23andMe fails in BRCA testing, say Invitae’s scientists
Current DNA tests usually do not check whole sequences of genes. Typically, companies such as 23andMe or Ancestry look at single variants (SNPs) known to be associated with tested trait in a population. The reasons are technological – it’s still much more cheaper and quicker to use an SNP array instead of a sequencer.
This approach puts special weight on the choice of SNPs. If a tested individual and a source population do not match, the DNA test can check wrong places in the genome, therefore missing variants important for that specific person.
That’s where the study began: 23andMe tests BRCA genes for SNPs present in Ashkenazi Jews. According to the analysis of Esplin team, presented at the ongoing conference of American Society of Human Genetics, 94% of non-Ashkenazi individuals will receive false negative results. This means that the DNA test misses nearly all dangerous variants in the general population. The results are even worse within specific ethnicities, rising to 98% in Asians and 99% in African-Americans.
The study was performed on data of 4,733 patients tested positively for BRCA in a clinical setting.
It should be noted that the authors have conflict of interest, as they are employed by a competing company called Invitae. The results were not peer-reviewed, but they are not surprising – not only for the human genetics community, but also for 23andMe, where their spokesperson comments:
It is patently wrong to state that 23andMe delivers ‘false negatives’ for variants that it does not test nor claim to test for. This is a false and misleading characterization of 23andMe’s test.Source of the quote
Effectively, 23andMe clearly states at the product website, that it tests for 3 variants present in Ashkenazi Jews.
One can ask – what’s the point of this DNA test for BRCA, if it does not detect mutations for 94% people?
Lead quote: https://www.ashg.org/press/201910-limitations-dtc.shtml
ASHG presentation: E.D. Esplin; E. Haverfield; S. Yang; B. Herrera; M. Anderson; R. Nussbaum (2019). PgmNr 235: Limitations of direct-to-consumer genetic screening for hereditary breast, ovarian, and colorectal cancer risk.