ODesign: A World Model for Biomolecular Interaction Design
Abstract
Biomolecular interactions underpin almost all biological processes, and their rational design is central to programming new biological functions. Generative AI models have emerged as powerful tools for molecular design, yet most remain specialized for individual molecular types and lack fine-grained control over interaction details. Here we present ODesign, an all-atom generative world model for all-to-all biomolecular interaction design. ODesign allows scientists to specify epitopes on arbitrary targets and generate diverse classes of binding partners with fine-grained control. Across entity-, token-, and atom-level benchmarks in the protein modality, ODesign demonstrates superior controllability and performance to modality-specific baselines. Extending beyond proteins, it generalizes to nucleic acid and small-molecule design, enabling interaction types such as protein-binding RNA/DNA and RNA/DNA-binding ligands that were previously inaccessible. By unifying multimodal biomolecular interactions within a single generative framework, ODesign moves toward a general-purpose molecular world model capable of programmable design. ODesign is available at https://odesign.lglab.ac.cn ,
Cite
@article{arxiv.2510.22304,
title = {ODesign: A World Model for Biomolecular Interaction Design},
author = {Odin Zhang and Xujun Zhang and Haitao Lin and Cheng Tan and Qinghan Wang and Yuanle Mo and Qiantai Feng and Gang Du and Yuntao Yu and Zichang Jin and Ziyi You and Peicong Lin and Yijie Zhang and Yuyang Tao and Shicheng Chen and Jack Xiaoyu Chen and Chenqing Hua and Weibo Zhao and Runze Ma and Yunpeng Xia and Kejun Ying and Jun Li and Yundian Zeng and Lijun Lang and Peichen Pan and Hanqun Cao and Zihao Song and Bo Qiang and Jiaqi Wang and Pengfei Ji and Lei Bai and Jian Zhang and Chang-yu Hsieh and Pheng Ann Heng and Siqi Sun and Tingjun Hou and Shuangjia Zheng},
journal= {arXiv preprint arXiv:2510.22304},
year = {2025}
}