NEW YORK (GenomeWeb) – Carrier screening tests are able to look for an expanding number of pathogenic mutations, beyond traditional testing for cystic fibrosis and other conditions, due to the dropping cost of next-generation sequencing. However, there is no standard expanded carrier screening test and the firms developing such tests each analyze different gene regions, which can lead to disagreeing results.
To get a better handle on potential variants that may be missed from these commercial tests, GenePeeks, which is developing a carrier screening test that analyzes all coding regions of 479 genes, analyzed 27 sperm bank donor applicants who had tested negative on an initial cystic fibrosis carrier screening test. The company also analyzed genomic regions in these individuals that are covered by commercially available tests from Counsyl, Good Start Genetics, and 23andMe. While the GenePeeks authors reported in a recent study published in Genetic Testing and Molecular Biomarkers that the commercial tests did not provide the same answer for each of the 27 donors, the firms questioned GenePeeks' methods and conclusions.
Lee Silver, co-founder and chief science officer at GenePeeks, told GenomeWeb that the study "goes against the notion that genetic testing is a replicable diagnostic." Even though the individual tests are "analytically accurate, it's a question of what variants are included."
The researchers collaborated with a sperm bank that had recently switched from screening all donors with a cystic fibrosis test to a more comprehensive test developed by Good Start Genetics, which analyzes 19 genes using NGS and also reports carrier status for alpha- and beta-thalassemia and spinal muscular atrophy with non-NGS methods. The sperm bank made available de-identified results from the cystic fibrosis test and from Good Start's GeneVu test.
The GenePeeks researchers then sequenced all 27 samples using Illumina's TruSightOne panel, which covers the coding regions of 4,813 genes. It then ran its own test, which analyzes the coding regions of 479 genes. The GenePeeks test is based on an assay originally developed by Stephen Kingsmore's laboratory when he was at the National Center for Genome Resources.
Because the GenePeeks researchers did not have access to Counsyl's or 23andMe's tests, the researchers performed what they called a proxy test, using bioinformatics to call variants at each locus typically covered by their respective tests. For the Counsyl test, the researchers analyzed variants in 98 genes, as indicated by version 1 of Counsyl's Family Prep Screen. To simulate the 23andMe test, they looked at 35 genes that are part of the firm's Personal Genome Service.
The original cystic fibrosis-only test did not find any carriers. Good Start's GeneVu test identified six disease carriers — one who was heterozygous for a Gaucher disease variant, one who was heterozygous for a novel variant predicted to be associated with Bloom syndrome, two carriers of SMA variants, and two alpha-thalassemia carriers.
23andMe's test identified just one carrier, who was heterozygous for a variant related to a metabolic condition known as MCAD deficiency. Counsyl's test identified the same individual as a carrier for MCAD deficiency, as well as the same person as a carrier for Gaucher disease that Good Start's test identified. Counsyl's test also identified a donor as a carrier for variants associated with multiple fetal abnormalities and intellectual disability and a donor who was a carrier for McArdle disease, a glycogen storage disorder. Neither the 23andMe nor the Counsyl test identified the two individuals that Good Start's test called as Bloom syndrome carriers, or the persons called as carriers for SMA and alpha-thalassemia, the researchers wrote.
However, according to Eric Evans, chief scientific officer at Counsyl, the researchers based their analysis on an older version of Counsyl's test that relied on a SNP chip, rather than its most recent NGS-based test, which it has been offering for the last three years, Evans said. The NGS test uses full-exon sequencing, he said, and "detects the variants noted in this study."
The GenePeeks team next analyzed the sequence data for all 4,813 genes on the TruSightOne panel, essentially finding variants in every single donor, Silver said. The study "shows the absurdity of using a targeted mutation panel," said he added.
However, Evans disagreed with Silver's conclusions, saying that the GenePeeks team's variant classification did not follow standard validated methods. For instance, he said, the GenePeeks researchers identified five samples, each of which had one of three different variants in the CYP21A2 gene, which is associated with congenital adrenal hyperplasia. However, that gene also has a highly homologous pseudogene. Two of the variants that the GenePeeks team called are likely pseudogene-derived, Evans said. "There is a considerable risk that these are false positives."
In addition, the GenePeeks team included variants that "should be reported as [variants of] unknown significance, rather than disease causing," he added, which incorrectly led them to conclude that all 27 donors were disease carriers.
Stephanie Hallam, medical director and vice president of laboratory operations at Good Start Genetics, agreed that even though the GenePeeks study focused on variants within the ClinVar database, as well as variants predicted to cause gene dysfunction, those shouldn't necessarily be reported to someone undergoing carrier screening because of their uncertainty. The "assessment of variants of uncertain significance is not appropriate for medical carrier screening where a specific couple is being assessed for carrier status," she said. "Clearly, there is a need to increase our knowledge about genetic variants and their effect, but this is best assessed by concerted clinical studies on affected individuals, or in research studies, not on carriers who do not manifest disease symptoms."
GenePeeks' Silver agreed that it would not make sense to disqualify all 27 donors from the sperm bank because they carried mutations in one of the 4,813 genes related to disease. And GenePeeks itself doesn't screen all those genes, but a subset of 479. Instead, he said, both the donor and the recipient should be screened, which would identify incompatible matches. The company offers such a service, called Matchright. So far, it has partnered with two sperm banks for its service and is looking to expand to egg banks, as well as to testing couples through fertility clinics, Silver said.
He said that the firm is taking such a comprehensive approach to testing in order to provide individuals using sperm banks and couples that get tested with a more definite result. "A lot of clients believe that donors are tested for all recessive mutations [in disease causing genes], when that is not the case," he said.
But Good Start, Counsyl, and 23andMe disagree. Good Start's Hallam and Counsyl's Evans said that many of the variants reported by GenePeeks are not established as being disease-causing. Also, a spokesperson for 23andMe disagreed with GenePeeks' conclusion that negative tests give patients a false sense of comfort.
The spokesperson told GenomeWeb in an email that its carrier status reports are "explicit with regard to what genes are tested and what variants are tested." With regards to Bloom syndrome specifically, the spokesperson said that user comprehension of its genotyping test report is greater than 90 percent, "including limitations of the test."
The GenePeeks study is not the first to demonstrate that most people are carriers of potentially pathogenic variants. For instance, researchers involved in the 1,000 Genomes Project reported a few years ago in the American Journal of Human Genetics that all of 179 healthy individuals contained about two variants on average in genes associated with disease.
Both Silver and Evans agreed that it would not make sense to exclude donors from sperm banks based on that information, since everyone would be excluded, and that it would be more important to know the carrier status of both the donor and the recipient.
Going forward, Silver said that GenePeeks plans to focus on screening both the donor and recipient and to expand its test from 479 genes to around 1,000. It is also conducting a pre-conception risk study in 200 prospective parents to compare its carrier screening test with others.