Let the poeple sing - nature

Let the people sing
vocalist of Gens du pays — scored Bella et al. (J. Acoust. Soc. Am. 121,
sing Gens du pays — an anthem of SNP sets consisting only of variants within study in a new sample (about 5,500 total cases of the linkage-disequilibrium-based studies genes3,4, GWA studies aspire to survey all vari- and controls) to test for significant associations. reported so far1,8 were able to replicate one ation in the genome, including non-coding The results now published refer only to the fol- known locus. On the other hand, the most sig- regions. Exhaustive surveys are not yet feasible, low-up of the most promising SNPs from the nificant new locus in both cases was identified but genotyping a subset of the known variable first stage: a more comprehensive second stage by only a single SNP, suggesting that even the dense marker sets employed in these studies Locations in the genome that are separated Sladek et al. identify strong associations provide insufficient coverage for detection by a small number of base pairs are often in for three novel loci (one detected in the link- of all important loci10. In certain populations ‘linkage disequilibrium’: that is, there is a age-disequilibrium-based marker set and two that are of recent origin and that have remained substantial association between the variants detected in both marker sets). Although more isolated, linkage disequilibrium is more exten- at the two loci. This phenomenon allows us loci may emerge from the complete two-stage sive than in the populations used in the GWA to survey variation across the genome fairly analysis, publication of these initial results studies so far11; the first GWA surveys in such precisely, simply by genotyping a subset of provides the opportunity for swift replication groups will be watched closely for evidence polymorphic loci. GWA studies rely on the (or not) by other research groups, using inde- that they permit more complete coverage using assumption that linkage disequilibrium ena- pendent samples, as exemplified by the case of comparable marker sets. Similarly, the results bles one SNP to act as a marker for associa- TCF7L2. Its association with type 2 diabetes tion to other sequence variants in that region. was reported last year2, and has already been using even larger numbers of SNPs, are much The GWA studies carried out thus far differ in replicated in at least 20 independent studies. In terms of the number and criteria for selection one- or two-stage studies, care must be taken A final observation is that for both of the of the genotyped SNPs: some use SNPs chosen that the specific definition of disease adopted in linkage-disequilibrium-based studies reported to be evenly physically spaced5–7, whereas oth- the follow-up (or replication) samples is com- so far (for type 2 diabetes1 and inflammatory ers8 choose SNPs to maximize the detection parable to the definition used in the original bowel disease8), the most significant novel of linkage disequilibrium, based on data from GWA samples. In this respect it is note worthy, associations were to a variant predicted to the International HapMap Project9. Sladek but of unclear significance, that Sladek et al. alter the protein product encoded by a gene et al. used a marker set based on HapMap link- used more stringent inclusion criteria in the (termed a non-synonymous coding SNP), and age-disequilibrium data, supplemented by a GWA sample than in the follow-up sample. thus possibly to have a strong functional effect. gene-centric SNP set; their combined marker The identification of a few significant dis- Furthermore, in both cases the disease is asso- set (about 400,000 SNPs) provides the highest- ease associations represents only one outcome ciated with the more common variant at these resolution survey of genomic variation of any of Sladek and colleagues’ study. GWA studies loci, suggesting that the less common variant should be evaluated primarily from an epide- may offer protection against developing the In deciding how many individuals to geno- miological standpoint, focused not just on what disease. That diseases are associated with such type, recent studies follow either a one-stage8 new disease-susceptibility genes they propose, common non-synonymous SNPs suggests that or a two-stage5–7 design: in the first case, sta- but on how they advance our understanding of these variants may have offered an evolutionary tistical significance is sought within the geno- the composition of genetic risk in the popula- advantage in previous environments. Clearly, typed sample (and other samples may be used tion. Sladek et al. take a first step towards such for replication); in the second, SNPs passing a understanding, presenting an evaluation of size of two. Nevertheless, these findings under- loose significance threshold in the GWA study what proportion of the disease cases can be score how GWA studies may not only deliver are genotyped in a follow-up sample, to seek attributed to variation in the loci they identify ‘new’ genes10, but permit advances in our significance. The choice of study design usually as significant in their second-stage analysis. As understanding of how human evolution has rests on estimates of the power to detect effects several additional GWA studies of type 2 dia- ‘made’ the diseases that are common today. ■ of a given magnitude that is considered realistic betes will shortly report their results, we may Nelson B. Freimer is in the UCLA Center for soon be able to estimate the number — and Neurobehavioral Genetics, Semel Institute for Sladek et al.1 adopted a two-stage design: in location in the genome — of the genetic vari- Neuroscience and Human Behavior, 695 Charles the first stage, they conducted the GWA study ants that are the main contributors to diabetes Young Drive South, Los Angeles, California in a total sample of about 1,400 cases and con- susceptibility, at least in some populations. trols (the largest sample in a GWA study so far). The results of the first GWA studies may also In the second stage, they genotyped the SNPs reveal the degree of genome coverage provided Chiara Sabatti is in the UCLA Departments of showing evidence of association in the GWA by the chosen SNP panels. Reassuringly, both Human Genetics and Statistics, 695 Charles

Source: http://www.mpblab.vizja.pl/documents/publications/Nature_2007_News&Views.pdf