Inference-time optimization for experiment-grounded protein ensemble generation

TL;DR

This paper introduces an inference-time optimization framework for generating protein conformational ensembles that better match experimental data, overcoming limitations of existing methods by optimizing latent representations and sampling from combined priors.

Contribution

The authors develop a novel inference-time optimization method that improves protein ensemble generation by optimizing latent space and sampling strategies, outperforming current guidance techniques.

Findings

Outperforms state-of-the-art guidance in ensemble diversity and data agreement

Improves physical energy and structural plausibility of generated ensembles

Reveals vulnerabilities in current design metrics related to model confidence

Abstract

Protein function relies on dynamic conformational ensembles, yet current generative models like AlphaFold3 often fail to produce ensembles that match experimental data. Recent experiment-guided generators attempt to address this by steering the reverse diffusion process. However, these methods are limited by fixed sampling horizons and sensitivity to initialization, often yielding thermodynamically implausible results. We introduce a general inference-time optimization framework to solve these challenges. First, we optimize over latent representations to maximize ensemble log-likelihood, rather than perturbing structures post hoc. This approach eliminates dependence on diffusion length, removes initialization bias, and easily incorporates external constraints. Second, we present novel sampling schemes for drawing Boltzmann-weighted ensembles. By combining structural priors from AlphaFold3 with force-field-based priors, we sample from their product distribution while balancing experimental likelihoods. Our results show that this framework consistently outperforms state-of-the-art guidance, improving diversity, physical energy, and agreement with data in X-ray crystallography and NMR, often fitting the experimental data better than deposited PDB structures. Finally, inference-time optimization experiments maximizing ipTM scores reveal that perturbing AlphaFold3 embeddings can artificially inflate model confidence. This exposes a vulnerability in current design metrics, whose mitigation could offer a pathway to reduce false discovery rates in binder engineering.