More on Race (Sorry)

Great Debates
1

I know a lot of people are tired of this already. But in the last few days I’ve seen a couple of studies that have a bearing on the issue - but they seem to directly contradict each other, and I wonder if our more learned posters might comment.

Race not reflected in genes, study says

Gene Study Identifies 5 Main Human Populations

From the perspective of this layman, the first study here seems to support the “races don’t exist” position, the second seems to support the “races do exist” position. The only thing I can think of is that perhaps Brazil might be more of a melting pot than other areas, so that races have intermingled to the point that they no longer exist there.

2

I’d like to know just how much margin of error the second study allows. How many people are in a grey zone or can’t be classified.

Are some people in 2 groups?
For example…

I’m pretty much a white American mutt. My ancestry comes mainly from “white” Europeans. My wife is “pure” Filipino.
So where would the second study classify our kids? Just to muddle the issue more, Filipinos had the Spanish on their island for quite some time. How do those markers get accounted for?

My bet is that the second study is overly broad and isn’t all that usefull in any practical way.
Just MHO of course…

3

Well, in the Brazilian study, they found that people who self-identify as white had, on average, 28% African genes and that people who self-identify as black had 52% African genes. So there’s some correlation. When that article says that

I wonder if that’s a quote from the scientists who performed the study, or a gross oversimplification by the reporter.

If you define “race” as the geographic origin of our ancestors, then apparently genetics can tell us that, with fairly high accuracy.

Using physical characteristics instead of genetic markers is less accurate. And it makes sense to me that, in populations where many people have mixed ancestry, physical traits would be even less useful for determining geographic ancestry.

You can define race just about any way you want to.

4

I just read the Science article and the News and Views. The New York Times really oversells the article.

The new Feldman article in Science samples polymorphism frequencies for different people in different populations. These populations were chosen because they are relatively isolated and relatively well defined on a geographic area. The people were chosen because they had been rooted on the land, usually up to 4 generations. So the people were the best representatives of a population sampled. They first find that all alleles of the polymorphisms are present in all populations. They next find that only 3-5% of genetic variation in humans is found between populations. 95% of genetic variation is found between individuals of the same population. So they use some serious genotyping and serious computer clustering to look at this 3-5% and are able to tell members of these clear populations apart. They find many genetically separable subgroups in some continents (4 in Africa), and no separable subgroups on others (Europe). Lastly, they find that there are no sharp cutoffs between geographic areas, rather large gradual transitions.

They suggest that this demonstrates little medical necessity of treatment or diagnosis by race, although they concede that self reported ancestry usually correlates with genetic ancestry unless the population is recently admixed. They suggest that this 3-5% which is different in the most extreme, most deliniated populations, is not suggestive of much genetic contribution, except in the case of evolutionary advantageous, ancient alleles such as malarial resistance genes.

If you would like, Izzy, email me and I will send you the Science News and Views, which gives a review of the article. It is very readable for anyone without a science background. It leaves one with a completely different opinion than the New York Times review.

The other study tried to correlate any type of genetics with skin color in a mixed population, namely Brazil. Here we run into the polar opposite of the first paper – take a mixed population and see how much of the genome is correlated with self-reported race. They look at 10 polymorphisms which can separate Portugese people in Portugal from African people in Africa. Lo and behold, none of these is correlated with skin color – the genomes thoroughly mix and everything is jumbled – once the two populations interact. There becomes no way to tell self-reported race genetically.

So both studies agree to a certain extent. Most people in the world are from relatively mixed populations. Among those that aren’t, upwards of 95% of the genetic variability occurs between members of a single population. In most cases, race is a social construct with no bearing in genetics. A person’s ancestry may be helpful, and it doesn’t hurt to ask as it may give you some clinical benefit of diagnosis (HbS and Duffy blood group for instance), but it probably won’t account for much. “Pure” populations can be separated genetically, but besides them, there is an enormous gray area where most of humanity falls.

5

edwino, your point in your post seems to be that in populations with mixed ancestry (or populations at geographic crossroads) it is no longer possible to determine ancestry, as it is mixed in many cases. This is not surprising - I alluded to the possibility in the final sentence of the OP. But what remains true is that the study does seem to be claiming is that there is some degree of consistency that maps to the “classical races” at least in situations where these races have not subsequently become intermingled. Posters such as Collounsbury have argued with great vehemence that a random person from Africa is no more likely to genetically resemble another random African than to resemble a random European. The study - both as summarized in the NYT and in the review that you kindly sent me - seems to contradict this.

