Brazilian and Portuguese comparative genetic data

[Silveira]

An interesting article published on the "Legion Europa" site:

www.legioneuropa.org/Racediv/brazil1.htm


Scientists Prove Race Does Not Exist?

By: MX Rienzi

The title of this essay - without the question mark at the end - was observed on the internet frequently in mid-December, in response to a genetic study by Brazilian researchers (Parra et al., "Color and genomic ancestry in Brazilians", PNAS 100: 177-182, 2003), summarized here:

www.laurushealth.com/HealthNews/reuters/NewsStory121620029.htm

which states that they saw, in Brazil, no clear correlation between "racial" physical traits and gene markers which can identify ancestral origins. Therefore, the race-deniers crow, this "proves" there is no such thing as (biological) race.

However, let us first quote Dr. Sergio D. J. Pena, one author of this study (see laurus health link above):

"First of all, let me stress that the conclusions of this study apply to Brazil, and should not be naively extrapolated to other countries."

Regardless of that admission, the authors in their discussion, and commentaries on the internet, use this study to attack the idea of relating "geographical ancestry" with traditional racial categories; in other words, attacking the biological race concept in general.

Now, someone who has read our essay "Race is a Myth?" as well as "Scientific Malpractice" should be quite skeptical of these race-denying claims, and, as we shall see, such skepticism would be richly justified.

The first clue of a problem with this paper is the fact that the authors, in the most shameless, subjective, and unscientific manner, openly display their political views on the subject of race in the introduction of their paper. They start with the following quote from S.L. Carter ("The Emperor of Ocean Park", Knopf, New York, 2002):

"You see leaders today, all over the world, doing it again!
Black, white, yellow, brown, people of every color
slaughtering people of every color! Because Satan is
always the same."

Very well. What exactly does that have to do with genetic science? That an ostensibly fair-minded scientific (not political activist?) journal like PNAS allowed such sentimental subjectivity in a published paper is absolutely stunning. What's next? Scientists submitting florid poems about their data? Love sonnets to their tissue culture cells? Off-color jokes about tumor samples? Is this science?

The authors then make the following comment:

"There is wide agreement among anthropologists and human geneticists that, from a biological standpoint, human races do not exist."

Really? What is "wide agreement?" Does every scientist agree with this statement? More to the point, how much of this "wide agreement " is motivated by the same unscientific and subjective political ideology that obviously informed the writing of this PNAS paper? How does this "wide agreement" coincide with the increasing realization that racial and ethnic categories are valid given that members of different "population groups" vary in their prevalence of certain diseases, response to medical treatments, and ability to accept bone marrow transplants from individuals of particular, other "population groups?"

To solidify the political orientation of this work, the authors then proceed to discuss Bill Clinton's opinions on race (!!), quote W.E.B. DuBois, and discuss the "widespread social prejudice" in Brazil as regards color and other aspects of physical appearance. After all of this, is it unfair to assume that this paper's authors are "liberals" on the subject of "race", and that, before doing the work of this paper, they already subscribed to the notion that "biological race does not exist?" Thus are the (to me) obvious biases of the authors (and it seems the reviewers and editorial staff of PNAS as well) revealed at the paper's beginning.

What about the data? The authors use 10 population-specific alleles (genetic markers) to assay for sub-Saharan/Negro "black" ancestry, what they refer to as the African ancestry index (AAI). Now, here is a crucial point. To determine whether these markers are useful, they first tested them in a comparison between 20 men from Northern Portugal and 20 men from sub-Saharan Africa. They observed that these markers demonstrate a sharp genetic difference between these two populations; they were highly genetically distinct. To quote the authors:

"There is no overlap between the two groups...A complete individual discrimination between the European and African genomes was obtained..."

