Saturday, January 17, 2015

The Omotic Cluster: A Horn Specific Cluster

The Omotic component which I sort of disseminated in this post seems to be unique in that its spread doesn't really leave the Horn of Africa.

The component for the time being peaks in Ari Blacksmiths (90 to 100%~) and after them in Ari Cultivators, these peoples essentially being speakers of an Omotic language for now of the Afro-Asiatic language family hence the naming of the component though one study (Hodgson et al. [1]) has opted to call it Ethiopic, noting its spread only in the Horn region.

You can observe its peaking in the Ari populations in Hodgson et al.'s ADMIXTURE run (names of the components) or even Shriner et al.'s (component that peaks in the Ari Blacksmiths) but here is perhaps a more detailed showing of the Omotic component's proportions within these two populations from an old run created by an ethnic Somali:

 His run is somewhat more adept at spotting Omotic admixture in populations than the runs in the studies as he removes the most inbred of the Ari and as one colleage points outs- :

  "I would guess if a component is based on shared genetics with your immediate family rather than a more ancient family ("population"), alleles that are not really characteristic of a particular region might show up in other regions, for reasons other than shared ancestry"

- this can perhaps prevent a skewing of results and show Omotic's spread better in the Horner populations (the majority of Cushitic & Ethiopian Semitic speakers) however I wold place his results for the Ari under some scrutiny. For one, his results maybe somewhat inflated. For one, his average in the case of the Cultivators at about 86%~ doesn't fit with their 17%~ Ethio-Somali admixture from Hodgson et al. and they show comparable levels in Shriner et al. in the case of Lowland East Cushitic; both these component's serve to map the Cushitic input in Ari Cultivators along perhaps with the Nilo-Saharan/ East African input in them in both studies that's also present in Horners outside of the Ethio-Somali mixed cluster/ component.

So while this run is quite adept it definitely has its discrepancies-> it's "West Eurasian"/ West Asian admixture scores at K=5 are definitely deflated for each of the populations and don't at all fit with most ADMIXTURE run's levels of admixture or the work of rather reliable studies like Gudrasani et al. or Pickrell et al. [2] [3] in the case of Horner admixture so I would advise a focus not on the exact percentages and numbers but more on the relative levels and climes of the admixture components (this group has more than that group and so on).

Now, what this run and studies such as Shriner et al. Pagani et al. & Hodgson et al. show us is that "Horners" (the Cushitic & Ethiopian Semitic populations that make up the majority of the Horn of Africa) or at least many of them have experienced gene flow from Omotics/ peoples genetically similar to Aris. It's clear that the component for the time being is tied to people who speak within the Omotic branch of the Afro-Asiatic language family. [4] [5]

For one, among "Horners" it peaks among the one group that actually speaks an Omotic language, it's total peak are Ari Blacksmiths and their Cushitic admixed kin the Ari Cultivators. So it can only really be assumed that the Omotics in the Horn (not necessarily Aris) could have somewhat genetically contributed to the populations specifically of Ethiopia (the very heart of the Horn region) however there seems to be no influence of Omotic in Somalia/ among non-Ethiopian ethnic Somalis.

Simply observe the levels of Omotic/ Ethiopic admixture among various autosomal DNA tested Horner ethnic groups/ populations: 

Horner Omotic Levels 1.0

However those are our levels with the outliers among each population set not removed. These are the results with the very few "outliers" removed:

Horner Omotic Levels 2.0

The "High" line just shows you the level of admixture the one person among each population set that had the most Omotic ancestry (the group's "High") with the "Low" line obviously doing the opposite. Ethnic Somalis from Somalia (North-Centralites if we're being exact) pretty much have none while their Ethiopian brethren also practically lack Omotic admixture with only a 2%~ negligible average whilst a good number of Ethiopian Somalis actually seemed to lack Omotic admixture as well.

The mixed component (a mix of seemingly non-ANE & WHG admixed West Asian ancestry + East African + Khoisan) however interestingly when we ignore Wolaytas who are Omotic speakers and thus arguably would be expected to have a large proportion of this component anyway; peaks mostly in Oromos.

However Oromos have many subgroupings among them with the Borana being the least West Asian admixed among all Horners tested so far, one group being very similar to Somalis in terms of admixture levels (Oromo B) & one being very Agaw (Beta Israel & Xamir Agaw) like, they're not exactly like Somalis who are a rather inbred and genetically "homogeneous" ethnic group.

After them it's Amharas alongside Beta Israels ("Ethiopian Jews") and then other than in Somalis it's at its lowest in Tigray-Tigrinyas as well as Xamir Agaws. Since there's in truth no genetic difference between Tigrinyas in Eritrea & Tigrinyas in Ethiopia (they're essentially the same ethnic group) and since one Eritrean Tigrinya is actually among the samples used in that Somali's ADMIXTURE run and he shows the same Omotic levels as Ethiopian Tigrinyas-> it can only be assumed the component has a spread in Eritrea as well among at least the Ḥabesha population there that makes up more than half the country.

How exactly Omotic spread to Agaw & Ethiopian Semitic populations (who were originally Agaw speakers given the Agaw substratum in their languages and the rise of Ethiopian Semitic at around roughly 800 to 1000 BC [6]) is unknown to me of course beyond the obvious being that they mixed with and interacted with Omotic peoples. Perhaps it was the ones to their west in regions of Ethiopia such as the modern "Beninshangul Gumuz Region".

The Gumuz are actually another population with Omotic input but interestingly; the Gumuz lack some of the heavier West Asian admixture in the Omotic component. I.e. Observe the "West Eurasian" (West Asian) admixture levels across East Africa from Pickrell et al. :

Note: The "Afar" in these studies are in fact Xamir Agaws.

