Jump to content

Talk:Human skin color/Archives/2013/February

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia


Specific Populations in Genetics Section

I have just reverted a change in the Genetics section which changed the generic "African/European/Asian" alelle frequencies to instead list frequencies of specific populations. Basically there are 3 genetic scenarios for skin colour - "Arican black", "European white" and "East Asian white". All populations on Earth are either one of these genetic sets or are a mixture of them. It may be techinically more correct to list the alelle frequency of individual populations, but particularly in cases where the skin colour is a mix of the primary variations, I feel that it confuses the issue of which genes cause which skin colour without adding anything useful. Tobus2 (talk) 07:17, 15 January 2013 (UTC)

I have no idea what you are talking about and I do not know what you base your "3 genetic scenarios" model on. What about dark-skinned Australians, Melanesians, or South Asians? What about light-skinned North Africans, about San Bushmen, about the relatively darker-skinned Innuits? What about the clinal distribution of skin colour phenotypes around the world according to the intesity of UV radiation? What about the various distribution patterns of alleles of different genes influencing skin colour among different local population groups? What you have reinserted are racial generalizations and distortions of source data. FonsScientiae (talk) 07:35, 15 January 2013 (UTC)
If you don't know what I'm talking about perhaps you should read and understand some of the papers referenced in this section. Basically, modern humans left Africa with dark skin and light skin has evolved twice in separate genetic circumstances (once in Europe and once in Asia). In the scientific material I've read these are generally referred to as the African, European and Asian skin colours. All populations on earth have one of these 3 genetic sets or are the result of a mix between two or all of them. In regards to the particular populations you mention:
  • Australians and Melanesians have the same skin colour genes as sub-Saharan Africans. ("African")
  • South Asians have a mixture of African skin colour genes and European ones (starting with pure African in Sri Lanka and gradually increasing in European alleles with proximity to Europe). Skin colour variation in South Asian is a result of European admixture. ("African/European")
  • North African has significant European admixture, so similarly light skin in North African populations is from the European skin colour genes. ("African/European")
  • I have no idea what skin colour genes San and other lighter-skinned sub-saharan Africans have - if you can find any research on it please let me know. They don't have any of the "European" or "Asian" alleles AFAIK. ("African")
  • Inuits and other Native Americans are recent descendants of East Asian populations and so have the same basic skin colour genes as East Asians. ("Asian")
Tobus2 (talk) 07:54, 15 January 2013 (UTC)
What you describe is your personal point of view. First of all, there are no African or European "skin colour genes". Every human has the same genes, what differs between individuals are the alleles of the genes. Secondly, for example dark-skinned Australians and Melanesians are genetically more related to Eurasian populations than to African populations and there is no study which proves that allele frequencies of genes which determine skin colour in these populations are similar to that of African populations (who themselves probably have high variation in skin colour determining alleles). Even some dark-skinned African populations are genetically more related to light-skinned populations than to other dark-skinned Africans (e.g. Ethiopians or Somalis). From what you write I assume that you do not have much knowledge about genetics or human biological variation. Finally, even if any of your points were true - which I doubt - you cannot change source information as it is and insert your personal pov instead. FonsScientiae (talk) 08:26, 15 January 2013 (UTC)
It's not just my personal point of view, it's the point of view of every researcher I've read on the subject - eg [1], [2] and [3] all talk about African, European and East Asians as the groups with distinct skin colour genetics. Other populations are considered derived mixtures of these 3 allele sets.
Yes you are correct, "skin colour alleles" is technically more correct than the lay-term "skin colour genes". My bad.
You are wrong about Melanesians and Aboriginals. Only a very tiny fragment of the genome controls skin colour (in Europeans maybe 5 sites out of the >3 billion) - it's possible to be more than 99% "European" and still have very dark skin if the other 1% happens to include the particular alleles that result in dark skin. Light skin evolved AFTER Aborigines and Europeans went their own ways, so the shared Aboriginal/European DNA has the original African alleles for skin colour and Europeans have mutated away from this in recent times. Furthermore both the Melanesian and Aboriginal genomes have been mapped - Melanesians (Papuans) can be seen in the UCSC Genome Browser (turn on Modern Humans in the Neanderthal section) (here's the results for the SLC24A5 SNP that produces white skin in Europeans [4]) and the Aborinal genome can be downloaded here [5]. Both show the same allele as African populations in all of the sites known to be responsible for light skin in Europeans and/or Asians.
From where you stand it may seem that I don't have much knowledge about genetics of human biological variation, but from where I stand the same could be said about you. Either way it amounts to name-calling and isn't going to help us arrive at a solution.