Now, isn't that interesting? European and African populations were found to be so highly genetically distinct that individuals from these populations could be easily distinguished with gene markers, with "no overlap" between the groups. No such thing as race? Well, you see the authors want to claim that there is no correlation between these markers and phenotypic "race" in Brazil. Before we get to that, one wonders why they didn't instead do a blind study (see more below) on the Portuguese and African samples. First, assign each subject a code number so that the researchers don't know if they are European or African. Then, determine the "geographical ancestry" via these markers, and THEN go back and see if this genetically determined ancestry correlates to "white" vs. "black" phenotypes. Hmmm...why not? Is it possible that such an experiment would not yield the politically convenient conclusions one can obtain by specific interpretations of the Brazil data?

The authors then looked at two Brazilian populations. First, 173 individuals from a rural community were studied. These people were classified "racially" based on a subjective evaluation of their "phenotype" by "two health care workers." As the authors state:

"...those with white skin, straight or wavy hair, prominent or upturned nose, thin or median lips, independently of hair or eye pigmentation were called white, whereas those with black skin, black curly hair, black eyes, and thick lips were classified as black."

Others, including subjects for whom there was "doubt or disagreement" were lumped together as "intermediate." The problems here are obvious. This is all very subjective (which the authors admit to in their discussion), particularly the dividing line between "white" and "intermediate." The authors claim that the nose and lip structures were compared to "models from a human morphologic (racial??) atlas", but what were the cutoffs? And isn't this still all the opinion of the two "health care workers?" Also, given that "black" phenotypic characteristics are, generally speaking, genetically "dominant" to "white" characteristics, it is an obvious foregone conclusion that "blacks" in a mixed population will exhibit significant "white" genetic admixture. This is certainly the case in the United States, where many individuals who are clearly "black" have significant degrees of European genetic heritage. As regards the "whites" identified in this study, one wonders just how "white" they really are. Would all, or even most, of these individuals be clearly identified as, for example, unmixed "North Portuguese", never mind as "Scandinavians?" In the United States there are some people who can, to the casual observer, "pass as white" - in the broadest non-specific use of that word - who nonetheless claim to be "African-American" or "Native American", and who indeed derive ancestry from those sources. And, of course, the "intermediate" category would be undoubtedly mixed.

What is the data obtained from these 173 individuals? Using two methods, the AAI (African ancestry) of the "whites" was determined to be 31-32%, of the "intermediates" 45-48%, of the "blacks" 51-52%. Because of the overlap in the AAI measurements of these individuals, and because of the clear evidence of admixture in all three sets of samples, the conclusion that some want you to make is that "there is no such thing as biological race." Lost in all of this is the admission, by the authors, that there was a statistically significant difference between the "white" and "black" samples (P < 0.01), although the authors stress that this difference is much smaller (surprise!) than the Portuguese European/sub-Saharan African genetic differential.

[Silveira]

So, what do we have so far? First, that Europeans (Portuguese) are highly genetically distinct from sub-Saharan Africans. Second, when a mixed population from Brazil (folks from the same village [!], who have been interbreeding with each other) is divided up - using subjective evaluations of "white"/"black"/"intermediate" - it is found that: 1) all three types have significant genetic admixture, and 2) that the differences in African ancestry between the "white" and "black" elements of this "one-class community" (authors' words) were significantly less than that of the unmixed founder populations of Europe and Africa - a finding which is obvious and should come as no surprise to anyone. It must be stressed that even though these 173 villagers were highly admixed, there was still a statistically significant difference in AAI between the "white" and "black" subjects, a finding which seems to me to be supportive of the race concept, especially given the conceptual problems with this study as discussed above. If the authors want to stress the fact that both "white" and "black" individuals in a highly admixed community have significant levels of genetic ancestry from the "other color group", very well and good, but all this in no way invalidates race, as the unmixed Portugal- African vs. Brazilian mixed "white"-"black" relative comparisons make very clear.

