Drug Discovery

Drug discovery is a complex, expensive, and time-consuming process. It involves the identification of a target, screening of a large number of molecules, and optimization of the selected molecules. The process of drug discovery can be accelerated by using machine learning and deep learning. I am interested in solving problems in drug discovery using deep learning. I am also interested in developing new deep learning models for drug discovery.

Graduate Thesis

My graduate thesis was on the application of deep learning to drug discovery. I worked on the the problem of building pharmacophore model for a given molecule (COVID-19 Mpro). The pharmacophore model of a molecule is a set of points in 3D space that represent the important features of the molecule. The pharmacophore model is used to screen a large number of molecules to find the molecules that are promising candidates for drug discovery. I used a deep learning model to build the pharmacophore model of a molecule. My approach contains following steps:

• Using a deep learning model to predict the binding affinity of a molecule to a target protein (Mpro).
• Using the predicted binding affinity to find the binding site of the molecule on the target protein.
• Using the binding site to build the pharmacophore model of the molecule.
• Using the pharmacophore model to screen a large number of molecules to find the molecules that are promising candidates for drug discovery and build a recommender system for drug discovery.

My built system is available at Link.
Account: Test
Password: test
Or you can register a new account.

Thesis PDF: Link
Thesis Presentation: Link
Thesis Code: Link

Publications

• Duc Q. Nguyen, et al. “Towards de Novo Drug Design for the Coronavirus: A Drug-Target Interaction Prediction Approach Using Atom-Enhanced Graph Neural Network with Multi-Hop Gating Mechanism” In Proceedings of 2022 9th Nafosted Conference on Information and Computer Science (2022). Best Paper Award Link