RESEARCH PROJECTS

Development of a physics-based deep learning network to calculate binding free energies of small protein–ligand complexes in an implicit solvent model. We build our model based on a deep learning framework called Deepchem, a Python library mainly used in drug discovery processes.

Deep generative models have recently been introduced for sampling the chemical space and discovering new ligands. Unlike experimental methods, the computational pipeline enables researchers to modify the generative algorithm quickly and effortlessly test the results from different perspectives. This paradigm shift opens up new opportunities to create a rich database of artificial structures and evaluate their synthesizability and viability.

Evaluation of the potential of a molecular mechanics generalized Born surface area (MM/GBSA) approach to estimate the binding free energy between the SARS-CoV-2 spike receptor-binding domain (wild type and mutants) and the human ACE2 receptor.

Adaptive steered molecular dynamics (ASMD) is a computational biophysics method in which an external force is applied to a selected set of atoms or a specific reaction coordinate to induce a particular molecular motion. This research proposes a novel method to guide ASMD with optimal trajectories collected from human experiences through experiments in virtual reality.

Optimization of atomic radii in an implicit solvent model to obtain close agreement with experimental results in terms of calculating electrostatic binding free energies. The underlying massively parallel optimization method, VTDIRECT95, runs on Virginia Tech clusters.

Introduction of a novel geometrical metric to identify protein binding pockets. The proposed metric is based on the von Neumann entropy of the weighted Delaunay triangulation of protein pockets.