Kerala Y Dna and miscellaneous

[Slayeroid]

Someone recently posted the qpAdm runs for 3 Nairs .

These remind me of models banned trolls on AG would use.

Afaik ,  IVC and Kurumba are 2 pops on the same cline so how can they be utilized as a left group?? Thats like me utilizing  Sintashta and Poltavka on the left pops. In Arjun Bidanda's paper which was focused on South Indian populations, these kind of models failed, infact and he used tonnes of samples .  Frankly I would imagine a Gujarati like Brahmin type source bringing Steppe related ancestry rather than straight up Andronovo.  

The old admix tools format was much nicer looking , looks like something off an 80s computer.

Pls give me a passing model of South Asians on qpAdm  using a  IVC like pop , Steppe and Bmac. All these simplistic modellings using a Tribal like pop , an IVC like pop and Steppe work perfectly fine and pass on qpAdm.

[vishankar]

Before I forget to pen my thoughts here....why cant the following be done with qp/adm?

model the nairs..as a two way mix....between

1) nambudiri brahmans and vellalas.....instead of gujarati brahmins and piramalai kallars?
2) see whether the nambudiri is BMAC influenced?( i suspect they are)....some user did mention that on AG
3) nambudiri brahmns and ezhavas.....(historically unlikely but a proto ezhava population could have admixed with nambudiris.


the first hypotheses is historically plausible.....if we feel a proto-vellala population admixed with nambudiris.....


Please pardon my ignorance of QP/ADM.....why cant we use those above mentioned populations???.....s it because they are not in David Reich's dataset???

[vishankar]

back to some more Y-DNA reports I got.....

1) nambiar( kannur nair nambiar)-R-Z2124.
2)Nair(Kollam)-R-Y6.
3) aNTHUR NAIR( NAIRS WHO ARE POTTERS BY OCCUPATION-J-M241.

[ESPLover]

this is a nair from kollam....his GEDMATCH oracle 3 population results.

nair 3pops

Regarding the need for BMAC or not .....the presence of Punjabi Arain n this three pop mix is interesting...the Arain are possibly a BMAC heavy population if i am not mistaken?

sample: MeeboMian:Arain Average 25
distance: 1.1848
8728/1459: 50
Gonur_BA_1: 30.5
Russia_Srubnaya_Alakul: 19.5

Yeah, they have quite a bit of BMAC-related admixture

[Slayeroid]

Before I forget to pen my thoughts here....why cant the following be done with qp/adm?

model the nairs..as a two way mix....between

1) nambudiri brahmans and  vellalas.....instead of gujarati brahmins and piramalai kallars?
2) see  whether the nambudiri is BMAC influenced?( i suspect they are)....some user did mention that on AG
3) nambudiri brahmns  and ezhavas.....(historically unlikely but a proto ezhava population could have admixed with nambudiris.


the first hypotheses is historically plausible.....if we feel a proto-vellala population admixed with nambudiris.....


Please pardon my ignorance of QP/ADM.....why cant we use those above mentioned populations???.....s it because they are  not in  David Reich's dataset???

This admixture of   Nairs using Guju Brahmin and Piramalai was done using g25 .A  lot of Nairs can be modelled as a mix between  a SI Indian landlord community plus  SI Brahmins , which is what essentially West coast  Landlord communities like Bunts and Kodavas score like.Offcourse Nairs have variation tho since many are fully Landlord like and many score identical to SI Brahmins.

[ESPLover]

Before I forget to pen my thoughts here....why cant the following be done with qp/adm?

model the nairs..as a two way mix....between

1) nambudiri brahmans and  vellalas.....instead of gujarati brahmins and piramalai kallars?
2) see  whether the nambudiri is BMAC influenced?( i suspect they are)....some user did mention that on AG
3) nambudiri brahmns  and ezhavas.....(historically unlikely but a proto ezhava population could have admixed with nambudiris.


the first hypotheses is historically plausible.....if we feel a proto-vellala population admixed with nambudiris.....


Please pardon my ignorance of QP/ADM.....why cant we use those above mentioned populations???.....s it because they are  not in  David Reich's dataset???

Yeah, Vellalar, Ezhavas and Nambudiri Brahmins aren't in the dataset.

On top of that, the standard errors will likely be quite high, due to these overlapping components between these groups

[Slayeroid]

Before I forget to pen my thoughts here....why cant the following be done with qp/adm?

model the nairs..as a two way mix....between

1) nambudiri brahmans and  vellalas.....instead of gujarati brahmins and piramalai kallars?
2) see  whether the nambudiri is BMAC influenced?( i suspect they are)....some user did mention that on AG
3) nambudiri brahmns  and ezhavas.....(historically unlikely but a proto ezhava population could have admixed with nambudiris.


the first hypotheses is historically plausible.....if we feel a proto-vellala population admixed with nambudiris.....


