Duke University said last week that it will sequence 4,000 individuals as part of a collaborative, $25 million effort to identify as many genes as possible implicated in epilepsy.
The project, dubbed Epi4K, aims to elucidate the genetic basis of epilepsy, which researchers hope will lead to new or better targeted treatments for the highly heterogeneous disorder.
The so-called "center without walls" includes researchers at Duke; the University of California, San Francisco; the University of Melbourne; and elsewhere. The initiative was awarded $25 million over the next five years from the National Institute of Neurological Disorders and Stroke.
According to Duke's David Goldstein, the project is divided into three core groups focused on administrative duties, phenotyping, and sequencing and bioinformatics, respectively. It will also involve several distinct sequencing projects spearheaded by investigators at a number of other institutions.
Goldstein leads the sequencing core at Duke's Center for Human Genome Variation, while administrative duties and overall leadership of the project are shared jointly between Goldstein, UCSF's Daniel Lowenstein, and Samuel Berkovic of the University of Melbourne.
The Epi4K collaborators proposed a sequencing plan involving "two primary designs," Goldstein told Clinical Sequencing News. One is a trio-based plan to sequence the genomes of probands with two severe childhood epileptic disorders — infantile spasms and Lennox Gastaut syndrome — for which Goldstein said there is a "general feeling" that de novo mutations play a relatively more important role.
"The other primary effort we're making is sequencing in multiplex families," he said. "The reason for that design is that we expect enrichment of strong genetic control in [those] cases."
According to Goldstein, an early planning meeting held last year by NINDS to discuss future directions for human genetics research in epilepsy featured a "pretty strong consensus" that the disease's heterogeneity meant it would be hard to make significant progress in small studies. As a result, there was agreement that it would be important to "get as much of the community together as possible [to] do something cooperatively."
He added that the participants also largely agreed that it would be preferable to do a sequencing study rather than a genome-wide association study. "We need to find as many genes that are securely implicated as risk factors as possible," he said, "as opposed to just signals of risk, which one gets from genome-wide association."
While Goldstein said that goals among Epi4K's many participants would surely vary, his hope is that even a small fraction of epilepsy genes that are identified would point to either a validated drug target or a druggable pathway.
Epilepsy is a therapeutic area "where drug development is actually not moving very well," he said, noting that there is "a tremendous need for new drugs." About seventy percent of patients have their seizures controlled by currently available medicines, but that means thirty percent don't, "and that proportion hasn't been changing much over recent years."
As a result, "we really need new directions for developing treatments, and I see [the data coming out of the Epi4K project] as possibly providing that direction," he said.
Meanwhile, the rarity of any one risk factor or mutation in epilepsy suggests that personalizing treatment will be harder than for more homogeneous diseases, he said.
Nevertheless, the more factors the center can identify, the more likely it is that "some will organize into common pathways that will have treatment implications," he added. "I think there is some reason to be hopeful that that will work."
Exome Sequencing for Trios
Goldstein said that the first distinct project in the overall Epi4K effort will be the trio-based component of the sequencing initiative. The collaborators plan to complete this first step, lead by Elliott Sherr at UCSF, by the end of the first year. According to the project's grant description, it will involve 500 subjects.
"We want to get that done fast," Goldstein said, "analyze and see what genes come out to give us an idea of how to proceed from there."
While the majority of Epi4K subjects will undergo whole-genome sequencing, Goldstein said, the initial analysis of the trios will be whole-exome sequencing. "We are hoping that some number of these cases will be explained through the analysis of the whole-exome data. Then, for the ones that aren’t explained, we can turn to whole genome after that," he said.
"Part of the reason to do that is that the costs, of course, are falling, so we're doing the whole-exome work earlier in the life of the grant and whole-genome later in the life of the grant."
Most subjects are already lined up for the first project as well as the following multiplex family project, according to Goldstein. Many have also already had their phenotypic information analyzed.
"We wanted to concentrate on already available material so it could move pretty quickly," he said.
A large number of samples come from a previous NINDS-funded effort by UCSF's Lowenstein called the epilepsy phenome/genome project. "They collected a great deal of great phenotypic information and DNA samples for a bunch of epilepsy patients," Goldstein said, and "many of those patients are [now] going to be sequenced in this effort."
For the multiplex families, he added, the subjects are already identified but phenotypic review is still needed.
Currently Duke's sequencing infrastructure comprises six Illumina HiSeqs, which are "easily sufficient to do all the genomes for this project," according to Goldstein.
He said the group has worked out a data release plan, and "virtually all" the genomes sequenced through the initiative will permit data release under "dbGaP-like" rules.
Last year Duke reported a comparative sequencing study of subjects with and without hemophilia and HIV as part of a larger project at the university to sequence "around 200 genomes and 500 exomes from various disease cohorts including schizophrenia, epilepsy, amyotrophic lateral sclerosis, and HIV." (IS 9/21/2010)
Goldstein said that the center's growing experience in sequencing disease studies has informed its plans for this new project, as well as its expectations.
"We have been sequencing now for a little while and have definitely learned from that experience," he said. "And what we've learned is that this is going to be really hard.
"I don’t mean that facetiously," he added. "It's clear that the traits we are studying have a very high degree of genetic heterogeneity — both high locus heterogeneity and high allelic heterogeneity."
In light of this, "where we have solid statistical evidence of genes that carry risk factors for these diseases, we are going to have to pursue this diligently until we have a large enough sample size and high enough quality data that the risk factors begin to emerge," he said.
"It's going to be a while."
Have topics you'd like to see covered in Clinical Sequencing News? Contact the editor at mashford [at] genomeweb [.] com.