23andMe and ancestry.com are accurately sequencing DNA and providing ancestry interpretations in line with peer-reviewed science.
2. Heritage tests are less precise if you don’t have European roots.
Human ancestry is extremely well-studied. These studies have included large numbers of genomes from many different ancestry groups. Certainly, the common ancestry groups among the likely readers of this article are well-represented and studied.
3. Your DNA says nothing about your culture.
From a statistical point of view, your DNA contains information about the ethnic groups to which your ancestors belonged. That in turn contains information about the culture in which your ancestors, and you, find themselves. If the author means "says nothing" in a sense other than "statistical information" then it would be interesting to learn what this sense is.
23andMe don't sequence DNA. They use extremely low resolution SNP chips. Such SNP chips are currently used in exactly zero real world (human) medical applications. This is because they are useless as a test with anything near acceptable diagnostic performance. Anyway, the point of the article is that the interpretation of the results may not be accurate, and that is a fairly generous way to put it.
> Such SNP chips are currently used in exactly zero real world (human) medical applications.
That's not true at all. There have been dozens of publications using that data and they have been instrumental in progressing our knowledge of genetic associations.
I have personally used 23andMe's array based data and resulting associations to investigate disease. It is priceless and hugely beneficial.
There is no medical test that uses these chips. You describe research into genetic associations - that is an entirely different matter.
Also, whether the type of information obtained from a 23andMe style test has any material impact on health is very much unproven, as outlined in their own terms of service, and the fact they were poleaxed by the FDA for overselling the utility of the test.
It tells you if you have the SNP it tests for in those genes. It doesn't tell you anything else.
Gilbert's syndrome is pretty common. The vast majority of people that have it are not bothered at all. There is also no proof that lifestyle interventions are in any way helpful. This is the case for basically everything that 23andMe tests for.
If you want to raise a proper counterexample, you would choose the BRCA1 gene. But again 23andMe's test doesn't tell you whether or not you have a BRCA1 mutation, it only tells you if have one of the SNPs they test for. If you don't get a positive result, and that is the vast majority of people, you aren't any better off. That's why in the clinic when someone needs a BRCA1 test, we sequence the whole gene using next generation sequencing technologies.
The majority of people are bothered by it. Having a 1/3 of the ability to detoxify certain things means stresses on other systems. Very few are saying my GS is making me feel foggy and tired but it does.
There are a lot of studies that show lifestyle influencing Gilbert Syndrome sufferers. The most common suggestion is to avoid drinking.
Not sure it's as completely useless as you believe.
Associations form the foundation for future reportable SNPs. Many of their tests are now FDA approved. Many of the variants they report have clear and drastic impacts on health.
They are also able to report carrier status for 40 other diseases which I am too lazy to copy and paste here.
Lastly they have dozens of risk alleles reported for parkinsons, CVD, high blood pressure, high cholesterol and many other diseases.
Being positive for any one of these tests would be a constitute a significant, life-altering diagnosis. Know the carrier status could also impact and help avoid bad outcomes in your children. All of these are provided by the basic SNP array and have real and immediate value and which you can use to help mitigate the development of future disease.
The FDA approval category is for a ‘genetic risk test’. This is not the same as an FDA approved diagnostic test. All it means is 23 can reliably detect what they claim to test for, and there is some evidence linking a genotype to risk.
You describe 3 types of tests:
1. Highly penetrant but rare disorders with severe clinical phenotypes.
- it’s difficult to argue that this is important. There is a long list of such disorders that are extremely rare. If you think it is so drastically beneficial to test for some, why not test for all of them? What is so special about the disease SNPs 23 do test for? Additionally, 23’s test doesn’t actually cover all pathogenic mutations - BRCA is the perfect example of this. They don’t even test the most common BRCA SNPs in the general population. If you have a franeshift indel you are out of luck. So you are testing imperfectly for in some cases extremely rare conditions.
2. Highly penetrant but common disorders without severe phenotypes
- G6PD etc. Well this is maybe justifiable for G6PD. Certain drugs are contraindicated in someone with G6PD such as chloroquine or rasbiuricase, and doctors test for G6PD deficiency in that setting. Otherwise I guess you should avoid fava beans. Hardly life changing, especially as the link between a SNP and actual enzyme function is not direct.
3. Polygenic conditions where genetic factors play a small role
- it has been said many times but knowing your relative risk is increased for condition X has very little impact. Inevitably it leads to lifestyle modifications you should make anyway. I don’t need to know my CVD risk is higher because of some SNP, because I live in a western country and my risk is already high. Furthermore, it is totally unproven that some kind of intervention, lifestyle or otherwise could reduce your risk of a neurodegenerative condition if you have certain SNPs. Who knows, drinking green tea and not eating carbs or whatever you decide to try could be harmful.
Regarding pre-natal screening, there are dedicated tests in that setting, and if you are serious you would definitely choose those for their comprehensive coverage and counselling services.
So yes, by all means test your SNPs if you want to, but touting ‘priceless/drastic’ health benefits is in my estimation severely overselling the matter. 23andMe seem to agree, because they don’t advertise any of these amazing benefits on their website. It’s all about ancestry tracing.
Yes I used sequence loosely. The SNP array chips they use (last time I checked) have sufficient resolution to assign chromosomal segments to ancestral populations using the standard hidden markov model techniques.
