Association studies performed, to date, have been restricted to candidate gene/region studies. As the number of SNPs in the public domain increases, there is a move to perform association studies with SNPs spanning the entire genome (whole genome by association [WGA]). The main argument for this approach is that it is hypothesis free; hence, no prior knowledge of gene function is required. No truly genome-wide studies have been published yet but many are planned. The first will probably use chip-based genotyping, for example using chips containing 100,000 SNPs (www.
Affymetrix.com) or bead-based methods (www.illumina.com). It is unclear how much of the genome will actually be captured but it is estimated to be in the region of 50%. Other companies are developing methods for higher throughput sequencing and it may be that whole genome sequencing will become standard in the future (e.g., www.curagen.com). Opponents of this approach point to the large number of false positives likely to be generated. This can be countered by setting stringent significance thresholds but, in turn, this reduces the power to detect modest effect sizes unless large sample sizes are used. Most argue for a multi-stage approach whereby only SNPs for which association is replicated are explored for functionality.
Although whole genome linkage analyses have had limited success, the data generated from such studies will continue to be of use in the search for important loci. While the statisticians develop strategies for the analysis of WGA data that make some correction for multiple testing without completely removing the possibility of detecting weak associations, it may be useful to focus analysis of such data on regions already known to be linked to the disease of interest. The DNA collected from families with multiple cases of RA will also prove to be a valuable resource. These cases tend to have a lower age at onset and more severe disease than their community based counterparts and are likely to have a higher 'dose' of genetic risk factors increasing the likelihood of detecting genes in these cohorts.
In considering the merits of linkage and association strategies it is clear that there is no simple answer as to which should be the method of choice. For the disease of interest it is important to take account of a number of factors, including feasibility of collecting multi-case families, the expected effect size of disease genes, the allele frequency of associated polymorphisms and the budget available. For major genes with significant effect sizes (e.g., HLA with RA) linkage analysis of small number of families or a modest case-control association study would both be successful. For disease alleles found only at low frequency in the population of interest, linkage would be powerful (e.g., the CARD15 gene and Crohn's disease) whereas a disease allele with high frequency would rarely be detected by linkage but could be found by an appropriately powered association study (e.g., the PPARG gene and type 2 diabetes)  (Fig. 2).
Current efforts are focused mainly around identifying susceptibility loci for complex diseases in an effort to gain a better understanding of the underlying disease pathogenesis and, ideally, to develop safer, more effective treatments. In the future, however, other potentially clinically useful applications arising from the identification of disease genes include firstly, the identification of severity genes so that interventions can be better targeted. There are now a number of inception cohorts of patients with early disease being followed prospectively across Europe and this should facilitate the separation of susceptibility from severity factors. For example, patients with early inflammatory or rheumatoid arthritis are being recruited in centres in the UK, The Netherlands and Sweden [1, 61, 62]. Secondly, it is clear that many of the rheumatic diseases are clinically heterogeneous and may in fact comprise a number of distinct conditions with a similar end-phenotype, in the case of RA, for example, symmetrical synovial joint inflammation. Once genetic susceptibility factors have been identified, it may be possible to use this information to perform phenotyping studies to identify these subgroups as has been done in Crohn's disease by analysis of CARD15 gene variants . Finally, knowledge of genetic susceptibility factors may influence response to or susceptibility to side effects from particular treatments. In this respect, the setting up of Biologics Registers in many European countries is encouraging and will provide an ideal opportunity to establish whether treatment response genes exist.
The Arthritis Research Campaign and The Wellcome Trust for their support of J Worthington and A Barton, respectively.
Was this article helpful?
It's time for a change. Finally A Way to Get Pain Relief for Your Arthritis Without Possibly Risking Your Health in the Process. You may not be aware of this, but taking prescription drugs to get relief for your Arthritis Pain is not the only solution. There are alternative pain relief treatments available.