This is a comprehensive pipeline for extracting biological meaning from spatial transcriptomics data. You feed it spatially variable genes, tissue type, and optional domain annotations, and it runs through eight analysis phases: gene characterization, pathway enrichment, domain comparison, cell-cell interaction prediction, disease/drug context, and literature validation. It's built around the "look up, don't guess" principle, querying databases like OpenTargets, STRING, HPA, and PubMed rather than relying on model knowledge. Works with Visium, MERFISH, and other spatial platforms. The skill is most useful when you have domain-specific gene lists and want to move beyond statistics into mechanisms and therapeutic opportunities. Includes an integration scoring system and evidence grading from clinical (T1) to computational (T4).
npx skills add https://github.com/mims-harvard/tooluniverse --skill tooluniverse-spatial-omics-analysis