I don't believe I am changing the source information. At worst I am interpreting it - for instance I refer to the HapMap CEU frequencies as "European", use the average of the JPT and HCB for "Asian" and the YRI for "African". I intentionally ignore admixed populations (North Africa, Latino, South Asian etc.) as they distract the reader from understanding each allele's effect and origin. This could possibly be considered to resemble wp:Synth. If this is your problem with the existing text then please say so and I will amend the text accordingly.
For the record, there are (at least) 2 errors in your revised text, firstly the ALFRED source you are using for SLC24A5 says "Jews, Ethopian" and is thus referring specifically to Ethiopian Jews, not "Ethiopians" in a general sense. Secondly you have misinterpreted the OCA2 frequencies - you quote the frequencies for the "T" allele which is the dark-skin variant, when you should be using the "C" allele which is the one the article is talking about.
As I've already stated I reverted the revised text because it introduces frequencies for admixed populations and thus confuses the effect of each gene. If you can clearly state (and if possible support with evidence!) the issues you see in the current text it will be easier for us to arrive at a solution we both agree with.
Tobus2 (talk) 10:53, 15 January 2013 (UTC)
I also suggest you read this paper Genetic Evidence for the Convergent Evolution of Light Skin in Europeans and East Asians which is a good example of the current scientific consensus on the issue. Tobus2 (talk) 13:32, 15 January 2013 (UTC)
As for the Ethiopians, please read this paper, it's not about Ethiopian Jews, but Amhara and Oromo ethnic groups who together make up the majority of the Ethiopian population.1 This source which you provide (2), using HapMap data, equates Utah residents of supposed North European ancestry with "European", Yorubas with "African", and Japanese of Tokyo and Han Chinese of Beijing with "East Asian" groupings and presents broad generalizations, while leaving out large geographic areas of the world. The study itself states "it is important to note that this simplified phylogeny and model only encompass some of the variation in human pigmentation" and "our simple model is also deficient in assuming broad population groups are uniform in terms of pigmentation mechanism and evolution". The second source you provide (3) does not define or explain how population groupings of "Europeans" and "East Asians" were achieved, or what ethnic groups they based their categorization on, but I assume that they use the same HapMap database. The same is true for your third source (4) which by way is an unreliable weblog article. You claim that "here's the results for the SLC24A5 SNP that produces white skin in Europeans 5", while the source does not even mention Europeans. The source have data on ethnic groups San, Han, French, and Yoruba; please avoid using wide generalizations to support your pov. Some genetic studies use undefined, confusing, and/or ambiguous geographical groupings.(Roberts, 2011) Better studies, like those published in Nature Genetics, require rigorous explanation why certain population groups are used and how classification was achieved.6. I would also suggest that you read this article on the conceptualization of human variation: 7.
But all these are only tangential to our original argument. What you do when you alter source material to reflect your personal pov, is against Wikipedia policies is SYNTH, OR, and NPOV. FonsScientiae (talk) 19:22, 15 January 2013 (UTC)
I agree, we are getting sidetracked. To help focus the discussion can you please answer the following two questions:
1. My reason for reverting your edit is that it introduces allele frequencies for populations with mixed African and European ancestry and this makes it harder for the reader to ascertain which alleles have which effect. I've expressed that twice previously but can't see anywhere where you've addressed that issue in your responses. Can you please tell me why you think it's important to include this information?
2. You've twice now suggested that I've altered the source to fit my personal POV but haven't provided examples of where I've done this - what parts of the text do you think is my personal POV and not reflected in the sources provided?
Although potentially off topic, I will now address the points you raised in your most recent response as they seem to indicate misconceptions on your part and these misconceptions may be the source of your issue with the text. There's no need to respond to these if you feel they are irrelevent, but please read them as it may help us reach an agreement.
  • Your comments on the 3 links I provided prove my point. You note that in each case the researchers assume an African/European/Asian model to group skin colour alleles in "broad generalizations, while leaving out large geographic areas of the world". This is exactly what I am trying to show you - the African/European/Asian model is not my "personal POV", it's the POV of scientists, researchers, academics and other experts in the field. If you look you will find many other examples of this model being used in human skin pigmentation research - it's the mainstream scientific consensus and as such should be reflected in Wikipedia.
  • The "unreliable weblog" I used as an example is by John Hawkes. He is a professor at Uni of Wisconsin and runs his own research lab doing research on Neanderthal/Sapiens DNA. The African/Asian/European model presented is what he teaches his students (the linked article is a synopsis of one of his actual classes) - hardly "unreliable" as a source to show the widespread acceptance of the model by experts in the field.
  • The results for SLC24A5 SNP I linked to does include Europeans, you even say so yourself in the list of populations you cite when saying that it doesn't. Where do you think "French" people come from?
