why your ancestry test probably isn’t that accurate
Ancestry tests are all the rage right now. Millions of people with vague family histories and gaps in their genealogical trees are getting them to find out who their ancestors were. Companies are promising to discover sinister diseases lurking in your genes. Doctors are talking about medication based on your genetic history. (And Aeromexico used them to troll Americans in a new ad.) My wife and I also had our genes tested to figure out who exactly we are and found out that she’s British and Irish, while I’m so Eastern European, my results came back with an Adidas tracksuit and instructions on proper squatting techniques. But just how accurate are those results? How much faith should you put in them? If you’re less than five percent something, does that mean a link to some ancestry you knew nothing about, or is it just statistical noise?
While it’s tempting to think we know enough about genetics to accurately place the history of your progenitors, the accuracy of these testing services depends on the data they have to calibrate their analysis. That data is generally proprietary, seldom shared, and any geneticist will advise you not to take it too seriously. Just to give you an idea of the kind of margin of error you might see, consider a recent experiment in which identical twins sent their DNA to five major ancestry testing services to receive wildly different results. Yes, the ancestry breakdowns for each service were roughly similar for each twin, but the actual ancestries were all over the place.
From personal experience, we saw similar inconsistencies in our results. While my results started off in Ukraine and Russia, with traces of random ancestors from as far as Persia and the UK, they were updated primarily to ancestors in Baltic states. My wife’s original report declared her 60% Scandinavian only to correct that to 0% and re-assign a similar percentage to Britain to her great chagrin because she was very excited about having Vikings for ancestors. As you can see, you shouldn’t exactly bet money on these results being right, and there’s quite a bit of disparity even within each service’s data set over time. And these inaccuracies are even more disconcerting when the genetic testing service offers a report on potential genetic disorders, especially ones requiring complex genetic mutations as predictive markers.
A big part of the problem is that the baseline genetic markers are extracted from people who self-report as being a certain ethnic group and nearly three quarters of them come from just three countries: U.S., UK, and Iceland. This means Asian and African genetic markers are insufficiently studied, and considering that African lineages are some of the most diverse on the planet, we’re potentially missing a lot of fascinating data and genetic relationships. In short, we lack the details of large parts of our genetic histories because we don’t have enough samples for a complete understanding of what denotes one ethnic group from another outside of some European territories.
Of course, this doesn’t mean that your genetic analysis is completely worthless. It can highlight some obvious genetic markers we understand well. But when it comes to complex patterns of heredity and inherited diseases and mutations, what we have today is just a fairly rough idea with very significant limitations that show up as errors and corrections in our ancestry tests. We’re going to need a lot more data, a lot more research, and more samples from all over the world to have a better understanding of what we’re seeing in the lab and why, and the first step is recognizing the gaps in our current knowledge.
See: Mills, M., Rahal, C., A scientometric review of genome-wide association studies, Communications Biology Vol. 2, Article No: 9 (2019), DOI: 10.1038/s42003-018-0261-x
