Antigen-specific Antibody Multi-modal Foundation Model for Functional Antibody Design

2026-07-22

AI Summary

Current protein language models struggle to accurately capture the epitope-level specificity of antibody–antigen interactions, hindering the design of functional antibodies. To address this limitation, this work proposes AAMFM, a multimodal foundation model that, for the first time, enables unified representation learning of antibody sequences and structures conditioned on antigen context—including geometric interfaces and epitope annotations. The model further incorporates a structure-informed Cal-DPO preference optimization mechanism to guide the generation of high-affinity antibodies. Evaluated on functional antibody design tasks, AAMFM achieves state-of-the-art performance, significantly enhancing the feasibility and practicality of antigen-specific antibody engineering.

Abstract

Antibodies are essential proteins that play a central role in immune recognition by binding specific antigen molecules. Although recent protein language models have enabled progress in single-chain protein modeling and generation, they often fall short in antigen-specific antibody design, where effective modeling requires explicit pairing between antibody and antigen, particularly at the epitope level. To address these limitations, we introduce AAMFM, an Antigen-specific Antibody Multimodal Foundation Model that learns unified representations of antibody sequences and structures conditioned on antigen context. AAMFM incorporates rich antigen information including geometric interfaces and epitope annotations via a cross-modal adapter, enabling joint modeling of antibody-antigen interactions in a shared latent space. To further guide the model toward functional relevance, we fine-tune AAMFM using Calibrated Direct Preference Optimization (Cal-DPO), leveraging preference signals extracted from a strong structural prior to align learning with binding-specific objectives. Extensive experiments demonstrate that AAMFM achieves state-of-the-art performance in functional antibody design, revealing its potential for antigen-specific antibody engineering. Our code is available at https://github.com/XL-S224/AAMFM.

Problem

• antigen-specific antibody design
• antibody-antigen interaction
• epitope-level modeling
• functional antibody design
• multimodal foundation model

Innovation

• multimodal foundation model
• antigen-specific antibody design
• cross-modal adapter
• epitope-level modeling
• Calibrated Direct Preference Optimization