Rajeshwar et al. (2026) Biophysical Journal

Allostery is the phenomenon whereby a binding event or covalent modification at one site in a protein modulates function at a distal site, thus changing a protein’s functional state. As such, it is a ubiquitous aspect of protein functional regulation. Computationally predicting allosteric states is important as part of the broader challenge of functional annotation, but it also has practical implications for drug development, as targeting an allosteric site often affords greater specificity compared with targeting an orthosteric site. This study introduces a machine learning approach to predict the allosteric functional state using the small G-protein KRas as the model system, due to its implication in many types of cancer and being well studied as a result with many x-ray crystallographic structures of KRas available with different mutations and ligands bound. Using structural and dynamical features that can be cast as images, namely interatomic distances, contact maps, covariance, and mutual information, supervised learning was performed using convolutional neural networks. Two pretrained convolutional neural network architectures, GoogLeNet and ResNet18, were fine-tuned to classify KRas into active or inactive states based on these features. Across training regimes, atomic contact maps emerged as the most effective structural feature, whereas linearized mutual information outperformed covariance in capturing dynamical correlations relevant to allostery. Models achieved significant validation accuracy, with atomic contact maps yielding up to 90% accuracy. The findings suggest that integrating global structural rearrangements and correlated motion patterns with deep learning can reliably predict protein allosteric states, offering a promising framework for understanding allosteric regulation and developing targeted therapeutics.

Significance: Proteins are molecular machines that exist in different functional states modulated by noncovalent or covalent interactions in a process known as allostery. Predicting how such interactions modulate the functional states of proteins has far-reaching implications from a basic-science and therapeutic standpoint. Predicting protein functional states computationally is challenging due to the vast range of timescales and dynamical and structural signatures involved, rendering their classification by a human difficult. However, such signatures can be cast as images amenable to classification by convolutional neural networks (CNNs). Therefore, we have developed CNNs trained on dynamical and structural features to classify a protein’s functional state., focusing on KRas, a small G-protein that has been experimentally characterized extensively due to its implication in many cancers.

Rajeshwar RT, Lagergren JH, Smith JC, Demerdash ONA. (2026) Allosteric Prediction Via Convolutional Neural Networks and Protein Structural and Dynamical Features. Biophysical Journal. DOI:10.1016/j.bpj.2025.12.011