India Autoimmune Trigger MapIn-silico flagship · zero clinical access requiredPRD v1.0 · hypothesis-grade
Which infections trigger Indian autoimmune disease — and in whom?
India sees 100M+ dengue exposures per year, plus chikungunya, typhoid, streptococcus, tuberculosis. The post-infection autoimmunity link is documented clinically but has never been mapped computationally at population scale with Indian HLA. Until now. Pick an Indian HLA allele and an infection. See ranked pathogen peptides that mimic self-proteins AND are presented by your patient's HLA.
Computational hypotheses requiring validation. Built from open data + frozen models, zero training, zero clinical access. The artifact that builds the network — collaborators with patient sera invited via /for-labs Track 6.
The killer interaction
Why this is the right flagship
Zero clinical access required
Built entirely from open data (UniProt, IndiGen, IMGT/HLA) and frozen foundation models (MHCflurry, ESM-2, Evo 2 — all commercial-clean).
Genuinely novel for India
The post-infection autoimmunity link is documented clinically but has never been mapped computationally at population scale with Indian HLA.
Demos spectacularly
'Pick HLA + infection → see self-peptides at risk.' Single interaction; investor-grade discovery surface.
Produces validatable hypotheses
Every mimicry hit is a ranked, testable prediction — precisely what a wet-lab collaborator wants to receive.
Dual artifact
bioRxiv preprint → collaborator magnet (validation cohorts). Interactive atlas → investor magnet.
Honest framing
These are computational hypotheses, not validated findings. Disclosure is the collaborator ask, not a weakness.
Computational hypotheses requiring validation. Open data + frozen models, zero training, zero clinical access. The artifact that builds the network. PRD v1.0.