posted
So many studies giving so many results and yet in all the cases I have seen very little archaeological evidence and fossil evidence to back up these percentages which leads me to be believe that ascertainment bias is playing a huge role in skewed results. Its been awhile since I have posted up on this so bear with me lol
Posts: 2595 | From: Vicksburg | Registered: Feb 2006
| IP: Logged |
posted
There's 2 main bias in those population structure studies imo: 1- The SNP ascertainment bias 2- Sample bias
Sample bias is when the people sampled (often as low as 20 people, often even lower) are not representative of the population they are labelled with.
For example can 28 Afar be representative of all the Afar? Maybe those 28 particular Afar are more or less admixed than the majority of the Afar. We don't know, but we know the sample size is very small.
For example, taking only 28 people would make an horrible election day poll survey.
Another example of sample bias is when the samples of 75-5 people are taken from the same location (the same clinic). Which is often the case.
For example, Afar all taken from the capital city or trading cities, which are often more cosmopolitan and thus more admixed.
Or Afar all taken from one town or region:
Are Afar in Djibouti the same as Afar in Ethiopia?
Or are Afar in coastal Djibouti the same as Afar in Djibouti interior?
To mitigate the effect of sample bias you need larger population samples taken from many different locations.
When you see studies always look at the sample sizes, try to consider at what location(s) the samples were taken, and you must take both the SNP ascertainment bias and the sample bias into consideration.
In general, we must never forger those are estimate. The numbers are not representative of the real situation since those are only estimate. They are representative of the samples. So always allow some room in your analysis. 20% can be 10% or 35% from the "same" population in other samples. So you can't see a number as being definite.
Posts: 2981 | Registered: Jan 2012
| IP: Logged |
posted
I've discussed the SNP Ascertainment Bias in another post:
The SNP Ascertainment Bias is mentioned, among other places, in the Pagani study called "Ethiopian Genetic Diversity Reveals Linguistic Stratification and Complex Influences on the Ethiopian Gene Pool". It is also discussed more generally in the linked study in my analysis below.
quote:The Semitic-Cushitic and North African populations showed the highest values of heterozygosity worldwide, which may reflect a combination of SNP ascertainment bias and the mixture of African and non-African components in these populations.
Usually African populations have a much higher level of heterozygosity than European or North African population. The effect of SNP ascertainment bias is a consequence of the SNP discovery process where a sub-set of European population are used. This translate in bias PCA analysis and Admixture software proportion of admixture (in favor of Eurasian SNPs/clusters). This SNP ascertainment bias is common in many SNP microarray. In general, all such analysis based on a pre-selected group of SNPs/microarray must always take this SNP ascertainment bias in consideration when analyzing population admixture and affiliation. Hopefully, in the future full genome analysis could end this bias toward non-African and Eurasian SNPs. Still those "admixture" and PCAs analysis have their use, but everything must be put in perspective and any bias taken into consideration. So when you see such admixture software analysis with all the K colors always consider this.
Bottom line, we're the one who must consider the effect of the ascertainment bias and sample bias into our analysis. Never see any numbers and studies as being definitive.
Posts: 2981 | Registered: Jan 2012
| IP: Logged |
quote:Originally posted by .Charlie Bass.: So many studies giving so many results and yet in all the cases I have seen very little archaeological evidence and fossil evidence
It may because a lot of the admixture in the Horn of Africa populations is more recent than most archaeological fossils studied. Also Horners are composed of various populations, speaking different languages for example, so you must take this into account too.
We can see more historical (trading, migrations) and cultural evidences in languages and religions.
For example, Ethio-Semitic speakers (Ge'ez , Tingrinya, Amharic, Tigre, Gurage, etc) are generally more admixed than their Cushitic counterparts. The entry of ethio-semitic language into the Horn of Africa by basic logic must postdate the existence and birth of the (proto) Semitic languages in the Middle East.
While still susceptible to sample bias, Y-DNA and mtDNA analysis don't suffer from the same ascertainment bias. Although they are affected by genetic drift. So uniparental haplogroups analysis must also be taken into considerations. In general, you must take into account all autosomal SNP, STR and uniparental data and always keep in mind those are estimates who may not be representative of the whole population they are labelled with.
In general, the Horn of Africa populations, as any borderline states in the world, are admixed with neighboring population at various degrees depending on which populations in the region you talk about.
Posts: 2981 | Registered: Jan 2012
| IP: Logged |
You have too much of an American mindset when it comes to old world genetics. In general old world populations that have not been recently formed have uniform autosomal DNA.
