets.zip (~97 MB) └── gene_sets/ ├── iPSC_oskm_driver_genes.csv ├── iPSC_oskm_driver_human_genes.csv ├── GMT_driver_genes_mouse.csv ├── TRRUST_mouse_OSKM_targets.csv ├── TRRUST_human_OSKM_targets.csv ├── TRRUST_GMT_targets_mouse.csv ├── nichenet_gr_signaling_net.csv └── nichenet_gr_signaling_net_human.csv Dataset descriptions GSE181451_Baseline_MEF.zip contains processed mouse Baseline MEF-to-iPSC data. It was used to model the transition from the earlier Baseline MEF source state to the iPSC target. GSE181451_Primed_MEF.zip contains processed mouse Primed MEF-to-iPSC data. It was used to model the transition from the transcriptionally distinct Primed MEF source state to the same iPSC target. GSE242423.zip contains processed human OSKM reprogramming data used in the cross-method driver-gene recovery benchmark. GSE297234_GM23815.zip contains processed OSKM reprogramming data from the young human donor GM23815. It was used for benchmarking and for comparing partial and complete reprogramming trajectories. GSE297234_GM00731.zip contains processed OSKM reprogramming data from the aged human donor GM00731. It was used to assess the cross-donor reproducibility of the partial-versus-complete reprogramming results. GSE98567.zip contains processed mouse GMT-induced cardiac reprogramming data used in the driver-gene recovery benchmark. The delivered factors were Gata4, Mef2c, and Tbx5. GSE237173.zip contains six processed lineage-pair objects used to rank candidate origins of the OSKM-induced aVEC population in mouse intestine. The candidate control populations are crypt-base columnar cells (CBC), enterocytes (EC), enteroendocrine cells (EE), Paneth cells (PC), transit-amplifying cells (TA), and tuft cells (TC). Each candidate is paired with the doxycycline-induced aVEC target and ordered by Slingshot pseudotime rooted at that candidate population. Gene resources gene_sets.zip contains the positive labels and prior regulatory resources used by FateDriver. iPSC_oskm_driver_genes.csv and iPSC_oskm_driver_human_genes.csv contain the mouse and human OSKM driver-factor lists, respectively. GMT_driver_genes_mouse.csv contains the mouse GMT driver-factor list. The TRRUST target files contain species- and system-specific direct target genes used as tier-2 labels in the positive–unlabeled ranking objective. nichenet_gr_signaling_net.csv and nichenet_gr_signaling_net_human.csv contain the mouse and human prior networks constructed from the union of the NicheNet gene-regulatory and intracellular signaling networks. These tables provide directed gene pairs, edge types, and edge weights. The mouse network also includes Ensembl gene identifiers. Data origin The processed objects were derived from the following publicly available GEO datasets: GSE181451: Mouse MEF-to-iPSC reprogramming GSE242423: Human OSKM-induced pluripotency GSE297234: Human dermal fibroblast OSKM reprogramming, including donors GM23815 and GM00731 GSE98567: Mouse GMT-induced direct cardiac reprogramming GSE237173: Doxycycline-inducible OSKM reprogramming of mouse intestinal epithelium. Users requiring raw count matrices or original sample metadata should obtain them from the corresponding GEO records and cite the original studies or data submitters.