Oh, dear. It appears that This Kingship wil not be returning to explain just how his odd site actually indicates the reality of race. It seems that he has been waylaid by a sojourn into explaining mud-bloods and similar purported phenomena.
Winston
As tomndebb pointed out, there is a big difference between the color scale and the race scale. Sure, we can divide contiguous wavelengths into different colors, we can draw definite lines around all of the colors. We can do the same with humans, if we categorize by any single category. If we do this, however, it has no correlation on genetics. It appears that even if we do this with all of the factors we associate with race – skin color, facial features, body size and shape – it still has no relation with genetics. People have had to scrap around for anything which can correlate genetics and geographic region. The Science paper which we have talked about is really the first to do this (divide the world up into 5 parts by genetics). I have pointed out my issues with the paper, and I agree with DSeid’s very good points as well.
Science (not the journal, the concept) does not operate in a vacuum and there is certainly political input into scientific matters. But, the good science always wins out, no matter what the politicians think. Evolution stands as fact even though I’m sure a good number of politicians would deny it. Race is much the same. Peer review is a very potent tool in stripping away political bullshit.
Anyway, I’m out for the next 10 days as I’m going on vacation. Everybody have fun.
Oh BTW Tars, if you follow the departed This Kingship’s link to his archived articles, you can see the Science article – it is linked under “Genetic Structure of Race.” It is a copyright violation, but I suppose it is not my, the SDMB, or the Chicago Reader’s problem.
NOTE TO MODS: If it is somehow illegal to link to copyright violated material, please delete this post.
i agree with DSeid (libraries in Illinois now??? Wow!! 
), nothing really spectacular except basic genetics, i.e., people who live in the same town are more likely to be related. Was interesting to find out there is a language called “She”, proof that women speak their own language…
Sorry, I may have misused the terminology here. I did not mean to imply that you’d have every group member with an identifiable gene that non-members did not have. Only that you’d have a group identifiable by (frequency of) genetic traits. (Similar to Ashkenazic Jews being a coherent group such that members are more apt to be susceptible to Tay-Sachs, even though most individual members are not).
If I could venture a bit into Analogy Territory here, imagine that we are discussing the cultures of various regions of the US. It is commonly said that there is a Southern Culture, which would differ from that of the Northeast, or the Midwest, or West. Here too, it would be true that the number of areas that have cultures is somewhat arbitrary - you could subdivide to as many areas as you’d wish. And there are surely many border areas that fall in between as well. And surely there is a lot of variation within groups as well - any individual Southerner might resemble a stereotypical Northeasterner more than he does a stereotypical Southerner. Still it would be fair to say that there is some degree of coherence to the Southern culture. All things being equal, it would be a good bet that your randomly chosen Southerner is more likely to be a Bible-thumping fundamentalist, for example, than your randomly chosen Northeasterner, even though there will be millions of cases that don’t fit this mold. And calling the deep South the Bible Belt might be a bit misleading if it implies that everyone down South fits this mold, but it does capture a reality that exists.
So the question that I see here is whether a similar situation prevails with regards to race. Is there some commonality that exists between groups large enough to inhabit entire continents? The impression I’ve gotten from you guys to this point is that there is not. (See for example here, and here). The article here seems to say otherwise - as summarized in the NYT review:
It now seems to me that you are also agreeing with this.
And edwino:
And Tars as well:
So I wonder if we can put this to rest here. Do you guys agree that a randomly chosen African is more likely to be genetically similar to another randomly chosen African than to a randomly chosen European? (And the same for the other groups.)
One other point about the article that seems to speak to the issue I brought up in the second paragraph of my second post here.
What this says to me is that the grouping of people by continent is not merely a matter of grouping together completely disparate groups that happen to live in the same geographic region, but reflects genuine similarities between the larger groups.
[QUOTE]
Do you guys agree that a randomly chosen African is more likely to be genetically similar to another randomly chosen African than to a randomly chosen European?
. . .
but reflects genuine similarities between the larger groups./QUOTE] I think that we (everyone in the ongoing discussion) may need to consider what these statements mean.
What is “gentically similar”? What are “genuine similarities”?
I’m not trying to be disingenuous. Let’s consider a different analogy. If we went to Houston, TX, Pittsburgh, PA, and Minneapolis, MN and examined particles found in people’s lungs, the folks from Houston would possibly show up with hydrocarbon deposits associated with oil refining, the folks from Pittsburgh would tend to have coal dust, and the people from Minneapolis would tend to have wheat pollen. We might be able to associate people according to the micro-specks we found in their lungs (analogous to the inert markers used in the study), but if we ignored those inert samples, we would be unable to identify anyone by region.
