OpenPsych forums

Full Version: [OBG] Nature of Race Full
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
Pages: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Wright (1978) wrote the following about Fst -

“The range 0 to 0.05 may be considered as indicating little genetic differentiation.
The range 0.05 to 0.15 indicates moderate genetic differentiation.
The range 0.15 to 0.25 indicates great genetic differentiation.
Values of FST above 0.25 indicate very great genetic differentiation.”

This is standard for taxonomy. Below I will just use % for simplification-

Lewontin (1972) reported 14.6% for blood group polymorphisms only, but this statistic combined local spatial populations with large.

You can only use the latter, i.e. for continental populations "Linnean races". Lewontin's data came out at 6.3%.

Actual Fst as a measure of overall genetic differentiation between human continental populations "Linnean races" is now under that. This is also discussed by Feldman & Lewontin (2008).

Rosenberg (2011) for example calculated 3.8% based on microsatellite loci

Betti et al. (2013) calculated Fst for post-cranial pelvis dimensions. They found only 3.0%.

The very low stats (which fall under Wright's "miniscule") are not discussed in your book.

Differences between your "Mongoloid" and "Caucasoid" are indeed trivial.
(2015-Sep-10, 01:42:59)Krom Wrote: [ -> ]Wright (1978) wrote the following about Fst -

“The range 0 to 0.05 may be considered as indicating little genetic differentiation.
The range 0.05 to 0.15 indicates moderate genetic differentiation.
The range 0.15 to 0.25 indicates great genetic differentiation.
Values of FST above 0.25 indicate very great genetic differentiation.”

The very low stats (which fall under Wright's "miniscule") are not discussed in your book.

I'll put my conclusion first, since it's obvious that you don't bother reading through:

So, for continental groups, using Mt/YDNA you get great differences. Using microsatellites and common methods such as Nei's or Weir and Cockerham's Fst, you get lower-end moderate ones. Using Jost's D or other similar measures which correct for Hs, you would get, by this scale, great differences. Using SNPs, the more relevant loci, you get upper-end moderate ones. Only by cherry picking -- AMOVA applied to microsatellites -- do you get the results you desire. How about that? And of course those results are misleading because (a) Wright discussed Fst, not AMOVA -- thus rendering problematic the use of Wright's Fst based scale, (b) using typical statistical scales, the 3.8% is out of < 30% max -- thus rendering problematic a straightforward percent comparison -- we actually are not dealing with percents, and © the whole genetic differentiation comparison is misleading, in context to discussions of quantitative trait differences, because between group differences are out of inter+intra individual ones.

As for phenotypic differences, it depends on which you look at. Some are larger than what you would expect given the SNP fst differences, indicating divergent selection e.g., Guo et al. (2014). Variation and signatures of selection on the human face. Journal of human evolution, 75, 143-152.

Some are about the same size e.g., craniometric and dental, indicating neutral variation. Some are smaller, indicating stabilizing selection.

My point would be, unless I am missing something, that the default prediction for quantitative trait differences would be ~ 2 x SNP Fst. And this is large by social scientific standards. Now I asked some, who had written on the topic, such as Henry Harpending, and they thought, from what I recall, that the above didn't sound too far off. But if you can find a flaw in the logic, let me know, I would appreciate it.

...

The longer version is:

In fact, I do discuss this issue extensively and I cite that very rule of thumb. I show both microsatellite and SNP differences -- see picks below. I discuss how differentiation values depend on both the method and especially the loci. I note that fst values are constrained by heterozygosity, thus the Nei's Fst of 0.055 found by Rosenberg et al. (not to be confused with AMOVA) is out of a maximum of 0.28. When you correct for heterozygosity using, for example Jost's D, you get a value of 0.20 (crudely 0.055/0.28). This is closer to the SNP difference, which is higher because SNPs have lower heterozygosity (around 0.30, correcting for Hs ~ 0.17). I noted:

"In short, species with high within population diversity will necessarily show low between population Fst and Fst analog values, regardless of the actual between population diversity as indexed by the number of shared alleles. (Readers are referred to Jakobsson, Edge, and Rosenberg (2013) for a discussion of this matter with regards to human populations.)...Thus, the low microsatellite Fst value between major human races is about what one would expect to find were one dealing with the subspecies of a species which had a heterozygosity value similar to that of the human species...."

