Antibody-Antigen Affinity Prediction with Chain-Aware Protein Language Modeling

Abstract

Motivation: Antibody-antigen affinity determines which antibodies advance in therapeutic discovery, repertoire analysis and affinity maturation, but experimental measurements are sparse relative to the scale of sequence libraries. Structure-based predictors can exploit interface geometry when reliable complexes are available, yet early discovery often requires ranking many heavy-light chain pairs against antigens for which no complex structure exists. Existing sequence-based models are scalable, but frequently compress heavy and light chains into a single antibody representation or concatenate antibody and antigen features, potentially obscuring the chain-specific and epitope-specific signals that drive binding. Results: We present AbAffinity, a sequence-only chain-aware three-stream architecture that maintains heavy chain, light chain and antigen as distinct streams. It integrates frozen ESM-2 embeddings with heavy-chain CDR-focused pooling, heavy-light self-attention, adaptive fusion and gated cross-attention, training only a compact interaction module. Under nested ten-fold cross-validation on SAAINT-DB, AbAffinity achieves Pearson r = 0.838 {+/-} 0.033, Spearman {rho} = 0.828 {+/-} 0.027 and RMSE = 0.738 {+/-} 0.078. AbAffinity maintains meaningful performance on stricter PDB-disjoint and antigen sequence-identity splits, and held-out complexes. Evaluation on external natural-complex benchmarks further supports consistent transfer. Partner-shuffling, attribution and uncertainty analyses support partner-conditioned prediction and biologically plausible residue-level hypotheses. AbAffinity provides a lightweight, explainable sequence-first framework for antibody triage when structural information is limited or unavailable.