Various reconstructions of ancient people based on available SNPs

Funny reconstructions indeed!
maybe a respectable number of SNP's but on what pop's sample, and how many by individual?
I'm astonished by the HG's as a whole and by CucuteniTripolye
 
The most ones are merges of one or more sites of one archaeological culture.
Some samples are not merges, the older ones like Cheddar Man or Kostenki. Salkhit is also only one individual.
Most ones must be merged, because single individuals do not contain enough SNPs.
They represent only one individual or one merge of many people from one culture, therefore is should be clear, that individuals can also look different, because people generally not look all the same.

For example i have now worked on samples that are labeled as from the Hittite era and others are labeled as from the Hattian era.
It seems they look different, the Hattians more middle eastern, the Hittites more EHG/Caucasus. But this can only be a coincidence and the region was generally more diverse in terms of population.

Kura Araxes



Devils Gate



Hittite



Hattian



Hun



Medieval Turk from Kazakhstan




The last ones for now.

And one thing should also be noticed: It is based on the available SNPs. The results for specific traits can/will be different, if more ones would be available.
 
20,000-Year-Old Link Between Brazil’s Indigenous People And Ancient Australians

University of São Paulo … Australian DNA is even more widespread in Indigenous South Americans than originally thought …

https://allthatsinteresting.com/ancient-australian-dna-south-america

Thank you very much for the article, Salento. So the Australasian DNA in the Brazilian indigenous is much more ancestral than one thinks and came from Beringia with the founding population. This information is very interesting.(y)
 
For people who like to store anthro stuff like me I made a new version with hair and beards, some faces reworked: https://i.ibb.co/MkrF56w/Collectors-Edition-All-Faces-Text-Beards-Hair.png

The female hair shape is based on the SNPs, but i should mention that predicting hair shape is not easy and not all SNPs are known today, this is much more complicated than eye or skin color. The SNPs that are causing straight hair in one ethnicity, doesn't do it always in another.
Generally there is no way to separate Afro Hair that is rare in Europeans today, from curly hair.
Straight Asian Hair can be predicted through the EDAR variant associated with it, but this is not 100%, if other ones for curl are present.
I used for Afro + Curly hair the SNP combination of the San and Pygmy, for Asian straight hair the EDAR variant, for Wavy my own and for straight the one of my Wife, because she has straight hair.

All faces in one file without hair and beard: https://i.ibb.co/v37zJ04/Collectors-Edition-All-Faces.png
 
Don't know why, but i cannot change the link to a newer version in the old thread, so new links to updated version here:

With hair and beards: https://i.ibb.co/tMYQD22/Collectors-Edition-All-Faces-Text-Beards-Hair3.png

Without Hair: https://i.ibb.co/ChspvFm/Collectors-Edition-All-Faces3.png

Funny reconstructions indeed!
maybe a respectable number of SNP's but on what pop's sample, and how many by individual?
I'm astonished by the HG's as a whole and by CucuteniTripolye

I guess the nice thing about doggerlands excersition there are more than 100 SNP's but in every case the same so a kind of level playing field. May be still not 100% but nevertheless.

I did some comparison with the ancestral populations.

With the 'super' (hg, neolithic, steppe):



And within those three ancestral populations the 'top' ones in my case Srubnaya, Michelsberg and Cheddar man.
https www google nl afbeeldingen

In bot cases my resemblance with 'super' steppe (85%) and Srubnaya even 91% resemblance!

In my region Single Grave Culture was most prominent, they came in about 2850 BC. And may be coincidence or not but there is may a connection with the (ancestral) population of Srubnaya.

We got this from SGC in Denmark, Carlos Quilles:

Gjerrild 5 (2284-2035 calBC), like the earlier Gjerrild 8 (2575-2345 BC), was a child (Infans I, roughly 2 years old), and both show an isotopic signature different from the older skeletons, clearly over their previously defined baseline for Denmark in Frei et al. (2019), suggesting that they were both recent migrants. The Y-chromosome haplogroup (see below) and radiocarbon date of Gjerrild 5 are clearly in line with the expected intrusion of Yamnaya-derived Jutish Beakers.

