Verified Agent Skill record

Bio Database Evidence

Unified biological database evidence skill for lookup, annotation, and evidence gathering across gene annotations, variant clinical significance, cancer mutations, GWAS trait associations, pathway mapping, target-disease evidence, protein structures, protein interaction networks, reference single-cell census queries, and cross-database biological ID mapping.

Tasks

Look up gene annotation, identifiers, RefSeq, Ensembl IDs, orthologs, VEP, GO, and genomic coordinates.
Retrieve variant clinical significance, VUS interpretation support, ClinVar review status, cancer mutations, and COSMIC evidence.
Gather GWAS Catalog trait associations, rs IDs, p-values, summary statistics, and genetic epidemiology evidence.
Perform pathway mapping, ID conversion, KEGG pathways, Reactome enrichment, disease pathways, and pathway evidence collection.
Collect target-disease association evidence, tractability, safety, known drugs, and Open Targets evidence.

Inputs

Biological entity types: gene, transcript, variant, pathway, target, protein structure, protein interaction, trait association, or reference cell population.
Query terms for biological database lookup.
Biological identifiers such as RefSeq, Ensembl IDs, UniProt IDs, or rs IDs.

Outputs

Evidence results preserving source names, query terms, access dates, identifiers, and API caveats.
Evidence tables for comparing multiple sources.
Notes on authentication, license, rate limits, or non-public access restrictions for a given source.

Limitations and checks

Not for full single-cell analysis, clustering, UMAP, marker genes, cell annotation, AnnData/h5ad container editing, or scVI/scANVI batch-correction planning (use scanpy instead).
Not for bulk RNA-seq differential expression (use pydeseq2 instead).
Not for BAM, SAM, CRAM, VCF, pileup, coverage, or region extraction as a primary file-processing task.
Not for deepTools signal-track processing and heatmaps.
Not for protein language models, embeddings, inverse folding, or protein-design workflows.
Not for constraint-based metabolic modeling, FBA, or metabolic-engineering simulation.
Not for BED/genomic interval embeddings, genomic-region ML, or gene regulatory network inference.
Not for FCS or flow-cytometry file parsing.