You are finding corelating traits and using those traits to justify "race"...
You've not established what racial populations are or how one gene would justify inclusion...
How on earth would a gene be a "racial" gene? OR What special combination of genes makes a race and how would we know of the myriad combinations which compose a race?
This shows either just more standard race-denier dishonesty (in an attempt to confuse people about correlation(s)), or an absolute lack of comprehension of what correlation is. Because I can't be certain which, and because there are other people reading and I can't be sure every single one of them has taken (and remembered) at least one statistics class (which I think should be a standard required math class at the high-school or junior-high-school level, just to better prepare people to not fall for nonsense arguments like yours here), I'll explain correlations and how they apply to this subject.
Correlation is the tendency for two things to vary together, so when one is present the other also tends to be present, and if the things being compared are numbers, then when one goes up the other also tends to go up. The opposite relationship, in which, when one is present, then the other tends to be absent, or one number goes up when the other number goes down, is a negative correlation instead of a positive one. If there's no particular likelihood either way, then there's little or no correlation, just a random distribution of the two things relative to each other. And these situations can be given numbers from -1 to 1 using a simple formula, if you can determine how many individuals have each of the possible combinations. For example:
Negative correlation, value -1: a two-chambered heart and sebaceous glands. When an animal has either of these traits, you can be sure that it doesn't have the other.
Positive correlation, value 1: a three-bone inner ear system and sebaceous glands. When an animal has either of these traits, you can be sure that it has the other.
Positive correlation, value not exactly 1 but closer to 1 than to 0 (among tetrapods): skin covered in scales, and reproduction by laying eggs in a flexible leathery shell. All reptiles and no non-reptilian tetrapods have the former trait, but some species give live birth (with ruptured shells coming out after the live juveniles do, so conventional reptile eggs are produced but not lain), and some mammal species still lay eggs.
Little or no correlation, value 0 or really close to it: Left-handedness and type B blood. People who have either of these traits are not significantly more likely or less likely to have the other one than people who don't.
============================
Now, its application to the genetics of human races...
Any large amount of data can be checked for internal correlations easily, and the gene sequences of a bunch of people add up to a large amount of data. All you have to do is find how many people with a given version of a gene also have a given version of any other gene. If people who have one are particularly likely to also have a certain other one, that's a high correlation between those two. If people who have one are particularly UNlikely to also have a certain other one, that's a strong negative correlation between those two. If people who have one are no more likely or less likely to have another, then that's no correlation.
And when you have lots of genes per person and lots of people's DNA sequences, you can dump the data in a computer and find all of the correlations among them, including the 0s as well as the 1s and -1s. So, what would you expect the results to be? Most human DNA is common to all of us, so most genes will have a correlation of 1 with most other genes, or really close to it. But our variation obviously lies in traits that we don't all have in common, so a study of the distribution of variation will obviously look at the remainder, the genes that lots of people have different versions of.
Now, what correlations would be expected of those variable genes, given that we're only looking at individual genes which, due to your dreaded recombination, are free to appear in any combination in any individual (in other words, not genes that are tied to each other on the Y chromosome or in the mitochondria)? Normally, there's no particular reason for any two genes to tend to appear together, or to tend to appear only one at a time as if avoiding each other. So most of the time, the correlation value you find will be near 0. But there's a catch when you compare people who are closely related to each other. For example, in my family, all of us can roll our tongues up like a taco shell by lifting the right and left sides but not the tip, and all of us can bend our elbows slightly backward. There are genes for those traits. In the general population, those aren't correlated, but if you sampled my family, the results would tell you that they are, positively, because we have recent common ancestors who just happened to have both. (It also means that in our family you would find a negative correlation between, for example, the backward-bending elbow gene and the NON-rolling tongue gene; people possessing that combination would generally have to be members of some other group, not this family.) In another family, other correlations between other traits would show up that don't apply in mine, for the same reasons: they have recent common ancestors with each other who weren't ancestors to us.
