Yue (Joseph) Wang and Jason Xuan are applying machine learning to understanding the complex interactions between environmental and genetic factors in common diseases.
Many diseases result from a complex interaction of multiple environmental and genetic factors and since the sequencing of the human genome, researchers have been gathering genetic data associated with common diseases. However, it is largely unknown how specific variations in a person's genes interact with each other to increase the propensity for any given disease.
Researchers have started using genome-wide association studies (GWAS), — hailed by Science magazine as the breakthrough of the year — to identify the risk associated with each factor. GWAS studies involve using arrays that can compare 500,000 factors against thousands of people.
GWAS studies generate huge datasets on these genetic factors," said Wang, who directs ECE's Computational Bioinformatics and Bio-imaging Laboratory. "To understand what combination of factors triggers these diseases and how they each contribute is an enormous computational task."
The diseases are caused by many small effects, which are hard to see in the data, he explained. The effects are not just complex, but also nonlinear. "Not everybody has all the effects. Taken together, however, all the small effects can equal disease."
The team plans to develop parallel and economics portfolio computing techniques to meet the computational challenges of the problem.
Wang and Xuan have received a grant from the National Institutes of Health (NIH) as a part of the Gene and Environmental Initiatives (GEI) to pursue the effort, jointly with Wake Forest University School of Medicine and Pennsylvania State University College of Engineering.
For more information, visit www.cbil.ece.vt.edu.