A recent New York Times article, featuring an interview with Duke geneticist David Goldstein, summarized Genome-Wide Association Scan (GWAS) studies as expensive and unable to identify disease-gene associations. While I agree that only a few SNP-disease associations have been identified to date, we should not forget about the exciting findings that have come out of this research. In the aggregate, about 90 cancer GWAS hits have been published in high impact journals as of early October 2009, including the 8q24, 11q13 and 17q24 regions for prostate cancer and the 8q24 and the 5p15.33 regions that have been identified in multiple cancers.
Given the complexity of diseases such as cancer, researchers are urged to view observed SNP associations as only a first step in understanding disease etiology. This is important because these SNPs may be genetic markers for other SNPs which may be driving host susceptability. Further, an individual's disease susceptability may be modified by exogenous factors, such as environmental exposures, occupational exposures, diet, infectious agenets, and other lifestyle exposures.
As we move into the age of full genome sequencing, researchers will be able to overcome the limitation of SNPs serving as genetic markers since data will be avaialble for the entire genome. This will not only allow researchers to identify the truly causual variant(s), but also to begin exploring SNP-SNP interactions, gene-gene interactions, pathway-based variation, and so on. Juxtiposing these data with exogenous exposure information will also allow researchers to start to understand the mechanisms of disease that may be highly dependent on environmental exposures, such as lung and bladder cancer. Until studies with whole genome sequencing are coupled with high quality exposure data, researchers should should view GWAS studies as the logical step, while expecting a range of genomic architectures underlying GWAS signals and that the model of a single gene resulting in a single outcome is less likely.