Hello There, Guest!  
 Previous 1 36 37 38 39 40 Next   

[OBG] Nature of Race Full

(2015-Sep-10, 17:38:21)Krom Wrote: You talk about cherry picking but you don't calculate a mean from all these studies. I take them all into account and there is more than above. A lot more genetic markers have very low values of under 5%. Obviously those traits like skin colour and limb have to be excluded because they are not neutral genetic/phenotypic markers.

I replied to an earlier point above.

Look, a Fst value of 10 to 15% is large and this is what you find with the SNP data, both in 1000 Genomes and HapMap. (Barbujani is fond of citing the unweighted 1000 Genome Fst values, reported in the original 1000 genomes report.*) For making inferences about quantitative differences owing to neutral selection, this is the preferred value:

"One valid solution is to estimate FST from neutral markers with mutation rates comparable to those of the loci underlying quantitative traits (e.g. SNPs). QST can also be compared to ΦST (PhiST) of amova, as long as the genetic distance among allelic variants used to estimate ΦST reflects evolutionary history: in that case, neutral divergence is independent of mutation rate. In contrast to their common usage in comparisons of QST and FST, microsatellites typically have high mutation rates and do not evolve according to a simple evolutionary model, so are best avoided in QST–FST comparisons."

Plugging in the SNP Fst differences into the equation for expected quantitative differences, gives you large differences. I discussed this in my paper. Now for phenotypic Fst (i.e., Qst values) you give:

(1) Pelvic shape (males) 2.6% Pelvic shape (females) 3.3% --
(2) Craniometrics is 14.6% as calculated by Relethford (1994, 2002) -- large
(3) Limb ratios are around 38% Fst. -- very great
(4) Skin colour is 87.9% Fst (Relethford, 2002) -- very great

All but (1) represent large to differences, as judged from the perspective of the social science. (See my table 4.12. Attached below.) And this is neglecting the fact that Qst is calculated as:

Qst = V G, among / (V G, among + 2V A, within),

with a 2 in the formula to account for diploidy, which means that expected quantitative trait differences:

2Fst/(1 + Fst) (Leinonen et al., 2013)

These are large numbers. (Oh yes, and I emailed Barbujani about this, but he didn't care to reply.)

Maybe you are just thinking of a different scale. But since we are talking about behavioral and medically relevant differences, the scale I am using is arguably the correct one, since this is the one people use in the biomedical and social sciences. For example, the SES variance between Whites and Blacks in the US is around 7% (medium effect size differences, measure depending). No one says, "Oh these don't exist, so there's no reason to study sociologically delineated race". Why? Because these magnitudes of differences are not typically interpreted as being irrelevant. Now maybe you disagree. Maybe you feel that sociological race classifications are empirically meaningless because most the of variance is within populations. Do you? Whatever the case, for better or worse I am following the common evaluative standards, which leads me to the expectation that there are "important" behavioral differences, where "importance" is evaluated according to common standards for judging inter-individual differences.

Now, that returns us back to the issue of "the" magnitude of genetic differences. There is of course no "the" because estimates vary according to the races under question, the loci analysed, and the method used. Since I am not particularly interested in continental level divisions -- for various reasons, some empirical some ideological -- I do not fixate on these differences. Rather, I look at differences between e.g., Europeans, West Africans, and Amerindians, which given my construction of the race concept and consistent with the biogeographical ancestry one, I can treat as three separate racial groups.

But for some reason you are interested in continental-level differences. In interpreting these, as I noted, you have to keep in mind mutation rates, both if you want to make interspecies comparisons or interindividual, between race ones.

serum proteins and red blood cell enzymes 9% (Latter, 1980)
X chromosome STR's Ramachandran et al. (2004) 4.9%
Alu insertions Romualdi et al. (2002) 8.9%
642,690 atDNA SNP's Li et al. (2008) 9.0%
HLA 7.0% Meyer et al. (2006)

Now you said, "you don't calculate a mean from all of these studies". That is correct and for good reason. It makes no theoretical sense to try to average e.g., short tandem repeats (microsatellites), which are picked specifically because of their high mutation rates, and e.g., SNPs, which have relatively low rates to try to get a "true differentiation rate", without even a clear idea about what you mean by "differentiation", or to combine values made using different estimators, different methods of loci aggregation, different methods of handling rare alleles, ect. If Barbujani in fact did this, I would have to guess that he was being intentfully misleading -- and if it was just you, that you don't have a good grasp of this issue.