(The possible “out” is if there are completely distinct sub-populations in the different geographic regions, e.g. the Mbuti Pygmies, San peoples et al) who have no connection to each other at all but are merely known to have lived in the same region. (I’ve brought up this issue previously here). But this does not seem to be indicated. And besides, there are not that many sub-populations, at least in Africa).

The review does suggest, as do you, that use of race in medicine may be limited. But reasons given are due to the possibility of non-genetic origins of population discrepancies of diseases, rather than the complete non-existence of race altogether.

One point to remember is that even in populations of mixed ancestry it is possible that some subgroups will have a larger percentage of ancestry from one geographic region than another (as noted above by heresiarch). So that although it will not be possible in such cases to determine ancestry of an individual person (the subject of these studies), the groups as a whole might differ on average, on the basis of their predominant region of ancestry.

6

The Science paper used a very set description of population. We (tomndebb, Collounsbury, Tars and others) have never disputed that for set, defined populations, specific alleles and distributions can be described. We just dispute that something like an overarching genetic similarity exists for some as simplistic as the “black race” which includes not only Africans but African Americans, black Brazilians, Caribbean people, and certain other populations with moderately dark skin.

I think if you were to look, any people not included in the exact “population subgroups” in the Science paper would be genetically indistinct. I think that the power to assign ethnicity by genetics is incredibly limited to the ability to assign one of perhaps 10 ethnicities (which represent relatively isolated populations) by genetics. I think that is the gist of the Science paper and I don’t know if more can be read into this. If taken like this, it agrees exactly with the PNAS Brazil paper.

7

I suspect in Brazil, as in the rest of Latin America except for Argentina and Uruguay, “races” are defined differently than in the Anglo-Saxon world. People who are brown skinned light mestizos, with have more or less European features (like wavy hair, narrow nose and lips) are often considered white. In America the same people might not be considered “white” by many people.

But even among people considered “white” by nearly everyone, there is evidence of widespread racial mixture, especially among white American southerners, Afrikaners, Russians, and other people who have had long standing contact with people of other geographic populations. And often, it is not easy to tell just by looking. A lot of people who are documented to be a fourth or an eighth “black” look completely white.

I have read in several books, papers, and journals references to a study in the 1950’s by an American anthropologist named Robert Stuckert . He somehow concluded that 21% of American whites had “black” ancestry withing the previous 4 generations. Presumably many of these partially African descended people were entirely “white” in appearance. I have never read the original, nor have I seen this study online (only further quotations which repeat these numbers). I wonder if there is anyone who ever corroborated or refuted this?

8

Probably best was a past Science article, GENETICS: Enhanced: Toward a New Vocabulary of Human Genetic Variation, Science 2002 298: 1337-1338. Should be accesible at http://www.sciencemag.org/cgi/content/full/298/5597/1337?maxtoshow=&HITS=10&hits=10&RESULTFORMAT=&fulltext=race&searchid=1040418285830_7043&stored_search=&FIRSTINDEX=0&fdate=8/1/2002&tdate=12/31/2002
edwino, you’d enjoy it.

Its point is that we mix up the definitions. Instead of closely defining what we mean by race in a particular context. As these snips illustrate:

And they link to this site, which has a good pro and con.
http://www.pbs.org/wgbh/nova/first/race.html

How I see it play out poorly, for example, is in reports of US Vital Statistics, where Black is all lumped together independent of SES and Hispanic is reported as a race, even thought there are Black and White Hispanics. And then people try to draw conclusions from this data. It becomes absurd.

9

Hispanic as a race? Wow.
I’m Puerto Rican. I’m dark. My brother & sister are white, my brother to the point that he has a hard time getting a tan without burning. My mom is dark, but not as dark as me, while my dad is so light that he too has a hard time in the sun if he hangs out up North for a while.
And just to really mix this up, my complexion is supposed to come from my dad’s side of the family. Don’t ask me if that’s true or not. Somehow I don’t think there’d be a scientific way to tell.