The authors then attempt to solidify their findings by looking at Brazil as a whole; they studied 200 unrelated Brazilian "white" men, mostly from large cities, and from the main geographical regions of the nation. They claim that these data match their findings with the 173 subjects fairly well, but that there are some "interesting regional differences." For example, those subjects from south Brazil, the region with "the highest levels of recent (19th and 20th century) European immigration...." had a "median AAI value" that was "three orders of magnitude" smaller than that from other regions that did not have this European influx - thus demonstrating that the region of Brazil considered "most white" is indeed the least African. The authors of course stress a different point - that this south Brazil sample still has a higher AAI than the European Portuguese sample. Now, here is the kicker. How were these 200 men determined to be "white"? Self-identification! Get it? These 200 subjects were not even evaluated via the subjective methods used with the previous 173 subjects; the only determination for these men is that they said they were "white." Well, given the "widespread social prejudice" in Brazil about color that the authors complain about, is it any wonder that anyone who can somehow "pass as white" will claim to be "white?" Isn't it a well known fact that everyone in Brazil who is not obviously "black" or Amerindian will claim to be "white", and even some folks obviously of mixed race claim this as well? How many of those "white Brazilian men" actually look like Vin Diesel, Harry Belafonte, Adam Clayton Powell, Jr., or even Colin Powell? And even with all of that, there was still a clear genetic imprint of "Europeaness" in the "whitest" section of Brazil! But, what the authors stress are differences between a sample of admixed "white" Brazilians and the unmixed European founding stock. Are they so desperate to prove their point?

The authors also claim that their autosomal data on AAI relates well to their previous mitochondrial DNA (gender-specific marker) data on Brazil. Such gender-specific marker data are useful in looking at ancestry at the level of the population, and if that is how it is presented, fine. However, what needs to be made absolutely clear is that such markers do not, and cannot, give any useful information on an individual's ancestry. If you go back 10 generations, the direct mitochondrial (matrilineal) precursor for any given individual is one out of a total of 1024 ancestors. Obviously, the overall autosomal genetic input (which determines racial phenotype) from that one ancestor (1/1024) is essentially near zero. The same applies for Y chromosome analysis for males. When looking at large numbers of people, a researcher can get a fairly reasonable snapshot of the genepool (or at least either the matrilineal or patrilineal halves) via such markers; but can make no conclusions whatsoever about individual subjects. Only a large panel of autosomal markers have the power to do that.

The authors, in their discussion, attempt to formulate a hypothesis as to why "color" (and I presume "phenotype" in general) is a "poor predictor of African ancestry" (their words) in Brazil. They put forth the (reasonable) idea that genes that encode for phenotypic traits such as color are a small fraction of the total genome. If mating takes place in an assortative ("like with like") fashion as regards color in Brazil (and the authors claim it does and give references) then those genes which encode for a particular phenotype can become more strongly "associated" in certain individuals, independent of their overall "geographical" ancestry. Thus, the link between markers of ancestry and "color" phenotypes would, as they say, "inevitably decay over time." They give an example of a European male mating with an African female. Their children who happen to look more European would be considered "white" (sic!, relatively speaking of course), while those who look more African would be considered "black." Given the "social prejudices" of Brazil, these "white" children would be more likely to mate with other such "whites", while the "blacks" would marry other "blacks." Over time, the authors theorize, this would produce a phenotypically "white group" and a phenotypically "black" group, who would have roughly equal amounts of actual African ancestry. Let us assume this is in fact true (although as stated, statistically significant differences were found between these populations in Brazil). What does it mean? That "race does not exist?" That "'racial' phenotypes are poor predictors of actual ancestry, as a general rule?" Hardly. All this means is as follows:

[Silveira]

1. Although phenotype, particularly in more unmixed populations, does have value, it cannot be used to precisely determine actual ancestry, especially in more admixed populations. Since phenotype can be very subjective, what is considered "white" by Brazilians in general, or by the two "health care workers" in particular, may not coincide with what others may consider to be "white." In other words, for some people (e.g., Brazil) "white" may be a relatively light-skinned person who does not look obviously purely Negro, Amerindian, or Asian, but for other people "white" may have a specific meaning describing European-derived people of more-or-less unmixed ancestry. Thus, both phenotype and genetic studies are needed if a more precise determination of "race" - a perfectly valid term analogous to "breed" for dogs - is required.