Racial admixture does not vary from individual to individual in most Northeast African ethnicities. It varies from ethnic group to ethnic group, not from individual to individual. Wild individual variance like you see in the new world is extremely rare and unlikely.
In general, if you take just one random autosomal sample from an old world ethnic group, you get the genetic profile of +90% of its members.
Posts: 424 | Registered: May 2011
| IP: Logged |
posted
^He thinks the "true" Afar, Oromo etc. have somehow not been sampled yet and all the samples so far are due to sampling biases. As for why all "True Afar" samples would conveniently "remain at large", no coherent explanation is ever given. Not that he needs one; conspiracy theories are known to require leaps of faith. If you think Xyyman is out there, wait til you read what this guy subscribes to.
This is a "True Afar", supposedly genetically different from his somewhat lighter skinned compatriots.
Posts: 8785 | From: Discovery Channel's Mythbusters | Registered: Dec 2009
| IP: Logged |
quote:Originally posted by Swenet: ^He thinks the "true" Afar, Oromo etc. have not been sampled yet. If you think Xyyman is out there, wait til you read what this guy subscribes to.
Oromos from all over the place have been sampled in various studies.
However, Oromos are the most problematic ethnic group in the Horn as they were only recently formed (1500s-1600s onwards).
Oromos from the Western and Central highlands are autosomally the same as Amharas (see Pagani et al. 2012). This study used Oromos from the Welega and Shewa provinces (areas near Addis). Oromos from the Eastern parts are autosomally the same as Somalis (see Lazaridis et al. 2013). This study used Oromos from the Bale province.
With most other ethnic groups like Tigrays, Somalis, Amharas etc.. the autosomal profiles are more uniform.
Posts: 424 | Registered: May 2011
| IP: Logged |
quote:Originally posted by Swenet: This is a "True Afar", supposedly genetically different from his somewhat lighter skinned compatriots.
OMG...
I thought he was somewhat scientific at first, but now I have to stop taking him serious.
Phenotype variance within ethnic groups has nothing to do with autosomal profiles.
An Amhara who looks like Haile Selassie or the Ethiopian soccer coach Sewnet Bishaw can have very similar autosomal genetics, but still look quite different.
Posts: 424 | Registered: May 2011
| IP: Logged |
posted
Afars are genetically homogenous despite the phenotype diversity.
Pagani and Lazaridis used unrelated Afars from different locations and their genetic results were the same. Chances of that happening to a recently mixed ethnic group are close to zero.
Posts: 424 | Registered: May 2011
| IP: Logged |
quote:Originally posted by Manu: Afars are genetically homogenous despite the phenotype diversity.
Pagani and Lazaridis used unrelated Afars from different locations and their genetic results were the same. Chances of that happening to a recently mixed ethnic group are close to zero.
Are you saying that therefore the Afar according to Pickrell are basically across the board 46% West Eurasian?
Ancient west Eurasian ancestry in southern and eastern Africa Joseph K. Pickrell et al. 2014
Posts: 42918 | From: , | Registered: Jan 2010
| IP: Logged |
quote:Originally posted by the lioness,: Are you saying that therefore the Afar according to Pickrell are basically across the board 46% West Eurasian?
Yes, he used genuine Afars.
The West Eurasian DNA in the Horn is very old.
Up to 23,000 years possible (that is before humans lived in the Americas).
posted
^That's where we differ. The autosomal component discussed in that link is equidistant between Eurasians and Africans. That doesn't gel with a wholesale origin in West Asia.
quote:Originally posted by Manu: An Amhara who looks like Haile Selassie or the Ethiopian soccer coach Sewnet Bishaw can have very similar autosomal genetics, but still look quite different.
Indeed. His "true Afar" obviously are included in the autosomal profiles he's in denial about because per Pagani et al 2012, 75% of their Afar sample have the ancestral SLC24A5 variant, which, according to various sources, accounts for ~30% skin colour differences relative to individuals who don't have it. In other words, these dark skinned Afar ARE a subset of Pagani's Afar, they are not "missing" from these samples.
When you get down to the genetic material of these populations, darker skin in Lowland East Cushitic speakers simply means MORE of the same Eastern Sahara component, expressed as E-M35 and various L3 clades uniparentally or as what Hodgson et al 2014 call the "Ethio-Somali" component. Somalis, Rendille and Borana are as dark-skinned as these people get and they don't have more West-African ancestry per se than lighter skinned Ethiopians, like the Amhara.