That demonstrates the crucial point here: among genes which come in multiple versions which are free to recombine, common descent causes correlation (values farther from zero), while separate descent causes a lack of it (values closer to zero), so
correlations reveal relatedness.
So, in a species which had truly never subdivided into separate populations, you know what to expect when checking gene correlations (among the variable genes) throughout the species: random zero and near-zero values, as all versions of all variable genes are just scattered all over the species with no particular combinations being particularly more likely or less likely (other than in closely-related families). However,
that is not what is found in humans. In us, although most of our variable genes do have little or no correlation with each other, there's a set of over a thousand in which some strong correlations do pop up. People with one version of one gene (such as for large jaws) are more likely to have a particular version of another one (such as for a longer narrower skull)... and a particular version of another one (such as for a high testosterone level)... and a particular version of another one (such as lactose intolerance)... in just a few groups of a few hundred such genes.
Those mass mutual correlations, in which not only is one correlated with another but both are also correlated with a third and all three of them are correlated with a fourth and all four are correlated with a fifth and so on, form just a few groups, which have their own internal correlations but don't match each other. In other words, an individual who has one of the genes in one such group of genes also usually has the rest in that group, but not any of the ones in another such group. It's exactly the same as with families, except that these "families" are much larger, there are only a few of them covering the whole world, and they've been this way for tens of thousands of years instead of just for a few generations, still without the effect ever dissipating away in all that time like it does between families by intermarriage.
In other words, correlations alone reveal a handful of human populations with common ancestry internally but mostly separate ancestry between them since they first migrated apart. That's just the way it is, no matter how anybody feels about it and no matter how ardently some might insist upon lying about it or trying to obfuscate it by burying these facts in distractions and irrelevancies. (Also, since the geographic origins of the individuals in these studies were known, these few large genetic groups can be mapped, and each one turns out to, by some amazing coincidence, just happen to originate in one of the regions associated with human races as already independently identified by phenotypes. Wow... it's almost as if there were some kind of mysterious connection between phenotype, genotype, and ancestral relatedness!)
============================
That not only exposes the multi-level ******** you were spouting when you claimed that correlations are just any old combinations you can draw from a hat and call a group in order to support racism, but also shows why the question of how one decides whether an allele gets included in such a group or excluded from it was also a ******** "question" (actually not a question but a false accusation): the alleles include or exclude
themselves solely by correlation or lack of correlation with the others.
Due to recombination we can't get every ancestral line... recombination erases those lines
It does no such thing. Every single gene you have is its own "line" just as much as your mitochondrial DNA is and your Y chromosome is if you have one. Each whole, intact chunk of DNA (whether it's a group of genes, as in the Y chromosome and the mitochondria, or a single one, as in all of the rest) is inherited from exactly one ancestor who inherited it from exactly one ancestor who inherited it from exactly one ancestor. On this, there's no difference at all between the propagation of a mitochondrial genome, a Y chromosome, and a single nuclear gene that's not on the Y chromosome. The fact that there are so many of those lines of descent and they intersect so much is part of what makes looking at this in terms of lines of descent irrelevant. Those countless lines with their countless intersections form a gene
pool, and when you're looking at what
gene pool someone comes from, no individual line in that pool matters on its own. That's why race deniers focus so much on just two lines out of the thousands, which is exactly the same as focusing on only two genes: they, or at least the charlatans they got their talking points from, know that discussing the whole gene pool honestly wouldn't let them reach the "conclusion" that they decided ahead of time to aim for. In statistics, obsessing over individuals plucked out of the set is never valid; it means you're not even paying attention to the set at all anymore.
So, tell me, what is a race?
This post answers the question already, but more importantly, why did you ask it? We all know that you know, not only because you speak English and it's a common English word, but also because you started a thread to argue that there's no such thing, and someone who actually didn't know what something is wouldn't start a thread to argue that it's not real; (s)he's start a thread asking what it is. It wasn't a genuine question; it was just another rhetorical debate gimmick.
See, this is what is so damn great about the dialectic. I would fall between you and Juno on this.
