Poster
in
Workshop: 2nd Generative AI for Biology Workshop
SOAPIA: Siamese-Guided Generation of Off Target-Avoiding Protein Interactions with High Target Affinity
Sophia Vincoff · Oscar Davis · Ismail Ceylan · Alexander Tong · Joey Bose · Pranam Chatterjee, PhD
Keywords: [ peptide binders ] [ discrete diffusion ] [ protein design ] [ Monte Carlo Tree Search ] [ Siamese networks ] [ fusion oncoproteins ] [ off-target avoidance ] [ DPLM ] [ biologics ] [ rare diseases ] [ multi-objective optimization ] [ AlphaFold-Multimer ] [ therapeutic peptides ] [ affinity prediction ]
Therapeutic molecules must selectively interact with a target protein while avoiding structurally or functionally similar off-targets. However, no existing generative strategy explicitly optimizes both target affinity and off-target avoidance. To address this, we introduce SOAPIA, a framework for the Siamese-guided generation of Off-target-Avoiding Protein Interactions with high target Affinity. SOAPIA generates de novo peptide binders by steering the generative process of a Diffusion Protein Language Model (DPLM) using a multi-objective Monte Carlo Tree Search (MCTS). Affinity is optimized via a pre-trained predictor, while specificity is enforced using a Siamese model trained with an adaptive Log-Sum-Exp Decoy Loss. This dual-guidance scheme enables Pareto-efficient exploration of discrete sequence space without gradient access. In benchmarks across 17 fusion oncoproteins, SOAPIA consistently identifies binders with strong affinity and high selectivity. For multiple clinically relevant targets, SOAPIA generated peptides that preferentially bind the fusion by engaging both its head and tail domains, while avoiding the wild-type counterparts. These results underscore SOAPIA’s promise for designing safe, specific biologics for fusion-driven cancers and other rare, currently untreatable diseases.