Structure-Aware Antibody Design with Affinity-Optimized Inverse Folding

2025-12-19

AI Summary

Antibody affinity optimization is time-consuming and costly, while existing protein language models exhibit limited predictive capability for this task. To address these challenges, we propose SimBinder-IF, a structure-aware antibody sequence inverse folding generative method. Its core innovation lies in the first integration of preference optimization into antibody inverse folding modeling: we freeze the ESM-IF structural encoder and fine-tune only the decoder, incorporating experimental affinity signals as supervision. This strategy reduces parameter count by 18% and substantially enhances affinity-directed generation. On AbBiBench, SimBinder-IF achieves a Spearman correlation of 0.410 (+55% over baseline), zero-shot cross-antigen generalization of 0.294 (+156%), superior Top-10 accuracy, and significantly improved ΔΔG prediction (error reduced from −46.57 to −75.16).

Technology Category

• Machine Learning: Imitation Learning & Inverse Reinforcement Learning
• Search and Optimization: Sampling/Simulation-based Search
• Natural Language Processing: Code Generation / Program Synthesis from Natural Language

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Motivation: The clinical efficacy of antibody therapeutics critically depends on high-affinity target engagement, yet laboratory affinity-maturation campaigns are slow and costly. In computational settings, most protein language models (PLMs) are not trained to favor high-affinity antibodies, and existing preference optimization approaches introduce substantial computational overhead without clear affinity gains. Therefore, this work proposes SimBinder-IF, which converts the inverse folding model ESM-IF into an antibody sequence generator by freezing its structure encoder and training only its decoder to prefer experimentally stronger binders through preference optimization. Results: On the 11-assay AbBiBench benchmark, SimBinder-IF achieves a 55 percent relative improvement in mean Spearman correlation between log-likelihood scores and experimentally measured binding affinity compared to vanilla ESM-IF (from 0.264 to 0.410). In zero-shot generalization across four unseen antigen-antibody complexes, the correlation improves by 156 percent (from 0.115 to 0.294). SimBinder-IF also outperforms baselines in top-10 precision for ten-fold or greater affinity improvements. A case study redesigning antibody F045-092 for A/California/04/2009 (pdmH1N1) shows that SimBinder-IF proposes variants with substantially lower predicted binding free energy changes than ESM-IF (mean Delta Delta G -75.16 vs -46.57). Notably, SimBinder-IF trains only about 18 percent of the parameters of the full ESM-IF model, highlighting its parameter efficiency for high-affinity antibody generation.

Problem

• Designing high-affinity antibodies computationally without expensive lab experiments
• Optimizing antibody sequences for stronger binding using structure-aware models
• Improving computational efficiency in antibody design while maintaining affinity gains

Innovation

• Freezes structure encoder, trains decoder for affinity optimization
• Uses preference optimization to favor stronger binding antibodies
• Achieves high parameter efficiency with partial model training