PDA

View Full Version : E-V32 and E-M329 among Ethiopian ethno-linguistic groups



Passa
01-16-2016, 11:10 PM
The 2011 Plaster thesis (http://discovery.ucl.ac.uk/1331901/) on Ethiopian genetic diversity provided us with a lot of uniparental data, though apart from an increased SNP resolution on limited samples, most of the data suffered from very low resolution, so I decided to try and get something worthy from the supplementary data (http://discovery.ucl.ac.uk/1331901/2/1331901.xls). Initially I wanted to predict the sub-clades of all E(xM2) haplotypes, though later I realised that there were many uncertainties, with two noticeable exceptions: E-V32 and E-M329, the haplotypes of which are easily identifiable. So here I provide the document in which I have written off the frequencies of E-V32 and E-M329 for each ethnic group analysed in the study, hoping that it's useful and interesting: https://www.dropbox.com/s/nza4g92qi6lxx40/E-V32%20AND%20E-M329%20IN%20ETHIOPIA.docx?dl=0

To show that my prediction is very accurate for these two sub-clades, I use high-resolution SNP data (http://ethiohelix.blogspot.it/2012/11/extensive-doctoral-thesis-on-ethiopian.html) on Maale from the same author. With a sample size of 69, Maale have 1.4% E-V32 and 44.9% E-M329. With a sample size of 119, my predictions on Maale based on Y-STRs are 0.8% E-V32 and 39.4% E-M329.

A distribution map of E-M329 is on the way.

Passa
01-16-2016, 11:22 PM
I want to add that the E-M34 thing in Omotics is pretty much real. I have predicted all E(xM2) haplotypes of several ethnic groups from the supplements and it looks like some of them have E-M34 as high as 55%.

Lank
01-17-2016, 12:03 AM
Very interesting work!

The groups where E(xM2) peaks in Ethiopia, according to Plaster; the Ganjule, Sheko and Gobeze (at or above ~90%, all Omotic speakers) "only" show 26.6%, 16.8%, and 10.6% M329 in your calculations. If this is accurate, it would be quite interesting to look at the non-M329 E lineages further.

Burji are estimated at 9.2% E-M329 here, and Konso 15.8%. In Hirbo's sample, they show 35-37% E-PN2(xM2,M35), which I had guessed would be mainly M329, although M281 or other unknown lineages are also possible.

I presume you don't have a way of estimating M281, or M123/M34? M281 would be particularly interesting as it's peaking in the Maale so far in the high res Ethiopian sequences, forms a deep branch in the PN2 lineage, and is a distant relative of M35.

Passa
01-17-2016, 12:26 AM
Very interesting work!

The groups where E(xM2) peaks in Ethiopia, according to Plaster; the Ganjule, Sheko and Gobeze (at or above ~90%, all Omotic speakers) "only" show 26.6%, 16.8%, and 10.6% M329 in your calculations. If this is accurate, it would be quite interesting to look at the non-M329 E lineages further.

Burji are estimated at 9.2% E-M329 here, and Konso 15.8%. In Hirbo's sample, they show 35-37% E-PN2(xM2,M35), which I had guessed would be mainly M329, although M281 or other unknown lineages are also possible.

I presume you don't have a way of estimating M281, or M123/M34? M281 would be particularly interesting as it's peaking in the Maale so far in the high res Ethiopian sequences, forms a deep branch in the PN2 lineage, and is a distant relative of M35.

Hirbo's Burji and Konso samples size is low though, plus it is unclear if they are from Ethiopia or Kenya.

Regarding M34 and M281, their prediction is not as safe as the ones for E-V32 and E-M329, mostly because there's some overlapping between E-V1515 and E-M34 haplotypes, but from what I were able to predict, quite a lot of Omotic groups show a notable frequency of E-M34.

Lank
01-17-2016, 12:33 AM
Hirbo's Burji and Konso samples size is low though, plus it is unclear if they are from Ethiopia or Kenya.
True. I just thought it made sense since they are neighbors, and both groups showed similar levels in Hirbo's data.

Perhaps it would be a good idea to also mark Mota man on your M329 map?

Passa
01-17-2016, 12:36 AM
True. I just thought it made sense since they are neighbors, and both groups showed similar levels in Hirbo's data.

Perhaps it would be a good idea to also mark Mota man on your M329 map?

Sure, I will do that. For the map I will use Plaster (my predictions included) and Hirbo data.

Agamemnon
01-17-2016, 02:37 AM
I want to add that the E-M34 thing in Omotics is pretty much real. I have predicted all E(xM2) haplotypes of several ethnic groups from the supplements and it looks like some of them have E-M34 as high as 55%.

Now that is particularly interesting.

Ebizur
01-17-2016, 03:59 AM
Now that is particularly interesting.
Yes. Wouldn't that mean that these E-M34 Omotic speakers must be patrilineally more closely related to many people in West Asia (many of whom are Semitic speakers) than to a majority of Berber speakers in North Africa?

