With an ever-increasing number of options for diagnostic sequencing, the decision of whether to do a gene panel, exome, or whole genome becomes complicated. At the Cambridge Healthtech Institute's Clinical Genome Conference in San Francisco last week, a number of experts in the molecular diagnostics and next-generation sequencing industry discussed the benefits and limitations of the various options.
Representatives from diagnostic companies looking to implement next-gen sequencing tests, clinicians using them in their practice, researchers, and representatives from industry weighed in on whether panels would continue to be relevant with the falling sequencing prices, the issues of incidental findings with exome and whole-genome sequencing tests, interpretation, and costs.
One issue brought up during a panel discussion was whether panels were still justified given falling prices and the move toward whole-exome sequencing by companies and academic labs.
"It's a big question about where things are going to go," said Robert Boles, medical director at Courtagen and clinician at the Children's Hospital of Los Angeles. "Soon it will be cost effective to use an exome on everyone," he said.
Currently, Courtagen offers diagnostic panels for mitochondrial diseases, including MtSeek, a 37-gene panel that analyzes the mitochondrial DNA genes, and NucSeek, a 1,197-gene panel for the nuclear genes involved in metabolism.
Boles has also implemented next-gen sequencing in his own clinical practice for patients with suspected mitochondrial disorders and said that the technology "revolutionized" his practice. "Patients that didn't have a diagnosis or only had a phenotypic diagnosis now have a molecular diagnosis," he said. And, "about half the patients have improved significantly because of the diagnosis."
While he said that the number of patients receiving a diagnosis was not surprising, he said he was surprised by how often there was a treatment available for the particular genetic alteration.
When deciding between panels and exome tests, Boles said that he will offer a panel if he has an inkling of what may be causing the phenotype and an exome if he has no clue.
"If you have a good phenotype, you might as well look at a small section," he said.
While Boles and others have had success in using exome sequencing to diagnose rare disorders, Andrew Grupe, senior director of pharmacogenomics in Quest's Celera business, expressed concern about using exome sequencing as a diagnostic tool.
Next-gen sequencing should be used to "follow up on clinical information that is known already," he said, rather than as a discovery tool.
Quest currently offers an HIV tropism test using a combination of Sanger and next-gen sequencing on Roche's 454 GS Junior platform, and the company has said that it is looking to develop other next-gen sequencing-based diagnostic tests for which it is evaluating the GS Junior, Illumina's MiSeq, and Life Technologies' Ion Torrent PGM machines (CSN 10/10/2012).
"If you do exome or whole-genome sequencing, you generate so much more information that is not related to the disease profile you are looking at," Grupe said.
While incidental findings unrelated to the disease cannot be avoided entirely, choosing to look at genes that are likely relevant for the patient's clinical presentation will minimize those, he noted.
Additionally, Boles said physicians in particular are wary of incidental findings and will be more likely to adopt panels over exomes or whole genomes. From his experience, he said that around 90 percent of physicians do not want incidental findings. "They don't want to explain 40 different mutations," he said.
Among patients, Boles said he has seen a split among those who want incidental findings, those who don't, and those who "don't know what they're getting into."
He said that one option in the future is to do a whole exome or whole genome but only analyze a set portion of that. Then, if that test comes back negative, the data is already available for another analysis. "Soon it will be cost effective to do an exome on everyone but only analyze a panel," he said. Physicians will pay for the analysis rather than the sequencing.
Christos Petropoulos, chief scientific officer at Monogram Biosciences, which offers a Sanger-based test for HIV tropism, said that insurance companies are already splitting up reimbursement between the sequencing and analysis portions. "In my space, we get paid one reimbursement fee for generating the sequencing and another for interpreting it," he said.
Petropoulous said that Monogram is currently looking to switch from Sanger sequencing to next-gen sequencing for tests that analyze drug resistance in HIV, HCV, and HBV patients.
Robert Klein, chief business officer at Complete Genomics, anticipated that a model of disaggregating the sequencing and analysis portions would likely be easier to implement outside of the US in countries with national payor systems. Nevertheless, he said that such a model makes sense. "Our view is to sequence everything and use software to get the exact panel you want," he said. "And then you can expand that over time."
However, such models of sequencing a patient's genome once, storing it, and analyzing a portion at a time also have ethical implications, said Grupe. While new research could lead to a diagnosis in a patient for whom exome sequencing previously failed, it will also lead to additional incidental findings.
"What are the ethical implications of storing patient sequence data?" Grupe questioned. "You now have new incidental findings that haven't been reported to the physician or patient because they weren't known at the time."
Another issue with whole-genome or exome sequencing is that those techniques don't always capture problematic regions, whereas targeted panels can be enhanced in such regions, said Pierre Antoine Gourrand, an assistant professor at the University of California, San Francisco School of Medicine, who is studying the use of next-gen sequencing for HLA typing. There are "problems with indels, GC-rich regions, paralogous genes, and repetitive regions across HLA loci," when using whole-genome sequencing. But with panels, baits can be optimized for such regions.
At least for the time being, the experts agreed that panels would continue to have a role in the diagnostic setting, particularly if implemented at individual hospitals and smaller laboratories.
Whole-genome sequencing, even though it's becoming cheaper and faster, still "requires huge amounts of computational power and support facilities," said Klein. "It's going to be a while before whole-genome sequencing is routine in individual clinical labs," he predicted.