23andMe did not, to my knowledge. FTDNA charges about $700 to fully sequence just the Y, so it seems like a full sequence would be cost-prohibitive.
Not sure I follow the logic there. There are certain markers where I (or the service) knows what they mean, but without testing, I wouldn’t know what my value was. The problem is that the health markers weren’t terribly useful for the most part. “You have 25% chance of having curly hair”. Either I’m in the other 75%, or you sold me some snake oil. Either way I’m not recommending this to others.
Looking at known markers is useful to the individual customer (to the extent that the known markers are for actually important things like breast cancer risk, rather than just trivialities like curly hair), but it does almost nothing for research. Like, OK, we know at least some of the genetic risk factors for breast cancer, and so we can screen for those specific markers, but if we don’t know the markers for some other disease, how will we ever find them?
OK, I knew that sequencing had gotten cheaper in recent decades, but it looks like it hasn’t gotten nearly enough cheaper. At least, not enough for it to be done by a for-profit business. That’s still cheap enough that you could get a heck of a database by throwing a few billion dollars at the problem, and that few billion dollars would probably give you a return well worth the price, in improved medical science, but the benefits would be enough removed from the people paying that nobody but a government would really be able to do it.
Full sequencing isn’t necessary for that kind of analysis. What you’ve described is a GWAS (genome wide association study), and can be done easily with the type of SNP chips that 23andMe and others use. This is pretty inexpensive to do, and is really not that interesting when done on a basic list of health issues. It’s almost certainly been done before, and probably in the UK Biobank with medical records rather than a self report questionnaire.
Deeper genetic information is only going to be interesting when trying to chase down the exact variation responsible for some trait.
Over the last 50 years genetics testing has become many orders of magnitude cheaper. 25 years ago that was the hard part. What has not become cheaper is collecting in-depth phenotypic data on people. Using data collected by anything from a trained interviewer to a medical doctor is probably going to cost more now than it did 25 years ago.
This is where things like the UK Biobank, that use existing medical records, are taking a useful shortcut. There was a path for 23andMe to collaborate with existing hospital networks, but I don’t know if that would be palatable to their customers. Plus, many of those hospital networks that are associated with universities have their own biobanks.
On the genetic side, one of the most useful things for the advancement of science would be to increase the genetic data collected from people of different genetic ancestries. Yet one more genotype from someone of white European descent isn’t of too much use. One more genotype from just about anyone of non-European descent is adding much more information.
That did occur to me, since self-reporting is likely to miss a lot of important conditions. One possibility I thought of was data on the cause of death for individuals in the dataset (or, for those still alive, their current age), because one category of genetic conditions we care very much about is those that cause death.
My family member was able to learn they had some abnormalities that could be watched and compensated for. Knowing this, they were able to learn when research was conducted that associated some of these abnormalities. Knowing this means they and others probably won’t die of the then-unknown thing that killed their parent. None of their doctors so far were aware of these genetic abnormalities, nor that they’ve been shown to constellate. Were it not for doing commercial genetic testing, my relative would likely still be unaware and having their body slowly destroyed.
As a person of Asian descent, that was part of the reason I wanted to submit data to 23andme. I spit into their vials several times over the lifespan of their service, and as their datasets and methodologies improved over the years, my reports also got more and more precise (though maybe not more accurate?).
My initial ancestry report when they first launched just said something like “99% East Asian”, but a few years after that, they were able to break it down into broad regions, and then a few years later, into a dozen or so specific countries (and parts of bigger ones, like China).
I didn’t keep as close tabs on the health side of things, although I know they had both opt-in surveys and a premium membership tier that had access to more health reports. I signed up for a month or two, but there was nothing particularly noteworthy, so I stopped.
I wish there was some international NGO nonprofit that could work on this stuff for the benefit of all, with open data. Sad that so much of it is squirreled away in proprietary databases.
So far it has been universities and national health and science research groups like the NIH in the US and MRC in the UK that have funded and done most of the research. Other non-governmental charitable groups like the Wellcome Trust have also been important supporters.
What needs to happen is support and education in lower and lower-middle income countries. Train and fund local scientists there to do their own research. Then their research and subject pools are combined with others worldwide to get a comprehensive look at human genetic diversity.
It would make sense to have some NGO coordinate lots of this, but it is also vitally important to make sure that locals are equal partners, and the leaders in their own regions.
Heres an opinion from a geneticist on the demise of 23andMe. He was not a fan.
The genius of its business model was not simply to get you to volunteer this personal data to a private company, but to persuade you to actually pay to give it to them. It then commercialised your DNA by selling it on to pharmaceutical companies, which would use it to develop drugs, ultimately for profit. It was the type of racket that a mob boss might look on and say: “And this is legal?”
