⚠️ Project under development

MuLAAIP: Multi-Modality Representation Learning for Antibody-Antigen Interaction Prediction

This repository contains the official implementation of MuLAAIP, a novel deep learning framework for predicting antibody-antigen interactions (AAI) by integrating 3D structural and 1D sequence data. Our approach addresses critical challenges in AAI prediction, including structural data scarcity, sequence-structure dependency modeling, and imbalanced label distributions.

📁 Benchmark Datasets

Dataset Summary

| Dataset | Type | Samples | Description |
|--------|------|---------|-------------|
| Wild-type/Mutant-type Affinity | Affinity Labeling | 1,191 / 1,742 pairs | Antibody-antigen binding affinity|
| Alphaseq | Affinity Labeling | 248k antibodies | Antibody-antigen binding affinity |
| SARS-CoV-2 Neutralization | Binary Classification | 310 pairs (228+/82-) | Neutralization activity labels |

All missing experimental structures were predicted using ESMFold (https://github.com/facebookresearch/esm).

📥 Data Acquisition

Download Instructions

1. Get Data:
   Baidu Cloud Link (Password: iuqs)

Installation

# Clone the repo
git clone https://github.com/trashTian/MuLAAIP.git 
cd MuLAAIP

# Install dependencies
pip install -r requirements.txt

Data pre-processing

(1) 1D Sequence Representation: use pre-trained protein (antibody) language models to process sequence data and obtain embeddings. For example, ProtTrans, ESM2, AbLang, AntiBERTy,BERT2DAb

python PLM.py

We have embedded and saved these sequences locally

(2) 3D Structural Representation: construct fine-grained structural graph.

python Dataset.py

Cross-validation

python train.py

Cite this work

@article{guo2025multi,
  title={Multi-Modality Representation Learning for Antibody-Antigen Interactions Prediction},
  author={Guo, Peijin and Li, Minghui and Pan, Hewen and Huang, Ruixiang and Xue, Lulu and Hu, Shengqing and Guo, Zikang and Wan, Wei and Hu, Shengshan},
  journal={arXiv preprint arXiv:2503.17666},
  year={2025}
}