And later:

"Whitlock (2008) explains the measure Qst: The calculation of QST for a trait requires two quantities: the additive genetic variance of the trait within a population (V A, within) and the genetic variance among populations (V G, among). For diploids, QST is calculated as:

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

For haploids, the same equation applies, but without the '2' in the denominator.[That '2' for the diploid case comes from the fact that the quantitative genetic variance among populations is proportional to two times FST (Wright, 1951).]


What is particularly relevant to the present discussion is Whitlock's last statement, since we are interested in predicting quantitative genetic variance from Fst, not comparing Qst to Fst. Amongst diploid populations, the predicted quantitative genetic trait variance is equal to 2Fst/(1 + Fst) (Leinonen et al., 2013). The 2 in the equation comes from the fact that roughly half of the genetic variation within diploid populations is within individuals....

Getting back to the main point, if we wish to estimate expected quantitative genetic trait variation it is often advised to avoid using low mutation rate genetic markers such as microsatellites, which, as discussed, have high Hs values and thus necessarily exhibit low Fst values. It is often advised instead to use SNPs, both because these markers do not tend to have very high Hs values and because SNP variation codes for typical quantitative trait variation (Edelaar and Björklund, 2011). Another way to look at this is to consider that the magnitude of (fixation index estimated) genetic differentiation varies by the class of loci analyzed, with part of this variation being attributable to loci variation in Hs (Jakobsson et al., 2013); for example, for humans, continental microsatellite, SNP, and mtDNA Fst values are typically around, respectively, 0.05, 0.12, and 0.20. Were one to try to infer the magnitude of genetically conditioned phenotypic variation from typical indices of fixation (e.g., Fst values), it would make sense to use the class of loci that most likely underpins the relevant trait variation. For example, since variation in single-nucleotide polymorphisms (SNPs) explains variation in many interesting polygenetic traits such as height and intelligence (for example: Yang et al., 2010; Davies et al., 2011), it would make more sense to attempt to infer magnitudes of genetic differentiation in these traits from SNP Fst values than from microsatellite or mtDNA ones." Blah blah blah ...
(2015-Sep-10, 00:58:25)Chuck Wrote: [ -> ]Would you agree that you can cut demes out of a breeding continuum? Or do you imagine that you need sharp barriers separating any two demes?

I keep encountering papers which say something to the effect of:

"In general, the variation in estimates of pairwise genetic distances explained by these effective distances is compared with that explained by geographic distances alone (that is, IBD). The latter is regarded as the most simple landscape genetic pattern that would be obtained even if there were no landscape effects and migration was thus only constrained by distance between demes (Spear et al., 2005; Balkenhol et al., 2009; Jenkins et al., 2010). This notion may have originated from spatially explicit simulation studies of IBD patterns, in which demes or individuals are usually placed in regular lattices throughout homogeneous spaces (Guillot et al., 2009; Epperson et al., 2010). Indeed, distance-constrained migration in such
models produces IBD patterns that are not influenced by any landscape elements. However, a heterogeneous landscape will not only affect migration probabilities between demes, but also the spatial arrangement of demes (that is, deme topology)"
http://www.nature.com/hdy/journal/v114/n...01462a.pdf

The authors seem to have no problem with discussing demes in context to isolation by distance i.e., a continuum... but, I imagine that you are more familiar with the literature, so I await to hear your reply.

My point was that if you are going to say that demes picked out of a continuum are necessarily sharply differentiated, you have to grant the same for races (as I define them).

If you don't wish to do the latter, you either have to say:
(a) demes can't be picked out of a continuum (interesting claim for which I would like to see support)
or
(b) demes need not be sharply different (as I originally argued)

Which will it be?
(2015-Sep-10, 03:07:49)Chuck Wrote: [ -> ]Only by cherry picking -- AMOVA applied to microsatellites -- do you get the results you desire. How about that? And of course those results are misleading because (a) Wright discussed Fst, not AMOVA -- thus rendering problematic the use of Wright's Fst based scale

I actually included a table with all published continental level micro Fst values I could find. Attached.
Of course IBD (correlation of genetic dissimilarity with geographic distance)-- each deme has minor overlap with the next in geographical space, for example 5% of a village, deme A, mate with the adjacent village, deme B. This forms a continuum. However the boundaries are not vague and there is sharp (but not complete) discontinuity. For example there is not 40% of A, mating with the adjacent deme. The gene flow is minimal. You can clearly see this for marriage distance analyses carried out in the 1960s and 70s-

"Neighbourhood Analysis: From the surname and marriage distance analyses, it seems reasonable to conclude that there was a high level of genetic continuity within a parish and that there would be some degree of isolation between parishes."
- Derek Frank Roberts, ‎Eric Sunderland (1973), Genetic Variation in Britain - Page 74

Those few marrying outside the parish, do so in the neighbouring parish.