And he wasn’t impressed by the bit telling you where your ancestors were from.
There is no method for identifying the geographical origin of your ancestors using genetics. Your ancestors about 50 generations back are from all over the world, and besides, biology does not bestow membership to a tribe or clan or people or country. What 23andMe was actually doing was comparing your DNA to that of other paying customers, and matching up where they live today, and inferring that you have ancestors in that location. It kind of works, but is mostly meaningless. When the data lets you believe that you are 37% German, or 18% Spanish, or whatever, it might feel fun, but of course there is no way of being 37% German.
This feels like the genetic version of the “there is no such thing as race” argument. Not trying to start a political flame war here, but really trying to understand their methodology here (and critiques of it).
My (extremely layman’s) understanding of this situation is that it depends on tracing mitochondrial DNA (among other data) to try to trace human descent backward across geographical areas in the past, not just based on where my genetic relatives currently live, but also based on growing archaeological evidence (“ancient DNA”, e.g. 1, 2, 3, 4 and others).
I don’t at all dispute that country borders, race categorizations, ethnicities, etc. are arbitrary groupings we apply. But to some extent, this feels like a cop-out version of the “species problem” applied to the question of human lineage. Yes, 23andme may be labeling descent based on modern groupings, and maybe that’s a contrivance that betrays the underlying science… but is the underlying methodology sound?
It seems to me that combining modern genetic surveys with fossil evidence should be able to provide some picture of how homo sapiens and our ancestors migrated across the world — and yes, back and forth, in different waves, with a lot of intermixing, etc. It may not be perfectly clean data, but nothing in archaeology and paleontology ever is, right…?
Is there something uniquely wrong about the methodology of tracing human lineage vs similar approaches in the animal/plant/fungus/etc worlds? It seems like cladistics (if that is even the right concept) is difficult work in any genus, and especially hard because humans are so good at moving and breeding and it’s not always easy to respectfully analyze human remains, etc… but does that mean the underlying science is somehow invalid?
The TLDR for me in all this is “23andme/Ancestry.com/etc. may not be perfectly accurate, but it is the best data science can provide to an everyday customer about their geographic lineage, at least at this price point”. Is that a fair assessment? I’m happy to be corrected here…
Edit: I guess I’m thinking of the underlying science as an animated heatmap of haplogroups found across the world, e.g. this gene cluster shows up more intensely in this part of the world, but drag the timeline and you can see it start to branch out across this continent, while this other haplogroup has remained relatively stable in time in this other area, blah blah.
While our labels for the areas might change over time, at the end of the day, we should still have a rough idea of how the frequencies of these genes change across longitude and latitude, regardless of what that area/people/culture was called at a particular point in history (thereby separating the geographical ancestral data from the cultural labels). Of course that won’t answer questions of nationality, cultural heritage, tribal membership, etc., but I think it’s still interesting (and useful) data…? If my understanding of it is correct.
Some billionaire should start a Kickstarter to buy it all and release it, anonymized, into the public domain… would be a great dataset for all researchers… and ne’er-do-wells. $200 mil seems like chump change for such a large dataset that took a couple decades to compile.
I’ll never understand why they bothered to test your DNA at all. Just write down some guesses based on your last name and your photos on social media and throw away the gross sample. People would just assume they knew what they were doing.
I think some of the services to evaluate the DNA of pet dogs have been accused of something like that - especially the ones that require a photo of the animal being tested.
It should be a lot easier for dogs, since dog breeds (at least, the genuine ones) are pretty genetically homogeneous.
People would figure out if they were just going by last names, since plenty of people have different ancestry on their mothers’ and fathers’ sides, and most folks know at least some of their ancestry already.
The Vikings weren’t just smash and grab reavers. They also settled in areas they captured. Dublin, for example, was likely founded by Vikings, while York came under Viking control as part of the Danelaw. So Irish and Yorkshire folk could easily be descended from Viking women.
What a shame. Through 23 & Me I was able to find both my biological parents and siblings. I’m very close with them. I didn’t pay the extra for the health info and I’m not that bothered about percentages of nationality and whatnot; I found out why I look the way I do and that is enough for me.
It’s not as volatile, since it’s owned by Blackstone, a big investment group that also owns Motel 6, Hilton, Legoland, SeaWorld, other theme parks, Bumble, etc.
But they were subject to their own privacy suit:
Your genetic data, once given to a private company, is going to be bought and sold like anything else. If you’re worried about privacy, don’t do it.
