In the decade since the Human Genome Project produced the first map of DNA sequences in the human genome, scientists throughout the world have combed through genome data to identify genes and gene variants that cause human disease. A new study suggests that researchers may need to broaden their search farther afield to pinpoint rare but powerful disease-causing mutations.
Researchers from two large genome research centers at The Children's Hospital of Philadelphia and at Duke University published a study today in the online journal Public Library of Science Biology (PLoS Biology), describing what they call "synthetic genome-wide associations."
"We believe our analysis will encourage genetics researchers to reinterpret findings from genome-wide association studies, which will also enable all of us to generate more meaningful diagnostic results for patients," said co-author Hakon Hakonarson, M.D., Ph.D., director of the Center for Applied Genomics at The Children's Hospital of Philadelphia.
Hakonarson and his colleague at Children's Hospital, Kai Wang, Ph.D., collaborated closely with the study leader, David B. Goldstein, Ph.D., of the Center for Human Genome Variation at Duke University. Both research teams had been working independently and simultaneously on a hypothesis that rare genetic variants had a larger role in disease than conventionally assumed.
When Goldstein presented his conceptual model last year to genetics researchers at the University of Pennsylvania, Hakonarson and Wang proposed a collaboration, subsequently supplying data from two genetic diseases-sickle cell disease and genetic hearing loss-that supported and validated the rare variant hypothesis proposed in the current paper.
To date, genome-wide association studies (GWAS) have detected many common gene variants associated with particular diseases, but those variants have shown only modest effects, accounting for a very small percentage of the genetic contribution to the disease.
"GWAS is a very powerful tool to identify disease genes, but for complex disorders, these common variants may not reflect true effect sizes," said Hakonarson. "We may need to look farther away from those common variants to find variants that are individually rare but have strong causative effects." The genetic variants being tested, also referred to as single-nucleotide polymorphisms (SNPs), are changes in a single chemical base of DNA that act as markers for a disease, without causing the disease.
In the current study, the researchers performed a computer simulation in which rare variants were distributed throughout 10,000 genotypes (models of DNA data simulating those collected from human study subjects). Their analysis yielded "synthetic associations"- statistical connections between the rare variants and the common variants that produced signals similar to those found in actual disease studies.
They then tested their approach on two large sample sets for well-characterized disorders, sickle cell disease and genetic hearing loss, in which causative genes were already known. They found a similar pattern of synthetic associations between rare and common gene variants. "Under conventional interpretations, GWAS found only modest contributions for associations with the gene that we know causes hearing loss," said Hakonarson. "Our study shows that conventional interpretation may undervalue the contribution of such gene variants in hearing loss, and we suggest that similar underrepresentation of effect sizes by common variants may occur in many other genetic disorders."
The usual assumption in GWAS is that disease-causing variants are located relatively close to the common variants that capture them (referred to as tagging SNPs). Researchers usually seek out causative variants that travel together with the common variant along the genome; in technical terms, the nucleotides are in relatively strong linkage disequilibrium. "Our study found the causative genes may be two to four times farther away than researchers tend to search, so their effect sizes are poorly captured," said Hakonarson.
Hakonarson and Wang are conducting follow-up studies, some in collaboration with Goldstein, to expand and refine the gene-hunting model using resequencing techniques. The immediate implications of this model, said Hakonarson, affect researchers more than clinicians. But eventually, he adds, this work may improve diagnostic evaluation for patients, furthering the goal of personalized medicine tailored to a patient's genetic profile. At the same time, technological advances in automated gene sequencing will enable researchers to work faster as well as smarter.
New insights on the association between genes, gut microbiota, and mental health