Say there are 999,999 parishes across Europe and West Asia (West Eurasia) -- it is totally arbitrary to aggregate these demes unless you have a sharp geographical barrier to create subpopulation structures. I'm saying these are rare, and don't divide continents.

Are you aware Rosenberg et al (2002, 2005) clusters are said to only capture 1.53% inter-population genetic variation. >98% is captured by IBD.
(2015-Sep-10, 04:36:06)Krom Wrote: [ -> ]Are you aware Rosenberg et al (2002, 2005) clusters are said to only capture 1.53% inter-population genetic variation. >98% is captured by IBD.

Now, Rosenberg (2005) said:

"When an additional binary variable B is added—equaling one if an ocean, the Himalayas, or the Sahara must be crossed to travel between two populations, and zero otherwise—R2 increases to 0.729. The regression equation is Fst = 0.0032 + 0.0049D + 0.0153B, where D is distance in thousands of kilometers....The effect of a barrier is to add 0.0153 to Fst beyond the value predicted by geographic distance alone."

The 0.015 is just the effect of their crude dichotomously coded barriers; this is not the percentage of the total variability between continental races (K=5). For that you have to turn back to Rosenberg (2002). Rosenberg (2002) gives an Amova of 4.3 between regions (not 3.6 as you said earlier!) for K=5 (Blumenbach races! Remember, Linnaeus had inconstant varieties!). But what does this 4.3% actually mean? For context, try:

.....................

Meirmans, P. G., & Hedrick, P. W. (2011). Assessing population structure: FST and related measures. Molecular Ecology Resources, 11(1), 5-18.

"For biallelic markers, this makes sure that FST is bounded between zero and one, with zero representing no differentiation and one representing fixation of different alleles within populations. For multiallelic markers, however, the maximum possible value is not necessarily equal to one, but is instead determined by the amount of within-population diversity (Charlesworth 1998; Hedrick 1999). The reason for this can be best understood by looking at GST, which is defined as (Nei 1987)....For highly variable loci, this can lead to a very small possible range of GST values. To illustrate this relationship, Fig. 1 gives the joint values of FST and HS found in the past 4 years in Molecular Ecology (expanded from Heller & Siegismund 2009; see also Table S1, Supporting information). Notice that the observed range of FST is always less than HS and that the range of FST becomes very small when HS is large. For example when HS = 0.9, a value that is commonly encountered for microsatellite markers, the maximum possible value of FST is 0.1. Such a value of FST is generally interpreted as representing a rather weak population structure. However, here it represents the case with maximum differentiation among the populations, meaning that the populations do not share any alleles at all."

But see also here: Verity, R., & Nichols, R. A. (2014). What is genetic differentiation, and how should we measure it—GST, D, neither or both?. Molecular ecology, 23(17), 4216-4225.

"By tracing GST to its origin – the parameter FST, and further to the inbreeding coefficient F upon which FST was built – we can identify the root cause of some of the disagreement in the literature around the measurement of population differentiation. We have found that there are two overlapping views regarding the definition of the probability of identity by descent, which in turn have rubbed off on our definitions of FST, leading to mutation dependent and mutation-independent versions. The criticism that GST is constrained is misplaced – at least if the task at hand is to estimate the mutation-dependent version of FST. If we wish to capture other aspects of the population history, such as the mutation independent version of FST, when the mutation rate is relatively high, then we will need to supplement GST with other measures that capture a different aspect of evolution. No single statistic can be informative about both parameters in this situation, as it is mathematically impossible for a single dimension to fully represent two.... Thus no statistic is differentiation, but some statistics can be used to infer differentiation. We find that mutation-independent FST is a sensible quantity to use as our definition of differentiation, although the general arguments made above are equally valid when applied to alternative definitions... Under the same model with a high mutation rate we found that GST is insufficient on its own to jointly estimate the true level of differentiation and mutation (Figure 5). Supplementing GST with either G’ST or D solved this problem, providing a distinct source of information that can be used to pull apart the confounded signals."

