Confirming rare variant-complex disease link
With a new focus on the use of high-throughput, high technology, next-generation sequencing to identify rare genetic variants that contribute to the development of complex diseases such as high blood pressure, diabetes and cancer, experts at Baylor College of Medicine have found that two methods of confirming associations have different advantages.
“You cannot believe an association which is only observed just once," said Dr. Suzanne Leal, professor of molecular and human genetics at BCM and senior author of a report that appears online today in the American Journal of Human Genetics. “There is always a chance that the finding is a false positive."
Second study backs up finding
Spending time and money following up a false positive result by performing functional studies, for example, would be a waste, she said. Therefore good science not only requires a well-designed initial study (stage 1), but also a well-designed replication (stage 2) study to help to determine whether the initial finding is a true positive.
At one time, researchers thought that large-scale, genome-wide genotyping projects that looked for associations between complex diseases and common variations would identify most of the gene changes that lead to the major diseases that plague man. After pursuing such studies, they now believe that rare variations also contribute significantly to development of complex disorders. Genetic studies are now focusing on identifying genes for which rare variants are involved in common disease etiology. Good scientific studies depend not only on initial findings but also on replicating initial results, and Leal and her colleagues are seeking the best way to accomplish this.
Many factors affect such determinations
Leal and Dajiang J. Liu (who is a graduate student in statistics at Rice University mentored by Leal), compared two possible methods of confirming associations in an independent stage 2 sample of diseased and disease free individuals:
-Variant-based replication, in which the variants from the sites within a gene identified as associated with the complex disease in the stage 1 study are genotyped in the replication sample and tested for an association with the complex trait.
-Sequence-based replication, in which the genetic code of the entire gene region implicated in the stage 1 study are sequenced and both known and newly identified variants are tested for an association to the complex disease in the replication sample.
Comparing two strategies
The technology used for the variant-based method is less expensive and involves “replication in the strict sense of the word," said Leal. “You only follow-up the variants found in stage 1 studies."
The sequence-based approach can find variants that were not observed in the stage 1 study, and initially seemed to be a more powerful approach.
However, when Liu and Leal compared the two strategies, they found that the sequence-based approach was only a little more powerful than the variant-based approach.
“Although we are not finding all the causal variants involved in disease etiology in the stage 1 study, those we do find if the gene is truly associated explain a large part of illness due to the particular gene," said Leal.
One caveat with the variant-based approach is that samples for both stage 1 and 2 must be drawn from the same population, said Leal.
“For rare variants, the spectrum is different among populations, even those of similar ethnic background, for example northern and southern European," she said. “If your stage 1 and 2 samples are being drawn from the same population, variant-based replication allows for a relatively inexpensive means of following-up a large number of genes and might be a good way to carry out your replication study."
Funding for this study came from the National Institutes of Health and the Keck Center Pharmacoinformatics Training Program of the Gulf Coast Consortia. Computation for this research was supported in part by the Shared University Grid at Rice University, which is funded by the National Science Foundation and a partnership between Rice University, Sun Microsystems and Sigma Solutions, Inc.