EvoFlows: Evolutionary Edit-Based Flow-Matching for Protein Engineering

2026-03-12

AI Summary

Traditional protein engineering approaches struggle to perform controllable and non-trivial sequence edits on template proteins while preserving their native-like properties. This work proposes a variable-length sequence-to-sequence modeling framework based on edit flows, which uniquely integrates edit operations with flow matching to jointly predict both the location and type of mutations by learning evolutionary trajectories among related proteins. By combining evolutionary information with a controllable editing mechanism, the method diverges from conventional autoregressive or masked language models. Evaluated on UniRef and OAS datasets, it demonstrates sequence distribution modeling capabilities comparable to state-of-the-art masked language models while significantly improving the generation of natural yet structurally novel protein variants.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Computer Vision: Diffusion Models for Vision
• Natural Language Processing: (Large) Language Models

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

We introduce EvoFlows, a variable-length sequence-to-sequence protein modeling approach uniquely suited to protein engineering. Unlike autoregressive and masked language models, EvoFlows perform a limited, controllable number of insertions, deletions, and substitutions on a template protein sequence. In other words, EvoFlows predict not only _which_ mutation to perform, but also _where_ it should occur. Our approach leverages edit flows to learn mutational trajectories between evolutionarily-related protein sequences, simultaneously modeling distributions of related natural proteins and the mutational paths connecting them. Through extensive _in silico_ evaluation on diverse protein communities from UNIREF and OAS, we demonstrate that EvoFlows capture protein sequence distributions with a quality comparable to leading masked language models commonly used in protein engineering, while showing improved ability to generate non-trivial yet natural-like mutants from a given template protein.

Problem

• protein engineering
• sequence-to-sequence modeling
• edit-based mutations
• variable-length sequences
• mutational trajectories

Innovation

• edit-based modeling
• protein engineering
• evolutionary trajectories
• sequence-to-sequence generation
• mutational paths