PDA

View Full Version : Take away for Y-STR data from Djibouti, Ethiopia, Eritrea and Kenya ( Iacovacci 2017)



Almagest
11-20-2017, 08:08 PM
It's not open access and I'm not sure if I am allowed to share it online so I will only comment here. Most of this will be regarding E-V32 and Somalis. You will probably need a basic understanding of Somali clans for the Somali bit.

Stats are as follows;
19893

The STR results were also included in the supplementary data.

1) DYS438=12 is a very good indicator for the SNP E-Y28701 (https://www.yfull.com/tree/E-V32/). YFull rates it 4/5 stars, and from FTDNA, 7/8 members who have tested downstream of V32 and are Y28701 have DYS438=12, the eighth member being DYS438=13. Crucially, none of the Y28701- members are DYS438=12.

2) DYS438 was included in the STR's tested. Unsurprisingly, 100% of Somali V32 in all regions was DYS438=11, and most likely Z813+.

3)Stats on DYS438=12, 13 by ethnicity as follows (brackets indicate DYS438=13);
Cunama 1/3
Nara 2/2
Tigray 5/6
Tigre 3/3 (1)
Ethiopian Jews 4/5 (1)
Oromo 1/8
Wolayta 0/1
Somali Ethiopia 0/2
Tigray Ethiopia 1/1
Somali Djibouti 0/9
Borana 0/3
Luhya 0/3
Maasai 0/3
Somali Kenya 0/3

This indicates that Somali/Southern Ethiopian/Kenyan and North Ethiopian/Eritrean V32 have a TMRCA of 5000 years.

4)I alluded (http://www.anthrogenica.com/showthread.php?12136-Is-V32-in-Arabia-ancient-or-recent&p=294892&viewfull=1#post294892) to there being no connection between Gulf V32 and East African V32 here, but that seems set to change with these STR results. I am now convinced that it is perhaps not as simple as I first thought. As far as I know, there are no downstream SNP tests for Ethiopian or Eritrean V32 yet. The next step is to compare the Gulf V32 STR's on FTDNA with the 16 DYS438=12,13 STR's results from Northern Ethiopia/Eritrea. It should be kept in mind that DYS438=11 does not mean Z813+, a few Gulf V32 have it but are V32* (Z813-, Y28701-). I am not sure if they belong to Y25511.

5)Djibouti Somali T-M70 has limited diversity (Dir 100% T-M70). I will do a simplified TMRCA dates of these and compare them to T-M70 in the Afar (5/20), Somali Ethiopian T-M70 (5) and H Yoonis T-M70 on FTDNA. Considering the Ethiopian Somali samples were deep into Eastern Ethiopia, Darood territory, I am surprised at the T-M70 % (5/10). T-M70 isn't usually associated with areas in which the Darood settle. It may be that T-M70 is not strictly limited to the Dir. However, the clan affiliation of the testers here is not mentioned. There is a Warsangeli T-M70 on FTDNA.

6) All 9 Somali Djibouti V32 samples were DYS393=12, only seen in the Isaaq so far for Somalis on FTDNA (4/4). The paper mentions that all V32+ tester claimed affiliation to Isaaq. It is very likely that most, if not all, were Sacad Muuse, like @drobbah. It is no surprise then that they match closely. TMRCA will follow, because they seem very close.

Lank
11-20-2017, 09:03 PM
Thank you for your analysis. It was very interesting.

This is why I personally put very little value on dating Middle Eastern African-affiliated lineages based on the wealth of high resolution data on Arabian Y-DNA lineages and their supposed divergence from (a small subset of) African lineages. There is way too little data on Africans. There is certainly a possibility that there are old V32 lineages in Arabia (we know M35 has a presence in West Asia since at least the Natufians, and there were back-migrations into Africa, likely bringing some M35 as well).

But this can't be concluded based on such an inadequate sampling of Africans, when we know there is significant, recent African admixture in Arabia. And it came from diverse groups of Africans, so a broad sampling strategy will be necessary to capture the diverse, recent African lineages that made it to Arabia.

Almagest
11-25-2017, 07:33 PM
Not sure if this is an accurate way of measuring things, but I decided to average the TMRCA and genetic distance from the modal haplotype, since they were a few samples that were skewing the results, one of which was a person that claimed to be Hawiye but still T-M70. He did not match the Dir T-M70, and the author made note of that.

