<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
</head>
<body>
<div class="" style="word-wrap:break-word">
<div class="">This is an announcement of Austin Clyde's Candidacy Exam.</div>
<div class="">===============================================</div>
<div class="">Candidate: Austin Clyde</div>
<div class=""><br class="">
</div>
<div class="">Date: Monday, August 22, 2022</div>
<div class=""><br class="">
</div>
<div class="">Time: 1 pm CST</div>
<div class=""><br class="">
</div>
<div class="">Location: JCL 298</div>
<div class=""><br class="">
</div>
<div class="">Title: Artificial Intelligence and High-Performance Computing for Accelerating Structure-Based Drug Discovery</div>
<div class=""><br class="">
</div>
<div class="">Abstract: Traditional techniques for discovering novel drugs are too slow for 21st challenges, from precision oncology to emerging global pandemics. The COVID-19 Pandemic demonstrated the unequivocal need for rapidly deployable drug discovery
capabilities as a matter of national biopreparedness and biosecurity. The challenge is, though, that drug discovery is an immense and complex interdisciplinary field drawing from cheminformatics, bioinformatics, biophysics, machine learning, and high-performance
computing among others. To accelerate the screening process of new molecules, researchers are applying developments from artificial intelligence (AI) to the problem; however, the direct application of traditional AI methods overlooks the essential complexities
of drug discovery, ranging from protein-conformation flexibility to unique statistical properties of virtual ligand screening. This dissertation presents a unique approach to AI for drug discovery based on tightly integrating insights from biochemistry and
biophysics, driving a more accurate and more interpretable drug discovery system, all while leveraging the same accelerating advances from AI. These cross-cutting contributions from AI and HPC workflows illustrate orders of magnitude speedup for computational
virtual drug screening, novel workflow designs for high-fidelity screening pipelines which are more accurate than traditional docking, new sampling strategies for exploring novel chemotypes, and complementary workflow analysis techniques which directly links
actionable and interpretable goals (such as the design of drugs) with quantitative cost functions. These methodological developments are realized in a case study discovering and validating a novel SARS-CoV-2 3CL-Main Protease inhibitor with a Ki of 2.9 μM
(95% CI 2.2, 4.0).</div>
<div class=""><br class="">
</div>
<div class="">Advisors: Rick Stevens</div>
<div class=""><br class="">
</div>
<div class="">Committee: <span class="" style="color:rgb(60,64,67); font-family:Roboto,Arial,sans-serif; font-size:14px; letter-spacing:0.20000000298023224px; white-space:pre-wrap; background-color:rgb(255,255,255)">Rick Stevens, Eric Jonas, and Arvind Ramanathan</span></div>
<div class=""><span class="" style="color:rgb(60,64,67); font-family:Roboto,Arial,sans-serif; font-size:14px; letter-spacing:0.20000000298023224px; white-space:pre-wrap; background-color:rgb(255,255,255)"><br class="">
</span></div>
<div class=""></div>
</div>
<div class="" style="word-wrap:break-word">
<div class=""></div>
<div class=""><span class="" style="color:rgb(60,64,67); font-family:Roboto,Arial,sans-serif; font-size:14px; letter-spacing:0.20000000298023224px; white-space:pre-wrap; background-color:rgb(255,255,255)"><br class="">
</span></div>
</div>
</body>
</html>