Kelin Xia (6/23/21): Topological data analysis (TDA) based machine learning models for drug design

Effective molecular representation is key to the success of machine learning models for drug design. In this talk, we will discuss a series of TDA-related models, including weighted persistent homology, persistent spectral models, and persistent Ricci curvature and their combination with machine learning models. Unlike traditional graph or network models, these persistent models characterize the intrinsic multiscale information and provide molecular representations that balance the data complexity and dimension reduction. Molecular descriptors can be generated from various persistent attributes and further combined with machine learning models, in particular, random forest, gradient boosting tree, and convolutional neural network, for drug design. Our models have been extensive tested on various databanks, in particular, PDBbind datasets. It has been found that persistent representations based molecular descriptors can dramatically improve the performance of learning models in drug design, and are significantly better than all traditional molecular descriptors.