English Version



Hypergraph based persistent cohomology (HPC) for molecular representation in drug design

主 讲 人 :刘祥    


地      点 :zoom


In this talk, i will discuss topological dataanalysis(TDA) and its application in drug design. Persistent homology, which isone of the most important tools in TDA, is used in identification,classification and analysis of biomolecular structure, flexibility, dynamics,and functions. Properties from persistent homology analysis are used asfeatures for learning models. Unlike previous biomolecular descriptors,topological features from TDA, provide a balance between structural complexityand data simplification. TDA-based learning models have consistently deliveredthe best results in various aspects of drug design, including protein-ligandbinding affinity prediction, solubility prediction, protein stability changeupon mutation prediction, toxicity prediction, solvation free energyprediction, partition coefficient and aqueous solubility, binding pocketdetection, and drug discovery. Further, I will discuss our recently-proposedweighted persistent hypergraph based machine learning model. Different from allprevious TDA-based models, we firstly use hypergraph to do the persistence.Then, we apply our weighted persistent hypergraph based machine learning modelto make the protein-ligand binding affinity prediction, one of the mostimportant tasks in drug design. We systematically test our model on threecommonly-used databases, including PDBbind-2007, PDBbind-2013 and PDBbind-2016.Our results, for all these databases, are better than all existing models withtraditional learning descriptors, as far as we know. This demonstrates thegreat power of our model in molecular data analysis and drug design.



发布时间:2020-10-30 15:51:18

版权所有bbin体育真人平台-平台首页    冀ICP备18011017号-3  

冀公网安备 13010802000630号

地址:河北省石家庄市南二环东路20号 邮编:050024 
XML 地图 | Sitemap 地图