BioResearch Agent — Biomarker Discovery Skill
Capability
End-to-end biomarker discovery: dataset loading (GEO GSE7621 or synthetic) → differential expression (t-test + Bonferroni) → hypergeometric pathway enrichment (KEGG/GO) → candidate ranking → volcano plot. Returns ranked candidate biomarkers, not just a narrative.
Run
bioresearch run biomarker --disease "Parkinson's disease"
Outputs (in outputs/biomarker/)
biomarker_deg_table.csv— differentially expressed genesbiomarker_top_candidates.csv— ranked candidatesbiomarker_pathway_enrichment.csv— enriched pathways / GO termsbiomarker_volcano_plot.png— volcano plotbiomarker_report.txt— structured report
Note
This skill dispatches to the framework's biomarker workflow. It adds no analysis of its
own; all statistics run in the workflow modules. Requires no LLM key.