Genomic selection in Nilo-Saharan and Niger-Congo speakers (Pre-print)

[blackflash16]

Pre-print: Genomic evidence for population specific selection in Nilo-Saharan and Niger-Congo linguistic groups in Africa (http://www.biorxiv.org/content/early/2017/09/10/186700)

Abstract:


Background: There are over 2000 genetically diverse ethnolinguistic groups in Africa that could help decipher human evolutionary history and the genetic basis of phenotypic variation. We have analysed 298 genomes from Niger Congo populations from six sub Saharan African countries (Uganda, Democratic Republic of Congo, Cameroon, Zambia, Ivory Coast, Guinea) and a Nilo Saharan population from Uganda. These samples were collected as part of the TrypanoGEN consortium project http://www.trypanogen.net. Results: The population genetic structure of the 298 individuals revealed four clusters which correlated with ethnolinguistic group and geographical latitude, that is, West African Niger-Congo A, Central African Niger Congo, East African Niger-Congo B and the NiloSaharan. We observed a spatial distribution of positive natural selection signatures in genes associated with AIDS, Tuberculosis, Malaria and Human African Trypanosomiasis among the TrypanoGEN samples. Having observed a marked difference between the NiloSaharan Lugbara and Niger Congo populations, we identified four genes [APOBEC3G, TOP2B, CAPN9, LANCL2, (iHS -log p > 3.0, Rsb -log p > 3.0, Fst > 0.1 Bonferroni p > 1.8x10e4)], which are highly differentiated between the two ethnic groups and under positive selection in the Lugbara population. Conclusion: The signatures that differentiate ethnically distinct populations provide information on the specific ecological adaptations with respect to disease history and susceptibility/ resistance; as demonstrated in this study where APOBEG3G is believed to be involved in the susceptibility of the Nilo Saharan Lugbara population to Hepatitis B virus infection.


This is more of a medical genetics paper but they did do a ADMIXTURE run and some dating work.

They refer to the Lugbara as Nilotic here but they're normally classed as Central Sudanic:


Ugandan Bantu and Ugandan Nilotic appeared to begin separating from other populations about 23 and 47kya, respectively and became isolated about 3kya but these estimations may be confounded by admixture.




At K=6 The Niger-Congo populations exhibited 17-60% 122 admixture with minor ancestries, whilst the Ugandan Nilo-Saharan population had 7% admixture with Niger Congo ancestries


ADMIXTURE:
18678

[Angoliga]

They refer to the Lugbara as Nilotic here but they're normally classed as Central Sudanic:




ADMIXTURE:
18678


Good observation -- the paper's use of the term "Nilotic" was probably used as a broad ethnic identifier to distinguish Ugandan Nilotes from their Bantu counterparts (UGN, UGB ). Personally, I wish linguists used another term than "Nilotic" to avoid this kind of confusion.

At a first glance it seems like mere semantics but based on linguistic classification, Nilotic (https://en.wikipedia.org/wiki/Nilotic_languages) and Central Sudanic (https://en.wikipedia.org/wiki/Central_Sudanic_languages) (Lugbara [Maracha], Aringa, Ma'di etc.) are completely separate major branches of Nilo-Saharan (https://en.wikipedia.org/wiki/Nilo-Saharan_languages) speakers with a starkingly contrasting genetic history. This seems to be apparent when Ancestry ADMIXTURE results (https://i.imgur.com/xJAUKLa.png)are compared between a Nilotic and Central-Sudanic Ugandan (Tracing African Roots blog (https://tracingafricanroots.wordpress.com/ancestrydna/african-results/north-east-african-results/)). Moreover, these two groups of Nilo-Saharan speakers entered the Great Lakes regions at different periods and through different migratory routes.


Ugandan Linguistic Map clearly distinguishing between the CS and Nilotic speakers:




https://i.imgur.com/7jOSgS9.png

Some observations/remarks:


It's puzzling why only one Nilo-saharan population was selected for this paper; an array of Nilo-Saharan speakers would've given the paper's title and conclusions more credibility and legitimacy
ADMIXTURE results for the Lugbara (UGN) are surprisingly pristine (blue=East-African African proxy?); most results for this population (*based on immediate Central-Sudanic Ugandan family members) are ~30-40% West-African/Bantu yet only minor traces (red/yellow) are shown on these K runs
Nearly half of the sampled populations have minor non-negligible Eurasian on K3 runs; is it fair to assume much of this ancient Eurasian admixture in Niger Congo-A speakers=North African ADMIXTURE?
the descriptions for the Yoruba and Esan (YRI,ESN) are incorrect, they've been conversely labeled on the ADMIXTURE chart :/
This is rather sloppy writing here: within the 2.0 Background section the paper omits classifying Central Sudanics as agro-pastoralists. The Lugbara are predominantly subsistence farmers and have been for centuries if not millennia; pastoralism along with fishing have been secondary for sometime.
If there's any merit to these dating methods (MSMC), I wonder if the West-African population decrease being contemporaneous with the Black Plague in Europe is actually valid -- if so, the explanation for this not occurring in East-African populations does seem to makes sense...



The brief population decline dated at ~1340CE by MSMC coincides with the timing of the Black Death (1343-1353), however time resolution is low and the decrease was only observed at a single time point. There is evidence of abandonment of multiple large settlements throughout West Africa around the time of the Black Death and there is speculation that this was caused by the disease (Chouin2015). The decrease at this time appears to have impacted the West and Central African Niger-Congo but not the East African populations. Both Bantu and Nilotic populations in East Africa were cattle keepers and pastoralists to varying degrees (Chritz et al. 2015) and the concomitant lower population density and mobile lifestyle may have made them less vulnerable than the more settled and urbanised West Africans to plague infection.

[Lank]

It's puzzling why only one Nilo-saharan population was selected for this paper; an array of Nilo-Saharan speakers would've given the paper's title and conclusions more credibility and legitimacy
ADMIXTURE results for the Lugbara (UGN) are surprisingly pristine (blue=East-African African proxy?); most results for this population (*based on immediate Central-Sudanic Ugandan family members) are ~30-40% West-African/Bantu yet only minor traces (red/yellow) are shown on these K runs
Nearly half of the sampled populations have minor non-negligible Eurasian on K3 runs; is it fair to assume much of this ancient Eurasian admixture in Niger Congo-A speakers=North African ADMIXTURE?
the descriptions for the Yoruba and Esan (YRI,ESN) are incorrect, they've been conversely labeled on the ADMIXTURE chart :/
This is rather sloppy writing here: within the 2.0 Background section the paper omits classifying Central Sudanics as agro-pastoralists. The Lugbara are predominantly subsistence farmers and have been for centuries if not millennia; pastoralism along with fishing have been secondary for sometime.
If there's any merit to these dating methods (MSMC), I wonder if the West-African population decrease being contemporaneous with the Black Plague in Europe is actually valid -- if so, the explanation for this not occurring in East-African populations does seem to makes sense...

Completely agree with your remarks!

I would say that the Lugbara look less admixed than they really are because they are the most "pristinely East African" reference in this dataset. These Lugbara samples probably still have significant Niger-Congo-related admixture, which could be teased out with a more diverse sampling. But then it would be interesting to find out how much was already present in the ancestral Central Sudanic population, in contrast to recent Bantu admixture. This would not be solved by merely adding a Nilotic South Sudanese reference.

And yeah, the Eurasian signal is mainly seen in Gambia and (Mande) Guineans, the northernmost samples, and those countries also have significant Fula populations. So it probably represents North African ancestry.