Neural Network–based approach for improving the evaluation of antibody–antigen docking poses

Edoardo Milanetti^1,2*Alessandro Meta^1,2Giancarlo Ruocco^1,2

• ¹Department of Physics, Faculty of Mathematics, Physics, and Natural Sciences, Sapienza University of Rome, Rome, Italy
• ²Istituto Italiano di Tecnologia Center for Life Nano- & Neuro-Science, Rome, Italy

The role of artificial intelligence (AI)–based approaches in computational biology and molecular biophysics has become increasingly central over the past decade; however, many challenges remain unresolved, such as the accurate prediction of protein–protein complexes, the complete solution of which would have a significant impact both on our understanding of cellular mechanisms and on the development of therapeutic and diagnostic strategies. Here, we present a protocol based on multiple minimal neural network (NN)–based approaches, trained on a set of carefully selected physicochemical features, to discriminate docking decoy poses (structurally distant from the experimental complex) from native-like poses (structurally close to the native conformation) within a specific class of biologically relevant protein–protein complexes, namely antibody–antigen systems in which the antigen is a protein. A specific version of the proposed method, trained on a set of antibody–antigen interface de-scriptors, some of which are derived from graph theory to capture the geometric complexity of intermolecular interactions, demonstrates the ability not only to distinguish native-like poses from decoys, as well as, more challengingly, to discriminate intermediate poses from native-like ones, but also to predict the DockQ score, a widely used metric for assessing docking pose quality. Notably, the ability of our NN-based approach, which relies solely on structural interface features, to estimate the DockQ was compared with the docking score provided by the HDOCK method, highlighting its potential as a valuable tool for improving the ranking of antibody–antigen docking poses.