Robust Biomolecular Complex Design Across Protein Conformational Landscapes

2026-09-27

AI Summary

This study addresses the failure of single-conformation protein design across multiple native conformations by proposing FlexEvo, a framework that enhances compatibility with unseen conformations through inference-time evolutionary adaptation without retraining or conformational ensembles. Core innovations include a model-agnostic evolutionary algorithm, FlexBox geometric partitioning, all-atom representations, and a Pareto selection mechanism enabling dual-objective optimization of geometric constraints from mono-state to multi-state regimes. Experimental results demonstrate that FlexEvo substantially reduces cross-conformation performance degradation from 47.8% to 4.4%, while introducing only 1.4 to 3.1 minutes of additional computational overhead per sample, thereby achieving both high efficiency and strong generalizability.

Abstract

Proteins populate conformational ensembles, yet structure-based biomolecular design typically optimizes candidates against a single target conformation. Consequently, a candidate that fits one state can lose favorable interactions or develop steric clashes when the target adopts another. We introduce FlexEvo, a model-agnostic evolutionary framework that adapts candidates once at inference time from a single target conformation to improve compatibility with alternative natural conformations unseen during adaptation, without retraining the source model or requiring a conformational ensemble. FlexEvo casts cross-state adaptation as geometry-constrained bi-objective optimization, balancing preservation of input-state interactions against robustness to plausible conformational perturbations. To limit the search space and reduce invalid structural edits, geometry-derived FlexBoxes define protected anchor regions, adaptable regions for local exploration, and forbidden regions for clash avoidance. A unified all-atom representation supports topology-preserving adaptation across diverse binder categories, while Pareto selection preserves nondominated candidates across the two objectives. We evaluate FlexEvo across multiple generation baselines and nine representative binder categories spanning diverse molecular sizes and structural topologies. FlexEvo reduces the category-balanced mean relative performance degradation from 47.8% to 4.4%, while adding only 1.4--3.1 minutes of adaptation per sample. These results establish single-state inference-time adaptation as a practical route toward robust biomolecular complex design across protein conformational landscapes.

Problem

• biomolecular design
• protein conformational ensembles
• conformational robustness
• steric clashes

Innovation

• inference-time adaptation
• bi-objective optimization
• conformational landscape
• FlexBoxes
• Pareto selection