There is another aspect to the study for which I will need to have edwino or Tars or DSeid correct me:
If I read the study correctly, the way the algorithm worked was to apply clustering to disparate elements. I think that this meant that if we looked at a range of markers labeled A to Z and found that markers A, B, C, D, E, F, G were all found mostly in Africa, this would indicate that there were people all through Africa who had ABC, ABD, ABE, etc. or BCD BCE, BCF, etc. and so forth. The computer program would note that any series of clusters involving the letters A through G would tend to appear in Africa, but that, within Africa, there would be smaller groups who shared A and F, but never included B or E. So the presence of combinations of A through G can be located geographically (just as markers on the Y chromosome and mtDNA have been used for founder effect population studies), but they can predict nothing more than geographic ancestry.
I dunno. It’s been my observation and personal experience that discussions regarding race hit raw nerves like no others, even when limited to the purely scientific aspects of it - and that includes discussions of evolution. This applies to all parties on all sides of the debate. No one should consider themselves exempt.
Tom, you have always been one of my favorite posters on these boards. You can usually be relied on for a reasonable, lucid, and articulate point of view, but on your posts regarding race and genetics I often find myself just going "Huh?” Your post about lung particulates and genetic markers is a good example.
Now, maybe I’m too stupid to grasp the subtleties of these arguments. Maybe my own raw nerves keep me from seeing the obvious. Maybe. Even so, I think it would be prudent for us all to recognize the political and emotional sensitivity surrounding discussions of race.
Finally there’s this:
Now, you know we’re gonna need a cite for that.
I mistyped the project that has been underfunded for political reasons: It was not the Human Genome Mapping Project but the
Human Genome Diversity Project.
I would never deny that the subject is sensitive or that discussions tend to revolve about issues that become political. Anything that touches on race will eventually become political in the U.S., simply because so much law was based on race for so long. However, that was not the point that I was making.
The sequence of the previous exchange was (in paraphrase):
The Kingship: The statement by sociologists was Politically Correct bullshit.
tomndebb: What “political agenda” is being served?
WinstonSmith:That there exists a political motive to minimize or eliminate race as a real category within humans is, to me, blindingly obvious
tomndebb: I do not see any effort to minimize the existence of race among the “Politically Correct.”
In other words, in two posts I encountered the notion that only politics (and probably politics on the Left) causes us to try to ignore or hide race. It has been my observation that the reverse is true. It is the Right (in the U.S.) that now makes an effort to talk about “color blind” initiatives and services while it was the Left that opposed funding for the HGDP while the scientist that was among the first to claim that race was invalid as a biological category was seeking funds for that project.
The principal motivator for the HGDP was Luigi Cavalli-Sforza of Stanford University who, in 1994, declared The classification into races has proved to be a futile exercise for reasons that were already clear to Darwin.. So we have a founder of the “no race” school of science looking to map ethnic/genetic diversity while the forces of Political Correctness oppose him.
(I am not claiming that all “PC” people embrace the notion of race. Certainly, when people such as Shockley, Rushton, or Murray and Herrnstein have gone to the popular press with claims that blacks (or some statistically significant number of blacks) are less intelligent than whites, it has been the PC people who have met them at the barricades and shouted them down. However, that is a function of the way the popular press operates and does not have a direct bearing on politics interfering with science. The scientific rejection of the pop theories of race have been handled outside the popular media and outside the halls of legislatures.) I am pointing out that in the context of the original exchange in this thread, the claims that “race” is being “suppressed” for political motives are actually an inversion of the historical record.)
As to the flaws in my analogy: analogies are always subject to poor presentation or misunderstanding. If you’d like it clarified, I’ll attempt it.
Let’s start at the top.
There clearly is concept of “race.” This concept classifies people into sociocultural groups based on a set of superficial features.
There clearly are human population groups that have been apart for many generations, based on migration patterns, based on geography, based on cultural features, etc. Members within these groups have great diversity. Members within one group are more closely related to each other than they are to members of another group, but the intragroup diversity still swamps any intergroup differences.
The differences and similarities between these groups is matter of legitimate scientific study. It tells of human migration patterns. It has clinical relevance.