In my paper, I believe I said: "[F]or example, for humans, continental microsatellite, SNP, and mtDNA Fst values are typically around, respectively, 0.05, 0.12, and 0.20...The median continental race SNP Fst value is said to be around 0.12 (Li et al., 2008; Campbell and Tishkoff, 2008; Elhaik, 2012; Bhatia, Patterson, Sankararaman, and Price, 2013)".

This sounds about right, though I guess we could conduct a more thorough review.

The point here though is that discussing average "differentiation" values is meaningless:

(a) If we wish to make an interspecies comparison to assess the magnitude of e.g., continental human differentiation relative to that between the formally recognized races of other species, then we would have to pick a specific class of loci and compare. Both Templeton (1998) and I did this and found that humans fell around the 15th percentile (low average).

(b) If we wish to use Wright's metric to judge the magnitude of continental-level differentiation then we would need to again use appropriate loci -- presumably biallelic, but we could check what he used.

© If we wish to determine the expected quantitative trait difference between e.g., continental races, then again we need to use the appropriate loci e.g., SNPs for common morphological/behavioral traits; perhaps others for biomedical ones. Presumably not high mutation ones such as STRs.

*In "Genes in space Selection, association and variation in spatially structured populations", the authors notes:

"On these scales, the 1000 Genomes Project Consortium (2010) reported pairwise FST of 0.07 between European and West African populations, 0.08 between East Asians and West Africans, and 0.05 between East Asians and Europeans. These numbers are unweighted averages over all sites, and are therefore lower (empirically, by a factor of about 2) then the corresponding weighted average"

Attached Files Thumbnail(s)
(2015-Sep-10, 09:21:20)Krom Wrote: “It is likely that each breeding population will prove to be genetically unique, so that all will be racially distinct in Dobzhansky’s terms. But this is not the general use of the concept of race in biology, and the concept has not in the past been associated with this theory of human diversity.” (Livingstone, 1962)

“There are undoubtedly no two genetically identical populations in the world; this has nothing to do directly with the validity of race as a taxonomic device. Unless we have defined exactly what we mean by this… differences between populations are population differences, nothing more.” (Hiernaux, 1963

Ok, but as said, natural division races are not equivalent to genogamodemes in the sense of breeding populations which differ in some genes. Do you grant this point or shall we keep playing pop up goes the weasel? For example social classes and religious groups would often be the latter, but would not typically constitute natural division races (unless they were sufficiently linebred). As discussed prior, I am not sure that Dobzhansky’s races were just genogamodemes -- that is I'm not sure that all of his genetically different Mendelian populations were races. Whatever the case this genogamodemes concept offers a nice contrast with the natural division one, which requires individual to be more overall related to other members of their groups than to members of other groups. And yes I agree that many genogamodemes would not make for good races.

Quote:Demes or meta-populations (groups of demes) with very few exceptions are not ethnic groups.

When I google "deme" and "degree of isolation, I get e.g.,:

"Instead, the population of the species is subdivided into a number of demes, which are isolated from one another to varying degrees, either by unfavorable habitat or simply by distance. Given some degree of isolation, it becomes highly probable, indeed inevitable, that some genetic divergence among the demes will occur". (Merrell, D. Ecological Genetics)"

"No extra operations are introduced as a results of distributing the populations into overlapping demes. The only factors that vary in these experiments are degree of isolation..." (Ebeling et al. "Parallel Problem Solving from Nature)

There is no clear degree of isolation, demes can overlap, demes can fall on a continuum, demes per se don't necessarily have genetic differences.

So I find this odd: https://books.google.com/books?id=yx8yBw...22&f=false

Nordgren's demes look just like my races: "Demes are groups of individuals who are more similar to each other than to any other individuals". And they look quite unlike the demes of others e.g., "Given some degree of isolation, it becomes highly probable, indeed inevitable, that some genetic divergence among the demes will occur". If you can find a meta-review of the "deme" notion send it along, because it looks like many are using the term to reference very different entities.

