I have always believed when it comes to DNA testing and “admixture” predictions (a.k.a. BGA or biogeographical testing) the proof of the pudding is when people who are “100%” from one particular ethnic background take the test. See how well their ancestry is being predicted or described and that already tells you a lot what you can expect for yourself 😉
On this page I will be posting the AncestryDNA results for Africans with confirmed ethnic background(s) within Africa. Unlike for Afro-Diasporans in the Americas or elsewhere these results can therefore be verified with known genealogy. This should be insightful on how reliable/predictive the various AncestryDNA regions can be and also how they might be interpreted.
Some results might appear to be “off” at first look for those not accustomed to how AncestryDNA or DNA testing in general works. Therefore please keep in mind the following disclaimers (for a more detailed discussion see the AncestryDNA Regions page):
- DON’T take the labeling of the regions too literally. The regional scores merely signal close genetic similarity to samples taken from the countries/regions after which they get named. They are not per se indicative of actual recent origins from those countries/regions. They could also imply very ancient ancestral connections (going back hundreds or even thousands of years) due to either migrations or shared origins.
- AncestryDNA is NOT meant to be taken as anything definite or final nor is it meant to be a substitute for your family tree. Rather it may be seen as a valuable starting point for exploring how your ancestry could be described regionally speaking.
The number of Africans on AncestryDNA used to be small but fortunately it has been increasing rapidly. Below some of their results are being featured. You can also see their breakdown in my spreadsheet via these links:
For other African AncestryDNA results see these pages:
- West African results (under preparation)
GHANA (Ewe from Peki/Volta region)
This is a very insightful even if perhaps counterintuitive breakdown for a Ghanaian person. The predominant score is afterall “Benin/Togo” combined with a smaller but still considerable proportion of “Ivory Coast/Ghana”. The socalled “Benin/Togo” region has been reported very frequently and also with high scores among African Americans and also West Indians. Often surprisingly so. I have no complete certainty about the ethnic background of the person whose DNA results are being shown above. However judging from his name and his family’s location in the Volta region of Ghana, nearby the Togolese border. And more specifically their hometown being Peki, a traditional Ewe state, this person could very well be an Ewe, an ethnic group living in eastern Ghana as well as southern Togo (see also this map).
If so this result could be very illustrative of how the country name labeling by AncestryDNA should not be taken as gospel. Almost all African countries have been colonial creations with borders cutting right through the homelands of ethnic groups. Also very closely related ethnic groups can often be found on both sides of the border. This is also the case for the Ewe people and the greater language group of the Gbe people they belong to. Also especially Togo and Benin have very artificially drawn borders (see Lower Guinea for more details).
To be sure valuable information can still be be extracted from this result. Just as long as you are aware of the relevant context! Actually Ancestry.com does a great job at describing this inevitable overlap across manmade borders by way of the additional information they provide when you click on each region featured in the Ethnicity Estimates (see also AncestryDNA Regions). However some people might not always be inclined to read the small print 😉 For more analysis and background information see:
The socalled “Ivory Coast/Ghana” region is capable of providing very accurate predictions of Akan ancestry. This is convincingly demonstrated in the screenshot above. Such a predominant score for just one single region is quite unique from what i have observed sofar, even among African results. This person of Akan background is being described as practically 100% “Ivory Coast/Ghana” if you ignore the miniscule Trace regions. Which is just what you would expect, even when such genetic statements are usually more complicated. It is important to realize that each of the nine African AncestryDNA regions have different predictive strenghts. Because of limitations in available sampling, differences in ethnic homogeneity found in each African region, distinctive population structure, overlap with other regions because of ancient migrations etc.etc. Especially the contrast with the “Nigeria” region is striking. Because of a greater genetic diversity found among Nigerians it turns out that additional regions (above trace level) are needed to describe their DNA, in particular “Benin/Togo” and “Cameroon/Congo”. See also this separate page featuring:
Even so, the socalled “Ivory Coast/Ghana” region also detects other types of ancestral origins from further west, beyond the Ivory Coast, into Liberia and even Sierra Leone. Especially for African Americans ancestral connections with Liberia and Sierra Leone would actually be much more likely than Ivorian ancestry, going by the Slave Voyages Database (see also this chart). Confusingly this is not mentioned by Ancestry.com themselves in their regional descriptions. However the results of an Ivorian (of mainly Akan background) as well as a Liberian and two Sierra Leoneans prove otherwise. I will post these results directly below. Judging from the Afro-Diasporan results again this region seems quite predictive and consistent. Jamaicans and other Anglo-Caribbeans are scoring the highest for “Ivory Coast/Ghana” followed by African Americans while it is less noticeable among Hispanic & French Caribbeans, in line with known slave trade history. See this chart and also:
IVORY COAST (7/8 AkanMalinké & 1/8 Krio from Sierra Leone/Gambia)
Judging from this 1 individual result the “Ivory Coast/Ghana” category seems to be VERY predictive also for someone of Ivorian descent! Just like the person from Ghana above this Ivorian is mainly of Akan descent, but not completely. The breakdown mentions 83% for “Ivory Coast/Ghana” but according to the range given it could be as high as 94%. Ancestry.com performs 40 separate estimates for each region and 83% would be the average of these 40 runs. 83% is also pretty much the same as the 86% scored by the “typical native” from Ivory Coast/Ghana according to Ancestry. This 86% being the median result of their 99 samples (see AncestryDNA Regions for more details).