The same goes for non-Cushitic speaking Ethiopian populations like the Hamer, who once were written off as "Bantus" due to their prevalence of what has been stereotyped as African features.
Posts: 8785 | From: Discovery Channel's Mythbusters | Registered: Dec 2009
| IP: Logged |
quote: archaeological evidence is not completely absent. During this time period, architecture in the Ethiopian culture of D’mt has an “unmistakable South Arabian appearance in many details” (19), although there is some debate as to whether these patterns can be attributed to large movements of people versus elite-driven cultural practices (19, 20). Additionally, linguistic evidence suggests that this time period was when Ethiosemitic languages were introduced to Africa, presumably from southern Arabia (21). It is perhaps not a coincidence that the highest levels of west Eurasian ancestry in eastern Africa are found in the Amhara and Tygray, who speak Ethiosemitic languages and live in what was previously the territory of D’mt and the later kingdom of Aksum.
Ancient west Eurasian ancestry in southern and eastern Africa Joseph K. Pickrell et al. 2014
The Tigray, 50.4% West Eurasian according to Pickrell (photo Rod Washington flickr)
Posts: 42918 | From: , | Registered: Jan 2010
| IP: Logged |
quote:Originally posted by Swenet: Indeed. His "true Afar" obviously are included in the autosomal profiles he's in denial about because per Pagani et al 2012, 75% of their Afar sample have the ancestral SLC24A5 variant, which, according to various sources, accounts for ~30% skin colour differences relative to individuals who don't have it. In other words, these dark skinned Afar ARE a subset of Pagani's Afar, they are not "missing" from these samples.
When you get down to the genetic material of these populations, darker skin in East Lowland Cushitic speakers simply means MORE of the same Eastern Sahara component, expressed as E-M35 and various L3 clades uniparentally or as what Hodgson et al 2014 call the "Ethio-Somali" component. Somalis, Rendille and Borana are as dark-skinned as these people get and they don't have more West-African ancestry per se than lighter skinned Ethiopians, like the Amhara. The same goes non-Cushitic speaking Ethiopian populations like the Hamer, who once were written off as "Bantus" due to their prevalence of what has been stereotyped as African features.
I don't think that SLC24A5 is the only gene that controls for lighter skin.
SLC24A5 frequencies (rs1426654-A allele freq)
Wolayta 62.5% Amhara 57.7% Tygray 57.1% Somali 55.0% Oromo 38.1% Afar 25.0% Ari cultivator 22.9% Ari blacksmith 11.8% Gumuz 7.9% South Sudanese 2.1% Anuak 0.0%
Table S7. - Pagani et al. 2012
The Wolayta and Amhara should be the lightest, and South Sudanese and Anuak the darkest.
The Afars seem to have the lowest levels of this ''light skin'' gene out of the Semitic-Cushitic groups. Despite this, they are more Eurasian than the Wolayta who have it at a higher level than them. So, I have my doubts on the accuracy of this gene.
Posts: 424 | Registered: May 2011
| IP: Logged |
quote:Originally posted by Manu: I don't think that SLC24A5 is the only gene that controls for lighter skin.
Absolutely agree
quote:Originally posted by Manu: The Afars seem to have the lowest levels of this ''light skin'' gene out of the Semitic-Cushitic groups. Despite this, they are more Eurasian than the Wolayta who have it at a higher level than them. So, I have my doubts on the accuracy of this gene.
Read Pagani. It's under positive selection in Ethiopia for whatever reason. Therefore, it doesn't have to correlate neatly with more "West Eurasian" ancestry. Most of the West Asian ancestry is very ancient as you said earlier (e.g. T-PS21), so the derived SLC24A5 percentages are bogus anyway (derived SLC24A5 is not that old that it would reflect old West Asian admixture in Ethiopia).
Posts: 8785 | From: Discovery Channel's Mythbusters | Registered: Dec 2009
| IP: Logged |
I have to say that you seem to be the most knowledgeable on African genetics relative to the other posters here.
Posts: 424 | Registered: May 2011
| IP: Logged |
posted
Can someone help me out here? I am new to genetics and I need a pat on my back... or "at a boy" stroking.
When you say West Eurasian you are talking Europe vs Arabia cf Ethiopia? I have never been to Africa so I have no idea what is an Afra vs an Oromo.
