Bryzant Labs' autonomous SMA discovery platform has completed more than 580 AlphaFold3 protein–protein structural predictions across its prioritized target hypotheses — p38/MAPK signaling, neuromuscular-junction signaling, the SMN assembly complex, and others — as part of a multi-method pipeline that also runs Boltz-2 and Chai-1.

These are computational predictions whose purpose is to prioritize biological hypotheses for experimental testing — not validated drug candidates. No target or compound on this platform has undergone wet-lab validation, and predicted interaction-confidence scores reflect model confidence, not demonstrated binding or therapeutic effect. We pursue, down-weight, or set aside hypotheses as the evidence warrants, and report negative or contradicted results alongside positive signals.