What problem does it solve? Researchers routinely need canonical gene lists — transcription factors, protein-coding genes, disease-associated genes — but typically rely on undated, unsourced lists passed around the lab. This Skill fetches each set live from the authoritative upstream source and writes a TSV with a sidecar .meta.json recording source URL, version, fetch date, and sha256 hashes, so every list is reproducible and citable. ## Core Features & Use Cases - Named gene sets: Transcription factors (Lambert 2018, AnimalTFDB, JASPAR motif-backed), protein-coding genes (Ensembl biotype or strict Ensembl ∩ HGNC/MGI ∩ GENCODE intersection), and union/intersection composites across human, mouse, and rat. - Ad-hoc disease-gene queries: Genes associated with a disease via Open Targets, GWAS Catalog, or OMIM, with optional score filtering — never literature search. - Genomic structural sets: Head-to-head (bidirectional promoter) gene pairs from Ensembl BioMart coordinates, filterable by chromosome and TSS distance. - Use Case: A reviewer demands a defensible background set for your GO enrichment analysis. Run one command to fetch the strict three-way protein-coding intersection, then cite the set name, Ensembl release, and source versions from the sidecar metadata in your methods section. ## Quick Start Ask the assistant to fetch the human Lambert 2018 transcription factor list and save it to a file, for example: "get me the human TFs from Lambert 2018 and save the list to /tmp/tfs.tsv".
