BioMatrix: Towards a Comprehensive Biological Foundation Model Spanning the Modality Matrix of Sequences, Structures, and Language

2026-06-20

AI Summary

This work addresses the challenge that existing biological foundation models struggle to natively unify cross-modal (sequence, structure, language) and cross-entity (molecules, proteins) representations, often relying on external encoders or task-specific output heads. The authors propose the first decoder-only architecture that leverages a shared discrete token space to jointly represent SMILES/SELFIES, protein sequences, 3D structures, and scientific text. Built upon Qwen3 (1.7B/4B), the model undergoes continued pretraining on 304.4 billion tokens using only next-token prediction, enabling native multimodal understanding and generation without auxiliary modules. Evaluated across six categories encompassing 80 tasks, the model achieves state-of-the-art or competitive performance on 77 tasks, marking the first demonstration of end-to-end, module-free unified multimodal modeling and cross-modal generation in the biological domain.

Abstract

We present BioMatrix, the first multimodal foundation model that natively integrates sequences, structures, and natural language for both molecules and proteins within a single decoder-only architecture. Existing biological foundation models pursue native multimodality and broad entity coverage separately: those that fuse multiple modalities under a shared objective remain confined to a single entity type, while those spanning multiple entity types either omit explicit structural modeling or rely on adapter-based designs in which the model cannot natively generate the very modalities it can read. BioMatrix closes this gap by mapping molecular sequences (supporting both SMILES and SELFIES notations), molecular structures, protein sequences, protein structures, and natural language into a shared discrete token space through a unified tokenization scheme, so that all modalities are consumed and produced uniformly under a single next-token prediction objective -- without external encoders, projection adapters, or modality-specific output heads. Built upon the Qwen3 language model (1.7B and 4B), BioMatrix is continually pretrained on 304.4 billion tokens spanning general and domain-specific text, sequence and structure views of molecules and proteins, and cross-modal corpora that interleave biomolecular entities with scientific text and link distinct entities through molecule-protein and protein-protein interaction data. After tuning on a comprehensive suite of downstream applications covering 80 tasks across 6 categories -- encompassing single-entity and multi-entity understanding and generation tasks across and within modalities -- BioMatrix achieves state-of-the-art or competitive performance on 77 out of 80 tasks, demonstrating that a single, natively multimodal generalist model can effectively match or surpass specialized approaches across a wide range of biological tasks.

Problem

• multimodal foundation model
• biological entities
• sequence-structure-language integration
• native modality generation
• unified tokenization

Innovation

• multimodal foundation model
• native modality integration
• unified tokenization
• structure-aware generation
• biological generalist model