Please pardon my ignorance of QP/ADM.....why cant we use those above mentioned populations???.....s it because they are  not in  David Reich's dataset???

Namboothiris are prolly identical to other SI Brahmins . Even if they have BMAC ( not confirmed ) ,it is not going to be significantly more when compared to other SI Brahmins.I have seen 100 + Si Brahmin kits and all of them for sone reason have the same variation.Some homogenisation seems to have across these Si Brahmin subsects and regions.

[vishankar]

Before I forget to pen my thoughts here....why cant the following be done with qp/adm?

model the nairs..as a two way mix....between

1) nambudiri brahmans and  vellalas.....instead of gujarati brahmins and piramalai kallars?
2) see  whether the nambudiri is BMAC influenced?( i suspect they are)....some user did mention that on AG
3) nambudiri brahmns  and ezhavas.....(historically unlikely but a proto ezhava population could have admixed with nambudiris.


the first hypotheses is historically plausible.....if we feel a proto-vellala population admixed with nambudiris.....


Please pardon my ignorance of QP/ADM.....why cant we use those above mentioned populations???.....s it because they are  not in  David Reich's dataset???

Yeah, Vellalar, Ezhavas and Nambudiri Brahmins aren't in the dataset.

On top of that, the standard errors will likely be quite high, due to these overlapping components between these groups

ok...LOL...IS there anyway we could contribute  to the data set???( wild thinking!))))..drop in an email to Reich))))?...seriously despite the possibility of high standard error....the exclusive need for the reich dataset...limits the results!
Thangaraj et al....in their paper on South West India
forgot about nambudiri brahmins ,ambalavasis etc....as well as AASI enriched Pulayas etc.....trying to decode kerala genetics without those groups is incomplete.

[vishankar]

Before I forget to pen my thoughts here....why cant the following be done with qp/adm?

model the nairs..as a two way mix....between

1) nambudiri brahmans and  vellalas.....instead of gujarati brahmins and piramalai kallars?
2) see  whether the nambudiri is BMAC influenced?( i suspect they are)....some user did mention that on AG
3) nambudiri brahmns  and ezhavas.....(historically unlikely but a proto ezhava population could have admixed with nambudiris.


the first hypotheses is historically plausible.....if we feel a proto-vellala population admixed with nambudiris.....


Please pardon my ignorance of QP/ADM.....why cant we use those above mentioned populations???.....s it because they are  not in  David Reich's dataset???

Namboothiris are prolly identical to other SI Brahmins . Even if they have BMAC ( not confirmed ) ,it is not going to be significantly more when compared to other SI Brahmins.I have seen 100 + Si Brahmin kits and all of them for sone reason have the same variation.Some homogenisation seems to have across these Si Brahmin subsects and regions.

Yes..your kONKANI Brahmin sample seems to reinforce that...in fact we can add Chitpawan brahmins to the SI brahmin genetic type...

[ESPLover]

Before I forget to pen my thoughts here....why cant the following be done with qp/adm?

model the nairs..as a two way mix....between

1) nambudiri brahmans and  vellalas.....instead of gujarati brahmins and piramalai kallars?
2) see  whether the nambudiri is BMAC influenced?( i suspect they are)....some user did mention that on AG
3) nambudiri brahmns  and ezhavas.....(historically unlikely but a proto ezhava population could have admixed with nambudiris.


the first hypotheses is historically plausible.....if we feel a proto-vellala population admixed with nambudiris.....


Please pardon my ignorance of QP/ADM.....why cant we use those above mentioned populations???.....s it because they are  not in  David Reich's dataset???

Yeah, Vellalar, Ezhavas and Nambudiri Brahmins aren't in the dataset.

On top of that, the standard errors will likely be quite high, due to these overlapping components between these groups

ok...LOL...IS there anyway we could contribute  to the data set???( wild thinking!))))..drop in an email to Reich))))?...seriously despite the possibility of high standard error....the exclusive need for the reich dataset...limits the results!
Thangaraj et al....in their paper on South West India
forgot about nambudiri brahmins ,ambalavasis etc....as well as AASI enriched Pulayas etc.....trying to decode kerala genetics without those groups is incomplete.

Nair samples can be added; the raw data can be converted to plink format and then to eigenstrat which can be added into the Reich dataset, tho the SNP overlap that commercial dna samples have isn't great with the 1240k dataset. That being said, if there are enough samples from raw data, that will compensate for the low overlap.

