By Kevin Davies
November 7, 2012 | SAN FRANCISCO—A team of computational biologists and clinical geneticists from the Brigham and Women’s Hospital in Boston has won the inaugural CLARITY Challenge to identify and present the putative mutations underlying the rare disorders of three children or infants who have received care at Boston Children’s Hospital and had their full genomes sequenced.
For one patient, sixth-grader Adam Foye, the contest ended a 10-year diagnostic odyssey marked by the indeterminate analysis of numerous candidate genes for his rare muscular disorder, known as centronuclear myopathy. In his case, the culprit turned out to be in titin, the largest gene in the human genome, which encodes a muscle protein.
The competition—CLARITY stands for Children’s Leadership Award for the Reliable Interpretation and appropriate Transmission of Your genomic information—was conceived by Isaac Kohane, David Margulies, and Alan Beggs at Boston Children’s Hospital/Harvard Medical School. 30 teams from academia and industry initially entered the competition, although seven withdrew before the final phase.
|Adam and Sarah Foye
The winning team—led by Shamil Sunyaev and assisted by investigators including Heidi Rehm (Harvard Partners), Daniel MacArthur (Mass General Hospital) and Mike Murray (Brigham)—receives the top prize of $15,000. Two teams were named co-runners up and win $5,000 apiece: the University of Iowa and a German team consisting of researchers from Genomatix, CeGaT (Tübingen) and the Institute of Pathology at University Hospital of Bonn.
The competition was launched in January 2012 and sponsored by Life Technologies and Complete Genomics, which also sequenced the genomes of the three patients and their parents. The results were announced at a press conference at the American Society of Human Genetics (ASHG) conference taking place in San Francisco this week.
Entries from the 23 registered teams were reviewed by a panel of judges chaired by Duke University geneticist Huntington Willard. The panel decided that the Brigham and Women’s entry featured the best combination of bioinformatics analysis and utility of its clinical reports for the three families, as well as the correct identification of the genetic basis for the Foye family’s disorder. The task of not only accurately interpreting the genetic basis of inherited diseases and cancer but also communicating those results in a medically meaningful and accessible way is considered a huge bottleneck towards the successful implementation of genomic medicine.
Of the two runners-up, the Genomatix team was considered to be the only group to correctly identify the putative mutation in all three families. The Iowa team drew credit for its approach to returning unexpected genetic results based on patient preferences and indicating regions of low confidence or coverage.
A “special mention” went to five other teams for particularly strong entries:
- The Clinical Institute of Medical Genetics (Ljubljana, Slovenia)
- The Research Institute at Nationwide Children's Hospital (Columbus, Ohio)
- Science for Life Laboratory (SciLifeLab) of the Karolinska Institute (Solna, Sweden)
- Scripps Genomic Medicine, Scripps Translational Science Institute
- SimulConsult (Chestnut Hill, Mass.) and Geisinger Health System (Danville, Pa.).
“The community took this challenge very seriously, and we had tremendous participation,” says Margulies, executive director of the Gene Partnership at Boston Children’s Hospital. “We got the best thinking from around the world, and it has moved us toward a consensus on how to report sequencing data for use in the clinic.”
“There was remarkable uniformity in the bioinformatics pipelines,” said Catherine Brownstein, involving the use of tools including BWA, GATK, Annotate, Filter, SIFT and Polyphen. A full account of the results will be presented in an upcoming paper.
For the Foye family, the CLARITY contest ended an interminable diagnostic odyssey. Adam Foye had been tested for a succession of candidates genes associated with his disorder, centronuclear myopathy. He also suffers from hearing impairment, another suspected inherited disorder.
Eight of the 23 CLARITY teams identified alterations in the gene for the muscular protein titin as the culprit of Adam’s muscular weakness. In addition, six teams revealed mutations in a gene called GJB2 as the likely cause of his hearing loss. Three teams identified both gene mutations.
