Skip to content

Data Sources

NCI Study Catalog

  • Dataset: CTRP study record.
  • URL: https://studycatalog.cancer.gov/#/dataset/7d356db1-e1de-5dcc-809c-97b4fbd7cb48
  • Access: Public Study Catalog metadata and direct CloudFront archives.
  • Validation: Required ZIP members and table columns are checked when source tables are read. Expected release counts are recorded below.

CTRPv2.0 Expanded Dataset

  • Version: CTRPv2.0_2015_ctd2_ExpandedDataset.zip.
  • URL: https://d1hp0ufb0xxisr.cloudfront.net/CTD2/Broad/CTRPv2.0_2015_ctd2_ExpandedDataset.zip
  • Format: ZIP containing tab-delimited cell-line, compound, experiment, curve, and raw dose-response files.
  • Role: Canonical sample, treatment, curve-profile, and raw dose-response source.
  • Expected counts: 6,171,005 rows in v20.data.per_cpd_post_qc.txt and 395,263 rows in v20.data.curves_post_qc.txt.
  • Processed products: base metadata, published curve profiles, and hash-partitioned dose-response chunks.

CTRPv2.1 Published Analysis Release

  • Version: CTRPv2.1_2016_pub_NatChemBiol_12_109.zip.
  • URL: https://d1hp0ufb0xxisr.cloudfront.net/CTD2/Broad/CTRPv2.1_2016_pub_NatChemBiol_12_109.zip
  • Format: ZIP containing AUC sensitivity, gene-expression, copy-number, feature metadata, and correlation tables.
  • Role: Published AUC companion table and molecular profiles.
  • Expected count: 326,148 rows in v21.data.auc_sensitivities.txt.
  • Expression interpretation: v21.data.gex_avg_log2.txt is an Affymetrix-derived microarray matrix. Its source value column is mrna_expression_avg_log2; the curated experiment is named microarray.gene_expression_log2.

CTRPv2.2 ACME Reference

  • Version: CTRPv2.2_2015_pub_CancerDisc_5_1210.zip.
  • Study: Seashore-Ludlow et al., Cancer Discovery 2015, 5:1210-1223.
  • Study Catalog file ID: 70c24374-121f-58c8-95e1-3b5df3680716.
  • URL: https://d1hp0ufb0xxisr.cloudfront.net/CTD2/Broad/CTRPv2.2_2015_pub_CancerDisc_5_1210.zip
  • Content length: 2,080,055 bytes.
  • ETag: 0408be36152af34b7e3337e3f632a76d.
  • Format: ZIP containing ACME hotspots, mutation annotations, protein targets, AUC sensitivity, and release metadata.
  • Scope: This archive represents the separate 2015 ACME analysis and is outside the integrated v2.0/v2.1 PharmacoSet scope.
  • AUC comparison: 260,496 rows and 253,433 unique compound-cell pairs cover 78.8% of the CTRPv2.1 pair set. Pair-level comparisons identify NCIH524/JHH6 and BFTC905/BFTC909 assignment inconsistencies. The 410 potentially valid pairs unique to this source have no matching v2.0 raw dose-response experiments for outcome recomputation.
  • Study-specific content: 40,471 ACME hotspot rows, 3,235 mutation-annotation rows, 64 cell-line annotation-feature mappings, and study-specific media and cluster dictionaries.
  • Compound coverage: all 481 compounds and all 468 protein-target assignments are represented in CTRPv2.1 compound metadata.

Sample And Treatment Annotation

  • AnnotationDB: REST API at annotationdb.base_url in config/pipeline.yaml, using /compound/many and /cell_line/many.
  • AnnotationGx: development ref fix/bioc-review-comments, installed by pixi run setup.
  • Query order: AnnotationDB first, followed by AnnotationGx PubChem and Cellosaurus functions for unresolved records or missing key fields.
  • Processed products: sample and treatment annotations under data/procdata/metadata/, plus compound and cell-line miss audits under data/procdata/annotation_audit/.
  • Reuse: Snakemake reuses completed annotation outputs. Forcing the annotation rule repeats the live service requests.

Gene Annotation

  • GENCODE: release 50 / Ensembl 116 / GRCh38.p14 GTF and Entrez metadata from https://ftp.ebi.ac.uk/pub/databases/gencode/Gencode_human/release_50/.
  • NCBI Gene: current gene2ensembl.gz and gene_history.gz from https://ftp.ncbi.nlm.nih.gov/gene/DATA/.
  • Access: GENCODE files are downloaded directly. NCBI files are streamed and filtered to human taxon 9606 before storage.
  • Resolution: NCBI Gene history normalizes Entrez IDs, then current gene2ensembl and GENCODE records resolve canonical Ensembl gene IDs.
  • HGNC metadata: GENCODE.HGNC.ID is read from the GENCODE GTF.
  • Processed products: the feature map under data/procdata/metadata/ and mapping/drop audits under data/procdata/annotation_audit/.
  • Public reference table: gene_annotation.tsv retains every CTRPv2.1 source gene, including records without a current Ensembl ID. Molecular assay matrices contain unique current Ensembl IDs.