There's ofc high quality academic samples from studies, someone with the knowledge to do so can convert those into eigenstrat format.

I guess you could try contacting Reich, but idk how helpful that will be.

But yeah, I agree, South Asian population is insanely under-sampled, which is a massive shame considering how many diverse groups there are in the subcontinent.

[parasar]

This paper (peer review not complete) hints at another possibility - that some populations had a different Iran-N source than from the InPe.
https://www.biorxiv.org/content/10.1101/...3.full.pdf

"ancestral Koraga diverged from the Andamanese ancestor carrying 66% of this ancestry (Figure 2 b).
The remaining 34% was derived from the ancestor of the 10,000-year-old Neolithic sample from Ganj Dareh, in the Zagros mountains of what today is western Iran. Notably, the Indus Periphery samples did not share any ancestry with the Koraga tribe ....

To corroborate this key finding, we included Paniya, a hunter-gatherer tribe, and Kuruba, a pastoralist tribe, in the model. Whilst Paniya formed a separate branch (Figure 2 e), the pastoralist tribe Kuruba turn out to have received their ancestral components from an Andamanese-ancestor-related population and the ancestor of the Indus Periphery samples (Figure 2 c, d, e). Koraga, on the other hand, remains unrelated to the Indus Periphery, although at a deeper level the Koraga ancestor shared 39% of his ancestry with the Indus Periphery lineage (Figure 2 d) ...

A recent study suggests that archaic DNA (Sarazm_EN) dating from the 4th millennium BC from what today is Tajikistan as the best proxy for Iranian plateau farmer-related ancestry 7 . We do not find any direct ancestry sharing between our hypothetical Proto-Dravidian and Sarazm_EN or with the Indus Periphery cline. However, the f3-statistics for our hypothetical Proto-Dravidian are similar to those of Sarazm_EN, the Indus Periphery cline as well as the ancient DNA dating from 9th–8th millennium BC in the Zagros mountains, i.e. Iran_N and Ganj Dareh_N, suggesting a pre-Neolithic common ancestor related to the ancient Caucasus hunter-gatherer component that diverged from the Andamanese huntergatherer lineage in the Late Pleistocene ..."

[pegasus]

Someone recently posted the qpAdm runs for 3 Nairs .

These remind me of models banned trolls on AG would use.

Afaik ,  IVC and Kurumba are 2 pops on the same cline so how can they be utilized as a left group?? Thats like me utilizing  Sintashta and Poltavka on the left pops. In Arjun Bidanda's paper which was focused on South Indian populations, these kind of models failed, infact and he used tonnes of samples .  Frankly I would imagine a Gujarati like Brahmin type source bringing Steppe related ancestry rather than straight up Andronovo.  

The old admix tools format was much nicer looking , looks like something off an 80s computer.

Pls give me a passing model of South Asians on qpAdm  using a  IVC like pop , Steppe and Bmac. All these simplistic modellings using a Tribal like pop , an IVC like pop and Steppe work perfectly fine and pass on qpAdm.

If you were on AG they were plenty,  but based of Kale's work nuanced models work best, I will take Kale's perspective on this over any Twitter or Reddit trolls. As I said a Triba

Here is one with Nagar Brahmins for example which ESPN ran, its a parsimonious model with more proximal sources rather than one which essentially is embellished distal one. Passing is one thing , reality is different. On formal papers a passable model is just 0.05 but the more than often thats not reality, I recall the Yamnaya models from 2015-17.

https://i.imgur.com/wsfrYR9.png

[ESPLover]

Someone recently posted the qpAdm runs for 3 Nairs .

These remind me of models banned trolls on AG would use.

Afaik ,  IVC and Kurumba are 2 pops on the same cline so how can they be utilized as a left group?? Thats like me utilizing  Sintashta and Poltavka on the left pops. In Arjun Bidanda's paper which was focused on South Indian populations, these kind of models failed, infact and he used tonnes of samples .  Frankly I would imagine a Gujarati like Brahmin type source bringing Steppe related ancestry rather than straight up Andronovo.  

The old admix tools format was much nicer looking , looks like something off an 80s computer.

Pls give me a passing model of South Asians on qpAdm  using a  IVC like pop , Steppe and Bmac. All these simplistic modellings using a Tribal like pop , an IVC like pop and Steppe work perfectly fine and pass on qpAdm.

Here:

I used an extensive set of right groups here:

[BMG]

Periodical Ancient Ancestry breakdown from illustrative dna.
   

Does 2% Canaanite indicate anything or is it excess ANF not covered by BMAC