PHOENIX (GenomeWeb News) – Whole-exome sequencing is inching its way into the clinic, and at the American College of Medical Genetics and Genomics annual meeting held here geneticists and clinicians showed how such sequencing could be used to find diagnoses for patients whose conditions had remained mysterious.
While whole-exome sequencing is not for all patients, speakers at a morning session noted that it could be used to help identify the source of disease for patients whose phenotype doesn't quite fit with known disorders, who show disease heterogeneity, and patients whose suspected disorder hasn't been able to be confirmed through other tests. These criteria closely follow the guidelines issued by ACMG last year.
However, whole-exome sequencing does not always resolve a diagnosis and, occasionally, further analysis of suspected pathogenic variants reveals that such variants are actually benign. Additionally, such tests occasionally turn up other, unrelated deleterious mutations.
At the Cleveland Clinic, which began offering whole-exome sequencing about a year ago and whole-genome sequencing shortly after that, Amy Shealy said that they've received results for seven patients. Four of those patients had deleterious mutations in a gene related to their phenotype, and all seven had variants of unknown significance that were in genes related to their phenotypes.
For example, Shealy, who is a genetic counselor in the Cleveland Clinic's Genomic Medicine Institute, recounted the case of an eight-year-old boy with a heart defect, intellectual disability, and other symptoms. In his exome, her team uncovered a nonsense mutation in SETBP1, which is associated with intellectual disability.
"Overall, our experience has been really positive," she said, adding that patients and their parents have largely been satisfied. Shealy noted that for patients who received a negative result, they could sometimes be at a bit of a loss as exome sequencing was really the last test that could be done.
Exome or targeted sequencing has also been used to isolate the cause of the retinal disease Leber congenital amaurosis. Xia Wang from Baylor College of Medicine presented his team's capture panel containing exonic DNA from 163 retinal disease genes, not just those associated with LCA. By applying that panel to a group of 189 patients with LCA for whom the genetic cause could not be determined, they found that 49 patients had mutations in known LCA genes while 24 patients had mutations in genes linked to other retinal diseases. They did not identify pathogenic mutations for the remainder of the patients.
For those 24 people with mutation linked to other diseases, seven people were re-diagnosed with a different disease, which actually better fit their phenotype, and five were confirmed to have LCA, despite having mutations in different genes. Wang notes that the initial clinical diagnoses have limitations, and molecular tests can help refine diagnoses.
"Molecular diagnostics should be an integral part of clinical diagnostics," he said.
As his findings indicated, there is also a range of phenotypes that can be associated with a particular disease. Indeed, in a separate presentation, Ambry Genetics' Layla Shahmirzadi noted that many of her exome-sequencing patients were not fitting into the typical phenotype associates with the condition. She and her team undertook a retrospective analysis of their first 50 whole-exome cases and found that few of those cases exhibited the most common features associated with the disease.
"We're definitely seeing a wider spectrum" of phenotypes for a particular disease through exome sequencing, Shahmirzadi said.
Additionally, other findings may be uncovered by using a whole-exome or whole-genome approach, and with the release of the ACMG guidelines Thursday morning regarding such incidental findings, there is an increased focus on which ones should be returned.
Cleveland's Shealy added that her group has found incidental findings in two patients that were medically actionable, namely GPD1L, which causes Brugada syndrome type 2, and KCNQ1, which causes long QT syndrome type 1.
Gail Jarvik from the University of Washington and her team are trying to get a handle on how common such incidental findings are. They gathered together 1,000 people, half of African descent and half of European descent, whose data were housed at the Exome Variant Server. They drew up a list of genes in adults that a group of 24 experts had agreed were actionable, and they also determined criteria to call a mutation pathogenic or likely pathogenic.
That group of 1,000 people was found to have 640 disease variants, according to the Human Genetic Mutation Database, but after filtering for unique variants and expert review of the primary literature regarding those disease variants, 20 people had findings that were actionable. The actionable genes found included BRCA1, MYCBPC3, CDH1, LDLR, among others.
Jarvik noted, though, that only three of those 20 people who had actionable findings were of African descent. This, she said, indicated that there is a dearth of data regarding deleterious variants in that population.
She added that the findings underscore a need to build a database of variant pathogenicity.
But just what is considered a pathogenic variant is also up for some debate as different centers and labs have different criteria. Heidi Rehm, from Partners Healthcare, noted in her presentation on the development of ClinVar, the database of human genomic variation, that there were differences in how disparate groups classified variants. She reported a 20 percent discrepancy rate, though that was mostly due to differences in a variant being called, for example, pathogenic versus likely pathogenic.
Still, Rehm added that there needs to be better classification standards. A working group at ACMG, she said, is focusing on doing just that.