Biomedical and Translational Informatics Laboratory

  • Incorporating Domain Knowledge into Evolutionary Computing for Discovering Gene-gene Interaction Conference Paper, Stephen D. Turner, Scott M. Dudek, Marylyn D. Ritchie, http://dl.acm.org/citation.cfm?id=1885031.1885074, PPSN'10, Berlin, Heidelberg, Springer-Verlag, 394–403, 3-642-15843-9, 978-3-642-15843-8, 2010, 2014-02-24 18:06:46, ACM Digital Library, Understanding the genetic underpinnings of common heritable human traits has enormous public health benefits with implications for risk prediction, development of novel drugs, and personalized medicine. Many complex human traits are highly heritable, yet little of the variability in such traits can be accounted for by examining single DNA variants at a time. Seldom explored non-additive gene-gene interactions are thought to be one source of this "missing" heritability. Approaches that can account for this complexity are more aptly suited to find combinations of genetic and environmental exposures that can lead to disease. Stochastic methods employing evolutionary algorithms have demonstrated promise in being able to detect and model gene-gene interactions that influence human traits, yet the search space is nearly infinite because of the vast number of variables collected in contemporary human genetics studies. In this work we assess the performance and feasibility of sensible initialization of an evolutionary algorithm using domain knowledge., Proceedings of the 11th International Conference on Parallel Problem Solving from Nature: Part I,