AI Insight
Researchers developed AIMS-Fold, a computational framework that combines AI-based protein structure prediction with experimental data from structural proteomics techniques (XL-MS and HDX-MS) to improve modeling of protein complexes. The method uses guided diffusion to incorporate experimental spatial restraints and solvent accessibility measurements during the structure generation process. AIMS-Fold outperforms existing computational models like Boltz-2 on challenging protein complex predictions, particularly for induced proximity targets relevant to drug design.
Why it matters
This approach addresses a critical gap in structure-based drug design by improving predictions for protein complexes that change shape upon binding, which is essential for developing antibodies, PROTACs, and other therapeutics that work by bringing proteins together. By integrating experimental data with AI models, the framework provides more accurate structural predictions for complex drug design applications.
arXiv:2605.26192v1 Announce Type: cross
Abstract: Protein structure generative models excel at predicting single protein static structures from sequence, but routinely fail to capture the correct conformational state of protein complexes, critical for protein design and induced proximity modalities such as antibodies and PROTACs. While structural proteomics techniques like Cross-Linking Mass Spectrometry (XL-MS) and Hydrogen-Deuterium Exchange (HDX-MS) offer valuable spatial and dynamic insights, integrating these sparse, heterogeneous measurements into these models remains an open challenge. Here, we bridge this gap by combining structural proteomics data with the rich biophysical priors learned by pretrained diffusion models. We introduce AIMS-Fold, an inference-time guided-diffusion framework that actively steers the generative sampling trajectory using differentiable physical potentials derived from XL-MS spatial restraints and HDX-MS solvent accessibility profiles. We demonstrate that these structural methods individually enhance predictive accuracy, and their integration yields synergistic improvement. Crucially, by leveraging these experimental restraints, AIMS-Fold achieves higher accuracy on challenging induced proximity targets than purely computational, unguided state-of-the-art models like Boltz-2. This establishes our framework as a powerful, integrative computational approach for the structure based drug design of induced proximity drugs. Evaluation code will be made publicly available upon publication.