While I was discussing the arbitrariness of defining human races on a forum, one individual contradicted me by presenting a means to objectively quantify human races. He explained that the number of human races is objectively quantified by using "Bayesian K-means clustering and determining the correct K by inferring the second order rate of change in the log probability of data between successive K values. The correct K seems to be 7 which also matches the phylogeographic criteria for race perfectly. Europeans, surprisingly, form a distinct race." I don't know if this helps or distracts, but here's some sort of diagram illustrating that K=7: http: // img399.imageshack.us/img399/9622/scimagtd6.jpg Unfortunately I lack qualifications in the field of mathemathics to verify these claims, so I need opinions of others on this.
This is complete nonsense. Bayesian k-means is a clustering algorithm, it has got nothing whatsoever to do with race.
That Bayesian k-means is a clustering algorithm is a fact that I already knew well. However, that it has got anything to do with race is not what has been argued.
That's what I argue. How do you suppose clustering people determines their race ?? Last time I checked race has to do with a certain amount of genetic variation.
Enmos, K means clustering can in theory do that if you clustered all the genetic information of all individuals and if you make an estimation of K (instead of an a priori assumption). The resulting clusters would show distinct subpopulations. The problem why it would not work is simply because we do not have an estimation of the overall genetic divergence of any given population. In principle you would have to have the whole genome information (or at least a sizable chunk of it) and make whole-genome-comparisons. To date we only have a limited number of markers. In other words, you could cluster the total population according to certain known allele (or SNP, or whatever) distributions but that's it. It is not a problem of the algorithm per se. What I suspect is that some allele frequency or microsatellite study has been perverted by certain people who want to believe that with the current knowledge race is a feasible concept.
I'm not sure why race ISN'T a feasible concept. Virtually every animal I've ever done population work on, that was geographically stratified, differed in their genetics related to their location in space. Humans would have to be astonishingly unique to lack genetic variation that corresponded with different populations. If race is defined as groups of humans that share some subset of their genome due to their geographt, then I see no reason why you need whole genomic sequencing to establish race. If this were the case, the field of population genetics would be prohibitively expensive. To determine race, that is, distinct subpopulations that are somewhat isolated from others, one need only look at loci that would drift. Sequencing highly conserved sequences would be a fruitless endeavor, especially considering the recency of modern humans. To say that race doesn't exist in humans is misleading. More accurately, what is commonly thought of as race isn't based well on the genetics of where people were, and where they came from.
Well, Roman, I'll tell you why people don't WANT it to be a valid concept: People today suffer from the Naturalistic Fallacy and they equate "Existence in nature" with "Good". If races exist, they think this would make racists ethical, which is nonsense.
As far as I know there is no biological basis to distinguish human subspecies, which is what race is.. subspecies.
There's plenty of biological basis to make up races- skin color is the most common, though the more discerning racist will discuss facial features and cranial structure. Whether or not these correspond to genetically distinct subpopulations has been shown to be mostly false. Jews and Arabs, for instance, are essentially the same genetic group, despite being considered different "races". Africans, which are usually just "black" are actually more racially diverse than Europeans, as they're from older populations that have had a longer period to diversify. Europeans, being a relatively new population, are all fairly related, having descended from a handful of common ancestors. There's also evidence that as little as 140,000 years ago, humans went through a population bottleneck, where there was just one female ancestor for all of us today- the mitochondrial Eve. Due to this, the variation between humans is quite low. In fact, the variation within "racial" subpopulations is as great as the variation between subpopulations. That is, two black men differ in genes almost as much as between them and a white person. By the way, race, as a term, has a complex and non-biological meaning. There is no real biological equivalent for race, as applied to human populations. Race doesn't refer to subspecies, it refers to subpopulations, in the strictest sense. Even then, many "subspecies" in the wild are little more distinct, genetically, than any two "subspecies" of human. For instance, the discovery that Jews & Arabs are essentially the same people came as something of a surprise, much like the discovery that many birds, previously classified as separate subspecies or even species, are not genetically dissimilar enough to qualify for either classification.
Bullshit. The efficacy of many drugs depend on race. The PC crap about there not being race was getting people killed until recently, when the drug manufacturers could no longer ignore the fact that some drugs work much better on some races than others. Race is not the same thing as subspecies. And race does exist. And this does NOT make racism OK, which is what the entire "No Race" movement is all about. But once again, the people that pretend to love equality do great harm by shying away from the truth and into their La-La land.
It's not bullshit. In biology race is the same thing as subspecies. I admit, that it could be that the threshold is crossed in humans.. I don't know what the threshold is. All I know is that biologists say humans are all the same subspecies.
You must have assumed that I'm arguing in favor of this clustering, even though it's revealed in the OP that this clustering method was presented to me by another individual, and that I needed to ask for your opinions on this clustering method, since I didn't have the qualifications in the field of mathemathics to verify or refute it myself.
Yes, I did assume that, sorry Please Register or Log in to view the hidden image! Clustering by geographical distribution is not a method you can use to determine whether or not humans have different races.
No, the efficacy of drugs depends on various genetic and epigenetic factors. eg. presence/absence of specific alleles, SNPs, STRPs, methylation status etc. Whilst genetic factors like alleles and SNPs can sometimes cluster in people of specific geographic areas and people of a certain appearance (thus giving the illusion of “race”), they rarely exclusively cluster in one specific people (thus indicating that there is no such thing as “race” from the point of view of genetics).
Huh, and here I thought biologists stayed away from biconditionals, given how many exceptions there are to presumed rules.
