Wednesday, September 1, 2021

We have discussed about various tools for researching autosomes

 We have discussed about various tools for researching autosomes. One of such a tools which is widely used in Academic papers is the qpAdm. It is similar to fst in a way that it works on alleles. It produce models of populations of various degree of probability called tail prob.

While ago Davidski produced one such an example. (see links in comments) But back then populations from Arslantepe or Alalakh were not available. Recently I had a discussion in Russian Molgen and a user from Georgia produced an alternative models. Their tail prob is as low as those from Davidski nevertheless they make more sense because they rely on more proximal pops.

It must be noted that qpAdm is not easily maniable as G25/Vahaduo. Finding there a good model is quite hard. More harder is to make the tail prob higher than 0.5.

Here an alternative model


Kura_Araxes_ARM_Kaps. 14.8%

ARM_LBA 63.8%

TUR_Arslantepe_EBA. 21.4%

tail: 0.112209

chisq: 16.856

And another with lower tail prob


Kura_Araxes_ARM_Kaps. 24.3%

ARM_LBA 59.0%

TUR_Alalakh_MLBA. 16.7%

tail: 0.0843576

chisq: 17.882

And this was the Davidski's model


Mycenaeans_&_Empuries2 0.233±0.041

Kura-Araxes_Kaps 0.767±0.041

chisq 18.422

tail prob 0.142151

All those models have poor quality. It's quite obvious now that for good modeling of Armenians an Urartian sample is mandatory. But it is also clear that rumours that Arm_LBA is a dead population are premature. It seems that it will be impossible to have a good model of Armenians without Arm_LBA also. (Lchashen-Metsamor). The real number would not be as high as 60% but not lower than 25-30% also.

No comments:

Post a Comment