I am also blown away by how the minor % of Nigeria seems to be pretty much in line even with this person’s 1 non-Ivorian great grandparent who is Krio, the descendants of ex-slaves from the USA, maroons from Jamaica as well as liberated slaves from all parts of West/Central Africa being settled in Freetown/Sierra Leone. It is known many of them were Yoruba so some minor Nigeria would fit well. The Trace Regions might also be explained by the very diverse Krio background but it could also just be noise, as their confidence level is lower than for the main regions.
LIBERIA (Dan or Gio, Mandé speaking ethnicity)
This person’s family is from a Mandé speaking ethnicity located in the border area with Ivory Coast and Guinea Conakry (scroll down for Liberia maps on this page). Interestingly only 17% of his ancestry is designated as “Mali” while “Ivory Coast/Ghana” is overwhelmingly predominant. So we could say that this Liberian person seems to be genetically similar to the result posted from Ivory Coast albeit with a noticeable “Malian” shift. Which is perhaps not very surprising given that Liberia and Ivory Coast are neighbouring countries with also overlapping ethnolinguistical groups. As there are no separate categories yet for Sierra Leone, Guinea Conakry or Liberia, it seems logical that any ancestry hailing from these parts will be described in terms of the existing regions. Somewhat early to speculate but i suppose for a Kru speaking person from the southern parts of Liberia their “Ivory Coast/Ghana” score might be even higher.
Even though the 17% “Mali” score might seem subdued, it is still showing up quite strongly when we take into account that 1) a “typical native” of Mali is only 39% “Malian” according to Ancestry (see also AncestryDNA Regions) 2) the Ivorian posted earlier has only 3% Mali while this mostly Akan descended person has also confirmed partial Malinké ancestry (=Mandé speaking group from Ivory Coast). So eventhough it shows up relatively minor, “Mali” might yet be a significant ethnic marker for this Liberian person. Still it is also intriguing to speculate what this may imply for the deeper ancestral origins of Mandé speaking groups in Liberia. It is known historically they were last to arrive in Liberia (Atlantic/Mel speaking groups being assumed to have settled first followed by Kru speakers from the east, see also this link). And they might have absorbed many previous inhabitants among their ranks by intermarriage and/or language shift.
Very fascinating to see the results of a person from Sierra Leone, as this country has not been included in any of the AncestryDNA regions yet. Even when many people from the Afro-Diaspora might have some degree of ancestry from this area. Because Sierra Leone does not have a category of its own it seems logical that neighbouring countries will show up to describe it in regional terms. And that is also what we are seeing above, for the most part anyways.
The main combination of “Ivory Coast/Ghana” and “Senegal” was to be expected therefore. It highlights yet again that the countryname labeling of the AncestryDNA regions is obscuring the way these regional DNA markers are also spread out across borders. We can now verify that “Senegal” can be found as far south as Sierra Leone, while “Ivory Coast/Ghana” goes all the way west beyond Liberia even.
I do not know the exact ethnic background of this person however the 15% “Nigeria” score seems to suggest some partial Krio or Recaptive lineage down the line. I have read that when the liberated Recaptives first arrived in Freetown some of them were taken as brides or apprentices by other ethnic groups further inland. Due to this unique historical circumstance (also valid for Liberia) samples from Sierra Leone should be scrutinized carefully because of the possibility of extensive inter-ethnic marriages. Because some of these unions involving either Krio or Recaptives might have taken place many generations ago already, some people might not even be aware and just self-identify as being from one single ethnic group.
SIERRA LEONE (Mende)
The main combination of “Ivory Coast/Ghana” and “Senegal” is quite similar to the result shown directly above for another Sierra Leonean. I had assumed that “Mali” might possibly also appear in noticeable proportions for Sierra Leoneans. But apparently not so for these two individuals. Compare for example with the 17% “Mali” of the Liberian person further above. Perhaps it is also correlated with ethnic origins. This person coming from a Mende background. This ethnic group is thought to have arrived relatively recent in Sierra Leone during the 1500’s/1600’s by way of Liberia (see this insightful article). Which would actually also explain why “Ivory Coast/Ghana” ended up as biggest region.