10

I’m trapped in Illinois for three weeks and do not have access to Science (unless a newstand sells it here, but i doubt it, i’ve never seen it for sale at one), so i cannot read the original article, but i will take edwino at his word.

11

But what you have disputed - at least some of you - is that these populations could be large ones. ISTR the assertion being made that there would be “hundreds” of populations if one were to divide the world down to the point at which we’d see genetic coherence. What I seem to be seeing from this article is that there would be significantly fewer.

To clarify your position here, do you agree - or at least agree that the Science article is claiming - that ignoring mixed populations such as African Americans, the world can be divided into five populations (based on geography) which will show some degree of internal genetic consistency?

12

No, it doesn’t say that at all. No claim of internal genetic consistency. No claim of “genetic coherence”. This article isn’t about “race” at all, it starts off with the unspoken acceptance that “race” implies a possibly (well probably) false answer to the questions being asked: how are human genes heterogenously distributed among different populations and what does that tell us about our history and what clinical significance might it have?

The study verfies that each population group is more varaible within itself than groups are between each other. This study utilizes a statistical tool that shows the echos of human migration patterns buried deep and subtley within the broad diversity of each group. A few snips from the Perspective article will hopefully clarify:

No sharp edged racial groupings…

The five clusters is arbitrarily fine,

While the piece does claim that clinical findings that appear to correlate with population clusters may actually be due to “social, economic, or discriminatory factors” it does not shy away from saying that these findings argue against using race as a proxy for clinically significant factors because

The two studies are consistent with each other. Entirely. I don’t know if I buy the total lack of utility of race as a clinical proxy for genetic susceptibility, given a lack of specific genetic markers identified, but the article argues strongly for just that point.

13

I think the paper claims that they have a way to put an individual into one of 5 geographic locations by genetics. Whether I agree with their findings is another deal. I think the column about the paper deals with the data in about the same way as me – with a grain of salt. I have looked hard at the data, and I must admit I cannot really understand what they did. I read the background data about their computer program and it did not help much.

They measured a large number of polymorphisms in a large number of people from the aforementioned well-rooted, relatively isolated populations. They look for the 2-5% of polymorphisms which differ between populations significantly. They calculate population allele frequencies at a bunch of these (around 400). Then, if you take an individual, measure their alleles, you can put them into a group based on probability matching to the composite allelic frequencies. How many groups you split the world into is the so-called K number.

They address the K number when K=2 to K=6. When K=2, most people fall closer to America versus closer to Africa. When K=5,. we get the result that things fall into a geographic type breakdown (East Asia, Eurasia, America, South Pacific, Africa). When K=6, the Kalash group of Central Europe splits off. I don’t think this paper would have been a Science paper if K=5 broke down Africa into North Africa/Eurasia versus the San, but when it falls along geographic boundaries then people pay attention.

So we have to address the data. I have three points. If their program works as they advertise, it is breaking down the allele frequencies into different K groups just by similarity, and blinded to region. So what they are looking at is admixture, pure and simple. They just have developed a really sensitive assay. If you have a sensitive measure of admixture, you will always break things down in a geographic fashion because obviously near populations mix the most. Which brings me into the second point.

The second point is an issue of noise. I don’t think I can do this justice, but I believe that whatever number you stick into your K value, what you find is of questionable physiologic relevance. I say this because I believe there is an acquisition bias in the system. Let’s simplify the assay. Let’s say you sample two populations in Africa and two elsewhere. You sample two related populations in Africa, and an Inuit and a Cherokee population in America. Well, if you have K=2, then obviously your line will be between Africa and America – the genetic similarity will bear it out. Expanding the study, I think the results are entirely dependent on which cultures you decide to sample. I think if you take a gradient of cultures and then one oddball culture, obviously your line will come between the two.

The last point is of physiologic relevance. We know that humanity was in Africa for around 5 million years. We know that the earilest humans may have started peeking out of Africa around 1 million years ago but perhaps as late as 100,000 years ago. So humanity has spent the longest in Africa, and many studies reveal that the genetic diversity (by rate of heterozygosis at loci, by haplotype size, etc.) is greater in Africa. When one finding comes along which uses a new strategy which is IMHO imperfect to purport that the Africa/Europe divide is bigger than the Africa/Africa one, I tend to take it with a grain of salt. Things should reflect physiologic relevance, or at least address how they fly in the face of what has been found before. This paper does neither.

