Cherry-picking DNA

It is a frequent source of perplexity for me when people read about scientific research, and because of their own biases (1) claim that it says something it clearly doesn't say, or else (2) deny it completely.  After all, as eminent astrophysicist Neil DeGrasse Tyson put it, "The good thing about science is that it's true whether or not you believe in it."

I ran into a particularly good (or appalling, depending on how you look at it) example of this yesterday over at Science Online, the news outlet for the American Association for the Advancement of Science, in article by Michael Price, entitled, "‘It’s a Toxic Place:’ How the Online World of White Nationalists Distorts Population Genetics."  Price interviewed Jedidiah Carlson, a graduate student in bioinformatics at the University of Michigan - Ann Arbor, about how the recent explosion in personal DNA analysis had been hijacked by white supremacists.

Carlson discovered the problem when he was looking online for a 2008 paper in Nature that analyzed hundreds of thousands of point mutations in people of various ethnic groups, and found that the paper had been linked in the notorious neo-Nazi site Stormfront.  Shocked but curious, he clicked the link, and found himself in a darker realm of genetic research -- using DNA evidence to support the bogus ideas that (1) races are little water-tight compartments except for cases of deliberate "race mixing," and (2) that people of western and northern European descent are superior to everyone else on the planet.

Norman Rockwell, The Golden Rule (1961)

Carlson started searching through Stormfront and other white supremacist sites, and found that this is an increasingly common phenomenon.  "People will grab figures from scientific papers and edit them in several different ways to make them look like they support the white nationalist ideology," Carlson said.  "For instance, in [the] 2008 Science paper, researchers published a figure with a plot inferring regional ancestry of dozens of different populations around the world.  Based on the genetic compositions of hundreds of individuals, the figure divided the populations into clusters that revealed patterns in their ancestral population structure.  So [people on the forums] take this plot and add some subtle text like 'The genetic reality of race,' with no context showing what the scientists were actually looking at, and ignoring the fact that there’s a continuum among the individuals.  Then they turn these images into memes and try to make them go viral."

They don't just cherry-pick data; they cherry-pick entire studies -- as long as there's some conceivable way to twist them around to support their ideology.  "They’re interested in anything that would reinforce traditional, discrete racial categories. Intelligence is probably the number one topic that they gravitate toward," Carlson said.  "And anything pertaining to history of human migrations, or things that play into traditional classifications of racial phenotypes like facial morphology or skin color.  There was a paper on lactose tolerance in Europeans and that turned into this weird viral YouTube trend where white nationalists were chugging bottles of milk, presumably to flaunt their European heritage."

I don't know about you, but that strikes me as a weird thing to be proud of.  "Look how well I digest milk" is not something you often hear people say.  I mean, my Louisiana heritage is probably why I love Cajun cuisine, wherein the Four Major Food Groups are onions, garlic, hot peppers, and grease,  but I have no desire to video myself eating a bowl of gumbo, and doing the Fists In The Air Of Victory afterwards.

Carlson himself has become something of a target, after his observations about the use of genetic research by white supremacists was the subject of an interview in The Atlantic.  It was a shock to him, however, to find how virulently they responded to his central claim, which was that the supremacists were warping the conclusions of the research to support their bigoted worldview, and ignoring any evidence to the contrary.  "When they finally saw it, the first few comments were actually rather celebratory, as they saw the article as evidence that the 'liberal, biased, Jew-controlled media' are nervous about the growth of white nationalism.  About me, there were comments like, 'He says he’s a grad student, but he’s probably never even seen a principal component analysis plot,' which is ironic because that’s about half of my dissertation.  And it was pretty alarming seeing my name on the site.  After that, I took a break from doing this work for a while for my own mental health."

His alarm is understandable.  These people are unstable, prone to violence, and usually well-armed.  It's not stopping Carlson, however, although he does acknowledge that fighting this kind of bias is an uphill battle at best.  "I don’t think engaging them directly will work," he says.  "In an argument between a logical person and illogical person, the logical person is always going to lose because the illogical person isn’t playing by the same rules.  The misappropriations and misinterpretations run so deep that you’ll just get shouted down and personally attacked, and you’re not going to change anyone’s mind.  But I think there’s growing recognition that we as scientists bear some responsibility for guiding the public interpretation of our work."

Of course, that's not easy.  You put the data out there, analyze it as rigorously as you can, state your conclusion as clearly as you can, and hope for the best.  The science deniers of the world will always find a way to get around it, either by claiming the data is faulty, the analysis is in error, the scientist(s) who did the research were paid shills and are trying to fool everyone for their own nefarious purposes, or (if none of these work) simply by ignoring the study entirely.  We've seen it over and over with climate change deniers and young-Earth creationists, both fundamentally anti-scientific views of the universe.  The same is true here; the white supremacists have their conclusion already figured out -- that they're better than everyone else based on their ancestry and skin color -- and the research needs either to fit that model, or it's rejected as "liberal, biased, [and] Jew-controlled."

The funny postscript to all of this is that when I did 23 & Me a few months ago, purely out of curiosity, I was honestly disappointed that my DNA didn't have any particular surprises.  My ancestry is primarily French, Scottish, Dutch, German, and English, and my DNA said that my ancestry is... French, Scottish, Dutch, German, and English.  I'd have been delighted if there'd been a random West African or Southeast Asian in there somewhere, as unlikely as that seems given my appearance.  Race is primarily a social, not a genetic, construct, as research by groundbreaking population geneticist Luigi Luca Cavalli-Sforza showed decades ago.  We're all mixtures, and if you go back far enough, we're all related.

So if you like to see races as neat little compartments with hard-and-fast boundaries, that's up to you.  But the bottom line is that you're wrong.  The view supported by science -- that the boundaries between ethnic groups are fluid, and almost all of us have diverse ancestry -- is true, as Tyson said, whether or not you believe it.

*******************************

This week's book recommendation is a brilliant overview of cognitive biases and logical fallacies, Rolf Dobelli's The Art of Thinking Clearly.  If you're interested in critical thinking, it's a must-read; and even folks well-versed in the ins and outs of skepticism will learn something from Dobelli's crystal-clear prose.