Autoimmunity as a protein-recognition problem.

Every step of autoimmune disease is one protein recognising another. An HLA protein presents a self-peptide. A T-cell receptor reads it. An autoantibody attacks. The frontier of open protein models (ESM-2 · MHCflurry · Evo 2) is pointed at the disease mechanism, not a proxy. For Indians the HLA proteins differ, so the presented self-peptidome differs, so the autoantigen repertoire differs — and almost no one has computed this.

What this engine returns: ranked autoantigen candidates (with PTM annotations), India-pathogen molecular mimicry hypotheses, and the open research wedges your patient phenotype maps to. Computation collapses discovery from years to weeks; it does not replace validation. Every output is a hypothesis requiring wet-lab confirmation.