2. In Brazil, regardless of self-identification, or subjective relative evaluations of "whiteness" vs. "blackness", the majority of the population are of mixed race, with "blacks" that have some European admixture, and so-called "whites" (actually "near-white" mulattos and mestizos) who have African and/or Amerindian admixture, but who can "pass" for "white" in a culture that creates very permeable and vague racial boundaries.

One can argue that both #1 and #2 reinforce, rather than refute, the idea of biological race, once the peculiarities of Brazilian history, culture and race-identification are taken into account. The case of Brazil no more refutes the Portuguese vs. African racial realities than the existence of German Shepherd-Bulldog hybrids (some of which may look more like one breed than the other) invalidates the existence of distinct dog breeds.

In summary, this paper does not, in any way, shape, or form, invalidate the biological race concept, and for anyone to make that claim is highly irresponsible to say the least. That a respected journal like PNAS published such a weak paper, and allowed such a politically-oriented introduction of the paper to remain intact, says a lot about today's scientific establishment - and none of it good. One wonders if, as regards issues dealing with human biology, we are now dealing with the same kind of entrenched, sociopolitically-motivated establishment that Galileo dealt with in his work on astronomy. If we let political correctness "inform" human-genetic scientific work, we are headed back to the days of the inquisition.

Finally, I would like to issue a challenge here. Let's 'cut to the chase' as regards that Brazil study. If the authors are really confident about their conclusions about the invalidity of biological race, then they should accept the following study, using available, off-the-shelf, commercially marketed technology:

They should get a large number of test subjects from throughout the world, from both more differentiated areas (Europe, sub-Saharan Africa, Northeast Asia, unmixed Amerindians, etc.) as well as from more admixed regions (Latin America, North Africa, Central Asia, Middle East-South Asia, Southeast Asia, etc.). They should then give all these test subjects code numbers (a "blind study"), and have their autosomal markers tested via the ABD 2.0 technology (DNA Print Genomics, Inc.), which can distinguish between the major racial groups (I assume the authors know of this technology). Determine the MLE's (Maximum Likelihood Estimate, best estimate of ancestry) for all test subjects. Then go back, identify the test subjects, and see if MLE ancestry can, or cannot, be matched to "phenotype" using: "racial" measurements, agreed-upon visual standards, and even self-identification. Their "null hypothesis" may well be that persons whose MLEs are 100% Indo-European/Caucasian ("White", European"), 100% African, 100% East Asian, or 100% Amerindian would be indistinguishable based on physical appearance/morphology, and that these 100% MLE "purebreds" would also be indistinguishable from 50:50 hybrids, or from any other two-way or three-way mixes of different proportions. We, on the other hand, would predict that all these groups WILL be easily distinguishable. Want to place bets who will be right? Why not do the test?

It is time to stop attempting to deform natural realities to match political ideologies, and instead accept the racial realities which exist and deal with them as best we can.

[Caipira]

Pena´s researches are very nice, but you need to remember two little things...

1- ) He tested white brazilians from mixed race places, specially. To represent southeastern brazilians, he choosed Minas Gerais state, instead São Paulo state ( São Paulo has 40 million people. Minas has 15 million. So, why MG ). If Pena had studied a little about Brazilian´s demography, he would see that more than 50% of Brazilian white population lives in São Paulo and southern states. With this information, Caucasoid DNAmt would grown from 39% to more than 50%.

