NEW YORK (GenomeWeb News) – The risk information provided by direct-to-consumer genetic testing companies are often inconsistent, according to an opinion article appearing online today in Nature.
Researchers from the J. Craig Venter Institute and the Scripps Translational Science Institute compared 23andMe and Navigenics DTC genetic test results for five individuals, focusing on more than a dozen diseases. The team found discrepancies in disease risk for seven of the diseases, leading them issue a set of recommendations for DTC testing companies and the genetics community as a whole.
So far, the team noted, roughly 1,000 or more disease-associated DNA variants have been identified. Although DTC testing is available for a range of traits and diseases, they added, the genetic underpinnings behind many complex diseases are unknown, leading some to question the benefit of DTC tests.
To explore this further the group compared test results from consumer genomics firms 23andMe and Navigenics, focusing on five individuals' risk for 13 diseases.
While they found that the accuracy of raw data for both tests was high and consistent from one test to the next, the team also noted that interpretation of these results often differed.
For four of the 13 diseases, the researchers got the same results from Navigenics and 23andMe. On the other hand, for another seven diseases the predictions from each company matched half or less than half of the time, potentially due to differences in the markers tested and the way populations are defined.
"DTC companies harness the same publicly available research to decide which markers to include, and for the most part, could use the same or similar markers," senior author Craig Venter and co-authors wrote. "Yet no disease has an identical set of markers between the two DTC companies because each company has its own criteria for accepting genome-wide association results into its relative risk calculation."
These differences can lead to disparate risk predictions when the markers in question don't have strong effects, the researchers explained. They also noted that some risk information is couched in a way that may be difficult for the average consumer to understand.
Based on their comparison, Venter and his co-authors put forth nine recommendations for DTC genetic testing — five for DTC companies and four for the community in general.
For instance, the team suggests DTC companies should report the genetic contribution of each of the markers tested. And, they say, companies should focus on high-risk predictions and pharmacogenomic markers determined through direct genotyping of markers that have strong effects.
Meanwhile, the researchers recommended that the community as a whole should keep track of the way consumers use DTC tests and how these tests affect their behavior. They also emphasized the need for doing prospective studies to test the clinical validity of DTC testing and to replicate disease-association studies in non-European populations.
Finally, when widespread sequencing becomes more common, the team said, the community will need to carefully weed out benign variants from those involved in disease risk.
"Our ability to identify variants from comprehensive sequence data will far outstrip our ability to characterize their biological effect," the group concluded. "However, accurate and complete reporting is a necessary predecessor to a precise functional understanding of genomic data for the consumer."