Is “race” the right word to use when identifying and studying these groups? Short answer: No.
The social construct that is “race” identifies the population groups and the shape of those groups in advance of the data. The data that exists shows, clearly, that the social construct of “race” poorly correlates with the real size and shape of those groupings. It imposes a conclusion upon the data that is based upon social and superficial characteristics, a conclusion inconsistent with the data, one that is false. “Race” has little utility in the scientific study of the genetic differences and similarities of human populations.
So let us take your example modified as the use of the term “race” would suggest: Are two “White” Americans more likely to be gentically similar than one “White” and one “Black” American? Does membership in that racial group tell us anything significant about these people’s genomes? When studying genetics on group levels, is “race” a useful construct?
The answer to the first question is - probably not. The two “White” guys may be from pure Icelandic ancestry for one guy and part Indian, part East African in a great grandmother, part Arabic, part Germanic for the other. The White/Black comparison may share a great grandparent or two. The diversity within “White” and “Black” is huge; the words tell us little about backgrounds and even less about genomes.
The second question - mostly no, but occasionaly yes. Certain alleles are more frequent in certain populations and certain clinical responses are more common in certain populations. Sometimes race is the best proxy we got for some as yet unidentified allele (or combination of alleles) … a poor proxy but better than nothing.
And the last question is answered with an emphatic no.
BTW, Welcome Winston!
OK, but these are questions that no one is asking in this thread. So I wonder if you could also take a crack at the actual question that I asked. Again:
Tom, I have failed to understand what you were driving at in your question about the meaning of “genetically similar” etc. Also the rest of that post.
The words “between groups” in my quote in the previous post should probably read “between members of groups”. Same for the final sentence of my previous post.
The question is essentially the same as this one:
I’ll do my best to answer, but it has little to do with “race”. And since your op was about “race”, I think that it is important to review what the difference is between that term and the term “population group” used by genetic scientists.
Well, every human has some similarity, so of course there are similarities acrosss the world, let alone a continent. All mnales will have a Y chromosome across the whole continent, for example. No, in the way I think you meant it. In today’s world continents are made up of peoples who migrated in at many different times and of many different origins. To the extent that these populations have intermingled, and were very similar to each other in the first place, they will have similarity across the continent. Now in the days of yore, sure. That is founder effect. For example, all Aborigines, I am sure, share some alleles, or at least cluster in alleles, compared to other population groups.
To the second, and not at all the same, question. I don’t know. I doubt it. Let’s work it out. 95 to 97% of genetic diversity is within groups and African populations (in toto) have a wider spread of that diversity that others. Think of that as a full card deck. The rest of the world’s populations have less diversity. Think of a card deck with all Aces, Kings, Queens, twos, threes, and fours removed. Now deal hands out. The Africans have a wider number of possible hands dealt. So any random two Africans are likely less similar genetically than any random two Europeans, any particular African has greater chance of coming from one of the statisitical tails, and the random European is more likely closer to the mean than the African is. So I think the answer is no. And the Science article doesn’t change that assesment.
That’s fine. But the answers to these questions may have implications for racial issues as well. Or not.
Of course. But the question here is whether they will have similarities that go beyond the mere fact that all are human. IOW, are peoples who inhabit (or originated from) the same continent more similar to each other on average than they are to people who inhabit (or originated from) other continents?
The second question is indeed pretty much the same as the first, though you’ve apparently misunderstood it. It was NOT about whether two Africans are more similar to each other than two Europeans are to each other, but rather about whether a random African is likely more similar to another random African than the same random African is to a random European. This has been the subject of some discussion in previous threads (I’ve linked to some of Collounsbury’s posts on the subject - I seem to recall this as being the thrust of lucwarm’s “average genetic distance between pairs” issue as well).
The Science article appears to have a great bearing on this, as per my earlier quote from the NYT summary. And the Science article itself seems to bear this out as well:
IOW, while this excess difference is small, there is a greater difference, on average, between people of different populations then people from the same population. You have some wriggle room here because the sentence does not specifically say that these are the large populations that span continents, but I think this is clear from the context of the article, which is dealing with just such populations.
Well, what does “genetically similar” mean? According to your restated (from months ago) question, I would guess (subject to a correction of my interpretation of the study) that the answer to that question is No.