Djbouti Somali TMRCA 865 years (23 Dir + 1 Hawiye) Average genetic distance from modal 1.16. The Hawiye is a genetic distance of 5 from the modal.
Djibouti Somali TMRCA 819 years (9 Isaaq). Average genetic distance from modal 1.

drobbah
01-29-2018, 02:00 AM
6) All 9 Somali Djibouti V32 samples were DYS393=12, only seen in the Isaaq so far for Somalis on FTDNA (4/4). The paper mentions that all V32+ tester claimed affiliation to Isaaq. It is very likely that most, if not all, were Sacad Muuse, like @drobbah. It is no surprise then that they match closely. TMRCA will follow, because they seem very close.
The only V32+ folks (Somali or foreigners) with DYS393=12 are a bunch of us Habar Awals,1 Habar Jeclo (Z813+) and 1 Kuwaiti (Kit# M6826).I have also seen a Sacad Muuse (Kit#68929) who took the 37 marker test and he also has DYS438 =11 which probably means we are Z813+ as you stated in the OP


P.S The majority of Isaaq Djiboutians are from my specific branch of SM called Jibriil Abokor so the TMRCA according to our abtirsi should be around 10-12 generations give or take.

Squad
01-29-2018, 04:04 PM
To be honest, DYS393 = 12 alone can't really tell you anything. Those few Somali samples from Djibouti are clearly descended from a very recent ancestor which happened to have DYS393 = 12 because the haplotypes are of a very reduced diversity. This ancestor probably lived very recently, some 300 years ago or less I'd say. Here it is DYS438 = 11 which would allow us to safely exclude E-Y28701 with a certainty of let's say about 90-95%. So yeah, they're most likely Z813+.

And Yfull didn't rate ''DYS438 = 12''. It rated the marker itself as being of a 4 star quality, based on its alleged mutation rate. However, what most (including Yfull) seem not to understand is that the quality of a certain marker isn't fixed it depends on the allelic value : the higher the value, the higher the chances for a mutation, thus the lower the quality. The reason why DYS438 is rated 4/5 is mainly because R-M239 is skewing the quality of this marker, due to the fact that it has DYS438 = 12 as a modal, which is a high value for this marker, resulting in a much higher mutation rate, same for E-Y28701. It is depending on the value that the quality should be judged, not on the marker itself. For instance, most haplogroups have DYS438 = 10 and if you check them out, you'll notice that DYS438 is rarely mutated, thus would be of a 5/5 quality, whereas in the case of R-M269 for example, there are many samples deviating from the modal and I would give DYS438 a rating of 4/5 or even 3/5 for this haplogroup. But overall DYS438 is very informative and in this case is a sharp indicator to distinguish between E-Y28701 and rest of E-V32, although some slight overlapping would be seen due to 11 and 12 especially being high values.

As for the study, it is very interesting, I might talk about a few observations later on.

drobbah
01-30-2018, 06:39 AM
There is a bunch of BY8150/BY8125 that have DYS438 =11 so there is no clear cut way to differentiate between V32 subclades via str results alone

Almagest
01-30-2018, 11:45 AM
There is a bunch of BY8150/BY8125 that have DYS438 =11 so there is no clear cut way to differentiate between V32 subclades via str results alone

Yes DYS438=11 will not tell you the if you are Z813 or BY8125 but so far DYS438=12,13 is only seen in Y28701.

drobbah
01-30-2018, 01:50 PM
Yes DYS438=11 will not tell you the if you are Z813 or BY8125 but so far DYS438=12,13 is only seen in Y28701.

Have you by any chance seen any African BY8150/25 On FTDNA yet?

Squad
01-30-2018, 02:00 PM
There is a bunch of BY8150/BY8125 that have DYS438 =11 so there is no clear cut way to differentiate between V32 subclades via str results alone

What I said : Here it is DYS438 = 11 which would allow us to safely exclude E-Y28701 with a certainty of let's say about 90-95%. So yeah, they're most likely Z813+. and But overall DYS438 is very informative and in this case is a sharp indicator to distinguish between E-Y28701 and rest of E-V32, although some slight overlapping would be seen due to 11 and 12 especially being high values.

What I meant was that, due to the ancestral for V32 at this locus being equal to 11, the mutation rate is higher than if it had a value of 10. In the case of Y28701, the mutation rate is even higher because the value is 12, hence why it is more common for Y28701 to have 11 or 13 than it is for the other V32 to have 10 or 12, which is itself more common than say E-V13 having 9 or 10. This is why there will inevitably be some overlapping between Y28701 and the rest of V32, much more than if both subclades respectively had 10 and 11 instead. This overlapping would be mostly created by the more unstable status of Y28701's high DYS438. Like I said, the higher the value for any locus, the higher the chance to mutate, thus the lower the quality !!

Almagest
01-30-2018, 03:19 PM
Have you by any chance seen any African BY8150/25 On FTDNA yet?
I haven't seen "African" V32 BigY on FTDNA expect the Somali guy. There is the Kenyan BY8150 on YFull though.