Exploiting the proteome to improve the genome-wide genetic analysis of epistasis in common human diseases

The completion of the Human Genome Project in 2003 and the HapMap Project in 2005 has broadened the spectrum of genetic research tools that allow researchers to conduct genome-wide association studies (GWAS) for detecting genetic variants that confer increased or decreased susceptibility to disease. However, detecting epistatic, or gene-gene, interactions in GWAS is a computationally daunting task given that the number of possible SNP combinations increases exponentially with the number of SNPs measured. Valid experimental interactions provide a biologically concise reason why an interaction may be detected statistically, and this work aims to address this computational challenge by utilizing biological expert knowledge derived from protein-protein interaction (PPI) information to guide GWAS. Initially, we review existing PPI databases and the ways that PPI information can be used as biological expert knowledge to facilitate genetic studies of common human disease. We explore a potential solution that entails the development of metrics based on the confidence score of PPIs extracted from the Search Tool for the Retrieval of Interacting Genes/Protiens (STRING). These metrics are evaluated on their ability to prioritize SNPs in a genetic dataset and thereby reduce the size of an analysis. We also determine if this biological expert knowledge is capable of improving the ability of our computational evolutionary system (CES) to detect genetic risk factors of common human disease. Subsequently, we develop a web-based tool with the capability to access PPI expert knowledge in STRING such that it can be incorporated in a genetic analysis using our methods and thereby addressing the computational challenges of detecting epistasis in GWAS.