[CS] [masters-presentation] Duan/MS Presentation/Nov 13, 2020

[Tricia Baclawski]

This is an announcement of Xiaotian Duan's MS Presentation.


https://uchicago.zoom.us/j/92663286758?pwd=SUVYUFVrU3JabTBGUEwzM0tuK0JYUT09
Password: 749613

------------------------------------------------------------------------------
Date:  Friday, November 13, 2020

Time:  9:00 AM

Place:  via zoom

M.S. Candidate:  Xiaotian Duan

M.S. Paper Title: Protein-Ligand Binding Affinity Prediction with
Uncertainty Quantification

Abstract:
Protein-ligand binding prediction is the centerpiece of drug discovery
and virtual screen- ing, and a relevant topic for many other fields in
biochemistry. However, the prediction of the binding affinity remains
a challenging task despite an extensive amount of work in the last
decade. The performance of the descriptor-based machine-learning
binding affinity predictors, which have been dominating the
benchmarks, has reached a plateau. This thesis surveys the problem
formulation, data sets, featurization, modeling of protein- ligand
binding affinity prediction, in order to provide insights on potential
directions. In addition, this thesis explores the possibility of
applying UQ (uncertainty quantification) methods to protein-ligand
binding affinity prediction. The results on the CASF benchmark have
shown that while no single UQ algorithm wins across all the comparison
metrics, there is a positive correlation between the predictive
accuracy and the uncertainty. On top of that, the predictive power of
NGBoost is marginally better compared to other benchmarked binding
affinity predictor or scoring functions. While there is much room for
improvement, incorporating UQ with binding affinity prediction is
capable of offering informative estimates of confidence bounds for all
predicted instances.

Xiaotian's advisor is Prof. Rick Stevens

Login to the Computer Science Department website for details:
 https://newtraell.cs.uchicago.edu/phd/ms_announcements#xduan7

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Tricia Baclawski
Student Affairs Administrator
Computer Science Department
5730 S. Ellis - Room 350
Chicago, IL 60637
pbaclawski at uchicago.edu
(773) 702-6854
/pronouns: she, her, hers/
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=