Lank
01-17-2016, 11:49 AM
M34 (TMRCA estimated at 15 kya by YFull) is a good candidate marker for Afroasiatic ancestry in Omotics. Omotic groups may mostly be an admixture between ancient Afroasiatic migrants from the north, and the Paleolithic locals (represented by Mota), who may have been dominated by M329. Y-DNA J and E-V1515 lineages may also have been involved, but we hardly know anything about them in Omotic populations. Then there are groups like the Wolayta, who have significant Cushitic ancestry as well.

Passa
01-17-2016, 01:59 PM
Distribution map of Y-DNA E-M329 (Mota is the tiny black dot in the >20% frequency area):
7336

Lank
01-17-2016, 02:26 PM
This (http://mbe.oxfordjournals.org/content/24/10/2180.full) paper has Y-STR data on Burunge, Datog, Iraqw, Hadza and Sandawe. Other than Hadza, they all show some E-PN2(xM2,M35).

Passa
01-17-2016, 02:46 PM
This (http://mbe.oxfordjournals.org/content/24/10/2180.full) paper has Y-STR data on Burunge, Datog, Iraqw, Hadza and Sandawe. Other than Hadza, they all show some E-PN2(xM2,M35).

Yeah right, I forgot about that! I am looking at the haplotypes, trying to discern their M329 status.

Passa
01-17-2016, 02:56 PM
Ok, comparing "M96*" Y-STR haplotypes of 6 Burunge (AA), 1 Datoga (NS), 5 Mbugwe (NK) and 4 Turu (NK) it is clear that most or all of them are V38(xM2,M329) and are similar to Gabonese haplotypes. I also think that likely V38(xM2,M329) would be resolved to multiple SNPs/sub-clades in future studies, based on what I have observed.

Passa
01-17-2016, 03:27 PM
I have updated the map, adding Qatar data reported in Abu Amero et al. (2009). I will have a look at the Saudi Arabia FTDNA project, data is divided by region there and I remember having seen very ancient haplotypes there, maybe there's some M329 too.

Agamemnon
01-17-2016, 05:48 PM
Yes. Wouldn't that mean that these E-M34 Omotic speakers must be patrilineally more closely related to many people in West Asia (many of whom are Semitic speakers) than to a majority of Berber speakers in North Africa?

I really agree with what Lank just said, namely that E-M34 is one of the best candidates for Afroasiatic ancestry in Omotic speakers. Now I know we had this discussion before but I think it really needs to be said: We should not expect a 1:1 correlation between genes and languages. While you are quite right in stating that these guys are somewhat more closely related to West Asian E-M34 individuals than they are to E-V65 or E-M81 Berbers from a patrilineal standpoint, in the same way R1b-Z2103 Armenians are more closely related to all the Yamnaya R1b-Z2103 samples than to R1a-Z93 Iranian speakers in neighbouring Iran and Kurdish areas. Does that necessarily mean R1a wasn't present in the PIE-speaking community (I know this isn't what you're saying)? I seriously doubt that. Odds are both R1a and R1b-M269 were to be found in the PIE-speaking community. Similarly, most of the major E-M35.1 branches we know of (M78, L19, V1515, M34) are over 13,000 years old, now does the association of a specific branch (M34) with the most basal branch of AA (Omotic) necessarily mean that all the other M35.1 branches weren't present in the PAA-speaking community? Once more, I think not, if anything their TMRCA estimates fit rather well with the break up date of PAA (~14,000 years BP). Moreover, Plaster also found that North Omotic groups had, by far, the highest J frequencies in Ethiopia, most of which is presumably J1. While this obviously leaves the door open for a potential involvement of J1 in the spread of AA at an early stage, an interesting detail I've noticed some time ago is that J1 and E-M34 often tend to show up together, this is another example of this IMO.

Lank
01-17-2016, 06:39 PM
Passa, I have not read anything about V38(xM2,M329) in Gabon. Could you share your sources?

And to add to Agamemnon's argument, I think one important factor is assimilation. Early on in the history of a language family, this could contribute to genetic phylogeny being discordant from linguistic phylogeny, as the neighboring groups would have carried related lineages. So if you had a mainly Z830 Proto-Afroasiatic group (this is just used as an example), they could have assimilated the neighboring V257 and V68 groups as they expanded, leading to somewhat confusing modern patterns. For instance, V1083, which is M78(xV22,V12) has only been reported in Sardinia and the Saho from East Africa, but it seems obvious (IMO) that it ultimately derives from North Africa in both groups, and just got caught up in the expansion of more successful lineages.

This can be compared with the Indo-Europeans, and different R1b/R1a clades that ended up successful in different IE subgroups just "by chance". You e.g. have the L664 R1a clade in NW Europe, which spread there with Corded Ware, but is divergent from the major European/Asian R1a lineages.

Passa
01-17-2016, 06:49 PM
Passa, I have not read anything about V38(xM2,M329) in Gabon. Could you share your sources?

I have interpreted the M96(xM33,M75,M2,M35) in that paper we talked about earlier as V38+.