Reducing Imprecision in Precision Medicine
Igor Jurisica, PhD, DrSc
Abstract
To fathom complex health and disease processes, we need to systematically integrate diverse types of information, including multiple high-throughput datasets and diverse annotations. However, thousands of potentially important proteins remain poorly characterized. Models are going to be only as good as the networks used to build them are comprehensive. Computational biology methods, including machine learning and data mining can help fill these gaps with accurate predictions, making disease modeling more comprehensive. In turn, they help to find hidden signal in data often rejected as noise.
Such models and integrative analyses help us develop new hypotheses, answer complex questions such as what factors cause disease; which patients are at high risk; will patients respond to a given treatment; how to rationally select a combination therapy to individual patient, etc. In turn, this enables data-driven precision medicine. However, AI-based prediction systems can have great cohort-based accuracy, yet fail on specific patient level. Thus, subgroup performance and individual prediction confidence must be assessed and understood in combination with model explanation, before translating AI-based prediction tools into clinical practice.
Lecturer
Igor Jurisica, PhD, DrSc is a Senior Scientist at Osteoarthritis Research Program, Division of Orthopedic Surgery, Schroeder Arthritis Institute and Krembil Research Institute, Professor at University of Toronto and Visiting Scientist at IBM CAS. He is also an Adjunct Professor at the Department of Pathology and Molecular Medicine at Queen’s University, and an adjunct scientist at the Institute of Neuroimmunology, Slovak Academy of Sciences. Since 2015, he has also served as Chief Scientist at the Creative Destruction Lab, Rotman School of Management, and since 2021 he is a scientific director of the World Community Grid.
His research focuses on integrative informatics and the representation, analysis and visualization of high-dimensional data to identify prognostic/predictive signatures, determine clinically relevant combination therapies, and develop accurate models of drug mechanism of action and disease-altered signaling cascades. He has published extensively on data mining, visualization and integrative computational biology, including multiple papers in Science, Nature, Nature Medicine, Nature Methods, J Clinical Oncology, J Clinical Investigations. He has been included in Thomson Reuters 2014, 2015 & 2016 lists of Highly Cited Researchers (http://highlycited.com), and The World’s Most Influential Scientific Minds: 2015 & 2014 Reports. In 2019, he has been included in the Top 100 AI Leaders in Drug Discovery and Advanced Healthcare list (Deep Knowledge Analytics, http://analytics.dkv.global).