University of New Mexico,
Keywords: Organic Cofactors, Binding Sites, Deep Learning
Summary:In structure-based drug design, identifying potential binding sites on proteins is a crucial step. Deep learning algorithms, when combined with existing knowledge on protein-ligand binding, can significantly improve the computational detection of ligands and cofactors binding cavities. In this talk, I will present a deep learning approach for efficiently detecting organic cofactor binding sites on proteins. Our method leverages the three-dimensional structure of a protein to create a surface point cloud. This point cloud is then combined with sparse voxels using an octree data structure and a feature vector comprising chemical and geometric details, which enables binary classification of candidate binding sites on the protein’s surface.