9 DNA Ancestry And Race

[MUSIC] >> Hi, I'm Jim, and I'm Eric. >> And I'm Joe, and this is Speaking of Race. It's been a minute since we recorded our last episode, and here we are. New year, new semester, but the last issues of science and race are only getting more potent. So last episode we tackled. Get it tackled. >> That's a pun. >> That's a pun. We tackled the myth that there's something biological that somehow makes black athletes superior to white ones. See now you get the pun. We talked about this because it's one of the first things that we as teachers of this topic get. When we tell people that race as we understand it is a cultural construct. And so when we advocate what we refer to as racial constructivism or the idea that race is constructed, we hear comments like, well, if race isn't biological, then why are black athletes so much better at whatever? >> So this episode of Speaking of Race, we're going to address another response that when we claim that race is not biological that we get from some people. That's a response that sounds something like, well, race must be genetic because I got a test from 23 and me. And it says that I'm 66% Romanian and 22% Italian and 6% British, but I'm 100% American. >> So if race is genetic, then how do companies like ancestry.com sell us these genetic testing kits? >> Well, they're pretty good at selling their excellent marketing companies. >> Recreational genomics is $100 million a year industry. >> What? >> Recreational genomics. >> Did we just call in a cool new term? >> I don't think you should be that excited. I don't think it's very widespread, but it's been out for a while. I think I saw a Washington Post article this summer that used that term recreational genomics. >> I like it. >> It's a huge industry. These people are good at selling things. Ancestry DNA says that they sold 1 and 1/2 million testing kits between Black Friday and Cyber Monday last year. Those are going at $70 to $100 a pop. So just in that four days, they really raked it in. >> That's a huge amount of money. >> Yeah, it is. Companies like this convince you to pay a fee, usually somewhere around $100 US to have some of your DNA analyzed to tell you about where your ancestors came from, or where your relatives lived today, or maybe what you're going to die from. And these tests are really popular and they're only starting to make it into mainstream advertising. We see them on the television all the time, nowadays. And so they will be increasing in popularity over the future and looking at what market analysts have to say about the people that make equipment for this industry. >> That's a good place to invest money because this is going to get bigger. They're going to make more money and they're going to be selling this to more people. They're making money hand over fist. >> Financial tips from Jim Benden. So anyway, today we're going to be looking at this question of why even if they are so very appealing, these tests are also so very inaccurate. And why that matters for the discussion we're having in this podcast about science and race in the 21st century. So I've seen these commercials, I swear they're like every five seconds on YouTube and Hulu. And I've been on the websites, these companies like 23andMe and Ancestry.com. It seems like they're doing some pretty cool things. One of the promises they make is that you have the potential to learn stuff that you just didn't know even about your own lineage. So one of the commercials that keeps coming up is the story of Kyle. >> The Kilt Guy. >> Yeah, yeah, the Kilt Guy. So his story is he was brought up to think he was German, he even went so far as he had some later hosting. The whole nine yards and then he comes to find out that actually he's Scottish. >> Thanks to 23andMe. >> And there are no Germans really anywhere in his family. That's what's hilarious, he's not German at all. And even though I guess Kilt's are pretty expensive, he goes all in and he gets a kill. So there's that story, there's other exciting stories. The post article I was referencing a second ago this past summer is all about. That's actually pretty interesting. It's about a woman who took an ancestry.com test. She believed that she and her siblings were Irish. Their father's last name was Collins and he swore he was Irish up and down. But then they learned that they weren't Irish at all. They had significant Ashkenazi genetic markers and those were from populations in Lithuania or Ukraine. In fact, they weren't even related to their cousins at all. So they talk about how tough that journey of identity is. But then the story ends with this line. I love this line. It is the truth after all. Oh, the truth. Well in anthropology we know that all kinship is cultural. So. Well, some of the stories are really exciting. And these services promise to help you uncover the real you and tell you who your real people are. And I admit, I've looked at their methodologies. It seems, I don't know, pretty sciency. They talk about microtomable DNA data. Based on family tree DNA dot com, they discuss the statistical model they use. They say that they're using a Markov chain Monte Carlo algorithm. So that sounds pretty official, right? I have no idea what that even means. It wouldn't make any difference if you did know what it means. There's much more to this story and I think we need to talk about how this industry of recreational genomics works because there's really two things going on here. There's the sales pitch, the kilts versus later hosens. And then there's the actual science. There is some biology being done behind the scenes, but what makes it saleable at all is that they tell stories. And what they're selling is the story. And the story actually has next to no science in it whatsoever. One