<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
</head>
<body style="word-wrap: break-word; -webkit-nbsp-mode: space; line-break: after-white-space;" class="">
<font face="Calibri, Helvetica, sans-serif" class=""><span class="" style="font-size: 14.666666984558105px;">This is an announcement of Matthew McPartlon’s Dissertation Defense</span></font>
<div class=""><span class="" style="font-family: Calibri, Helvetica, sans-serif; font-size: 14.666666984558105px;">===============================================</span>
<div class="ContentPasted0" style="font-family: Calibri, Helvetica, sans-serif; font-size: 14.666666984558105px;">
Candidate: Matthew McPartlon</div>
</div>
<div class="ContentPasted0" style="font-family: Calibri, Helvetica, sans-serif; font-size: 14.666666984558105px;">
<br class="">
</div>
<div class="ContentPasted0" style="font-family: Calibri, Helvetica, sans-serif; font-size: 14.666666984558105px;">
Date: Monday, May 22, 2023</div>
<div class="ContentPasted0" style="font-family: Calibri, Helvetica, sans-serif; font-size: 14.666666984558105px;">
<br class="">
</div>
<div class="ContentPasted0" style="font-family: Calibri, Helvetica, sans-serif; font-size: 14.666666984558105px;">
Time: 9:00 am CST</div>
<div class="ContentPasted0" style="font-family: Calibri, Helvetica, sans-serif; font-size: 14.666666984558105px;">
<br class="">
</div>
<div class="ContentPasted0" style="font-family: Calibri, Helvetica, sans-serif; font-size: 14.666666984558105px;">
Location: JCL 223</div>
<div class="ContentPasted0" style="font-family: Calibri, Helvetica, sans-serif; font-size: 14.666666984558105px;">
<br class="">
</div>
<div class="ContentPasted0" style="font-family: Calibri, Helvetica, sans-serif; font-size: 14.666666984558105px;">
Remote Location: <a href="https://urldefense.com/v3/__https://www.google.com/url?q=https:**Auchicago.zoom.us*j*95179647999*pwd*3DQ1BwTndCTDF0MjhnZUFhOVNMbi9UZz09&sa=D&source=calendar&ust=1684785610252984&usg=AOvVaw3_Q2IwAa6KfKHdAwrN6uLu__;Ly8vLz8l!!BpyFHLRN4TMTrA!4xAcY7OaukpCFsdYiIyDzeLFt0P5KoejonxyJPCy30Ujt74DHn9Q1OQ6ZqpEk0mmyBemLWxW4mjsJt_tiDlLLL38oP6dt_Rf$" target="_blank" class="" style="font-family: Roboto, Arial, sans-serif; font-size: 14px; letter-spacing: 0.2px; white-space: pre-wrap;">https://uchicago.zoom.us/j/95179647999?pwd=Q1BwTndCTDF0MjhnZUFhOVNMbi9UZz09</a></div>
<span class="" style="color: rgb(60, 64, 67); font-family: Roboto, Arial, sans-serif; font-size: 14px; letter-spacing: 0.2px; white-space: pre-wrap;">Meeting ID: 951 7964 7999 Passcode: 917327</span>
<div class="ContentPasted0" style="font-family: Calibri, Helvetica, sans-serif; font-size: 14.666666984558105px;">
<br class="">
</div>
<div class="ContentPasted0" style="font-family: Calibri, Helvetica, sans-serif; font-size: 14.666666984558105px;">
Dissertation Paper Title: Accelerating Protein Design with Deep Learning</div>
<div class="ContentPasted0" style="font-family: Calibri, Helvetica, sans-serif; font-size: 14.666666984558105px;">
<br class="">
</div>
<div class="ContentPasted0"><font face="Calibri, Helvetica, sans-serif" class="" style="font-size: 14.666666984558105px;">Abstract: </font><span class="" style="white-space: pre-wrap; font-size: 15px;">The human proteome comprises tens of thousands of proteins,
each tailored for a specific</span></div>
<div class=""><font color="#000000" class="" style="font-size: 15px;"><span role="presentation" dir="ltr" class="" style="box-sizing: border-box; white-space: pre-wrap;">function by the selective pressures of evolution. The field of protein design seeks to
develop</span><br role="presentation" class="" style="box-sizing: border-box; white-space: pre-wrap;">
<span role="presentation" dir="ltr" class="" style="box-sizing: border-box; white-space: pre-wrap;">proteins with new or enhanced functions at will, ultimately bypassing the evolutionary clock.</span><br role="presentation" class="" style="box-sizing: border-box; white-space: pre-wrap;">
<span role="presentation" dir="ltr" class="" style="box-sizing: border-box; white-space: pre-wrap;">In this thesis, we introduce general machine-learning methods for accelerating protein design,</span><br role="presentation" class="" style="box-sizing: border-box; white-space: pre-wrap;">
<span role="presentation" dir="ltr" class="" style="box-sizing: border-box; white-space: pre-wrap;">focusing on modeling protein structure.</span></font></div>
<div class=""><font color="#000000" class="" style="font-size: 15px;"><br role="presentation" class="" style="box-sizing: border-box; white-space: pre-wrap;">
<span role="presentation" dir="ltr" class="" style="box-sizing: border-box; white-space: pre-wrap;">First, we propose an approach for fixed-backbone design (Chapter 2), the problem of</span><br role="presentation" class="" style="box-sizing: border-box; white-space: pre-wrap;">
<span role="presentation" dir="ltr" class="" style="box-sizing: border-box; white-space: pre-wrap;">designing primary sequence and side-chain rotamers for a given backbone conformation.</span><br role="presentation" class="" style="box-sizing: border-box; white-space: pre-wrap;">
<span role="presentation" dir="ltr" class="" style="box-sizing: border-box; white-space: pre-wrap;">Whereas classic approaches formulate sequence and rotamer design tasks separately, we offer</span><br role="presentation" class="" style="box-sizing: border-box; white-space: pre-wrap;">
<span role="presentation" dir="ltr" class="" style="box-sizing: border-box; white-space: pre-wrap;">a single approach to predict both simultaneously. To realize this, we develop a deep neural</span><br role="presentation" class="" style="box-sizing: border-box; white-space: pre-wrap;">