  • Both the articles you suggested I read at the end of your comments are about race, not skin colour. Perhaps this is your problem, you think the African/European/Asian model for skin colour variation is a race-based paradigm? If this is your concern then perhaps I can rephrase the introductory paragraph to make it clearer that we are talking about the inferred geographical origin of the alleles, not a reference to a supposed race of people.
  • The link you posted in regards to the Ethiopan reference is about variable drug response and has nothing to do with skin colour or the reference you used in your edits. The reference in your edits is an ALFRED page that gives an allele frequency for "Jews, Ethiopian", described as "Unrelated Ethiopian Jews collected by Prof. B. Bonne-Tamir as part of the repository of samples on Israeli populations" in the sample details.
Tobus2 (talk) 00:13, 16 January 2013 (UTC)
While your sidetrack arguments are erroneous and unfounded, I am going to concentrate on the issue at hand.
1. You think that every population is a mixture of African and European alleles which – even for genes influencing skin colour - is fallacious. There are no “African” or “European” alleles. African, European, or any populations of large geographical areas, have variegated skin colour genotypes and phenotypes. Even if your arguments were true, the sources does not support your surmise, and altering the sources to fit your arguments – without direct evidence from the same source – would be WP:OR, WP:SYNTH.
2. You changed edits, where I directly referenced source material. 1 2 3 4 Your changes are personal POV because you seem to think along the lines of your “3 genetic scenarios” and “mixture” model which are not presented in the sources, and because you alter original source information according to this model, instead of stating sources directly. FonsScientiae (talk) 04:33, 16 January 2013 (UTC)
Thanks for replying to my questions, although you didn't really answer them. You still haven't shown me:
a) Why you think it's important to include the reverted information, and,
b) Which parts of the existing text you think are my personal POV and not in the sources
In response to your "answers":
1. I've provided 4 references that show widespread use of an "African/Asian/European" model of skin colour alleles in genetic research and academia. You have repeatedly stated that this isn't true but have not provided any evidence to support it. Can you please provide reliable and relevant references to support your claims - perhaps some recently published papers that identify skin pigmentation alleles but don't assume this 3-way model.
2. Your edits added material from the sources that is irrelevant and confusing to the topic being discussed. I haven't altered source information, I've filtered it to those parts relevant to explain the known genetics of skin colour. If you think the removed material is relevant can you please state your case for it, with references to recent on-topic research.
From your responses you seem to be hung up on the use of geographical terms to refer to the alleles - I should make it clear that "European" in this sense doesn't mean that the allele is only found in Europe, nor that all individuals in Europe have it. It simply means that the allele is believed to have originated in an ancient population that lived in what is now Europe - it's only called 'European' to differentiate it from the other known evolutionary path to light skin that scientists believe originated in ancient Asia and from the ancestral alleles that look like they originated in Africa. The "African/Asian/European" nomenclature relates to the location of the allele's origin and not to any modern human populations (although there is obviously going to be some degree of overlap). My reason for reverting you edits is that the effect of each allele can be more clearly seen if we only quote allele frequencies for populations that descend from a single one of these origins. Including populations that descend from more than one of these origins will mask the effect of the allele.
Tobus2 (talk) 06:37, 16 January 2013 (UTC)
You keep repeating your previous arguments, so I am going to be concise. You say "your edits added material from the sources that is irrelevant and confusing to the topic being discussed." It may be irrelevant and confusing to you, but who are you to decide what is irrelevant or confusing to others? "I haven't altered source information, I've filtered it to those parts relevant to explain the known genetics of skin colour." No, you removed the exact population groups used in the sources and replaced them with your own terms. That is exactly WP:SYNTH and WP:OR. Furthermore, what you describe as the "known genetics of skin colour" is your personal opinion as well as other parts of your hypothesis. Today, genetics behind skin coloration is barely understood, especially regarding different ethnic groups. None of the your shared sources claim the contrary. Please read my previous posts again and Wikipedia Policy on original research. FonsScientiae (talk) 04:55, 19 January 2013 (UTC)
Again, zero supporting evidence from you. I've already provided 4 references that show I'm correct, here's some more
So you are either wrong, or you know better than the top scientific researchers in the field... no offense, but I know who I'm going to believe.
Tobus2 (talk) 11:38, 19 January 2013 (UTC)
None of your references has to do with anything of our initial argument (and neither they prove your pov). What you do is against Wikipedia policy WP:OR. FonsScientiae (talk) 22:17, 24 January 2013 (UTC)
I disagree. Our original argument is whether allele frequency counts for admixed populations hinder or help an understanding of the known genetic components of human skin colour. All these articles show that an African/European/Asian model is the widely accepted consensus among experts in the area. Including frequency counts for populations with a mix of the skin colour alleles (such as North Africans and South Asians) is like including green and orange when trying to explain what makes yellow - ie. irrelevent and confusing. I'm happy to hear your reasons for including such populations if you feel their inclusion would benefit the article.
