Genome-wide association studies: promises and pitfalls.

1 Department of Neurology, Singapore General Hospital, Singapore 2 National Neuroscience Institute, Singapore 3 Duke-NUS Graduate Medical School, Singapore Address for Correspondence: Dr Eng-King Tan, Department of Neurology, Singapore General Hospital, Outram Road, Singapore 169608. Email: [email protected] Genetic testing is an important means to confi rm the diagnosis of an inheritable disease.1 For this to be feasible, genes that are associated with the disease need to be identifi ed. Hunting for the genes that cause or are associated with a particular disease is a challenging task. Traditionally, mapping causative genes or “genetic susceptibility loci” involves intensive linkage analysis in many families containing clusters of the disease of interest. Genetic association study helps to unravel gene variants that shows an association with a defi ned disease trait.2-4 This association study is usually done using a casecontrol methodology. However, an association does not necessarily equate to a causative effect of the gene variant. Furthermore, the relation between genetic variants and disease status depends on many confounding factors, such as the pattern of physical proximity of the gene variants (linkage disequilibrium) (also described as the non-random association of alleles at two or more loci) in the population; the type of cases and controls; ethnic differences; and selection bias in recruiting the subjects. This “non-rocket science” approach has yielded many observations that by and large could not be consistently replicated. With technological advances and the completion of the Human Genome Project, it has become easier to search for genetic associations of common diseases. In recent years, genome-wide association studies (GWAS), sometimes referred to as whole genome association (WGA) studies, have generated considerable interest among investigators in genetic epidemiology.2,5 The fundamental approach in GWAS is nothing new, as this is essentially a genetic association study, albeit with a much more intricate, and arguably more reliable, approach. Earlier genetic association studies frequently involved the analysis of a single or a few gene variants of selected candidate genes of interest.2-4 However, with an estimated 30,000 genes in the human genome, and possibly million of both common and rare gene variants, it may be unrealistic to expect investigators to strike a “jackpot” and uncover susceptibility or causative genes easily. In GWAS, the analysis involves examining variants across the wholegenome (i.e. the entire DNA sequence) for genetic variations that may associate with certain traits or diseases. Modern chip technology can allow the interrogation of up to 1 million single nucleotide variants (SNPs) at one go.