I have a hard time basing a lot of proclamations on this study. First of all, I don’t understand the visceral nature of the computer algorithms used, nor do I really have the background to find out. Second, I believe that the central finding of this paper, that genetics can be used ot break people down into 5 geographic areas, is prone to acquisition bias which is cannot be controlled for. Nor IMHO is there a better way to get the data. Third, the paper contradicts many other findings in the field and does not address it. Lastly, go back to the 95% number. If no alleles are private except for selected ones, what relevance does breaking the world up into 5 groups based on 2-5% of their genomes? I think very little if any.

I will definitely use and cite this paper – there are many interesting things here and the data are bound to arise again. I just will not be changing my worldview until something really convincing comes along.

14

Dangit, edwino, now i want to read the paper even more!!! Why didn’t i send in the subscription card yet?? Gah!!!
First thing i am doing Jan 7th at work is grabbing that issue…

15

Tars, you know we do have these things called libraries here in Illinois …

16

**The Brazilian researchers looked at one of the most racially mixed populations in the world for their study, which found there is no way to look at someone’s genes and determine his or her race. Brazilians include people of European, African and Indian, or Amerindian, descent. **

There you have it, they admit the population tested was already mixed.

**“There is wide agreement among anthropologists and human geneticists that, from a biological standpoint, human races do not exist,” Sergio Pena and colleagues at the Universidade Federal de Minas Gerais in Brazil and the University of Porto in Portugal wrote in their report, published this week in the Proceedings of the National Academy of Sciences.

“Yet races do exist as social constructs,” they said. **

Then please explain the mountains of genetic evidence here: http://www26.brinkster.com/archived/
**They found 10 gene variations that could reliably tell apart, genetically, 20 men from northern Portugal and 20 men from Sao Tome island on the west coast of Africa. **

Yes, i could have told you that. Portuguese are hybridized with blacks, as are brazillians.

**But the genetic differences did not have anything to do with physical characteristics such as skin or hair color, the researchers found.

They next tested two groups – 173 Brazilians classified as white, black, or intermediate based on arm skin color, hair color, and nose and lip shape, and 200 men living in major metropolitan areas who classified themselves as white. **

There you have it, they admit that the ‘‘races’’ tested were simply mixed brazilians of differing skin tones, unlike comparing a pure white and a black.

**They used the 10 genetic markers that differed between people from Portugal and Africa, but found little difference among anyone in their study. **

Yes, like i have already stated, they tested mixed populations.

**To their surprise, they found maternal DNA suggested that even the “white” people had, on average, 33 percent of genes that were of Amerindian ancestry and 28 percent African.

This suggested European men often fathered children with black and Indian women.

“It is interesting to note that the group of individuals classified as blacks had a very high proportion of non-African ancestry (48 percent),” they wrote.

“In essence our data indicate that, in Brazil as a whole, color is a weak predictor of African ancestry,” they concluded. **

Yes, in BRAZIL colour is a poor indicator of ancestery.

**“Our study makes clear the hazards of equating color or race with geographical ancestry and using interchangeably terms such as white, Caucasian and European on one hand, and black, Negro or African on the other, as is often done in scientific and medical literature.”
**

What politically motivated hogwash. Check this out: http://www25.brinkster.com/humanraces/calc/

The difference between n.euros and the san of africa is 234 genes on average, and it is a single gene which determines the sex of a foetus.

17

And what exact politics are motivating whatever it is that you are claiming is hogwash?

I’m quite open to the possibility that there is some sort of vague association that we can identify with the word “race,” but I have not seen any persuasive evidence for it at this time. Are you suggesting that “white, Caucasian and European” can be equated? And that “black, Negro or African” can be equated? So where do you place the Asians of the Middle East (traditionally Caucasian (and even white if one ignores the claims of some political groups), but certainly not European). How about the peoples ranging across the North of Africa from Egypt, through Libya, Tunisia, Algeria, and up to the Atlas mountains? They are clearly African, but they are hardly Negro–many, in fact, would be classified as Caucasian under the older norms.

