Degree
Doctor of Philosophy (PhD)
Department
Computer Science & Engineering
Document Type
Dissertation
Abstract
This dissertation introduces an advanced computational pipeline aimed at accelerating drug discovery for both antiviral and antibiotic agents. This research spans multiple projects, including the development of eVir, a machine learning tool that repurposes existing drugs to target viral infections like COVID-19. eVir analyzes interactions in the human protein network, using graph embedding techniques and Siamese networks to predict drug efficacy. Laboratory studies confirmed that compounds such as Mebendazole show significant promise against SARS-CoV-2, both in vitro and in vivo. Additionally, this work introduces a fragment-based approach to antibiotic drug design using tools such as eFilter, eSynth2, and eMolFrag2. These tools decompose organic compounds into fragments, synthesize new compounds, and predict pharmacokinetic properties to address the growing threat of antimicrobial resistance. Experiments validated eFilter’s ability to recreate historical discoveries, such as penicillin derivatives, and identified new hybrid molecules with enhanced therapeutic potential. The study also demonstrated that the molecular decomposition and synthesis performed by eMolFrag2 and eSynth2 is a sound and complete process, reconstructing known antibiotics and exploring novel chemical spaces. By leveraging AI and machine learning, this work showcases how integrating computational techniques with medicinal chemistry expertise can lead to the discovery of novel antiviral and antibiotic candidates, providing a cost-effective and scalable solution to global healthcare challenges.
Date
11-1-2024
Recommended Citation
Bess, Adam, "Graph-Based AI Approaches for Revolutionizing Drug Discovery: The DeepDrug Approach" (2024). LSU Doctoral Dissertations. 6649.
https://repository.lsu.edu/gradschool_dissertations/6649
Committee Chair
Mukhopadhyay, Supratik
Included in
Artificial Intelligence and Robotics Commons, Biochemical and Biomolecular Engineering Commons, Bioinformatics Commons, Biotechnology Commons, Data Science Commons, Medical Biochemistry Commons, Medical Biotechnology Commons