Anyways, this is what I like to focus on concepts not terms. What was a Krom-demes, again? If you wish to say that demes need a very high degree of reproductive isolation ("more or less" is pretty ambiguous), how do you justify a cut off and what do you call less isolated groups ... incipient demes?

Quote:Ethnic groups (e.g. "Germans") are mostly arbitrary spatial constructs like races, e.g. "Caucasoid".

I don't imagine that most would define a German strictly spatially e.g., such that a "German" could never exist outside of "Germany". With the amount of immigration, I also don't imagine that "Germans" in the political sense even constitute and "ethnic" group in the anthropological sense of groups of individual with a shared identity, history, heritage... Though, there are ethnic Germans and some e.g., ethnic Germans belong to a different natural division race than some political/ethnic Chinese. Caucasoid is easier since this isn't an ethnic identity or a political construct. One can stipulate that Caucasoids refer to members of the West Eurasian biogeographic ancestry group in which case they would constitute a genetic construct (biological race, my sense). They don't form, as typically used, an obvious spatial construct (spatial population) since "Caucasoid" usually refers to Whites in Australia but not Black African in Europe. Would they be a social construct -- well even if they formed a taxa subspecies they might be, so this is a pretty non-informative category. Maybe a cluster-morph?
(2015-Sep-10, 18:33:31)Krom Wrote: Despite the standard scientific/textbook definition of a (gamo)deme is a more-or-less isolated breeding group/mating circle.

Why not just type "gamodeme"/"deme" and search Google books. Its not hard.


You missed the point as usual. (I wasn't discussing degree of isolation here, but location on a continuum.)

(1) Demes can be cut from an isolation by distance breeding continuum. True or false?
(2) If you want to say that these demes which fall on a breeding continuum are "sharply differentiated", then you must also say that my discretely conceptualized (non-overlapping) races cut from a continuum are likewise "sharply different".

I am fine with that. Just acknowledge the point. It's odd that you can't see the symmetry.
(2015-Sep-10, 18:48:03)Krom Wrote: Ignoring genetic differentiation for a moment, races are intergenerational aggregates of demes.

Are you saying that Krom-races are these or that Chuck-races are? That is, who's concept are we discussing now? Or are you arguing that this is what "race" generally should mean (general vesus narrow concepts), given historical usage? If the latter, what's your justification for this particular broad concept? And how isn't this definition "revisionist", given that the "deme" concept is relatively new? And isn't this thread concerned with my race concept? Or are you now arguing that mine mismatches with what race, broadly understood, really means (according to Krom)? Clarify please!

Quote:As Templeton explains, races must be "geographically circumscribed". This doesn't just mean arbitrarily marking boundaries across space as you are doing for "Caucasoid", "Negroid", "Mongoloid". Your races are not "natural" and don't exist anywhere but in your mind.

As Templeton explains... Templeton couldn't even distinguish between a morphological rule of thumb (the 75% rule) and an imaginary genetic rule. Moreover, he couldn't recognize the difference between taxa subspecies and races. As I asked prior, if races = taxa subspecies, how can there be nested races? But now -- according to you, at least -- Templeton's taxa subspecies -- he doesn't discuss race in a more general sense (e.g., "race" versus "geographic race") -- must be (prescriptively) "geographically circumscribed". If this is generally the case, why, I ask, do formally recognized races of plants and animals which are transported around the world to zoos in heterogeneous locations retain their trinomen? Whatever the case, the modern taxa subspecies / "geographic race" concept is mostly an appendage of Mayr's biological species concept. So we can simply turn to Mayr for guidance on this murky issue:

"When dealing with human races we must think of them as the inhabitants of the geographical regions in which they had originated... A human race consists of the descendants of a once isolated geographical population primarily adapted for the environment conditions of their original home country... But, as is illustrated by the success of Europeans and Africans and Asians in all parts of the world, any race is capable of living anywhere." (Mayr, 2002 "The biology of Race and the Concept of Equality")

Mayr's geographic races -- and therefore BSC ones -- hardly unconditionally necessitate geographical circumscription (they just have to had descended from geographically circumscribed populations). Which is why non-human races in zoos spread around the world are treated as they are.

But of course maybe Templeton's taxa subspecies based on Templeton's cohesion species concept are different. If so then, by his idiosyncratic concept there perhaps could be no human taxa subspecies. But who cares?
The *ideal* deme/meta-population as a unit of study (population genetics) has 100% in-group mating, 0% out-group mating (Wright's "panmictic unit").

