Skip to main content
Premium Trial:

Request an Annual Quote

Duke Center Looking to 'Significantly' Expand Clinical Exome Sequencing Following Pilot Project


Duke University researchers are working to further develop the clinical exome sequencing program at that institution in the wake of a pilot study suggesting the approach can help find causal mutations in a significant proportion of difficult-to-diagnose cases.

"What we're trying to do now is very significantly expand [clinical exome sequencing] at Duke," senior author David Goldstein, director of Duke's Center for Human Genome Variation, told Clinical Sequencing News. "What I would like to be able to do here is offer this to all the children who come through Duke who don't get a genetic diagnosis — that's the ambition."

In the Journal of Medical Genetics this week, Goldstein and colleagues from Duke and the University of Michigan outlined the successes and limitations they encountered when they did exome sequencing on parent-child trios to try to diagnose unidentified genetic conditions in a dozen individuals.

The team found what they believe to be disease-causing mutations in six of the 12 cases through that pilot effort. The analysis also uncovered mutations that seem to explain macular degeneration symptoms in a seventh individual, though her condition involves other clinical features believed to stem from additional genetic contributors.

Based on the results of the pilot effort, Goldstein and his colleagues have been enrolling additional families for diagnostic exome sequencing through a sequencing clinic set up through Duke's Center for Human Genome Variation. The clinic meets every few months to enroll additional patients.

The sequencing is not being offered in a clinically billable manner and is, instead, being done with research dollars, according to Goldstein. He said funds allocated to the effort so far should allow for clinical exome sequencing on around 50 more trios, though the goal is to eventually offer such services much more widely.

Currently, patients who are referred to the clinic with an apparent genetic condition that cannot be diagnosed in other ways are evaluated by Duke clinical geneticist Vandana Shashi, a co-first author on the Journal of Medical Genetics study, to determine whether they meet all of the criteria for clinical sequencing.

As in the pilot study, families selected for sequencing receive genetic counseling both before and after sequencing and have the option of having information on strong causal mutation candidates passed on to their physicians, if and when they are detected.

High-throughput whole-exome sequencing has been finding favor as a diagnostic tool for commercial clinical labs, such as Ambry Genetics, EdgeBio (CSN 5/2/2012), and GeneDx (CSN 10/19/2011); and academic centers, including Baylor College of Medicine (CSN 11/16/2011), Washington University School of Medicine's Genomics and Pathology Services (CSN 2/29/2012), the University of California at Los Angeles (CSN 3/7/2012), and the Emory Genetics Laboratory (CSN 3/21/2012).

At Duke, researchers have been testing the clinical exome sequencing waters for the past couple of years, using high-throughput sequencing to try to diagnose 12 individuals with unexplained conditions involving a range of phenotypic and clinical features.

The patients, who were enrolled through Duke's Center for Human Genome Variation, did not carry chromosomal abnormalities that could be detected using Affymetrix 6.0 arrays. Nor did they carry genetic alterations that could be detected by conventional genetic tests suggested by their physicians.

For each of the parent-child trios, researchers isolated genomic DNA from blood samples, captured coding sequences with the Agilent SureSelect Human All Exon 50Mb kit, and sequenced these exomes to an average of 71-fold coverage on the Illumina HiSeq 2000.

The researchers then interpreted the exome data not only by comparing the genomes of cases to those of their unaffected parents, but also by comparing these exomes to sequence data from controls and data from the University of Washington's Exome Variant Server, or EVS, database.

Once they had exome sequence information for each trio, the team sifted through the data to look for variant patterns in the children that would correspond to de novo mutations or possible disease-related variants that had been inherited in an X-linked or recessive fashion.

For the latter, researchers focused on variants that were heterozygous in both parents and homozygous in the child as well as situations in which children had inherited different alterations to each allele of a given gene, known as compound heterozygosity.

"The high-level approach is to look for very high-penetrant genotypes," Goldstein explained. "The idea is to take all of the obvious genetic models that might be responsible in this kind of a setting."

From there, they looked at which of the variants that conformed to one of the plausible genetic models also landed in a gene that might fit with the phenotype observed in each child based on knowledge of Mendelian disease genetics.

In some instances where functional consequences of variants were not clear from previous studies, researchers did their own functional follow-up to get a better sense of the potential pathogenicity of these changes.

The team then sent samples away for testing in a CLIA-certified lab to verify any apparently causal mutations before the families met with the team's genetic counselor and clinical geneticist to learn about the results of the analysis. Information on likely disease-causing variants was passed along to the child's physician when permitted by the families.

