Mining beyond the exome

The premise of GWAS is the "common disease-common variant" hypothesis, which posits that common diseases are, at least partly, associated with DNA sequence variations or polymorphisms present in more than 1-5% of the population. It turns out that most allele frequencies battle to reach the 5% detection threshold of commercial genotyping arrays and the "common disease-rare variant" hypothesis is gradually taking precedence over its counterpart [2]. Hence, aiming for the rare variants using whole genome sequencing for example is one first step into the right direction [3]. A further step is to deliberately include synonymous polymorphisms among the genetic variants considered in association studies. Although largely disregarded, synonymous polymorphisms are about twice as numerous as non-synonymous ones [4] and are often found responsible for altered protein structure, function and expression level [5]. Accordingly, a considerable list of disease-associated synonymous polymorphisms is already available [5] and there are more to be found. Besides single nucleotide polymorphisms (SNPs), variation can also be structural: multi-kilobase genomic regions can be inserted or deleted (copy number variation, CNV), or they can be moved (copy neutral variation), within (inversion) or between (translocation) chromosomes [6,7]. Structural variants have already been shown to contribute to disease phenotypes [8,9], but with the help of high resolution GWAS purposely designed to detect them, there are undoubtedly more discoveries ahead [6,7].Variants can adopt different forms but they can also occur in different locations throughout the genome. When given the choice between (quasi) random SNPs and SNPs located in coding regions (gene-centric approach), choosing the latter is the safer bet [10]. However, the fact that more than 80% of the risk-associated variants identified so far fall outside of the coding regions suggests that there is a third option, namely the non-coding regions of