Identifying differences, if real, between the San and the Swedes is only relevant if we can find a border that separates the groups. If the differences shade slowly across a difference of genetic spectrum from one point in the world to another, we still find ourselves with the problem of drawing arbitrary lines to indicate race.

18

I am sorry, but I read and reread that article and can only come away with a resounding “And … the point is?”

Look at enough neutral markers and people who have been rooted in a similar area with shared ancestry for generations will cluster together. So big friggin’ deal. That is a surprise? They also may have some superficial characteristics that look alike. And …? No duh. The more completely and longer a particular population has been seperated from other populations the more various neutral bits will be subject to drift. What is so earth-shattering about that? Different Pygmie groups have been seperated from each other for many generations so they cluster differently from each other on neutral alleles subject to drift … and that’s supposed to qualifiy as “race”?

This is supposed to imply something about alleles that are not neutral? Why? This group of handpicked populations chosen with foreknowledge of their rootedness for generations, chosen because it was known that their reported heritage was fairly indisputable, is supposed to imply something about the reliabilty of everyone else’s self-reported group mebership? Sorry, but that just doesn’t seem to follow to me.

BTW, This Kingship, I visited your link, but fail to find mountains of evidence supporting the biological concept of race there. I find the articles under discussion highlighted. Perhaps you’d care to point out which items there you think makes the case. Maybe you’d even like to start by saying what you think would qualify as “race” biologically, and list what you think those races are? 'Cause you see, I do think that there are population groups that differ in the frequency of various alleles, some of which may have clinical significance, I just don’t think that “race” doesn’t, per se, correlate with how this data segragates out.

19

“And what exact politics are motivating[…]”
That there exists a political motive to minimize or eliminate race as a real category within humans is, to me, blindingly obvious, I might even find the motivation commendable, however in the end, any such motivation (I’m sure we all agree) much not lead to the suppression of free inquiry or debate, or indeed to excessive manipulation or falsification – as such tactics will invariably backfire, and in the end we’ll all lose. But this is sadly what I often have observed, as soon as we turn to the question of race.

I certainly find it very dubious that the term race can be applied in any meaningful way to differentiate humans, however I find the nearly ritual ridicule shoved upon any dissenting voice quite disturbing, since I am of the opinion that there are issues that have not been adequately addressed – issues that, on the surface, appear intimately linked to precisely races. While a former believer, I have lately grown increasingly vary of all the evidence that purport to prove race is non-existing or irrelevant, since I often find the arguments so blatantly political in nature and the science shallow and quite unable to hold up the grand conclusions, this together with a widespread knee jerk vilification of everybody with a different viewpoint has now lead me to the belief that this is still very much an open question that remains to be answered; for this many will label me a racist.

“[…]drawing arbitrary lines to indicate race”
There is no clear border between red and green, if you are claiming that red and green are irrelevant categories please reconsider that before you come to the next stoplight. Likewise you could set up an evolution tree, and claim there is no clear border between a human and a tadpole since there are many races in between that have more or less in common with one or the other. Few things in this Universe are black or white (heh!); to insist there can be no categories because any boundaries must be arbitrary may be logically correct but hardly an interesting point to make. Identifying differences are relevant, if for no other reason, then at least because we’re human (black or white) and examining our own navel is our very favourite past time.

Ps. I just accidentally googled past, and don’t really know what thing you have going here, so excuse me if I have hijacked the thread or repeated something stated a zillion times. Also I’m not English the so grammar and splelling may be somewhat experimental.

20

I’m sorry. I do not find it blindingly obvious. In fact, I find the opposite to be true. The people who speak most forcefully for recognition of race (either culturally in the context of Affirmative Action in the U.S., or biologically among advocates of “racial pride”) tend to be the most political and least science-based groups. The Human Genome Mapping Project was seriuously underfunded because people feared that the evidence might point away from racial identities.

As to the color spectrum, I can, indeed, find bands of lightwaves of specific frequency that will be clearly identified as being in the red “family” and not in the “yellow” family, and so quite distinct from the green “family” of colors. I know of no place where I can draw such a line among humans with their mad interbreeding.

The issue has been contentious on this MB, (with as much of it relating to philosophy and semantics as biology), but all the evidence that I have seen from science in the last 15 years has argued to me against race as an objective reality.