I have slowly been collecting 23andMe results of Greeks with Peloponnesian ancestry and have found it super strange to see how many of them are given trace amounts of SSA, EA, and SA. To me, this makes no sense whatsoever, but I thought I'd share some examples. These surely can not be taken seriously or literally, but I do not know why the 23andMe algorithm continues to do this for many. Any ideas?
---
---
---
---
---
---
---
---
---
---
---
---
---
---
---
---
---
---
---
XXD (01-23-2021)
See, I told you most Peloponnesians Ive seen have low SA ancestry. I'm not sure why these results would not be real. Romani people are very endogamous, but many would have assimilated in sedentary Greek populations in the last 1100 years. Such intermarriages are even mentioned in literature e.g. Karagatsis.Originally Posted by Markos
I have slowly been collecting 23andMe results of Greeks with Peloponnesian ancestry and have found it super strange to see how many of them are given trace amounts of SSA, EA, and SA. To me, this makes no sense whatsoever, but I thought I'd share some examples. These surely can not be taken seriously or literally, but I do not know why the 23andMe algorithm continues to do this for many. Any ideas?
---
---
---
---
---
---
---
---
---
---
---
---
---
---
---
---
---
---
---
This is especially true for Arcadia, where up to 3% Y chromosomes belong to haplogroup H.
So low level admixture definitely did happen.
distance: 1.656775
Ancient Greece/Balkans: 48.8
Early Slavic: 24.5
RUS_Maykop_Novosvobodnaya: 12.7
Levant_Megiddo_IA: 10.2
IRN_Shahr_I_Sokhta_BA2: 1.7
MAR_Taforalt: 0.8
CHN_Yellow_River_LBIA: 0.7
Yoruba: 0.6
I don't think it's likely or else this would be well-known information and not something found out suddenly by 23andMe trace ancestry, which literally changes for most people every update. Doesn't make sense!
XXD (01-24-2021)
Who knows, to be honest. However, as you have seen, I score 0.2% SA in 23andme and I really DO have full Romani matches on 23andme. When I ran the Sahr i Soktha sample as a proxy for SA ancestry, many Peloponnesians score about 0.1 to 3% of this ancestry.
Yes, sure, it could be because of the West Asian ancestry that exists in both Greeks and Romani. However, when you have 3% of the Arcadian population carrying Romany Y chromosomes, it is evident that low level admixture did happen. This also explains why more than 35% of all the G25 modern Greek samples have 0.2% to 3% Shahr i Sokhta/Kanjar/Paniya/Gupta admixture (roughly same percentages no matter what group one uses).
distance: 1.656775
Ancient Greece/Balkans: 48.8
Early Slavic: 24.5
RUS_Maykop_Novosvobodnaya: 12.7
Levant_Megiddo_IA: 10.2
IRN_Shahr_I_Sokhta_BA2: 1.7
MAR_Taforalt: 0.8
CHN_Yellow_River_LBIA: 0.7
Yoruba: 0.6
discreetmaverick (01-25-2021), xripkan (01-25-2021)
I'm not buying it, lol. I totally get what you're saying but I just know these results will change in the next update for most people. There's got to be a better explanation.
It is not only the Peloponnese samples.Many other Greek samples in G25 showing exotic components.Even the Cretan samples,some of them require something North African in their modelings.
Btw this Laconian sample is coming like this.I am not sure if some of the samples are low-res or some kind of bad quality.
Target: Greek_Laconia:638
Distance: 2.3220% / 0.02322050
93.0 Greek_Laconia
3.2 Sardinian
2.0 Colla
1.8 Montenegrin
Target: Greek_Laconia:638
Distance: 2.0676% / 0.02067645
15.8 DEU_MA_o
10.8 UKR_Srubnaya_MLBA
10.0 UKR_Trypillia
9.8 TUR_Isparta_EBA
9.4 VK2020_RUS_Pskov_VA
9.0 TUR_Tell_Kurdu_EC
6.6 AZE_Caucasus_lowlands_LC
6.4 GRC_N
5.6 Kura-Araxes_ARM_Kalavan
4.2 DEU_LBK_UW
3.6 UKR_Chernyakhiv_Shyshaky
3.4 USA_colonial_period
2.4 CHL_Fuego_Patagonian
1.6 TUR_Kumtepe_N_low_res
1.4 Levant_PPNC
There is no meaningful or substantial "exotic" ancestry in the Peloponnese. Those figures are all noise.
Markos (01-25-2021)
That is the most realistic conclusion, in my opinion. Though, I don't know how to 'justify' the noise components either.
Just looks like calculator garbage to me. There's no other logical explanation that I've heard/seen so far.
Individuals might very well have some kind of recent, exotic ancestry at very low levels that shows up that might have come via admixture with populations admixed with populations admixed with populations(...) or maybe be expressing more ancient/diffused streams of ancestry more strongly than others due to random inheritance over generations but at 0.X%, I'm not sure how reliably we can say anything either with only this kind of method (the categories in something like 23andme are a different matter, likely even worse especially when splitting closely related populations into groups and depends on the samples they use to create the categories, so extra issues). It might be just adding tiny amounts of things that aren't really there because the more appropriate sources haven't been used while trying to model them etc. That's probably more likely with individuals (and distances of the resulting model are generally greater for them) than with populations.
I think a comparison with surrounding populations might be instructive too, e.g. if the geographic pattern that emerges is a bit odd compared to what we might expect, or a look within populations e.g. does that kind of ancestry pop up consistently or just among a few of the samples. For example, in G25, if you include Yoruba in a generally ultimate-source appropriate model, it consistently pops up among the Andalusians in G25 (except only 1) at low levels while it shows up in a small minority of Cretans. Or if you include Nganassan, only a couple of Irish out of 85 get it (or a similarly low number of the much more eastern Poles for that matter) but half of Estonians do at low levels, which probably makes it more likely that something is there.
Last edited by DFSTFD; 01-25-2021 at 09:22 AM.
Posting Permissions
Reply With Quote