Protein mutation analysis improved by separating structure and properties
RipplePLM: Structural and Property Decoupling for Protein Mutation Effect Generation
Computational Engineering, Finance, and Science
Summary
Figuring out the effects of mutations in proteins helps scientists understand biology and design medicines. Existing methods often mix all mutation information together, missing how structure and chemical properties separately impact the effect. The authors created RipplePLM, which organizes mutation data into nearby and distant parts of the protein's structure and adds expert knowledge about chemical properties. This approach gives clearer, more accurate descriptions of mutation effects in plain language. Experts found RipplePLM's explanations more biologically relevant than earlier models.
What this means in practice
- •For biotech engineers: Generate detailed natural language summaries of protein mutations that distinguish structural and biochemical impacts to guide experimental design.
- •For computational drug developers: Create interpretable mutation effect reports to support drug discovery by better understanding mutation consequences on protein function.
Authors
Liuzhenghao Lv, Yuyang Liu, Yuyang Gao, Li Yuan, Yonghong Tian
Abstract
Protein mutation effect generation asks a model to describe the functional consequence of a point mutation in natural language. Existing protein-to-text systems typically encode mutation information into undifferentiated representations, overlooking the organization of mutation-induced evidence across structural and biochemical factors. We propose RipplePLM, a mutation-aware generation framework centered on Direct-Distal Cross-Attention (DDCA). By constructing a residue-level Mutation Perturbation Field from pre-trained protein language models, DDCA leverages predicted contact maps to organize mutation representations into two pathways: the mutation site's immediate contact neighborhood and its multi-hop distal context. To complement this structural decomposition, we further introduce the Property Latent Chain (PLChain), which injects expert-guided supervision of biochemical property changes (e.g., thermostability and optimal pH) into the LLM hidden-state pathway through latent property tokens. On MutaDescribe, RipplePLM improves over mutation-specific baselines on temporal and structural splits; under a matched-backbone comparison, average structural-split ROUGE-L increases from {22.23} to {35.65}. Expert evaluation further shows a higher proportion of biologically accurate or relevant descriptions than the mutation-specific baseline. Additional ablations, representation diagnostics, and low-$N$ fitness regression experiments further support the effectiveness of the learned mutation-aware representations. Code: https://github.com/Lyu6PosHao/RipplePLM.