June 10, 2026 – Protai, an AI-driven drug discovery company, today announced two complementary preprints that together outline a new approach to structure-based drug design for induced proximity therapeutics. The first introduces AIMS-Fold, a structural-proteomics-guided generative AI framework for protein complex modeling. The second demonstrates the AIMS™ platform through the design of a potent, bioavailable, in vivo–validated KAT6A degrader. Together, the works highlight the use of structural proteomics as an input to generative AI systems, aimed at improving modeling of protein complexes relevant to PROTACs, molecular glues, and other induced proximity modalities.
AIMS-Fold integrates experimental structural proteomics data directly into pretrained AI structure prediction models. Unlike leading approaches such as AlphaFold 3 and Boltz-2, which rely primarily on sequence-derived and evolutionary signals to infer static structures, AIMS-Fold incorporates experimental measurements to guide generative sampling toward conformations consistent with biological and biophysical constraints.
The system incorporates two experimental modalities: XL-MS (Cross-Linking Mass Spectrometry) spatial restraints and HDX-MS (Hydrogen-Deuterium Exchange Mass Spectrometry) solvent accessibility profiles which are translated into differentiable physical potentials applied during inference. The first preprint, “Co-folding model guided by structural proteomics”, outlines how this approach allows the model to better reflect multi-state and flexible conformations in protein complexes, which are often central to induced proximity biology. Across benchmark evaluations, AIMS-Fold showed improved accuracy on challenging protein complexes and induced proximity targets compared to unguided state-of-the-art models, suggesting that integrating structural proteomics data can improve performance in settings where sequence-only models are limited.
The second preprint, “Leveraging AI and structural proteomics for rational design of a KAT6A degrader,” demonstrates the application of the AIMS™ platform to a PROTAC discovery program targeting KAT6A, a clinically validated epigenetic regulator for the treatment of solid tumors. By integrating structural proteomics with computational modeling, this study shows how different AIMS™ modules address various aspects of PROTAC optimization. This approach enabled a more efficient development timeline and reduced costs, resulting in a best-in-class degrader with optimized potency and bioavailability.
“At the core of this work is a shift in how we think about structure prediction for drug discovery,” said Eran Seger, CEO and Co-founder at Protai. “Rather than relying only on static, sequence-derived representations, we are integrating experimental structural proteomics directly into generative models, allowing us to guide predictions toward more biologically meaningful conformations.”
About Protai
Protai is an AI drug discovery company unlocking the therapeutic potential of protein complexes through structural proteomics. Unlike standard AI models that rely on static predictions, Protai's AIMS™ platform maps dynamic protein interactions using proprietary experimental data, to model the function of protein complexes in the disease native state. Protai is leveraging this engine to advance an innovative internal drug pipeline, led by a best-in-class KAT6A degrader for the treatment of breast cancer and other solid tumors.
