OXFORD U.K., July 16, 2007— Researchers at the Wellcome Trust Centre for Human Genetics, led by Lon Cardon, have applied new, more powerful statistical methods for locating rare genetic variations that lead to common complex diseases, such as diabetes, heart disease, osteoporosis and bipolar disorder. The ability to find these rare variants may help researchers better understand the mechanisms behind complex diseases and suggest new targets for therapy.

Cardon, of the Wellcome Trust Centre for Human Genetics in Oxford and the Fred Hutchinson Cancer Research Center in Seattle, together with other statistical genetics groups in Cambridge and Oxford, United Kingdom, have been using Affymetrix 500K Mapping Arrays to perform whole-genome scans to find important changes to human DNA that may make a difference in diagnosis or prognosis for some patients. These whole-genome array studies can provide significantly more information than previous linkage studies because the array gives researchers the ability to quickly conduct fine-scale scans of the genomes of thousands of people, improving the statistical power to detect new genes.

“I think 2007 is the year of the whole-genome scan,” said Cardon. “If you think about the number of real discoveries that we had in the entire history of human complex traits genetics, we are probably going to exceed that by an order of magnitude this year alone. That should give us plenty of purchase to move forward.”

Cardon’s group and the Wellcome Trust Case Control Consortium recently completed the initial analysis phase of a project comprising more than 25 scientific groups throughout the U.K. who joined together to study 17,000 individuals using Affymetrix 500K Mapping Arrays. Their aim was to identify common genetic variants for seven complex diseases.

Cardon recently spoke with Wei-Min Chen, a postdoctoral fellow at the University of Michigan Center for Statistical Genetics whose work includes performing statistical analysis of data from whole-genome studies.

Expectations for whole-genome studies
Chen: How have whole-genome studies improved the statistical power to identify susceptibility genes for complex traits and diseases compared to linkage studies?

Cardon: I don’t think that it is fair to directly compare association studies and linkage studies. Linkage studies don’t aim to detect effects of single nucleotide polymorphisms or to find specific variants. The objective of such studies is to identify broad regions that harbor genes. As such, they have very different assumptions and different goals than association studies. I think both strategies are needed. If you wanted to compare the two approaches, linkage studies actually have been more successful than association studies in terms of monogenic disorders.

Recent linkage studies of nuclear families and sib-pairs have not been successful in identifying gene regions. They have given a lot of ambiguous results and caused a lot of confusion in the field. Statistical power is doubtless a big part of the problem. Five hundred sib-pairs are not enough to identify the types of genes that we are looking for in common complex disorders. In addition, the design lacks the precision necessary for studying the conditions inherent to these disorders.

Whole-genome studies are proving their mettle right now. For instance, the macular degeneration studies yielded great results and wonderful advances for the field. But, in terms of setting a model for future studies, they have presented some real challenges. One hundred cases and 50 controls are not going to provide enough statistical power to generate strong results, no matter

how many markers we genotype. In order to benefit from the power of whole-genome association studies, we need to collect enough samples and evaluate them properly.  

Chen: If we assume that most comparative traits have at least one underlying gene that can explain, say, 5 percent of the total phenotypic variance, will whole-genome association studies enable us to identify at least one associated genetic variant for most complex traits?

Cardon: Whole-genome association studies by themselves guarantee absolutely nothing. We can do whole-genome association studies just as poorly as we’ve