Speaker: Chad Myers (University of Minnesota)
Title: Discovering biological networks by integrating diverse genomic data
Abstract: Understanding protein function and modeling biological networks is a key challenge in modern systems biology. Recent developments in biotechnology have enabled high-throughput measurement of several cellular phenomena including gene expression, protein-protein interactions, protein localization and sequence. The wealth of data generated by such technology promises to support computational prediction of network models, but so far, successful approaches that translate these data into accurate, experimentally testable hypotheses have been limited. I will discuss key insights into why we face this imbalance between genomic data and established knowledge and present computational approaches for addressing these challenges. Specifically, I will present a Bayesian framework for integrating diverse genomic data to predict biological networks. I will describe the machine learning methods as well as important data visualization features that support a public, web-based system for user-driven search of network predictions from genomic data. Using our approach, we correctly predicted new functions for approximately 100 genes in yeast. I will discuss experimental validation for these predictions as well as our recent efforts to use predicted network models for directing large-scale genetic interaction screens.