The Gumuz as you can see in that Somali's run and Hodgson et al.'s run; have a significant amount of Omotic/ Ethiopic input. Although they just don't have the levels of West Asian admixture you would expect from peoples with their Omotic levels. Ari Blacksmiths who are practically 95 to 100% Omotic are about 15-16%~ West Asian however the Gumuz are just about relatively 1-2%~ in Pickrell  and actually show no West Asian admixure in that Somali's run at K=5 however as I said; his West Asian admixture levels are quite deflated, no doubt due to a different method for identifying such admixture which seems to have inflated the African levels in many groups and deflated their Eurasian levels.

Nevertheless their levels of West Asian admixture are a bit too low for peoples possessing their proportions of Omotic/ Ethiopic input. It's the same effect in Hodgson et al. as you might have noticed where they barely even show "Arabian" the way Ari Blacksmiths and Cultivators do before the formation of the Ethiopic component at K=11/ 11.

This low level West Asian influence as a friend brought to my attention as well as suggested; might mean that the Gumuz interacted with Omotics long before these Horner groups did. Likely at a time when Omotics had far less West Asian input?

The Omotic cluster/ component's mixed nature ultimately makes the Gumuz the most East African population to be tested so far. By East African I am referring to the "Nilo-Saharan" cluster in Hodgson et al. , a component that is also often dubbed East African (EA) in other ADMIXTURE runs and studies whilst I and some others I'm familiar with used to often call it "Ancestral East African" (AEA). The Gumuz are by all accounts predominantly made up of this component as all of their non-Omotic derived ancestry is EA, the rest which comes with Omotic is a small West Asian influence that is practically negligible along with a slightly non-negligible Khoisan influence which comes with Omotic/ Ethiopic.

Ethnic Somalis lack input from Omotics very likely because Somalia or at least North-Central Somalia very likely never had an Omotic presence for Somalis in those areas to interact with. Nevertheless one sub-cluster of Somalis; Sab Somalis who have their own sub-languages such as Tunni, Jiido, Garre & Maay were all identified to have a strong Oromoid influence in their languages, one Somali linguist who actually had a good part in establishing them as totally separate languages from from Standard/ "North Somali" even argued that their Oromo influence goes as far as being a substratum. Their languages also lack pharyngeal sounds... A phenomenon found in Oromo & Sidamic languages (they lack them as well) however Somali, Saho & Afar (fellow "East Cushitic" languages) are very heavy on pharyngeal sounds.[6]

One ethnic Somali on also encountered a half Ogadeni Garre who was straddling in a PCA plot (Principal Component Analysis/ a cluster like this one) between him and Borana Oromos whilst the guy had no Niger-Congo input or anything of the sort that would "pull him south" of standard inbred Somalis like the chap on This all might mean that these Sab Somalis who are very often traditionally sedentary farmers or agro-pastoralists could very well have Omotic input via Oromo populations similar to the Borana who for one have Omotic admixture; which would give Omotic a certain spread in Somalia as well, though there are some small Ḥabesha and Oromo minorities all over the country anyway, some of whom have been around for centuries.

The Sab are counted under the "Raḥanweyn & Digil":

Beyond the Horn of Africa; I would assume it could also be found in parts of Kenya where the Borana live. Other than that; it's a generally Horn specific cluster, showing only extremely low "noise" levels in non-Horn populations.

Reference List:

4. Genome-wide genotype and sequence-based reconstruction of the 140,000 year history of modern human ancestry, Shriner et al.


1. A good part of Diriye's book is based on some outdated information on history, anthropology and so on (using terms such as "Hamitic") and I wouldn't make much of any of it beyond his comments on linguistics (concerning those sub-languages of "Macro-Somali") albeit I'm skeptical as to whether or not what these languages have is a clear influence from Oromoid or actual susbtratums.

2. If you'd like some ethnic maps of the Horn of Africa, knock yourself out: [Eritrea] , [Ethiopia] , ["Greater Somalia"] , [Djibouti 1, Djibouti 2]

3. Make of Kitchen et al.'s linguistic ideas about Semitic what you will but it is generally accepted that the Ethiopian Semitic languages arose after a single introduction of South Semitic at around 800 to 1000 BC with no prior evidence (inscriptions etc.) of them before that. The Ethiopian Semitic languages carrying an Agaw/ Central-Cushitic substratum (showing that the speakers once spoke Agaw languages) is also a well-known fact.

4. To their credit, Shriner et al. actually did also find that Somalis lack Omotic admixture as you can see via their supplemental.Whilst showing that Ethiopian Somalis have a certain amount albeit their levels for them are arguably quite inflated and probably skewed even further by two "outliers" among the Ethiopian Somali samples who had some pretty high Omotic levels.


That Somali's run for the Ari populations (in terms of Omotiv levels) are likely pretty solid at least in the sense that he was onto the right idea about removing the inbred samples. In fact an upcoming study with its abstract shown at a recent conference (have a look) seemed to be focused on this exact same subject and it aims to show that Ari Cultivators are essentially a less inbred edition of Ari Blacksmiths.

As a result Ari Cultivators probably to some extent have less Omotic due to being less inbred however the "Ethio-Somali" (Cushitic influence) in them seems to indeed be real, for one it's noticeable in how they are notably more West Asian than their Blacksmith counterparts. 

One should also note that Omotics such as the Ari have a rather distinct Y DNA profile that you can see in Plaster et al. (Variation in Y chromosome, mitochondrial DNA and labels of identity in Ethiopia) or have it more comfortably simplified for your viewing at the Ethio Helix blog. I'd highly recommend that blog btw.