of the things that we need to deal with here is the use of the term ancestry. When we think of ancestry, it seems like a really simple, late term. But even if you're thinking of something as broad as the continent of origin as an ancestry focus, this term is very dependent upon what time period is being considered. And our DNA tells different stories. If we go back 60,000 years, all three of us sitting in this room are of African ancestry. If we go back 20 to 30,000 years, our ancestry would be based in southwest Asia. And if we look at 3,000 years ago, our DNA would most likely be located somewhere in Western or Northwestern parts of Europe. Well, at least that's what we're told by the DNA story. There's also another thing about this term ancestry. We hear this a lot from genomic scientists today. And what's important is that in 2008, there was a big federal meeting held in the US where a lot of the genome scientists got together. And one of the things that they discussed was finding terms other than race to try and describe the patterns that they wanted to talk about. And they came up with the term geographic ancestry. And you find this very widespread throughout genomic work starting in about 2008 and right through to the current day. They want to use the term geographic ancestry because they're talking about race and ethnicity, but they're trying to do it using a term that won't get them in trouble with other media sources. It's like the new PC term for talking about race without having to say race. It is the genomic PC term for race. It absolutely is. Well, we could just stop the episode there. Now we know why it matters for race. Exactly. Done. Check, check. Okay. But one of the other things about ancestry is that we tend to think of it in terms of a family tree. We tend to think of it that it's very neat and, you know, the trunk divides into branches and branches, divide into smaller branches. We sit out on the tip of a branch and our ancestors are closer to the trunk and our farther back ancestors are even closer into the trunk of the tree. But the recent ancestry of our species has much more in common with the bowl of ramen noodles, as my friend Jonathan Marx has said, then it does a neatly divided tree trunk with discrete branches that get smaller and smaller. The more we learn about the history of our species, the messier the tangle of noodles grows. So is it safe to say that this recreational genomics and broader patterns of looking at DNA ancestry is comparative and perhaps even inexact? That's completely safe to say. The few companies that actually take the time to present any kind of detail about their techniques as to how they estimate ancestry from your DNA sequences point out in their position papers that the estimates of ethnic ancestry vary depending on the method used and the reference samples used for comparison. We should say right here that in the show notes to this episode, we're going to link to the papers that reveal the methodology for what we can find anyway for ancestry DNA, 23 and me and family tree DNA. You should definitely check these out. And then if you have comments or questions, please do post them on our Facebook page and we'll do our best to answer those. Yeah, so what does this difference in methodologies used mean? Well it means that one of the things that happens and this has been pointed out many times in media experiments is that you're going to get different estimates of what your ancestry is if you send your samples to different companies. In other words, they take the same sequence of DNA, put it through their black box and they come out with a different set of estimates about where your ancestors live at some point in time. The other thing is that you will also find the estimate of your ancestry changing even if you stick with the same company because their sample grows and changes and the people that you're being compared to changing. And so all of a sudden, you're less Lithuanian and more Romanian. Those kinds of shifts are going to happen as the sample increases at the company that you're staying with. Okay, so I think that this means we do need to get a little bit into the weeds and talk about the methodologies these companies use. Now, listeners, you don't have to get a degree in genomics to understand this part. Just hang in there. There's an industry group called the International Society of Genetic Genealogy and they surveyed five major recreational genomics companies. And what they found was that all of these companies use between 600 to 700,000 individual spots in the DNA. Each of ourselves contains around 3 billion pairs of nucleotides that go to build up our DNA molecules. But these companies are looking at roughly 2-100 of a percent of our DNA that they feel is informative. These spots that they're looking at are called single nucleotide polymorphisms or SNPs for short. Nucleotides are the building blocks of the DNA and they're looking at differences between different individuals. So, can we be certain that each company is actually examining the same set of SNPs? No, in fact, we know for a fact that they are examining different SNPs and each company has its own algorithm for what's informative and what they should find. Weird things happen when you compare multiple companies looking at very simple, even physical features. One firm told the geneticist Adam Rutherford that he had a gene for red hair while another firm told him he didn't. Both companies were looking at single nucleotide polymorphisms in his MC1R gene. That's the Melana-Cortin-1 receptor gene and this gene plays an important role in hair pigmentation but there are many different mutations in it that result in a red hair phenotype. So, what happened was these companies examined different ones of the