Supporting data and code for Prioritizing gut microbiota-derived metabolites and host targets in ulcerative colitis through network and single-cell analyses

Associated manuscriptPrioritizing gut microbiota-derived metabolites and host targets in ulcerative colitis through network and single-cell analyses CreatorsGuilin Ren; Junhao Wu; Bo Fu; Jin Wu; Yingchao Liu. ScopeThis record provides the supporting data, archived scientific records and current analysis/figure-source files for a computational study of gut microbiota-derived metabolites and host targets in ulcerative colitis. The study combines network prioritization, donor-level reanalysis of public intestinal single-cell data, matched virtual gene deletions and a crystal-referenced comparison of archived docking poses. The complete inventory contains 18 files: 11 numbered supplementary files, three source-data archives and four documentation/checksum files. Additional_file_1-11 correspond one-to-one to the journal Online Resources ESM_1-11. The matched journal and repository files have identical contents. Numbered supplementary filesAdditional_file_1.xlsx contains 500 shared predicted metabolite targets and source annotations. Additional_file_2.xlsx and Additional_file_3.xlsx contain archived property/drug-likeness and toxicity predictions, respectively. Additional_file_4.pdf contains the complete extended methods and results, Figures S1-S21, Tables S1-S2 and 53 supplementary references (77 pages). Additional_file_5.pdf provides the current file inventory, journal/repository filename mapping and source-data access information. Additional_file_6.txt contains self-contained Python extraction/visualization source and the original 15 primary mode-1 docking inputs, coordinates, topology and scores. Save a copy with a .py extension to extract or reproduce the archived local views. Current Fig. 5 source and the crystal-reference comparison are in Additional_file_11.zip. Additional_file_7.csv retains all 20,351 original gene-lineage expression results, including nonsignificant findings. Additional_file_8.csv contains 78,253 cell records with donor labels, annotations and exported expression values in two separate coordinate systems: the rectal immune UMAP and the author-discovery epithelial t-SNE. Blank feature fields indicate that a feature was not exported, rather than measured zero expression. Additional_file_9.txt contains archived single-cell, independent-cohort model and plotting source with processed inputs and verification records; save a copy with a .py extension to extract the files. Additional_file_10.csv retains all 83 independent-cohort primary, paired, sensitivity and secondary-readout outcomes, with their separate statistical families and unavailable-test statuses. Additional_file_11.zip contains current plotting scripts and plotted inputs for Figures 5, 6, 8, 9, S20 and S21, the crystal-referenced PPARgamma geometric comparison, unchanged archived docking inputs, official PDB/CCD provenance and native PyMOL scenes. The comparison retains all 15 original docking scores and supplies 42 minimum ligand-residue heavy-atom distances. The 3-Angstrom crystal-ligand neighborhood describes spatial overlap and is not a binding criterion. Internal documentation lists dependencies, reproduction commands and file checksums. Source-data archives and reproducibilitySource_data_network_enrichment.zip preserves recovered networks, GeneMANIA and enrichment exports, gene-set reconciliation and graph/source verification records. Source_data_rectal_singlecell.zip preserves public-data provenance, metadata, donor counts/results, formal sensitivity analyses and the GSE116222 PPARG subset. Source_data_virtual_perturbation.zip preserves primary/sensitivity networks, candidate and matched-factor deletion results and formal diagnostics. Historical figure numbers in archived source paths identify provenance; current focused figure source is in Additional_file_11.zip. README.txt explains the data scope and interpretation. CHANGELOG.txt records the update. FILE_MANIFEST.csv records file roles, sizes and SHA256 values, excluding itself and the checksum file; checksums.sha256 covers every other record file, including the manifest. Public source studiesGSE125527: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE125527GSE116222: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE116222GSE214695: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE214695SCP259 author-discovery objects: https://github.com/cssmillie/ulcerative_colitis The original investigators generated these datasets. SCP259 analyses use the author-shared discovery subset of ten healthy and seven UC donors, rather than the complete atlas. Original large count matrices remain available from the source studies; this record does not duplicate the complete upstream atlases. Interpretation and reuseNetwork prioritization, differential expression, virtual-deletion sensitivity and structural geometry represent distinct evidence layers. The original hub-lineage primary family contains 20 tests; each independent primary family and the separate paired family contain 25. Secondary lineage-wide FDRs and virtual-deletion model FDRs remain distinct. Complete negative, nonsignificant and nonestimable results are retained. Broad epithelial pseudobulks do not separate subtype composition from within-subtype expression changes. FOSB/ZFP36 are shared network-sensitive readouts, rather than established candidate-specific downstream effects. Original numerical model results and archived pose coordinates are preserved. No new wet-laboratory experiment, sequencing run, docking run or molecular-dynamics trajectory is included. Docking scores and geometric proximity do not establish binding affinity, receptor activation or anti-inflammatory efficacy. Some historical network/enrichment inputs and parameters remain unrecovered, as specified in the manuscript. Third-party data and software retain their original terms, and included code licenses are preserved. Repository DOIhttps://doi.org/10.5281/zenodo.23619150