2- ) Finns have about 60% of Mongolic Y chromossomes. But it doesn´t mean that they´re mixed-race types. Y chromossomes or DNAmt are just TWO lineages and can´t represent the whole genome, saotome198. If we see the whole genoma, Krieger tests would be more interesting than Pena´s test. Krieger showed that white northeastern Brazilians are, in fact, 70% white. Pena found about 1/3 of non-Caucasian DNAmt between white southern brazilians. But if we look to the WHOLE genome, Lisker found about 92% of European descent in the same people. The avarege white Brazilian European heritage should be 80%. So, looking to the genome as a whole, white brazilians have about 20% of admixture. This number is very close to lots of European countries:

I. Mongoloid and Negroid Ancestry
Asian lineages were much more prevalent than sub-Saharan lineages, detected even in the Southern and Western groups, mainly because Mongoloids entered Europe as migrants long ago, while Negroids came only recently as slaves and thus in smaller numbers with less access to natives for intermarriage. Northeastern Europe was by far most affected by the Asian ancestry, which is correlated with Finno-Ugrian migrations from the Arctic and Siberia occurring c. 4000 years ago, and possibly also with later invasions. Indeed, it comprises a principal component of genetic variation in Europe.

NATIONALITY ADMIXTURE
Lapps 47.4%
Finns 30.0%
Lithuanians* 23.5%
Estonians 22.7%
Latvians* 16.0%
Russians 13.2%
Turks 5.9%
Ukrainians* 5.5%
Swedes 5.5%
Bulgarians 4.9%
Portuguese 4.6%
Norwegians 3.5%
Czechs* 3.0%
Danes 2.9%
Frenchmen 2.9%
Poles* 2.5%
Italians 2.1%
Spaniards 2.1%
Slovakians* 2.0%
Greeks 2.0%
Armenians* 1.5%
Germans 1.3%
Belarusians* 1.0%
Yugoslavians* 1.0%
Icelanders .6%
Britons .4%
Scots .3%
Irishmen

* D.N.A. mt

So brazilian white people are more "pure" than some european types like Lapps, Fynns, Lithuanians and Estonians... Interesting, isn´t it?

Oh, Silveira, you talked about black brazilian´s genetic composition... Remember that Pena tested black people just from a small town in Minas Gerais called " Queixadinha ". This village can´t represent blacks from Salvador, for example. Blacks from Ribeirão Preto have about 82% of African DNA mt and 96% of African Y chromossomes, acording Hortolini... Unfortunally, Pena won a lot of atention from non-scientific magazines. According Mark D. Shriver, the Penn State geneticist who is a leading expert on racial admixture in the U.S., Pena had such an obvious political agenda that it was hard to take seriously. It appeard to be very clear when he sayd at PNAS magazine:

"You see leaders today, all over the world, doing it again!
Black, white, yellow, brown, people of every color
slaughtering people of every color! Because Satan is
always the same. "

Should such a sentimental statement be published in a respected scientific magazine ?? What do you think, Silveira ??

Caipira !!

[Vitor]

God is great?
"ALLAHU AKHBAR "


And I say... indeed.

God Is the only God, He is not only the God of muslims, he is everybody GOD!
Understand that?

[Valery]

2Caipira

>Lapps 47.4%
>Finns 30.0%
>Lithuanians* 23.5%
>Estonians 22.7%
>Latvians* 16.0%
>Russians 13.2%
>Turks 5.9%
>Ukrainians* 5.5%
> ...
>* D.N.A. mt


Is it correct to combine autosomal data with mt? If you lack appropriate and reliable data on this topic do not post anything such a mt/Y, because NE-Eur and Latin "mestizations" are completely different in demography!

Brasilians are really pure Europeans, with one exception: about 30% descend in direct maternal line from Amerinds and black slaves, as you have just written.

[Octavivs]

Dr. Pena is a froud. His work is totally manipulated. He commited serious errors in his works. He distorted everything with political motivations. His research abou the Whites in Brazil are a joke.

[Graeme]

I have not heard of Dr Pena. Why is he a fraud? Please elaborate.