SSnet: A Deep Learning Approach for Protein-Ligand Interaction Prediction

Computational prediction of Protein-Ligand Interaction (PLI) is an important step in the modern drug discovery pipeline as it mitigates cost, time, and resources required to screen novel therapeutics. Deep Neural Networks (DNN) have recently shown excellent performance PLI prediction. However, highly dependent on protein ligand features utilized for DNN model. Moreover, current models, deciphering how determine underlying principles that govern not trivial. In this work, we developed a framework named SSnet utilizes secondary structure information proteins extracted curvature torsion backbone predict PLI. We demonstrate by comparing against variety currently popular machine non-Machine Learning (ML) models using various metrics. visualize intermediate layers show potential latent space proteins, particular extract structural elements model finds influential binding, which one key SSnet. observed our study learns about locations where can bind, including binding sites, allosteric sites cryptic regardless conformation used. further biased any specific molecular interaction extracts fold critical Our work forms gateway general exploration structure-based (DL), just confined protein-ligand interactions, such will large impact research, while being readily accessible de novo designers standalone package.