If Rosenberg and company can find clusters of markers that are associated with Africa and clusters of markers that are associated with other continents, that indicates that certain markers have grouped themselves geographically. However, if we need to associate five separate clusters of markers to arrive at a geographic origin, we cannot predict that any individual will have any specific set of clustered markers.
(emphasis added)
It would seem to me that they are saying that any of the clusters may appear anywhere, but that a statistical analysis will show relationships in frequency of clustering among populations. They started out with 377 loci and needed 150 associations in order to identify the major world groups. I’m still open to correction from the true science guys as to my interpretation, but that seems to me to be saying that they are not claiming that an African will have all of the 150 associations, but that the clustering at level K in their algorithm will display more frequent associations among those 150 associations in Africa. So any individual from Africa may have a totally different set of associations from any other individual in Africa. Given that they began with the most genetically coherent groups that they could find (“HGDP-CEPH Human Genome Diversity Cell Line Panel”), they were able to use statistical methods to rate frequencies within those coherent groups.
Does that lead to a conclusion that those groups are, internally, “more similar”? It does not yet suggest that to me.
Suppose an individual, Aa, has some constellation of clustered loci to indicate a geographic origin. (And that does not mean sharing all the same markers, but only some group in a statistically reported frequency.)
Suppose that among the 93% of intra-group genetic variablility, the alleles for individual Aa are more similar to an individual from a different geographic group, Be, and less similar to an individual from Aa’s geographic group, Ce.
Are Aa and Ce “more similar” because they share a statistical frequency for the geographic loci, despite their other differences?
We already knew that we could use Y-chromosome and mtDNA analysis to identify parentage–and that parentage is clearly geographically realted. The variability among the rest of the genetic structure, however, appears to me (subject to correction) to overwhelm the similarities we find on either the parental genes or on the clustered frequencies.
Tom,
I believe you may have misunderstood the term “cluster” as used in the article. It does not refer to a “clusters of markers”, but a cluster of people, i.e. people that are grouped together by virtue of tending to share similar markers. The sentence you highlighted means that any individual can be only partially associated with the “pure” form of any given grouping, and may be also partially associated with another group. This is increasingly relevant as fewer cluster are used (i.e. lower numbers of K). So that if K = 2, for example, then the pure form of the groupings are given as Africa and America - a European would be somewhere in the middle and would identify with both groups to some extent, with the coefficients summing to 1, as noted.
Beyond this, the basic gist of your post seems to be - and correct me if I’m misstating your position here - that the 93% intra-population variability will overwhelm the 3%-5% inter-population variability. So that an “Aa” will be less similar to a “Ce” from the same population than to a “Be” from a different population - even though the first couple share common geographic markers, the dissimilarity between the Aa and the Ce overwhelms whatever similarities they have.
This seems like a simple error to me. It goes without saying that any individual might be more similar to someone from a different population - that’s why I keep saying “on average”. Of course in your example an Aa will be more similar to a Be. But the real point is that since we are assuming that both populations have the same percentages of "Aa"s, "Be"s, and "Ce"s (since these are assumed to not vary by population) there will still be on average a greater difference between members of different populations. Because while in your case the Aa is closer to the Be from the other population than to the Ce, he is closer still to the Be from his own population.
Again, fill me in if I’ve misstated your position. Also, you might want to comment on the quotes I’ve cited from the NYT article and the Science article, which seem to support what I’ve said in plain English.
I’m not sure I agree with your perception of the use of cluster, but you have correctly understood the rest of my comments.
However, supposing a similarity “on average,” where does that leave us in regards to defining race? Can some individual be of one race or another “on average”?
If we take the mtDNA marker from Africa that Bryan Sykes found in the Brighton family with a clear genealogy going pack to the Domesday Book, does that mean that that family is more nearly African than British? Obviously, whatever other differences and similarities they have with the surrounding Britons, they have a clear connection to Africa. Does that outweigh the other differences?
Averages apply to groups, not individuals. Again, individuals can vary.
I’m not trying to establish any Nuremburg Laws here, so I don’t see why these are issues. As noted previously, there are many people of mixed ancestry. Still, people can vary based on how much of their genetic code is inherited from their ancestors in one region or another. If someone was for whatever reason obsessed with categorizing the Brighton family they would have to run them through the model and see where they fall out. But either way, this does not mean that a general statement could not be made about Europeans as a group. Again, I don’t see the relevance here.
