Target Prioritization

Takes a gene list from differential expression and builds multi-source dossiers by hitting UniProt, OpenTargets, PubMed, Human Protein Atlas, and ChEMBL in parallel, then re-ranks by a composite score weighing druggability, disease genetics, tissue specificity, CRISPR essentiality, gnomAD constraint, and expression context. Ships tuned for autoimmune/T-cell work but you can retarget by editing a few disease-term arrays. Outputs a markdown report with per-gene rationales (Claude fills those), a sortable CSV, and raw JSON audit trails. Honest take: this is opinionated scaffold code, not a black box. The weights are exposed in YAML and the fetchers are standalone scripts you'll probably want to tweak once you see what signal matters for your program.

Install to Claude Code

npx -y skills add agents365-ai/365-skills --skill target-prioritization --agent claude-code