Quote: Read Pagani. It's under positive selection in Ethiopia for whatever reason. Therefore, it doesn't have to correlate neatly with more "West Eurasian" ancestry.Posts: 12143 | From: When you have eliminated the impossible, whatever remains, however improbable | Registered: Jun 2007
| IP: Logged |
posted
Eurasians are split up in 3 broad ''races''
West Eurasians (also known as ''Caucasoids''). East Eurasians (also known as ''Mongoloids''). South Eurasians (also known as ''Australoids'').
The Eurasian ancestry in Northeast Africa is only affiliated with West Eurasians.
Although haplogroup M1 is still a mystery... It may have been spread by people who were Australoid, but Australoid autosomal admixture is not found in Northeast Africa.
Posts: 424 | Registered: May 2011
| IP: Logged |
posted
Sarcasm ?? The English Language.....forget it
-------------------- Without data you are just another person with an opinion - Deming Posts: 12143 | From: When you have eliminated the impossible, whatever remains, however improbable | Registered: Jun 2007
| IP: Logged |
Who knows? You should stick around. Readers need a different viewpoint from what they've been spoon-fed here. Lots of leaving out and covering up inconvenient data. I'm out for now. Good luck for when Amun comes back and the floodgates of pseudo-science will open up.
Posts: 8785 | From: Discovery Channel's Mythbusters | Registered: Dec 2009
| IP: Logged |
quote: archaeological evidence is not completely absent. During this time period, architecture in the Ethiopian culture of D’mt has an “unmistakable South Arabian appearance in many details” (19), although there is some debate as to whether these patterns can be attributed to large movements of people versus elite-driven cultural practices (19, 20). Additionally, linguistic evidence suggests that this time period was when Ethiosemitic languages were introduced to Africa, presumably from southern Arabia (21). It is perhaps not a coincidence that the highest levels of west Eurasian ancestry in eastern Africa are found in the Amhara and Tygray, who speak Ethiosemitic languages and live in what was previously the territory of D’mt and the later kingdom of Aksum.
Ancient west Eurasian ancestry in southern and eastern Africa Joseph K. Pickrell et al. 2014
The Tigray, 50.4% West Eurasian according to Pickrell (photo Rod Washington flickr)
This is where I draw the line, there is evidence for an influence from ancient Ethiopia into Arabia as well, teh influences went both ways, and yet the text assumes all geneflow was one way from Arabia to Africa.
Posts: 2595 | From: Vicksburg | Registered: Feb 2006
| IP: Logged |
quote:Originally posted by .Charlie Bass.: This is where I draw the line, there is evidence for an influence from ancient Ethiopia into Arabia as well, teh influences went both ways, and yet the text assumes all geneflow was one way from Arabia to Africa.
@Charlie,
The majority of the West Eurasian ancestry in the Horn of Africa is from areas near Egypt and the Levant.
Not Arabia, only a small portion of the recent Semitic ancestry is (less than 10%).
Most of it is from Egypt/Fertile Crescent according to Pagani and Hodgson .
Posts: 424 | Registered: May 2011
| IP: Logged |
posted
^^ DNAtribes are not academic. They are a company.
Actual studies show that Horn Africans have more ancient ancestry from Egypt and the Med basin than from Arabia.
Posts: 424 | Registered: May 2011
| IP: Logged |
quote:Originally posted by .Charlie Bass.: This is where I draw the line, there is evidence for an influence from ancient Ethiopia into Arabia as well, teh influences went both ways, and yet the text assumes all geneflow was one way from Arabia to Africa.
@Charlie,
The majority of the West Eurasian ancestry in the Horn of Africa is from areas near Egypt and the Levant.
Not Arabia, only a small portion of the recent Semitic ancestry is (less than 10%).
Most of it is from Egypt/Fertile Crescent according to Pagani and Hodgson .
How does that make sense. Egypt is not West Eurasia
""areas near Egypt and the Levant."
^^ name these areas please
Posts: 42918 | From: , | Registered: Jan 2010
| IP: Logged |
quote:Originally posted by the lioness,: How does that make sense. Egypt is not West Eurasia
Egyptians are racially akin to West Eurasians. Including their ancient counterparts.
Oh boy... this is not going to fly here.
The genetic difference between Egyptians and Palestinians is very small.
Egyptians are genetically much closer to Syrians than to the Habesha, Beja etc..
Did DNA Tribes Amarna or Rameses III reports have an ulterior motive? Weren't they using JAMA data?
The other DNA testing company DNA consultants? They cited a strongest correlation between Sudanese Copt DNA and ancient Egyptians and Sudanses Copts are recent coverts
Are these companies all wrong?
