Reviews can be read, not computed; databases can be queried, not reasoned with. This map gives biomedical knowledge coordinates and road networks — so every intervention forecast can be re-tested.
SteeraMed Map helps organize human-state and intervention hypotheses. It does not diagnose disease, prescribe treatment, or replace clinical trials.
Three entries for three needs: explore the map, run the agent, read the methods
Demo pages use public or de-identified data and do not constitute medical advice
Disease reclassification → Patient stratification → Drug repurposing → Target discovery
Re-cluster traditional disease diagnoses into biological subtypes using multi-omics features.
Stratify patients within subtypes by state-deviation magnitude and steerability metrics.
PPI proximity + compound-target network generates candidates; retrospective recall validation.
Network centrality + perturbation propagation screens high-leverage nodes, outputting testable hypotheses.
The current version generates candidate hypotheses, not causally validated conclusions
v0.2 Paper-Aligned is a research preview: maps and workflows organize hypotheses and rank candidates. The current release is rated evidence tier 2 (network inference) overall — see Methods & Boundaries. It does not constitute diagnosis or treatment advice.
Evidence anchor — SteeraMed Bench: 332 modules × 1,916 small-molecule drugs × 23 disease categories; null results for AI-nominated modules are reported as-is.
