NPR’s muddled take on scientific racism and direct-to-consumer genetics

NPR’s science blog Cosmos & Culture has a post up about a new book on scientific racism and population genetics, particularly in connection with personal genetic ancestry reconstruction like that offered by 23andMe and other “direct-to-consumer” (DTC) genetic testing outfits. Now, these are important and complex issues, and we have some interest in them right here at TME, so it was only sensible to take a look. But the post, in which biological anthropologist Barbara King reviews the new book Is Science Racist by anthropologist Jonathan Marks, starts off with a premise that took me aback:

“If you espouse creationist ideas in science [Marks writes] you are branded as an ideologue… But if you espouse racist ideas in science, that’s not quite so bad. People might look at you a little askance, but as a racist you can coexist in science alongside them, which you couldn’t do if you were a creationist. Science is racist when it permits scientists who advance racist ideas to exist and to thrive institutionally.” This is a strong set of claims, and Marks uses numerous examples to support them. For example, a 2014 book by science writer Nicholas Wade used genes and race to explain, as Michael Balter put it in Science magazine, “why some people live in tribal societies and some in advanced civilizations, why African-Americans are allegedly more violent than whites, and why the Chinese may be good at business.”

Huh. So the first example of racist science is Nicholas Wade, who is … not actually a scientist. And whose 2014 book inspired a panel of 139 evolutionary geneticists (many of them who did research Wade cited in the book) to sign a letter to the New York Times saying that Wade was wrong That was after the book racked up blistering takedowns, pans, criticism, denunciation, elaborate multi-part debunkings, and condemnations from people who actually work with the kind of data Wade cited — including right here on this very blog. This is science “permitting scientists who advance racist ideas to exist and to thrive institutionally”?

(The other example the blog post gives of a racist scientist tolerated by the mainstream, psychologist Phillipe Rushton, I have literally only ever heard about in the context of condemnations of his theories on racial differences in intelligence.)

To be clear, there is unreconstructed racism in science, from structural biases in education and career progression all the way through to James Watson, and science has often been a tool for reinforcing society’s racism. Nevertheless, I can’t think of a clearer case of a scientific field reacting against racist misuse of its data than evolutionary geneticists telling Nicholas Wade “you know nothing of our work.” We don’t want a cookie for that — writing against Troublesome Inheritance was an ethical responsibility for scientists who study genetics. But it would be nice not to have Wade characterized as one of our respected colleagues.

The post goes on to make the (to most evolutionary biologists) uncontroversial point that human racial divisions are a social construction imposed on biological variation, and pivots from this to its central point, DTC genetic ancestry testing. Ancestry analyses have always been a major offering of consumer genetics services, and they’ve become more central since the FDA cracked down on the unsupervised use of DTC tests to estimate disease risk. Ancestry analyses assess the probability that a person’s genotype (potentially across many genetic markers, or at each marker individually) came from one or another of multiple pre-determined population clusters — if the comparison is made to a comprehensive enough sample, it can pinpoint the part of the world where your ancestors came from.

“How do they come up with numbers? [Marks writes in an e-mail to King] They take DNA from people from disparate regions and compare yours to theirs. The numbers reflect a measure of your DNA similarity to those of the divergent gene pools. How do they calculate it? Don’t know; the algorithms are protected intellectual property. Are they accurate? About as accurate as looking in the mirror.” … “Sociologists find that customers make sense of the results, and ignore the nonsense. For example, I’ve come out 95 percent Ashkenazi Jewish (not a geographical population, but a gene pool with its own minor genetic idiosyncrasies due to history) and 5 percent Korean. A good scientific question would be: +/- how much? 15 percent? 10 percent? Is my 5 percent Korean ancestry the same as 0 percent Korean ancestry? “Scientific answer: Yes. Corporate answer: Wouldn’t you like to know? “So there is sense, but it blends into nonsense, and may be difficult to distinguish them.”

This is the real crux. Ancestry assignment methods are publicly accessible in the scientific literature, but the detailed approach used by a company like 23andMe is not presented alongside an ancestry report, and would be challenging for a layman to understand if it were. Helping non-scientists understand the methods underlying the genetic test results and how to interpret them is an ongoing challenge that is in no small part responsible for the FDA scrutiny of DTC testing. And DTC testing companies do not necessarily have much incentive to help customers wade into those complications — or even to make sure their results are based on comprehensive reference data. In that context, it’s easy — maybe inevitable — for customers to misunderstand the uncertanties around the contents of an ancestry report, and in a society where all too many people are ready to believe that race is a real biological thing and racial disparities are caused by the laws of genetics rather than our social and political choices, those misunderstandings don’t help.

You might get the impression, from King’s post, that ancestry analyses based on genomic data are fundamentally flawed — but the fault here is not in the data or analyses. It’s in how their results are presented to the public. Many of the scientists who use those methods in research also work to help improve that presentation. See, again, all the responses to Wade’s book — or consider that in the last few years, every time I’ve talked with colleagues about teaching genetics, someone mentions understanding ancestry analysis as a learning goal. Human genetics courses now often incorporate DTC data collection into the curriculum, and in a classroom, personal ancestry analysis can be placed in proper context. Given the raw genetic data, which many DTC services provide, the analysis can even be performed more rigorously and transparently than a commericial service provider might do.

The history of racism in the U.S. and around the world makes it vital to provide careful context for every study of human genetics, from a nationwide study of variation to a mail-order do-it-yourself kit. Biology does have a mixed record on providing that context (or indeed making things worse — see, again, James Watson). But evolutionary geneticists are, in fact, wrestling with this history.


About Jeremy Yoder

Jeremy Yoder is a postdoctoral associate in the Department of Forest and Conservation Sciences at the University of British Columbia. He also blogs at Denim and Tweed, and tweets under the handle @jbyoder.
This entry was posted in politics, STRUCTURE and tagged , , . Bookmark the permalink.
  • Michael Stewart

    I’ve just seen notices of this book, the topic of which seems to be Marks’ hobby horse. 23andMe published an account of their matching algorithm and there are sites that will analyze their data using principle component analysis (e.g., Stanford University), etc.. Not all algorithms are secret, therefore, but the results do depend on the reference samples. There is such a history of fraudulent and obfuscatory arguments from pro and anti race ideologues that it would be better to drop the term. It is a fact, however, that there are ancestral groups that can easily identified by suitable statistical analyses of a small number of genetic variants and some of them have lived in distinct enviroments for an extended period of time — the pygmies and the bushmen, for example. The question whether there are genetically influenced differences in aptitudes and behavior between such groups will in time be answered by behavioral genetics researchers. I’ve noticed that race obfuscators find it convenient to ignore behavior genetics, but that will no longer be possible in a decade or two, I think.

  • Pingback: No, I don’t write for the Genetic Literacy Project (and I never will) |()