RosettaSearch: Multi-Objective Inference-Time Search for Protein Sequence Design

2026-04-18

AI Summary

Current single-pass decoding approaches for protein sequence design struggle to simultaneously achieve high fidelity in both structure and sequence, limiting design success rates. This work proposes a novel multi-objective search framework that, for the first time, integrates large language models (e.g., o4-mini, Gemini-3) as generative optimizers and incorporates vision-language models to provide multimodal feedback based on structural images. By synergistically combining RosettaFold3 for structure prediction with ProteinMPNN and LigandMPNN for sequence generation, the framework enables a controllable balance between exploration and exploitation. Evaluated on 400 suboptimal sequences, the method improves structural fidelity by 18%–68% and increases design success rates by 2.5-fold, demonstrating strong generalization across independent test sets and de novo backbones.

Technology Category

• Search and Optimization: Learning to Search
• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: Language Grounding & Multi-modal NLP

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Economics, Online Markets and Human Computation: Uses of LLMs and GenAI for marketplace design, bidding, and strategic interactions
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

We introduce RosettaSearch, an inference-time multi-objective optimization approach for protein sequence optimization. We use large language models (LLMs) as a generative optimizer within a search algorithm capable of controlled exploration and exploitation, using rewards computed from RosettaFold3, a structure prediction model. In a large-scale evaluation, we apply RosettaSearch to 400 suboptimal sequences generated by LigandMPNN (a state-of-the-art model trained for protein sequence design), recovering high-fidelity designs that LigandMPNN's single-pass decoding fails to produce. RosettaSearch's designs show improvements in structural fidelity metrics ranging between 18\% to 68\%, translating to a 2.5$\times$ improvement in design success rate. We observe that these gains in success rate are robust when RosettaSearch-designed sequences are evaluated with an independent structure prediction oracle (Chai-1) and generalize across two distinct LLM families (o4-mini and Gemini-3), with performance scaling consistently with reasoning capability. We further demonstrate that RosettaSearch improves sequence fidelity for ProteinMPNN-designed sequences on \textit{de novo} backbones from the Dayhoff atlas, showing that the approach generalizes beyond native protein structures to computationally generated backbones. We also demonstrate a multi-modal extension of RosettaSearch with vision-language models, where images of predicted protein structures are used as feedback to incorporate structural context to guide protein sequence generation. The sequence trajectories generated by our approach can be used as training data in sequence design models or in post-training and will be released along with the code and datasets upon publication.

Problem

• protein sequence design
• structural fidelity
• multi-objective optimization
• inference-time search
• sequence fidelity

Innovation

• RosettaSearch
• multi-objective optimization
• protein sequence design
• large language models
• structure-guided generation