Alan Beggs, director of the Manton Center for Orphan Disease Research at Boston Children’s Hospital, has recently identified mutations in the titin gene in 4 other patients (out of 28) with centronuclear myopathy using exome sequencing. But “it’s an enormous gene,” says Beggs, “and to sequence it individually by hand would have taken nine months in the lab at a prohibitive cost. That’s why genomic sequencing is such a revolutionary technology.” Indeed, it would have cost more just to sequence titin alone than doing the whole genome sequence.
Adam’s mother, Sarah Foye, attended the press conference holding a framed photograph of her son. She offered emotional thanks to the competing teams, organizers and sponsors of the CLARITY contest. Her son’s diagnostic odyssey has lasted 11 years, she said. “It is thanks to the advent of the technology of whole-exome sequencing that we can arrive at an answer,” she said. “This is a huge success for our family.
She added: “It doesn’t mean we know the treatment now, but it’s pointing us in the right direction and we can cross other possibilities off the list.”
The second patient, Liam Burns, died less than 2 weeks after birth suffering from irregular cardiac rhythm. Seven teams fingered mutation in a gene called TRPM4, although this does not necessarily explain the structural heart defects that affected Liam and several other family members.
The third patient also suffers from a muscle-weakening disease called nemaline myopathy, but here the findings were equivocal. Seven genetic variants were reported by two or more teams as potentially deleterious, while four were judged worthy of further investigation, including variants in two genes not previously associated with this disorder. Researchers at the Manton Center are conducting further studies on the TRPM4 gene and the putative variants in this third case.
The leader of the winning team, Shamil Sunyaev, is a computational biologist and associate professor in the Division of Genetics, whose research applies bioinformatics approaches to problems of evolutionary genetics and population genetics. One of the tools Sunyaev’s group has said it would use is a program called Spliceman, developed by William Fairbrother and colleagues at Brown University, to identify the likelihood of single-base variations causing gene splicing abnormalities.
Beggs said that the eight finalists each had a well integrated bioinformatics pipeline. The successful teams typically had highly multi-disciplinary teams, including computational scientists, bioinformaticians, medical genetics, and disease-domain expert biologists. On the other hand, some groups neglected to consider pattern of inheritance.
The winning Brigham team, said Beggs, provided “the very best clinical report—a short 2-3 page report patterned after a typical clinical test [report],” providing information on interpretation, pitfalls, a short methods description, and appropriate references of the tested mutations that other researchers and clinicians could follow up. “That was put together with a strong informatics pipeline,” he said.
The University of Iowa team (runner-up) gave “a very nice report” presenting information on the regions of sensitivity and/or low coverage, down to the base position. The Iowa team was also credited for providing its consent form as part of the entry, which allows patients to set preferences such as whether they wish to discover incidental findings.
The Genomatix consortium was the only team that “got the right answers” for all three patients, said Beggs, while stressing that “right answers” was in quotation marks. The team identified the titin mutations in Adam as well as another genetic abnormality that explained Adam’s hearing loss.
“You need proper bioinformatics but sensory hearing loss is not part of [Adam’s muscle disorder]. For a clinical diagnostics lab, you don’t have that expertise on hand,” said Beggs. The best groups brought that insight either through data mining or medical expertise. “Genomatix was the only team we feel identified all the probable causes of all genetic conditions,” including a gene for family #3.
While Complete Genomics provided their own list of variants for the sequenced genomes, Life Technologies provided only raw SOLiD data. “That was part of the challenge,” said Beggs. “Which groups could process those data?”
Heidi Rehm, a clinical lab director at Harvard Partners and a member of the winning Brigham group, said the competition had been an “incredible experience.” For ten years, she has struggled with the division between physicians, clinical laboratory, and the research wing. “This challenge was an incredible experience to break down those barriers and allow those teams to work together in collaborative ways.”
As a result, her group had established much better communication channels, evening co-hiring a new employee with Sunyaev “to bridge those teams.”
One of the judges, Elaine Lyon (ARUP Utah) said that the judges liked parts of many of the reports entered. “People had unique ways of presenting complex data,” she said. “This is not simple. How do we convey that and put it in a written document?”
The CLARITY team plans to hold a second competition for cancer genomes in 2013.