<span role="presentation" dir="ltr" class="" style="box-sizing: border-box; white-space: pre-wrap;">network that effectively leverages backbone coordinates. By exploiting backbone geometry,</span><br role="presentation" class="" style="box-sizing: border-box; white-space: pre-wrap;">
<span role="presentation" dir="ltr" class="" style="box-sizing: border-box; white-space: pre-wrap;">we can efficiently represent atomic microenvironments at the coordinate level and ultimately</span><br role="presentation" class="" style="box-sizing: border-box; white-space: pre-wrap;">
<span role="presentation" dir="ltr" class="" style="box-sizing: border-box; white-space: pre-wrap;">avoid discrete rotamer sampling. This results in more robust designs and accurate quality</span><br role="presentation" class="" style="box-sizing: border-box; white-space: pre-wrap;">
<span role="presentation" dir="ltr" class="" style="box-sizing: border-box; white-space: pre-wrap;">estimates for downstream tasks.</span></font></div>
<div class=""><font color="#000000" class="" style="font-size: 15px;"><span class="" style="box-sizing: border-box; white-space: pre-wrap;"><br role="presentation" class="" style="box-sizing: border-box;">
</span><span role="presentation" dir="ltr" class="" style="box-sizing: border-box; white-space: pre-wrap;">Next, we introduce a framework for flexible protein-protein docking (Chapter 3), the</span><span class="" style="box-sizing: border-box; white-space: pre-wrap;"><br role="presentation" class="" style="box-sizing: border-box;">
</span><span role="presentation" dir="ltr" class="" style="box-sizing: border-box; white-space: pre-wrap;">task of determining the structure of a protein complex given the unbound structures of its</span><span class="" style="box-sizing: border-box; white-space: pre-wrap;"><br role="presentation" class="" style="box-sizing: border-box;">
</span><span role="presentation" dir="ltr" class="" style="box-sizing: border-box; white-space: pre-wrap;">components. Traditional docking methods are limited by their reliance on empirical physics-</span><span class="" style="box-sizing: border-box; white-space: pre-wrap;"><br role="presentation" class="" style="box-sizing: border-box;">
</span><span role="presentation" dir="ltr" class="" style="box-sizing: border-box; white-space: pre-wrap;">based scoring functions, inability to accommodate conformational flexibility, and failure to</span><span class="" style="box-sizing: border-box; white-space: pre-wrap;"><br role="presentation" class="" style="box-sizing: border-box;">
</span><span role="presentation" dir="ltr" class="" style="box-sizing: border-box; white-space: pre-wrap;">incorporate information on binding sites. To address these challenges, we propose an end-to-</span><span class="" style="box-sizing: border-box; white-space: pre-wrap;"><br role="presentation" class="" style="box-sizing: border-box;">
</span><span role="presentation" dir="ltr" class="" style="box-sizing: border-box; white-space: pre-wrap;">end approach that can model conformational changes and target specific interactions while</span><span class="" style="box-sizing: border-box; white-space: pre-wrap;"><br role="presentation" class="" style="box-sizing: border-box;">
</span><span role="presentation" dir="ltr" class="" style="box-sizing: border-box; white-space: pre-wrap;">significantly reducing computational time. As one of the pioneering deep learning methods</span><span class="" style="box-sizing: border-box; white-space: pre-wrap;"><br role="presentation" class="" style="box-sizing: border-box;">
</span><span role="presentation" dir="ltr" class="" style="box-sizing: border-box; white-space: pre-wrap;">for this task, we uncover key determinants underlying our success and provide important</span><span class="" style="box-sizing: border-box; white-space: pre-wrap;"><br role="presentation" class="" style="box-sizing: border-box;">
</span><span role="presentation" dir="ltr" class="" style="box-sizing: border-box; white-space: pre-wrap;">insights for future research. Finally, we highlight our method’s generality by extending it to</span><span class="" style="box-sizing: border-box; white-space: pre-wrap;"><br role="presentation" class="" style="box-sizing: border-box;">
</span><span role="presentation" dir="ltr" class="" style="box-sizing: border-box; white-space: pre-wrap;">simultaneously dock and co-design the sequence and structure of antibody complementarity-</span><span class="" style="box-sizing: border-box; white-space: pre-wrap;"><br role="presentation" class="" style="box-sizing: border-box;">
</span><span role="presentation" dir="ltr" class="" style="box-sizing: border-box; white-space: pre-wrap;">determining regions targeting a specified epitope</span></font></div>
<div class="ContentPasted0" style="font-family: Calibri, Helvetica, sans-serif; font-size: 14.666666984558105px;">
<br class="">
</div>
<div class="ContentPasted0" style="font-family: Calibri, Helvetica, sans-serif; font-size: 14.666666984558105px;">
Advisors: Jinbo Xu & Andy Drucker</div>
<div class="ContentPasted0" style="font-family: Calibri, Helvetica, sans-serif; font-size: 14.666666984558105px;">
<br class="">
</div>
<div class="ContentPasted0" style="font-family: Calibri, Helvetica, sans-serif; font-size: 14.666666984558105px;">
Committee Members: Jinbo Xu, Andy Drucker, Janos Simon</div>
<div class=""><br class="Apple-interchange-newline">
</div>
<br class="">
</body>
</html>