And the connection with Srubnaya:

Without finding full-fledged R-V1636 samples in Yamnaya, we cannot say where the parent hg. R-V1636 (ancestor of all modern descendants) diversified, since there are important regions around the Caucasus without a proper geographical and chronological sampling, but I’d say that the steppes to the north of it are the most likely origin. If the Srubnaya individual I0233 from the Caspian-Aral region, of a possible R-V1636 subclade, is confirmed with other samples, a Steppe origin will be all but proven.

Isn't it the case that the Steppe phenotypes are in fact quit dominant in the modern Northern European populations?. According to Maciamo these were the phenotypes among Yamna (more or less congruent with Steppe?):

According Kurts (1984, p.90), people of the Yamna culture consisted of three distinct phenotypes corresponding to the relatively recent blend between three populations. The dominant type was tall, dolichocephalic, with broad faces of medium height. The second type was more robust with high and wide Proto-Europoid faces. The third type was more gracile, with narrow and high faces of the East Mediterranean type.

I belong tp the second type. But I guess most of the (Northern) Europeans belong to one of those three categories. Correct me If I'm wrong.
 
Everybody belongs to one of three facial types, if one will sort them in this concept ;)

Betaface has some disadvantages like many face recognition engines. This includes not guessing ethnicity and gender of people right, using beards and hair styles as similarity indicator, or facial expression. It is good at comparing individual traits like nose and facial bone shape, but is far from being able to say if someone is Finnish or Basque. It also can do strange guesses by crossing the ethnic barrier. For example someone with a very board nose can be rated similar to Africans, or fat faced people and Trisomy21 can be rated as East Asian, because of their small eyes.

Facial recognition software is programmed by people, so it will reflect their own ratings and beliefs.

The same problem occurs with the older system of metrics.
Once there was a metric calculator for physical skull shape by Dienikes. It was able to calculate your skull shape in the old terms of Coons racial classification by measuring many facial variables. I measured many people in my social environment and they all came out different.

After all it stays a point of view which traits one or the machine/program is interpreting as more important as the other.

But there are many other things to consider why measuring ethnicity/race is so complicated. For example some alleles are dominant over others and some are recessive.

Lets imagine you would have two parents of completely different ethnic background. Many trait alleles would be heterozygote AG or TC for example. A simple PCA tells you that you would have to look 50% 50% like each parent. But in reality you don’t, because some traits are recessive and wont show up on your phenotype at all.

People want to know what ancient phenotype they have. But this is much more complicated then tracing ancestry. You can look more similar to a person that is not directly related to you, than to you own grand parents. But a machine can easily tell you today that they are your grand parents and not the person who looks more similar to you, but is not directly related to you.

Different people would rate the same person may not as the same ethnicity, when it comes to continental classifications.
But there are also people who cannot see a difference between a Papuan and a Nigerian, or cannot see difference in a Japanese and a Mongolian.

Different ratings of similarity are also present in other things:

https://www.eupedia.com/forum/threads/35849-Is-your-brain-more-Western-or-more-Eastern-part-1
 
@Doggerland correct me please if I'm wrong.

But imo this is the essence what you have done.

Collect more than 100 SNP's that are involved in the facial phenotype. Based on that you made a facial reconstruction. Those facial reconstructions can be compared.

Ok those 100 SNP's will not be enough for a 100% reconstruction nevertheless I'm convinced that the foundation -compare with SNP's- is quit 'objective'.

I guess this gives the opportunity to compare the faces in a better way than the old classifiers. I guess they were basically more biased.
 
@Duarte … I watched a documentary where they DNA tested a Brazilian tribes, … they carry some extra Aboriginal Australians Ancestry, … or so they say :)

according to pics, some tribes show these ties through their phenotype too. It was noticed by old anthropologists.
 

This thread has been viewed 7109 times.

Back
Top