Sunday, June 28, 2009

From genome-wide association studies to molecular biology   posted by p-ter @ 6/28/2009 05:10:00 PM

One of the rationales advanced for the identification of common alleles that confer modest risk to a disease via genome-wide association studies is that these associations will lead to biological insight into the disease. Two papers published today represent an important first step towards this goal for a variant associated with colorectal cancer.

Like many polymorphisms associated with complex diseases, the one investigated in these studies does not fall within a gene--this particular variant falls hundreds of thousands of bases away from the nearest gene. It does, however, fall within a non-coding element that is conserved across millions of years of evolution, suggesting that it is functional. These studies show that, indeed, the SNP falls in a binding site for a transcription factor, and that the two alleles have different binding affinities for that factor. Additionally, one of the studies shows that the genomic region containing the SNP loops over and makes physical contact with the nearest gene (MYC, a known oncogene), supporting the hypothesis that the SNP affects its regulation.

These studies raise more questions than they answer, of course. None of the studies find an association between the SNP itself and steady-state MYC expression in cell lines. My guess is that, like many transcriptional enhancers, developmental-time-point-specific manner. An important direction now is to determine when that important time point is.