As an operational unit, the % of in-group mating cannot deviate too much from 100% because then you do not have relatively clear-cut breeding or mating boundaries (let's say "natural" gene pools), but arbitrary spatial populations-

"If any set of organisms is legitimately called a “population,” selection and drift processes become purely arbitrary, too. Moreover, key ecological variables, such as abundance and distribution, depend on a non-arbitrary way of identifying populations." (Millstein, 2009)

"Suppose, for example, that we were to deploy the most permissive of the definitions identified above, where a population is taken to be “a group of individuals of a single species”—a definition that implies that any arbitrary set of conspecifics can be taken to be a population. The problem with this definition is that it also renders selection and drift—processes that occur within populations—arbitrary." (Millstein, 2009)

"In short, if any (gerrymandered or otherwise) set of organisms is legitimately called a “population,” our ascriptions of selection and drift are purely arbitrary. Indeed, there simply would be no fact of the matter about whether selection was operating or not." (Millstein, 2009)

Millstein is not concerned with reproduction (sexual mating) but interaction, but note how she attempts to non-arbitrarily define a population:

"[I]n this context would mean that population consists of the organisms whose rates of interaction are much higher among each other than they are among other organisms... Because the interactions aren’t necessarily reproductive, however, the continuity isn’t necessarily genealogical (as it is on some definitions of species)." (Millstein, 2009)

"If the rates of interaction with local groupings are not significantly higher than the rates of interaction among local groupings, it is a patchy population, a term coined by Harrison (1991)." (Millstein, 2010)

"Patchy populations" are just arbitrary spatial populations, not demes/meta-populations where in-group mating (for Millstein: interaction) is "significantly high[er]".
Now, the lowest % of in-group mating/reproduction I have ever seen in the literature for a deme is still 69.1% (30.9% out-group mating). This is from Templeton (2006) who gives an example of isolation-by-distance for Japan:

Marriages (sample: 2200)-

Within village: 69.1%
Neighbouring village(s): 19.1%
Outside of neighbouring villages(s) but within same district: 6.4%
Outside of district but in the same state: 2.9%
Outside of state, but in the same country: 2.5%
Outside of country: 0%

Obviously outside of country would not be 0% with more data, but it would be negligible (<1%), especially since Japan is an island or linked chain of islands.
I take issue though with that figure, and yet 69.1% is the lowest I've seen. As Millstein states, this is fuzziness but not arbitrariness. For example no one has considered a deme with only 51% in-group mating (this is far too low, and would mean in-group mating is only slightly higher than out-group).
(2015-Sep-11, 05:27:39)Chuck Wrote:
(2015-Sep-10, 18:33:31)Krom Wrote: Despite the standard scientific/textbook definition of a (gamo)deme is a more-or-less isolated breeding group/mating circle.

Why not just type "gamodeme"/"deme" and search Google books. Its not hard.


You missed the point as usual. (I wasn't discussing degree of isolation here, but location on a continuum.)

(1) Demes can be cut from an isolation by distance breeding continuum. True or false?
(2) If you want to say that these demes which fall on a breeding continuum are "sharply differentiated", then you must also say that my discretely conceptualized (non-overlapping) races cut from a continuum are likewise "sharply different".

I am fine with that. Just acknowledge the point. It's odd that you can't see the symmetry.

(1) Yes, but demes have significantly high in-group mating. This contrasts to your race concept where you cannot show a high or sharp level of (genetic) discontinuity, or ignoring a genetic threshold - lets just use in-group mating like deme. Obviously since we aren't talking about complete isolates, there is a continuum.

(2) No, because they aren't. As an example, where is the sharp mating boundary-line as a spatial/geographical barrier between "Caucasoids" and "Mongoloids"? There is no ocean, or other sharp barrier separating West and East Asia. This is why if you look in old race books, no one could agree where to place "Indians" and "Central Asians", nor even "Turks".
Your races, the continental ones at least are "patchy populations" (Harrison, 1991) meaning arbitrary spatial populations, not aggregates of demes.
 Previous 1 36 37 38 39 40 Next   
Forum Jump:

Users browsing this thread: 4 Guest(s)