Latent Dynamics · 2026-11-11
ProTEUS: A mutation-resolved world model for closed-loop protein evolution
Yikun Zhang, Minghao Guo, Guangxue Xu, Peter Yichen Chen, Wojciech Matusik, Wengong Jin, Jingjing Chai, Kaiyi Jiang
AI-assisted summary based on the paper. Consult the linked source for methods and limitations. Benchmark numbers appear after editor verification.
TL;DR
ProTEUS is a mutation-resolved world model for closed-loop protein evolution. Pretrained on deep mutational scanning, it learns a joint latent representation of sequence, evolutionary, and structural information with MutationFormer, where mutations are actions and predicted outcomes guide exploration. It also uses a fast physics-constrained all-atom predictor, and refines via closed-loop experimental feedback. The authors report improved mutation prediction and >10× activity over wild type in two engineering campaigns.
Why it matters
Efficiently exploring large protein mutation spaces is a key bottleneck in protein engineering, where experiments sample only a small fraction of sequence space. ProTEUS aims to connect virtual prediction of mutational outcomes with experimental learning to guide evolution toward desired phenotypes, using mutation-resolved latent actions and fast physics-constrained predictions to enable closed-loop multi-objective engineering.
Method
- Pretrained on deep mutational scanning assays; learns a joint representation of sequence, evolutionary, and structural information using MutationFormer, a mutation-biased attention architecture.
- Represents variants as states in a joint latent space; defines mutations as actions and predicts their functional consequences.
- Develops a fast, physics-constrained all-atom predictor with explicit geometric constraint enforcement; uses closed-loop experimental feedback to refine predictions for protein activity.
Abstract (from arXiv)
Protein engineering requires navigating rugged fitness landscapes, where small sequence changes can produce large functional differences, but experimental screening can only sample a small fraction of sequence space. Thus, a central challenge is to efficiently traverse the large combinatorial space of mutations in silico. We introduce ProTEUS, a mutation-resolved world model for closed-loop protein evolution. Pretrained on deep mutational scanning assays, ProTEUS learns a joint representation of sequence, evolutionary, and structural information using MutationFormer, a mutation-biased attention architecture. Variants are represented as states in this joint latent space, and mutations define actions whose functional consequences are predicted. To support exploration of many mutants, we developed a fast, physics-constrained all-atom predictor that utilizes explicit geometric constraint enforcement to achieve >10 fold increase in speed than AlphaFold3 with guaranteed physical validity. Lastly, ProTEUS refines its prediction and understanding of protein activity through closed loop experimental feedback to guide evolution toward a desired phenotype. ProTEUS surpasses established models in both zero-shot and supervised fine-tuning mutation prediction tasks. We experimentally demonstrate its utility in two closed-loop multi-objective engineering campaigns on aldehyde dehydrogenase and levansucrase, yielding variants with >10 fold higher activity than wild type. ProTEUS provides a world-model framework that connects virtual evaluation of mutational outcomes with experimental learning for efficient protein evolution.
Related papers
- CellCanon: An Agentic Cellular World Model for Prediction, Explanation, and Observation-Driven Updating
- VCLMU: Mechanism-Centric Virtual Cell World Modeling for Perturbation Response
- JANUS: Enhancing Long-Horizon Autonomous Research with Co-Evolving World Models
- The Delta Gap: Cellular World Models for Perturbation Response Prediction Are Far Harder Than Global Correlation Implies
- ALDER: Discovering the Laws of a World by Acting in It
