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

Source

• Source name: arXiv
• Source kind: paper
• Authors / organization: Xiaoliang Shi, Zichen Wang, Runze Ma, Zhongyue Zhang, Shuangjia Zheng
• Venue: arXiv: q-bio.BM
• Published: 2026-07-22
• Collected: 2026-07-23
• URL: http://arxiv.org/abs/2607.20057v1
• PDF: https://arxiv.org/pdf/2607.20057v1
• DOI: UNKNOWN

Plain-Language Brief

Researchers introduce a new model called AAMFM (Antigen-specific Antibody Multimodal Foundation Model) for designing functional antibodies. The model learns unified representations of antibody sequences and structures that are conditioned on antigen context. It incorporates rich antigen information, including geometric interfaces and epitope annotations, via a cross-modal adapter. To ensure the model focuses on functional relevance, it is fine-tuned using Calibrated Direct Preference Optimization (Cal-DPO). Experiments show that AAMFM outperforms existing methods in functional antibody design.

What the Document Says

• The paper introduces a new model called AAMFM for antigen-specific antibody design.
• AAMFM learns unified representations of antibody sequences and structures conditioned on antigen context.
• It incorporates rich antigen information including geometric interfaces and epitope annotations via a cross-modal adapter.
• The model is fine-tuned using Calibrated Direct Preference Optimization (Cal-DPO).
• Experiments demonstrate that AAMFM achieves state-of-the-art performance in functional antibody design.

Key Concepts Mentioned

• Antibodies
• antigen-specific antibody design
• AAMFM
• cross-modal adapter
• Calibrated Direct Preference Optimization (Cal-DPO)
• functional antibody design

Claims Made by the Document

• AAMFM is an effective model for antigen-specific antibody design.
• AAMFM achieves state-of-the-art performance in functional antibody design.

Evidence Used

• Extensive experiments demonstrate that AAMFM achieves state-of-the-art performance in functional antibody design.

Methods / Approach

• Introduction of AAMFM
• Learning unified representations of antibody sequences and structures conditioned on antigen context
• Incorporation of rich antigen information including geometric interfaces and epitope annotations via a cross-modal adapter
• Fine-tuning using Calibrated Direct Preference Optimization (Cal-DPO)

Limitations Stated by the Authors

UNKNOWN

Limitations Noted by the Agent

Agent interpretation — not from the source.

The paper does not discuss any limitations of the AAMFM model.

Open Questions

• What are the potential applications of AAMFM beyond functional antibody design?
• How does the performance of AAMFM compare to other models in non-functional antibody design tasks?

Notes for Future Q&A

May include interpretation. Labeled where it does.

The paper focuses on the application and performance of AAMFM in functional antibody design. However, it would be interesting to explore its potential applications in other areas and compare its performance with other models in different tasks.

Raw Abstract / Excerpt

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.