Genomic Data

This section documents the large precomputed genomic knowledge bases that iris adapters read from — their data model, coverage, and provenance — independent of any single use case. Where Reference documents iris' own classes, and How-to guides document tasks, this section documents the external data itself: what's actually in the files on disk, so you can reason about coverage and trust before wiring an adapter to a pipeline.

Each page pairs a data catalog (schema, categories, scale, provenance) with a showcase of the iris adapter(s) that read it — proof that the domain model in Reference → Adapters can actually load and normalize the source, not just a description of the files.

Source	Coverage	iris adapter(s)
uvar	Every SNV + enumerated repeat-region indel in GRCh38, ~8.9B rows, precomputed population frequency / ClinVar / in-silico scores	UvarVariantStore, UvarAnnotator
dbSNP	698 contigs (25 usable), reference variant IDs (rsID) + COMMON flag, dbSNP157/GRCh38	DbSnpAnnotator
CADD	Genome-wide SNV deleteriousness (RawScore/PHRED), GRCh38 v1.7, ~691M rows on chr1 alone	CaddAnnotator
MCPS	141.8M rows, allele frequencies (global + AFR/EUR/MEX) from a mixed WGS/WES/array cohort	McpsAFAnnotator
dbNSFP	~59 aggregated in-silico predictor scores per coding/near-coding variant, dbNSFP5.0a/GRCh38	DbNsfpAnnotator