Posts: 42918 | From: , | Registered: Jan 2010
| IP: Logged |
posted
There have never been any proper autosomal studies on Ancient Egyptians.
Just some vague haplogroup (lineages) stuff that is not as concrete as autosomal full genome DNA.
My bet is that once full genome analysis is performed on Ancient Egyptians they will be predominantly West Eurasian, and less Sub-Saharan than modern Egyptians.
Posts: 424 | Registered: May 2011
| IP: Logged |
quote:Originally posted by the lioness,: Are you saying that therefore the Afar according to Pickrell are basically across the board 46% West Eurasian?
Yes, he used genuine Afars.
Pickrell used the same samples of 12 Afar from the Pagani study.
The sample of 12 Afar in the Pagani study were taken from the Wag Hemra zone which is a Amhara region of Ethiopia and Amharic are Semitic speakers. If those 12 Afar are from the Ahmara region, there's a good chance they are admixed to some higher degree with Amhara people. Possibly making them not perfect representative of Afar from other countries and regions.
It's the same for any such small samples size.
The location of Afar samples in the Pagani and Pickrell study can be seen in Table S1 of the Supplemental Data of the Pagani study (Ethiopian Genetic Diversity Reveals Linguistic Stratification and Complex Influences on the Ethiopian Gene Pool).
In the Supplemental Data you can see the sample of 12 Afar were taken from the Wag Hemra zone in Ethiopia not the Afar region
Ethiopian Jews were from Isreal. Somalis were from Garissa, Kenya. Oromo were from Southeast Ethiopia.
Posts: 424 | Registered: May 2011
| IP: Logged |
posted
^^^The Lazaridis study didn't have the "same results" and the study only used a sample of 5 Afar (iirc).
Posts: 2981 | Registered: Jan 2012
| IP: Logged |
quote:Originally posted by Amun-Ra The Ultimate: ^^^The Lazaridis study didn't have the "same results" and the study only used a sample of 5 Afar (iirc).
They did have the same results.
They were slightly more Eurasian than the Somalis and on par in ''Eurasianess'' as the Ethiopian Jews or Habesha in general.
Same results.
Stop denying the truth.
Posts: 424 | Registered: May 2011
| IP: Logged |
quote:Originally posted by the lioness,: Amun-Ra, why did you bring up the Afar in the first place in the other thread?
It was to show the phenotypic diversity of East African people and African people in general.
Some people try to make Horn Africans some sort of Caucasian/hamitic race but while many are highly admixed with Eurasians like Semitic speakers in Ethiopia, others are much less admixed.
Posts: 2981 | Registered: Jan 2012
| IP: Logged |
quote:Originally posted by Amun-Ra The Ultimate: It was to show the phenotypic diversity of East African people and African people in general.
Some people try to make Horn Africans some sort of Caucasian/hamitic race but while many are highly admixed with Eurasians like Semitic speakers in Ethiopia, others are much less admixed.
The difference in Eurasian ancestry is at most 5%-15% between Semitic and Cushitic Ethiopians
Not a huge difference. They all plot near each other. The admixture predates Semites.
Only when you throw in the Gumuz and tribal Omotics does the difference gets substantial.
Posts: 424 | Registered: May 2011
| IP: Logged |
quote:Originally posted by Amun-Ra The Ultimate: ^^^The Lazaridis study didn't have the "same results" and the study only used a sample of 5 Afar (iirc).
They did have the same results.
They were not the same.
The proportion of Eurasian admixtures in the 2 studies were not the same.
The Lazaridis had them less admixed with Eurasian (maybe around 25%-30% at K=20).
Posts: 2981 | Registered: Jan 2012
| IP: Logged |
posted
^^^Unfortunately you can't just pick and choose the K you like the most. That's not how it works.
Anyway the Lazaridis study only use 5 Afar, which can't constitute the basis for a good population structure analysis.
Posts: 2981 | Registered: Jan 2012
| IP: Logged |
posted
My point still stands. Between the 2 studies only 17 Afar were tested. Not enough to be considered representative of Afar as a whole.
This is a graph of various Y-DNA in Ethiopian populations:
No populations has the same population structure. For example some population have larger Eurasian J Y-DNA component than others. While other populations have a much larger African E Y-DNA component than others.
Admittedly the Eurasian component is mostly female mediated in Eastern Africa.
Posts: 2981 | Registered: Jan 2012
| IP: Logged |
UBBFriend: Email this page to someone!
, supposedly genetically different
Printer-friendly view of this topic