Project Description
In this project, we will attempt to develop a algorithm to predict the binding mode and potency of covalent ligands. We will start by a set of particular targets. These targets include Bruton's Tyrosine Kinase, Kirsten rat sarcoma viral oncogene homolog (KRAS) and Cathepsin. Initially, our focus will be on BTK which causes isolated growth hormone deficiency type III. Currently, there are no efficient inhibitor or drug for BTK as the existing ones suffers from toxicity.
Tasks and Responsibilites
The student will first use AutoDock Vina to dock the existing ligands found in the crystal structures and will then use Schrodinger covalent docking suite to reproduce the binding modes in these PDBs. Next, they will then explore different parameters with these docking suites, primarily the box size to dock fragments of the ligands. Once that successfully achieved, they will explore that how we can connect those fragments with chemical scaffolds to reproduce the binding modes. The student will study in the interaction of each of these ligands with the surrounding residues and define them as features and will then develop a simple machine learning algorithm to encode these features. Finally, we will run this machine learning program to predict the binding mode of an unknown protein-ligand complex.