ResearchNot part of the commercial buyer journey

Research capabilities

These modules exist in the codebase and are usable, but they are deliberately not surfaced in the primary buyer journey. A serious clinical claim needs its own validation programme, and putting six of them on the homepage under one "LIVE" label weakens trust in all of them.

Commercial front doors are MedSafe and PetriDish Autoimmune.

Research modules

PetriDish Treg

Guideline implemented

Treg-epitope prioritization for antigen-specific tolerance programs. Rammensee anchor-motif scoring against IEDB substrate.

artifact: T1D · RA · MS candidate lists

PetriDish Design

Guideline implemented

Generative hit-design with real RDKit chemistry (BRICS, QED, Lipinski, Tanimoto novelty). Five hard go/no-go gates before wet-lab handoff.

artifact: CRO-ready candidate queue

PetriDish Cardio

Guideline implemented

Cardiovascular risk simulation with six Monte Carlo engines; rng-seed reproducible.

artifact: Population-level risk deltas

PetriDish Discover

Guideline implemented

GNN hypothesis generation, structurally quarantined from the fact graph — hypothesis edges never merge with facts.

artifact: Ranked hypothesis edges

PetriDish Lens

Guideline implemented

ANA-IIF pattern recognition → reflex genotype routing across 29 ICAP AC-codes.

artifact: Reflex-genotype recommendation

IndoVarClassifier

Guideline implemented

Grant-research variant classification pipeline with live Ensembl VEP + gnomAD + ClinVar APIs and IndiGen frequency layer.

artifact: Sourced classification per variant

IndoAutoImmune v2

Guideline implemented

HLA precision medicine — hard benchmark gate; SJS/TEN discipline.

artifact: Population-risk report

None of the above are "clinically deployed." Each will move up the validation ladder — retrospectively evaluated → prospectively evaluated → clinically deployed — as its own study is completed. See /methods for the current state per module.