In six of the 12 children tested, investigators found mutations that were considered likely disease causing. There were some suspicious variants in the six individuals for whom they did not find strong causal mutation candidates, though those alterations were less conclusively linked to disease.

Though he said that the ability to interpret variation in the genome remains "a little bit of an art," Goldstein said the focus on Mendelian genes makes it possible to get a reasonable degree of confidence when trying to make diagnoses from sequence data.

"In a study like the one we've just done, I would be quite confident that most of the likely genetic diagnoses we have are correct, but it's certainly possible that one or two are not," he said. "And that is part of doing this work and one of the challenges for this field — to try to correctly communicate the degree of confidence when you decide to communicate the results."

For instance, one of the children diagnosed in the study was found to have a de novo truncating mutation in TCF4, a gene that's mutated in individuals with Pitt-Hopkins syndrome.

In retrospect, researchers realized that the patient did have some of the features normally associated with the condition but had not originally been tested for TCF4 mutations because she lacked other Pitt-Hopkins symptoms such as hyperventilation and epilepsy.

Another two children had mutations in EFTUD2. It was not clear whether alterations in that gene were consistent with the children's clinical features until the gene was implicated in a very recent American Journal of Human Genetics study involving children with similar phenotypes.

"Clearly there will be more diagnoses over time," Goldstein said, explaining that the ability to interpret information in whole exomes and genomes will improve as centers amass sequence data on more and more clinically phenotyped individuals.

For his part, Goldstein is keen to see the adoption of a central database where researchers and clinicians can deposit both clinical and genetic information to help with future efforts to diagnose individuals using sequence data and make connections between different mutations and phenotypes.

"There are tens of thousands of children like this born every year," he said. "As we analyze large numbers of those, we will be able to find new connections and be able to move beyond already known Mendelian disease genes."

At Duke, co-author Shashi and her pediatric medical genetics colleagues have started putting together a clinical database for patients similar to those included in the current study, Goldstein noted, explaining that the researchers hope to add in a genetic component to the database in the future.

Given privacy and other concerns, he said the genetic information linked to each clinical case may initially be limited to a handful of variants of potential interest rather than complete exome or genome sequences.

He also stressed the importance of having access to parental sequence data, especially since the "vast majority" of apparently causal variants found in the pilot study relied on comparisons between each patient's exome and the exomes of his or her parents.

"For conditions like this, where you're dealing with relatively high-impact mutations, having the parents is all important," Goldstein said.

For the current study, the consumables cost associated with sequencing each individual was on the order of around $1,000 per exome, or $3,000 per trio, and it took roughly six months from when researchers received patient samples to when they could offer an opinion on whether any of the genetic changes they saw appeared to be causal.

The team plans to continue using the Illumina HiSeq 2000 for clinical exome sequencing for the foreseeable future and is still aiming for an average of 70-fold coverage per exome — a depth that Goldstein said has so far proven useful for distinguishing authentic de novo variants from false positive variants.

At the moment, the Duke team is primarily sticking with whole-exome rather than whole-genome sequencing for its initial diagnostic efforts, owing to the price and time difference between the two.

Even so, the researchers are taking something of a "two-tiered" approach at the moment, moving on to whole-genome sequencing for some families who do not receive diagnostic information after clinical exome sequencing.

"The price difference is large enough that that's the way we prefer to run it," Goldstein said. "But eventually we will certainly transition to doing this to whole genomes."

The Scan

Octopus Brain Complexity Linked to MicroRNA Expansions

Investigators saw microRNA gene expansions coinciding with complex brains when they analyzed certain cephalopod transcriptomes, as they report in Science Advances.

Study Tracks Outcomes in Children Born to Zika Virus-Infected Mothers

By following pregnancy outcomes for women with RT-PCR-confirmed Zika virus infections, researchers saw in Lancet Regional Health congenital abnormalities in roughly one-third of live-born children.

Team Presents Benchmark Study of RNA Classification Tools

With more than 135 transcriptomic datasets, researchers tested two dozen coding and non-coding RNA classification tools, establishing a set of potentially misclassified transcripts, as they report in Nucleic Acids Research.

Breast Cancer Risk Related to Pathogenic BRCA1 Mutation May Be Modified by Repeats

Several variable number tandem repeats appear to impact breast cancer risk and age at diagnosis in almost 350 individuals carrying a risky Ashkenazi Jewish BRCA1 founder mutation.