Comparing Statistical and Data Science Approaches to the High-Dimensional Problem of Traditional Genome-Wide Association Studies

Join the GW Department of Statistics and the Data Science Program for a joint seminar featuring Assistant Professor of Data Science Angelica Walker.

Abstract:
Genetic association analysis is the process of assigning single nucleotide polymorphisms (SNPs), also known as mutations, to specific phenotypic traits observed in a population of individuals. As technology has improved, these analyses can detect more of the 12 million SNPs that are known to exist within the 3 billion base human genome. Due to biological processes such as recombination, these SNP matrices have a high degree of local multicollinearity. Combined with the high dimensionality of these datasets, it is increasingly difficult to process and infer interactions between SNPs with a known phenotype using traditional methods like linear regression. This seminar will focus on some of the statistical and data science approaches used to discover meaningful insights by assigning these mutations to phenotypes of interest.

About the Speaker:
Angelica Walker, PhD, is an assistant professor in the Data Science Program at George Washington University. She earned her PhD in data science and engineering with a focus in health and biological systems from the University of Tennessee, Knoxville, and Oak Ridge National Laboratory's joint program the Bredesen Center. Her research is focused on developing and applying explainable artificial intelligence tools to very large biological datasets. This includes topics such as: improving tree based methods, producing efficient high performance computing implementations, creating and analyzing predictive gene expression networks, performing multi-omic network analyses and improving traditional Genome Wide Association Studies.