A single-cell atlas linking intratumoral states to therapeutic vulnerabilities across cancers

The Therapeutic Cancer Cell Atlas (TCCA) is an integrated single-cell multi-omics resource comprising 1,839,242 cells (1,089,024 malignant + 750,218 non-malignant) from 853 samples across 36 studies and 34 cancer types. The atlas integrates five complementary data layers across all cells/samples: Expression: Single-cell RNA-seq with batch-corrected embeddings (scANVI). Clinical: Curated sample metadata. TME: 12 microenvironment archetypes from immune/stromal composition. Genomic: Copy-number variation profiles for genetic subclones (SCEVAN). Therapeutic: Drug sensitivity predictions and therapeutic clusters. Functional: 43 recurrent transcriptional metaprograms (NMF). File organization (13 files): tcca_expression_metadata_lvl2.h5ad (~ 23 GB) — Integrated atlas with ALL data layers: expression, clinical metadata, cell annotations, TME archetypes, genetic subclones, therapeutic clusters, embeddings, and metaprogram enrichment scores. Standalone files (12 files, ~2 GB) — Individual tabular extracts for targeted analyses without loading the full atlas: cell-level metadata (`obs`) from the AnnData object (1 file), source studies (1 file), clinical data (1 file), TME (2 files), CNV profiles (2 files), therapeutic predictions (3 files), metaprograms (2 files). See README.md for detailed file descriptions and usage example.