NEW YORK (GenomeWeb) – About 40 percent of variants reported in raw data from direct-to-consumer genetic tests may be wrong, according to a new analysis.
There's a growing interest among the public for DTC genetic tests, with companies like 23andMe, Ancestry.com, Family Tree DNA, and others providing services. While such tests are not diagnostic, some DTC firms will give customers their raw genotyping data, with a disclaimer. Consumers' doctors might then send them for additional testing if their DTC results indicate they harbor a disease-linked variant.
Researchers from Ambry Genetics analyzed how accurate those DTC test results are. As they reported yesterday in Genetics in Medicine, they found that two out of every five samples sent to them for confirmation testing were false positives.
"Such a high rate of a false positives in this particular study was unexpected," first author Stephany Leigh Tandy-Connor, a genetic counselor at Ambry, said in a statement. "While DTC results may lead to healthy changes in lifestyle or diet, these could also result in unwarranted emotions, including anxiety when someone obtains unexpected information, inaccurate information, or disappointment when receiving a lack of comprehensive diagnostic analysis."
For this analysis, the Ambry researchers queried their internal database to identify patients who had been referred to them following DTC genetic testing. They uncovered 49 cases that came to them for confirmation testing between January 2014 and December 2017, largely for testing of cancer genes. For slightly more than half the cases, the variant of interest was provided to Ambry on a requisition form or in a clinic note, while slightly more than 25 percent were from a DTC or raw data report, and the rest from a third-party interpretation service.
Ambry's testing relied either on Sanger sequencing or next-generation sequencing followed by Sanger sequencing. The researchers noted that the DTC firms use a range of methodologies.
Overall, 60 percent of the variants Ambry analyzed confirmed the DTC results, but the rest were false positives.
All CFTR and all MEFV variants — which are linked to cystic fibrosis and familial Mediterranean fever, respectively — were confirmed, the researchers reported. Meanwhile, for BRCA1/2 variants, the pathogenic Ashkenazi Jewish founder variants were confirmed in all 13 cases, as were four other variants, but eight other BRCA1/2 variants were false positives.
Most of the false-positive calls — 94.1 percent — were of cancer-linked genes, the researchers reported. In some of these cases, the variants were misclassified in the raw data or by the third-party interpretation service.
For instance, the researchers noted that eight variants in five genes — ATM, BRCA1, BRCA2, COL3A1, and COL5A1 — had been deemed to impart increased risk in initial testing, but that Ambry and other clinical labs consider those variants to be benign. The variants, they added, are present in the general population at a level that's too high to link them to disease.
The study, Tandy-Connor noted, was small, with a cohort of less than 50 people. But she said that could be in part because most people don't seek confirmatory testing. "People may assume that they are being provided accurate medical grade testing, so understandably do not go to the trouble and expense of seeking confirmation," she said.
This study's findings highlight the need for such confirmations, she noted.