Explore the application of persistent homology in drug discovery through this 40-minute conference talk. Delve into structure-based DHFR drug design, Rips filtration, and barcode interpretability. Learn about chemical compounds as finite metric spaces, filter functions, and parameter choices. Discover how Mapper visualization and Support Vector Machines (SVM) can be utilized for classification in the drug discovery process. Gain insights into potential improvements and intriguing ideas in this field of applied algebraic topology.
Anthony Bak - Spaces of Shapes Persistent Homology for Drug Discovery
Applied Algebraic Topology Network via YouTube
Overview
Syllabus
Intro
Problem Context
Why do virtual screen at all? High throughout screening HTS
Problem Complexity
Structure-based DHFR drug design
Overview
Persistent Homology. Rips Filtration
Persistent Homology. Barcodes
Barcode Interpretability and Engineering
Persistent Homology: Functions
Persistent Homology, Different Metrics
Chemical Compounds as Finite Metric Spaces
Filter functions
Some Parameter choices
Visualization and Discovery: Mapper via Ayasdiris
Machine Learning: Functions and SVM
SVM for Classification
Summary
Improvements
Intriguing ideas
Acknowledgements
Taught by
Applied Algebraic Topology Network
