The genome-wide association studies (GWASs) have successfully identified a number of single nucleotidepolymorphisms (SNPs) associated with melanoma risk. However, the traditional GWASs focus only on marginaleffects of individual markers and have incorporated external functional information only after identifying robuststatistical associations. It is often lack power to detect relatively small effects conferred by most genetic variants.The pathway-based approaches, which evaluate the cumulative contribution of the genes within biological pathways,may help collect the modest signals contained in the GWAS data and identify biological pathways in the etiology ofdisease on a pathway level. As used in the traditional pathway approaches, assigning SNPs to their physicallocation may introduce many false positive associations from multiple testing on non-functional SNPs and miss-annotate SNPs that regulate the expression of genes in distance. Instead, defining the expression quantitative traitloci (eQTLs) and assigning them into the genes they regulate may help functionally annotate SNPs and increase theenrichment of functional variants in the pathway analysis. Among numerous SNPs with weak associations in GWAS,the selection of those from the identified pathways by using the pathway analysis integrating the genetics of geneexpression and the GWAS data for replication may increase the likelihood of targeting true signals than by chance.The integration of eQTLs of liver and adipose tissues into the pathway analysis for the GWAS of type 2 diabetes hassuccessfully identified several novel disease-related pathways. More recently, a GWAS on global gene expressionof the skin have systematically generated skin eQTLs. However, no pathway analysis has been conducted formelanoma GWAS. The goal of the current proposal is to systematically assess the associations of biologicalpathways with melanoma risk by applying this new approach to melanoma GWAS in the discovery stage and tovalidate specific loci associated with melanoma risk within the identified pathways in the replication stage. Mediationanalysis on potentially intermediate phenotypes will also be conducted to investigate the etiological contribution ofthe identified pathways/SNPs. We plan to use a nested melanoma case-control study of 420 melanoma cases and2,284 controls in two large, well-characterized cohorts, the Nurses Health Study (NHS) and the HealthProfessionals Follow-up Study (HPFS) in the discovery stage, and use a melanoma case-control study of 1,804melanoma cases and 1,026 controls from the MD Anderson Cancer Center in the replication stage. All the casesand controls have been previously genotyped on Illumina chips. The pre-existing GWAS genotype data gives us acost-effective opportunity to apply the new approach to melanoma research. Our proposed study would be the firstto combine the genetics of gene expression and functional classification of genes as prior information to melanomaGWAS. Findings from this study will pick up the genetic variants with modest effects from the GWAS data, which willutilize the GWAS data to a greater extent and provide new insights into the etiology of melanoma.
Using a new pathway-based approach that integrates the genetics of gene expression of the skin and themelanoma GWAS; we will systematically assess the associations of biological pathways with melanoma risk;and investigate the etiological contribution of the identified pathways/SNPs using mediation analysis onpotential intermediate phenotypes. This innovative work may identify genes/pathways that might be missedusing the conventional GWAS approach; and provide new insights into the etiology of melanoma; which maylead to better strategies on melanoma prevention and treatment.