Izzy,
Sorry if I wasn’t clear, but I did answer your question. Put another way, the wider diversity of the African population I think, makes two Africans less likely to be similar at any particular locus, than one African and one European. There is nothing in Rosenberg’s study that makes me change that assessment. Nor in your quote, since you specified the African population, which is that special case of significantly greater diversity than other population. Still, to really answer you’d need hard numbers. But my WAG is that the average propotion of gentic differences on all alles (not just neutral ones, but ones more likely to be conserved) is wider between unrelated individual within the African population than the difference between individuals in different populations. Now if you want to compare, say Eskimos and Aboriginines, then sure. Between someone from Japan and a Peruvian Indian, sure. Or even one Western European and one Asian vs. the same in each. But I’d not significantly raise my odds over the coin flip in any of those cases. Nothing that qualifies as a continental commonality.
And Izzy, while the term “cluster” was relating to population clusters, the cluster were formed based upon groups (or patterns or clusters) of neutral alles. Not individual alleles. Tom’s understanding is the same as mine after many rereads.
I still do not understand the significance that you are placing on this, and what this has to do with “race” as a scientifically meaningful concept. Now if this study looked a group of Black Americans who had resided for several generations on the same continent, and compared it to White Americans with the same generational length of rootedness, and found that they clustered differently, then it might have some significance … but even then very little. And I believe that such is the case. Certain alleles ocur more frequently among Blacks than Whites, and visa versa. But the predictive value of finding a particular version of an allele based on self-reported race makes it a poor proxy for actually testing for the allele.
But, since you are asking, I will state that sure, looking at poulation groups as your unit of analysis, you can find differences. These differences correlate with human migration patterns, with length of seperation and with magnitude of seperation. They might even overlap some with some racial identities … or not. Or cross over racial identities and cluster more according to geography and number of generations rooted in the same locale. Or not.
And again, so? The point is?
Well I picked Africa at random. The point would apply to Europe and Asia, if you want to say that Africa is a special exception.
But even for Africa, I’m unsure if you might not be making a conceptual mathematical error. I would think that if Europe and Africa had identical distributions with the sole difference being that the tail of the African distribution is larger, that the difference on average would be about the same. This because while it is true that for a randomly chosen African who is not in the tail of the distribution, the maximum potential distance to another African is larger than the maximum potential difference to a European, for those Africans who are in the tail of the distribution the number of Africans who are close to them is significantly greater than the number of comparable Europeans. I thought about testing this out on some simple distribution and testing for E(X[sub]1[/sub] – X[sub]2[/sub]), but decided not to bother because this is all assuming that that the European distribution falls somewhere in the middle of the African distribution – it may be somewhere at the end or off the spectrum entirely, in which case your position fails completely. So your WAG is pure speculation – as you say, you would need hard numbers.
I don’t see where this contradicts anything I’ve said – I believe what you refer to may be what is meant by the term “multilocus genotypes” in the article.
Well again, I’ve never that race is a scientifically meaningful concept – I seem to recall specifically disavowing this in previous threads (I am happy to let scientists decide what is meaningful to them). What’s the point? Does it have to have one? I am not about to do anything different as a result of any conceivable conclusion to anything in any race thread. But the same could be said of all sorts of threads. Sometimes knowing what is or might be the truth about an issue is of significance or interest to people, and this particular question has come up before and assertions have been made that are apparently contradicted by this new study. So I brought it up.
But if you specifically want to know of a larger implication, I would suggest that there may be an implication with regards to the hoary question of whether observed differences between groups might be attributable - in part – to genetic factors. To the extent that groups are considered to differ – on average – genetically between each other, this raises the possibility that these or similar differences might actually manifest themselves in some physical way.
Now I am aware that you have attempted to distinguish (in your fourth post to this thread) between neutral markers and coding DNA, based (apparently) on the presumption that genetic drift is a greater factor with regards to the former. But I would also point out that the founder effect would have the same impact for both types, and the effects of similar climates would have the opposite impact. The authors of the article certainly felt that their conclusions would have implications for coding DNA, with their speculation about the clinical uses of the information.
Now the impact of such a finding on a mixed population, e.g. African-Americans would be a smaller one. Still, to the extent that a subgroup has a larger or smaller percentage of ancestry from a particular group, it might be presumed to share to a larger or smaller percentage that group’s (average) differences.
Again, this is not to imply anything about what you can or cannot determine about individual members of a mixed group based solely on their ethnic self-identification.
On further reflection, I retract this. The rest stands.