And here: Should I use FST, G’ST or D?

"Over the past few decades researchers have increasingly used microsatellites, due to their high level of variability and the relative ease of development and scoring in non-model systems. However, now that next-generation sequencing is getting more affordable, sequence-based markers can be assessed throughout the genome (e.g. using RAD sequencing). As we move back towards such low-mutation-rate markers as SNPs, FST becomes easier to assess reliably. On the other hand, FST and other current methods are all designed to assess one or a few markers at a time, and genomic approaches just apply these methods thousands or tens of thousands of times for markers throughout the genome. One can look for outliers, calculate means, etc., without really taking full advantage of the data. For instance, I have seen bi-modal or skewed distributions of FST and other summary statistics; clearly means and standard deviations can be misleading in these cases. My hope is that new methods for assessing divergence will focus not on individual loci but on many markers throughout the genome....
The theoretical maximum of FST = 1 can only be reached if each subpopulation is fixed for a single unique allele. If there is variability within any subpopulation, the maximum FST is (1 – HeS). Unfortunately, this limit to the maximum FST is often overlooked. The maximum value for FST is the smaller the more variable a marker is, and the effect can be especially dramatic for microsatellites, which often exhibit high HeS (over 0.9, in which case the maximum FST is only 0.1). In the extreme (yet possible) scenario of two subpopulations completely divergent (i.e., not sharing a single allele), but both with HeS approaching 1 (i.e., all individuals are expected to be heterozygous because of high allelic diversity), FST becomes meaningless, as its theoretical maximum is then 0 (see Fig. 1 in Jost 2008 for a graphical representation)"

...................

I'm not 100% sure how Amova -- which is a F/Gst Analogue -- works, so I will just use F/Gst values, which were about 0.05 (rounding down). The clearest interpretation would be: (a) the between regional diversity accounts for 5% of the total diversity, (b) given an upper limit of 28% for these loci. © This is equivalent to a 20% mutation independent differentiation (measured in allele sharing) (e.g., Jost's D). Now claim (a) is fine so long as you understand what F/Gst is measuring and recognize its dependency on mutation rates. But (a) is not fine if you wish to make a global claim about between population differentiation and base it on high mutation rate loci. On this point, It is notable the Wright (1969) developed Fst for biallelic markers, which as Meirmans and Hedrick note actually have a range of 0 to 100. So it's not clear how well Wright's scale transfers to multiallelic markers, which are constrained by Hs. On this point, Nolan Kane notes (correctly in my opinion):

"For many situations, certainly, this can be quite problematic – for microsatellites with high heterozygosity, maximum GST is often 0.1-0.2! Clearly, in these cases Wright’s (1978) guidelines are entirely misleading, when he states that values ranging from 0-0.05 indicate “little” genetic differentiation; 0.05-0.15 is “moderate”, 0.15-0.25 is “great”, etc. This is only plausible for biallelic cases, and in other situations we cannot rely on such simple rules of thumb. (Should I use FST, G’ST or D?)"

So, to summarize, yes, Human CT microsatellite F/Gst is low to moderate. But this is expected given the high heterozygosity of the markers. (Indeed, my regression line for subspecies Hs by G/Fst, showed that the human micro G/Fst was what one would expect for a subspeciated species with the same level of Hs!) More generally, since mutation independent measures of differentiation give estimates which substantially diverge from those given by mutation dependent ones, it's difficult to interpret the situation in terms of "general" differentiation. Also, since Wright's rule of thumb was based on biallelic Fst (which has a practical range of 0 to 100), it's not clear to what extent it is applicable to multiallelic Gst, specifically when using high mutation rate microsat.

