Machine-learning dissection of congenital heart disease genetics
Katherine S. Pollard, PhD, Director of the Gladstone Institute of Data Science and Biotechnology, Investigator at the Chan Zuckerberg Biohub, and Professor in the Department of Epidemiology & Biostatistics and Bioinformatics Graduate Program at the University of California San Francisco
Abstract: Most families with congenital heart defects do not carry damaging variants in known risk genes. We will show how predictive modeling can be used to identify variants that might otherwise be overlooked. This includes structural variants that alter three-dimensional genome folding and combinations of rare protein coding variants in genes that are not individually significant. These results demonstrate that machine learning is an emerging tool for risk prediction and prioritization of loci for functional characterization to understand disease mechanisms.
Application of ML to cardiovascular medicine: unlocking the diagnostic potential of the ECG
Peter Noseworthy, MD, Cardiac Electrophysiologist and Professor of Medicine at Mayo Clinic, and Director of the Mayo Clinic Heart Rhythm and Physiologic Monitoring Laboratory
Abstract: We have spent several years developing and applying various AI algorithms that run on routine 12-lead ECGs. We have found that these algorithms can (1) scale human capabilities and (2) unlock hidden patterns in the data that expand our interpretive ability beyond human capacity.