By Kevin Davies
April 15, 2009 | Two prominent geneticists agree to disagree on the merits of the hugely popular genome-wide association studies (GWAS) in dueling commentaries in the pages of the latest issue (April 16) of the New England Journal of Medicine (NEJM). The studies are published online today.
In the past couple of years, thanks to plunging microarray costs and greater international collaboration, GWAS have come to dominate the pages of the leading journals, as scientists successfully pinpoint scores of gene loci associated with complex diseases including diabetes, heart disease, mental illness and cancer. Despite their success and popularity, Duke University geneticist David Goldstein believes their usefulness is limited.
“After many years in which the study of complex human traits was mired in false claims and methodologic inconsistencies, genomics has brought not only comprehensive representation of common variation, but also welcome rigor in interpreting statistical evidence,” he writes. That’s the good news. But here’s the rub. “Most reported associations reflect real biologic causation,” says Goldstein. “But do they matter?
Goldstein, who is director of the Center for Population Genomics and Pharmacogenetics at the Institute for Genome Sciences and Policy at Duke, was the subject of a profile in the New York Times last year, written by veteran science writer Nick Wade. The Duke geneticist came to prominence while working in London on the population genetics of Ashkenazi Jews and the Y chromosome. Goldstein pulled no punches in the Times interview, expressing dismay at the limited impact of GWAS in explaining the genetics of complex diseases such as mental illness and diabetes.
“David Goldstein submitted his paper to the Journal presumably because of our focus on understanding disease etiology (in addition to the treatment of disease),” explains NEJM Deputy Editor Bette Phimister. “It provides much food for thought.” Phimister decided to solicit a different point-of-view from another leading geneticist, Joel Hirschhorn (Children’s Hospital, Boston), because she says, “we are persuaded that GWAS are providing useful insights into disease susceptibility and we wanted to give our readers a balance of opinion.”
Hirschorn points out that during the past two years, GWAS have identified more than 250 genetic loci harboring common variants (more than 5%) that are reproducibly associated with polygenic traits. “This explosion represents one of the most prolific periods of discovery in human genetics,” counters Hirschorn.
In his NEJM article, Goldstein says he does not mean to imply “criticism of the strikingly successful efforts to represent common variation and relate it to common diseases.” But he says, “It’s hard to justify the current enthusiasm for conducting genome scans using ever-larger cohorts after study of the first several thousand subjects has identified the strongest determinants among common variants… If effect sizes are so small as to require a large chunk of the genome to explain the genetic component of a condition, then no guidance would be provided: in pointing at everything, genetics would point at nothing.”
Once the first group of genetic variants has been identified, subsequent variants invariably produce ever smaller effect sizes. At that point, “there are probably either no more common variants to discover or no more that are worth discovering,” Goldstein says. He cites examples from recent gene quests in type 2 diabetes, where the most robust association (TCF7L2) has a very modest sibling relative risk. In studies on height, where a score of variants account for just 2-3% of height distribution in the normal population, Goldstein estimates that conservatively there may be 1500 more variants to be discovered! “If common variants are responsible for most genetic components of type 2 diabetes, height, and similar traits, then genetics would provide relatively little guidance about the biology of these conditions, because most genes are “height genes” or “type 2 diabetes” genes.”
Rather than invest more and more money on ever larger GWAS for diminishing returns, Goldstein suggest that the priority should be to study rarer variants of larger effect. “Effectively searching the full human genome for rare variants will require not only sequencing capacity, but also thoughtful selection of the most appropriate groups of individual genomes to resequence and thoughtful evaluation and prioritization of the many rare variants identified.”
Hirschorn, who is also an associate member of the Broad Institute and a former postdoc with director Eric Lander, agrees that most GWAS hits have modest predictive power, and so typically explain only a small fraction of the disease or trait’s heritability. “But, the success of genome-wide association studies is not tied to prediction. If we only identify new pathways underlying disease, these studies will have a tremendous impact,” he says.
Hirschorn acknowledges that new insights obtained from gene variants do not necessarily guarantee a smooth transition into clinical practice. Then again, the transition from elucidating the structure of cholesterol to the development of statins took a century and three Nobel Prizes.
Hirschorn concludes: “Each discovery of a biologically relevant locus is a potential first step in a translational journey, and some journeys will be shorter than others. With a more complete collection of relevant genes and pathways, we can hope to shorten the interval between biologic knowledge and improved patient care.”
David B. Goldstein, "Common Genetic Variation and Human Traits." NEJM April 16, 2009.
Joel N. Hirschhorn. "Genomewide Association Studies — Illuminating Biologic Pathways." NEHM April 16, 2009.