red hair SNPs within this gene and so they came up with different findings because he had the mutation of red hair. He had the mutation at one spot and not at another. And there's also the issue of sampling, which is a much more basic question than what we're talking about here with the genomics. How do these companies get the samples that they use to compare your DNA to other people and say, "You must be Scottish but you're not German." And here is what a Scott-ish person looks like genetically. And so, not surprisingly, we've learned that most of these companies samples are made up of people who have paid to do the test in the past. Each of them, I believe, also has access to larger sort of starter samples. But what happens is, especially as these companies get older and older, they've all been around for what a decade now at this point, at least. Or a larger and larger proportion of their sample comes from people who've paid to do the test. And what that means is that most of these samples are very highly skewed towards wealthy Euro-Americans who have the resources and the money and the curiosity in the first place to take these tests. There's an article I really love that I'll make sure to link in the show notes by a woman named Uni Hong. She's Korean by descent. And she wrote this great article in 2016 about how she took a 23-and-me test and came up with very, very little Korean ancestry for her. And when she dug down and actually got ahold of some of the folks at 23-and-me, it turned out that their sample consisted of 76 Korean people. 76 people intended to represent an entire population. And in her case, what had happened was simply that the various SNPs she had in her DNA didn't match up with the 76 people that she was being compared to. And so it picked up no Korean ancestry whatsoever. There's a similar phenomenon going on in Great Britain today. In fact, most residents of Great Britain don't show as having any British heritage or very little British heritage if they send their samples to ancestry DNA or 23-and-me, these American companies. Because of course, our British genes come from the near-do wells that were the outcasts that ended up in the Americas and formed the basis of the British American population. And the genes are different. So that's a good point about the population samples that they use. I also would want to point out what I think is a historical problem. So computer algorithms will compare your DNA with the DNA of other people who are alive in the world now or have been recently sampled. We don't have DNA for famous people from one century ago, let alone obscure people from many centuries ago. So in the best cases, these companies are really telling you about how closely you resemble people living today, not really your ancestors at all. That's right. And again, going back to that Bola Noodles. In addition to that, the more we learn about the incredible amount of movement and inner breeding that has happened throughout human history and really accelerated over the last few thousand years, the more problematic the whole concept of geographic ancestry becomes because it relies on the genes that they sample today representing where populations were 500 years ago, a thousand years ago, 3000 years ago, 5000 years ago. And it doesn't work that way. We also know from mathematical modeling that the family trees that we construct rapidly begin to collapse upon one another. So that all folks that are European share a common ancestor only about 3500 years ago and beyond that, all of the pedigrees begin to collapse for everyone in the world today about 10,000 years ago. This is because each of us has to have two parents and each of our parents have to have two parents. And the thing is you don't have to go back very far before you have generated more ancestors than existed in the history of our species. You don't have to go back very many generations at all before you have dramatically outnumbered the number of people. And what that means is that when you go back 3500 years, there's a person alive in the European subcontinent that is in the ancestry of every European today. And when you go back 10,000 years, there's an individual alive who has left DNA segments in everyone on the planet today, 10,000 years. This is the part where I make a snide remark about how we're just one big happy human family. Yeah, it's a small world after all. Yeah. Oh, Jim Singh's podcast is complete. So anyway, not to pile on more criticism, but why not? Another thing that's problematic about these tests is that we don't actually know how each company that's doing recreational genomics goes about making the comparisons that they do. You know, they're for profit companies. They have proprietary algorithms and they keep them relatively secret on purpose. About a year ago on the show, the doctors, a set of triplets who became famous once they were featured in Playboy had a public reveal of their DNA ancestry. First, a series of their SNPs were tested to document that the women were genetically identical, or in other words, monozigotic triplets. What this means is that they all come from a single fertilized egg. So they all have exactly the same DNA sequence throughout their entire genomes. Next, their ancestry estimates were done by the same company that documented that they were monozigotic. And when their ancestry was revealed, some of the triplets, sisters varied by as much as 100% in their ancestry ethnicity. Meaning they had nothing in common whatsoever? No meaning that some had twice as much French in German as others. One of the women was 11% French in German and a sister was estimated to be 22% French in German. The estimates of their ethnicity, throughout, were varied. And this makes no sense because you're looking at exactly