HUNTSVILLE, Alabama (GenomeWeb) – Whole-genome sequencing was able to diagnose nearly 30 percent of children with unexplained developmental delay or seizures, with subsequent re-analysis boosting the numbers of diagnoses made.
As part of the Clinical Sequencing Exploratory Research (CSER) consortium, a HudsonAlpha Institute for Biotechnology-led team performed whole-genome sequencing on about 600 children and their parents to try to pin down the cause of their condition. CSER projects, funded by the US National Institutes of Health, aimed to see how genomic sequencing could be used in clinical settings.
More than three quarters of probands they did diagnose genetically had de novo mutations, said Greg Barsh, a faculty investigator at HudsonAlpha, during a session of the 2018 Genomic Medicine Conference being held at the institute in Huntsville, Alabama, this week.
However, Barsh noted that "we are missing some stuff" among patients.
He and his colleagues recruited 536 probands for their study. The probands all had unexplained developmental delay or seizures that more conventional molecular diagnostic tests were unable to diagnose. Through trio sequencing, they aimed to identify causative variants as well as secondary findings.
After variant calling, Barsh said they had between 3 million and 4 million variants that they then filtered based on their annotations, population frequency, and inheritance patterns to cull the list down to between 10 to 100 variants per case. Those then went through manual review and a variant review committee that weighed in on which variants to return as pathogenic.
For 155 of the probands, Barsh and his colleagues were able to identify a variant that was pathogenic or likely pathogenic. Of those, most were de novo mutations, but about 12 percent were autosomal recessive and 7 percent were X-linked.
Overall, 12 percent of the probands had a variant of uncertain significance, and Barsh said most of those were de novo mutations as well.
In a separate talk on secondary findings, HudsonAlpha's Kelly East, a genetic counselor at the institute, said 8.6 percent of participants in this study received secondary findings of either the 59 conditions on the American College of Medical Genetics and Genomics' list or carrier status results for cystic fibrosis, sickle-cell anemia, or Tay-Sachs disease.
But, Barsh noted in his presentation, sequencing wasn't catching everything in their patient cohort. There were undiagnosed patients with severe symptoms, who have a number of affected family members where it "smells genetic," he said, adding this means there is room for improvement.
For instance, Barsh noted that the Combined Annotation Dependent Depletion (CADD) algorithm, developed by HudsonAlpha's Gregory Cooper and the University of Washington's Jay Shendure, helps automate the process of determining which variants are likely to be deleterious. He added that the tool is genome-wide — includes non-coding sequences — and is extensible as new annotations are made.
Their findings also underscore the importance of data sharing, according to Barsh. He said that many of the VUS results they had seemed promising as an explanation of patients' symptoms, but hadn't been reported by another group, leaving the evidence for it weak.
He said something like Baylor-Hopkins Center for Mendelian Genomics' GeneMatcher, which enables people to ask others if they've seen something like their VUS in other patients, could help. For instance, they found another group with 10 additional patients in four countries with de novo mutations in EBF3, similar to what they'd found in two patients at HudsonAlpha, strengthening its association with intellectual disability.
Barsh also said that as knowledge improves over time, going back to re-analyze data could shift variants from one column to another. Fifteen percent of the pathogenic or likely pathogenic variants they found were identified after the initial assessment, he said. Of these, 23 variants were upgraded, while seven were downgraded
As part of a follow-on Clinical Sequencing Evidence-Generating Research (CSER2) program, he and his colleagues will be conducting whole-genome sequencing of newborns in nurseries in Alabama, Mississippi, and Louisiana, a project dubbed SouthSeq. Through that initiative they aim to develop and test the effectiveness of sequencing on diagnosing infants in the NICU, especially at hospitals representing minority and rural populations.