Luckily, the above is somewhat tangential to the immediate discussion, which concerns expected quantitative differences. I say luckily, since the issue above is too complex for me to easily convince you of the point, given your stubbornness on the topic.. I say that it's tangential, because unless I am mistaken, you are arguing that as (micro) Fst is low, typical quantitative differences must also be so. Conveniently for me, rules of thumb for making these inferences have been outlined. (references cited in my paper.) And they note that when doing so, when assessing the expected trait differences owing to neutral divergence (the default magnitude of differences), one should use low mutation rate markers of the type which likely underlie the genetic structure of the trait, such as SNPs, for which, amongst humans, there happen to be moderate to large differences, depending on the groups discussed. I elaborated on this in my paper -- section 4 -- and so I won't rehash all of the points made. What I said, though, definitively refutes these types of silly arguments which you are reiterating.

Your move.
“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)

Mayr (1970, p. 82): “The local population is by definition and ideally a panmictic (randomly interbreeding) unit. An actual local population will, of course, always deviate more or less from the stated ideal”.

"[I]t is a bad idea to let the self-identified social groups serve as gatekeepers in population genomics research... they cannot function as proxies for demes.” (Juengst, 1998)

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

Ethnic groups (e.g. "Germans") are mostly arbitrary spatial constructs like races, e.g. "Caucasoid". Exceptions include: the Amish, Samaritans and Kalash people.

The Amish are a meta-population made of two demes: “the Lancaster County and Pennsylvania, Amish, represent one deme, the Holmes County, Ohio, Amish, another.” (Hostetler, 1993)
Quote:So, for continental groups, using Mt/YDNA you get great differences. Using microsatellites and common methods such as Nei's or Weir and Cockerham's Fst, you get lower-end moderate ones. Using Jost's D or other similar measures which correct for Hs, you would get, by this scale, great differences. Using SNPs, the more relevant loci, you get upper-end moderate ones. Only by cherry picking -- AMOVA applied to microsatellites -- do you get the results you desire. How about that? And of course those results are misleading because (a) Wright discussed Fst, not AMOVA -- thus rendering problematic the use of Wright's Fst based scale, (b) using typical statistical scales, the 3.8% is out of < 30% max -- thus rendering problematic a straightforward percent comparison -- we actually are not dealing with percents, and © the whole genetic differentiation comparison is misleading, in context to discussions of quantitative trait differences, because between group differences are out of inter+intra individual ones.

No, you get the very low figure I posted by calculating the mean from virtually every study and using all neutral genetic/phenotypic markers. Overall it is under 10%. See Barbujani (2013) who has compiled 10 studies and calculated an overall mean: 9.0%. There are though many more markers, and the figure I calculated is under 9.0% and is closer to 5%.

Pelvic shape (males) 2.6%
Pelvic shape (females) 3.3%
http://www2.zoo.cam.ac.uk/manica/ms/2013...LoSONE.pdf

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)

Craniometrics is 14.6% as calculated by Relethford (1994, 2002).

The traits that show strong climatic selection obviously have very high Fst.

Limb ratios are around 38% Fst.
http://bhap.artsrn.ualberta.ca/images/up...S_2013.pdf

Skin colour is 87.9% Fst (Relethford, 2002) based on reflectance spectrophotometry.

Yet compare to dental non-metric from Hanihara, which is 11%.

http://onlinelibrary.wiley.com/doi/10.10...2/abstract

Metric is lower.

Dermatoglyphics (e.g. finger ridge counts) is 20%.
http://www.ncbi.nlm.nih.gov/pubmed/9074746

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.
Quote:demes need not be sharply different (as I originally argued)
[Image: picard1_normal.jpg]

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 can "argue" anything you want, but you aren't an authority to change basic definitions and then bizarrely lie claiming your re-definition is the correct definition (when a mere google search shows you are wrong).

http://www.merriam-webster.com/dictionary/gamodeme
Ignoring genetic differentiation for a moment, races are intergenerational aggregates of demes (its highly unlikely a race would be only a single deme with no subpopulation structure). So for you to show your races are based on "naturalness", you have to demonstrate the spatial (geographical) boundaries of your races are relatively clear-cut, and not vague or arbitrary.

"This definition requires that a subspecies be genetically differentiated due to barriers to genetic exchange that have persisted for long periods of time; that is, the subspecies must have historical continuity in addition to current genetic differentiation. It cannot be emphasized enough that genetic differentiation alone is insufficient to define a subspecies." (Templeton, 1998)
http://www.unl.edu/rhames/courses/curren...pleton.pdf

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. Yes, 20+ pages of debate, but you've not shown anything otherwise.
Pages: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40