the same set of SNPs. So what this says is that it depends upon where you sit in the computer queue, having your DNA compared to other people, and that will make your outcome differ by as much as 100% from someone with exactly the same DNA as you. There's a reason that these tests are labeled as recreational. They don't pass the kind of scientific muster that we think they should, or that people think they do. So I think we've done a pretty good job of pointing out some of these ugly flaws and the methods of these companies. I want to bring up an additional reason why I do not like recreational genomics. So I think what it does is it monetizes and actually reinforces our fascination with this idea, which I think is a pretty pernicious one, that it's really just our genes that make us who we are. So the one, actually, it's really a point, a tenth of a percent difference in the genetic material between any given two people is somehow more important than our 99.9% sameness. And that has implications for how we think about all kinds of human difference, maybe especially race. So there's a couple of examples right now. In fact, just this month in Kansas, one of the state lawmakers, he had to apologize because he made this claim that marijuana was outlawed back in the 1930s, because African Americans were using that drug and were genetically less able to handle the potency of the drug. So ancestry tests make us think that these cultural categories have real biological meaning, which is exactly what this legislature is. was referring to. It turns out, it's a fun story, but it turns out the connection between marijuana and African populations was deliberately constructed by a guy named Harry Anselinger who was the first commissioner of the federal Bureau of Narcotics, which interestingly enough was housed in the Treasury Department. Anselinger made this connection between drugs and race and jazz music so that the government would ban it rather than what they wanted to do, which was legalize it and then tax it, which is why it was housed in the Treasury Department. So his personal racism became biologized as an attempt to justify using the authority of science. Yeah, and I totally agree with that. I think that it does a couple of other things as well. This idea of or this fascination that we have with our own ancestry. Because these tests are done largely by Euro-Americans, I think there's also this sort of reinforcement and fetishization of the other, so to speak. I've seen this happen a lot of times. We do an ancestry test in the race class that I teach here at UA primarily so that we can spend time doing exactly what we have done in this episode, which is picking apart why they don't work. That class is for the most part comprised of students of European descent, but not entirely. And I can't tell you how many times students have gotten their results back and been like, "Oh, I'm just British." Or, "I'm just Scottish." Oh, I was really hoping I was going to have like a secret Asian identity, you know, and so there's there is this sort of reinforcement of the exoticization of things that are not what most people taking these tests are, which is Euro-American. What counts as good exotic versus bad exotic? Has there been that discussion in that class? No. So, for instance, would East Asian be acceptable, but South Asian not be acceptable, or would Middle Eastern not be acceptable, but North African be pretty cool, or is like Native American the hot thing to be, but not South American? Oh, everyone was super up in arms this past semester about Native American ancestry. So, a lot of people said, "Oh, but there's this story in my family," or right, of the great ant, twice removed, who was Cherokee. Is everybody Cherokee? Yes, probably. No. If you go back 10,000 years. Yeah. And there are some people who are really upset by the fact that the ancestry they thought they had that was Native American did not show up at all on their tests, and we had to keep sort of reassuring them and saying, "Just because it doesn't show up on your 23 and me test doesn't mean that it doesn't exist." And that was actually a really great point of entry for talking about why these tests are total bunk. So, the bottom line is we have this 21st century science that is really being monetized, really being turned into a profit machine via these stories that they make up about people's ancestries. But when you look at it closely, they have the same problem that Samuel George Morton had in the middle of the 19th century. It didn't matter how precise his measurements were or the statistical techniques that he may or may not have used to show the differences between skull sizes. He was using a bad set of cultural categories to compare his data. And we've got exactly the same thing going on here. They may not call them race, they may call them ethnicities, they may call them geographic ancestries, but they're doing the same thing. They have very precise biological science coupled with highly complex statistical techniques that produces garbage because the categories they're comparing are garbage. So, you know, save your hundred bucks and buy a nice map of the world and throw darts at it if you want to get a better idea about where you come from or where your ancestors are. You'd get almost as good a set of results doing this. That's a really good point. If you want to know more about that Morton story, go back a few episodes. We actually talk about crenometry, polygenism and monogenism. That's where that fits in. That's right. I'm Jim. I'm the physical anthropologist. I'm Eric. I'm the historian of science. And I'm Joe, the cultural anthropologist. And this is Speaking of Race. Thanks for listening. Be sure to visit our Facebook page and give us your comments. [Music]