In regards to your accusation of original research, you should note that the current reference used for all the allele frequencies quoted in the article is the HapMap database. On it's home page this database states that it contains "DNA from populations with African, Asian, and European ancestry"[6]. I don't believe that referring to the HapMap CEU population as "Europeans" in this context amounts to original research, it is data from "Utah residents with Northern and Western European ancestry from the CEPH collection"[7] that HapMap has intentionally selected to represent European DNA. Likewise I see no problem referring to the JPT ("Japanese in Tokyo, Japan"[8]) and CHB ("Han Chinese in Beijing, China"[9]) populations as "Asian" nor with referring to YRI ("Yoruba in Ibadan, Nigeria"[10]) as "African". Within this particular context of discussing the African/Asian/European model of human skin colour genetics, these references are appropriate, relevant and consistent with 3rd party research and the source of the data.
Tobus2 (talk) 11:56, 26 January 2013 (UTC)
When the HapMap study uses the words African, Asian, and European in the link you have provided, it refers to the broad continental regions from where the ethnic samples were collected, not the genetic homogeneity of skin coloration between various ethnic groups within these continents. This, by the way, would be impossible using HapMap data, as it has only examined 11 ethnic groups and, as now, is incomprehensive for any broad conclusions.
The HapMap source found the following allele frequencies for the ASIP gene among different ethnic populations: 85% Japanese (Tokyo), 80% Gujarati Indians, 76% Italians (Tuscani), 70% Han Chinese, 9% Kenyans (Luhya in Webuye), and 6% Nigerians (Yoruba in Ibadan). (source) You altered this information to „roughly 80% in Europeans, 75% in Asians and 20–25% in Africans”.
First, if you try to make an average for the Asian ethnic groups (Japanese, Gujarati Indians, and Han Chinese) you get the same „roughly 80%” as you calculated for European populations, and not 75%. Second, by generalizing by continent, you completely ignore that allele frequencies for the ASIP gene show more similarity e.g. between Italians (Europe) and Chinese (Asia), than between Chinese (Asia) and Japanese (Asia) populations. Thirdly, you won’t get 20-25% for dark-skinned African ethnic groups (9% Luhya, 6% Yoruba), which indicates that different alleles of the ASIP gene are strongly related to skin coloration.
The same is true for the other piece of information you have altered. It found the following frequency distributions of the allele: 88% for Gujarati Indians, 87% for Han Chinese, 73% for Japanese in Tokyo, 68% for Tuscan Italians, and 24% for Luhya in Webuye (Kenya), and 15% for Yoruban in Ibadan, Nigeria. While it is true that the frequency distribution was the highest in the three Asian populations (Gujarati, Chinese, and Japanese), the changed version completely ignores that the frequency distribution is more similar between Japanese (Asia) and Italian (Europe), than between Gujarati or Chinese (Asia) and Japanese (Asia) populations.
And finally, even if your calculations were accurate, you are doing original research by generalizing from the few examined ethnic groups to big continental groupings. This is inappropriate and is against Wikipedia policies.
FonsScientiae (talk) 04:26, 13 February 2013 (UTC)
I'm not sure where you're getting your ASIP allele frequencies from, the page I linked to has different (but close) values to those you quote. "Asian" in this context means "East Asian", so you shouldn't include South Asians in the average (they are mixed-European/African in terms of skin colour alleles). Although your are technically correct in saying Japan has a higher value than China and is closer to Europeans, none of the referenced papers discuss this, instead they group Japan and China in the "East Asian" population. Making a correlation between European and Japanese frequencies would be original research.
I have reworked the section to clarify the 3-way model that is given in the research. I've replaced all of the HapMap population frequencies with frequencies reported in the papers themselves, with the exception of the ASIP allele. The only paper I found giving an allele frequency for it was confusing ("According to the National Center for Biotechnology Information (NCBI) website, there is a large difference (approximately 50%) in allele frequency of variants of rs2424984 between Blacks and non-Blacks"[11]) - they mean 50 percentage points (ie 25% to 75%) not 50% higher frequency (which would be 25% to 37.5%). In the article I've said "about 3 times more common in non-African populations than in Africa" (ie 25% to 75%) and referenced HapMap as the source. Hopefully this will appease your concerns about misrepresenting the data.
Tobus2 (talk) 11:11, 22 February 2013 (UTC)
Retrieved from "http://en.wiki.x.io/w/index.php?title=Talk:Human_skin_color/Archives/2013/February&oldid=565001368"