Project Summary/Abstract High performance computing (HPC) has become an indispensable resource for scientific research in diverse fields including deep learning, bioinformatics, and molecular simulation. In the parent award (Grant No.: 1R16GM146633), Dr. Negin Forouzesh (PI) proposed to design, develop, and test a Physics-Guided Neural Network (PGNN) model for improving the accuracy of protein-ligand binding free energy using implicit solvent models. Recent results on more than 300 protein-ligand complexes demonstrate that the proposed PGNN model can successfully improve the “accuracy” of the pure physics-based model. In addition, the “interpretability” and “transferability” of the model have been boosted compared to the purely data-driven model. In this proposal, the PI requests the purchase of a workstation to (1) run molecular dynamics (MD) simulations, (2) train deep learning models, and (3) run a massively parallel implementation of an optimization algorithm. The requested storage is for saving protein-ligand structures, MD trajectories, source codes, student theses, and manuscripts. The new equipment will enable the student assistants to run HPC simulations efficiently and will improve the quality and extent of research in the Biomolecular area at Cal State LA.