An intriguing score is the 4% “South Central Hunter-Gatherers”, even when being minor it is still at an elevated level when compared with results from Afro-Diasporans. The first Sierra Leonean result shows the exact same amount actually. Leaving aside the possibility of Congolese Recaptive ancestry i suppose it might signal DNA traces of absorbed Pygmy-like people once living in the Upper Guinean rainforest zone. This region also showed up for the Ivorian result above with 2%, proving that it is not exclusive to Southern and Central Africa as implied by the name.
I personally find the rather high “Benin/Togo” the most surprising component of this breakdown. I like to emphasize i am just speculating out loud here because this person has two confirmed Mende speaking parents. Still i have read it is quite normal for Sierra Leone to have inter-ethnic marriages occurring also in the past. And who knows perhaps this outcome is the result of such intermarriage from a couple of generations back, possibly involving a Krio of ultimately Yoruba, Ewe or Beninese background. Which resulted in a 20% “Benin/Togo” score , just like the previous Sierra Leonean received a score of 15% “Nigeria”. Either way it goes to show we shouldn’t underestimate the genetic diversity for Africans themselves as they are also often reflecting ethnically mixed backgrounds and ancient migrations if you go back far enough.
CAPE VERDE (Santiago island)
In a certain way Cape Verdeans can also be considered to be Afro-Diasporans, despite being located in West Africa. Afterall their mainland African ethnic roots are multiple and mostly resulting from the Atlantic Slave Trade. Plus generally speaking they will show some non-African admixture to varying degree. Nevertheless AncestryDNA testing has been very consistent with Senegal showing up as first region for nearly all Cape Verdean results i’ve seen sofar (see this sheet) and Mali usually coming in as second main region as is the case for the result above. This seems to be a quite solid confirmation of the overwhelmingly Upper Guinean origins for Cape Verde’s African ancestry (for more details see “Cabo Verde Raizes Na Africa“). Also i suppose Cape Verdean results can be seen as validating the prediction accuracy of the “Senegal” region which seems to pick up on quite distinct DNA markers not only from Senegambia but also Guinea Bissau. There’s usually also a variety of sometimes surprising Trace Regions showing up for Cape Verdeans but with reduced confidence level. For more analysis of Cape Verdean AncestryDNA results follow this link.
The person who took this test also made this excellent youtube video:
SAUDI ARABIA (Fula/Fallatah)
This person is not African but Saudi Arabian rather as her family was born there and she herself as well. However she is of Fula descent on both sides. Fula communities can be found all across the interior of West Africa, a.k.a. the Sahel region (see this page for maps). They can even be found all the way east in Sudan or like this person in Saudi Arabia, where they are also known as Fallatah or Fellatah (see this page for more info on Sudanese Fula or also this Facebook page for Saudi Fula). And the amazing thing is that despite centuries of migration all across the Sahel this person’s original Fula origins from Senegambia are still clearly reflected in this DNA profile! The original homeland of the Fula according to most historians would be the Senegal river valley and the above breakdown seems to confirm this theory very nicely judging from the clear majority of Senegal + Mali + North Africa %’s. It is an (extra detailed) corroboration of what was obtained recently in a DNA study about Sudanese ethnic groups, where the Sudanese Fula samples were also found to be distinct from the others. See Dobon et al. (2015, fig.3) or also this blogpost for more in depth analysis.
There’s a couple of striking things about these results for me:
- The Fula have been migrating extensively for many centuries and so it can be expected they have intermarried with other ethnic groups along the way to some degree. In fact this is known to have ocurred for especially urbanized and agricultural Fula speakers in Senegal itself (the Toucouleur), the Fulakunda or socalled Fula Preto in Guiné Bissau and the Hausa-Fulani in northern Nigeria. However the strictly nomadic branches of the Fula have always had the reputation of being much more endogamous and these results seem to be clearly in support. Still it’s only reflective of the family history of this single person and individual variation among Fula might still be greater in other cases.
- The relatively high North African proportion of 16% is in line with previous findings of possibly partial Berber or related origins for the Fula people. See also “The Fulani have an old “Berber” (?) element”. This element contributes to making Fula genetics quite distinct overall. Interestingly minor North African %’s have also been appearing for many Afro-Diasporans. For Hispanic Americans there’s a great chance it could also (in part) be derived from their Iberian/Canarian ancestry, however for others a small North Africa % could very well be traced back to a Fula ancestor! Although of course nothing’s set in stone and you will need additional clues to confirm 😉
- The pronounced Senegal + Mali percentages have been very useful in highlighting this person’s ultimately Senegambian origins. However for Afro-descendants you can’t make the sweeping generalization that any amount of Senegal or Mali would automatically be correlated with Fula origins. Afterall there are many other ethnic groups from Upper Guinea and Mali who might score similarly high for Senegal or Mali. In fact the previous breakdown of the person from Cape Verde looks very similar to this Fula breakdown. But given their long history of intensive creolization and intermixing of various people from across Upper Guinea it’s unlikely Cape Verdeans would be of exclusively Fula descent (see this overview). However a partial Fula contribution, significant even, is not be ruled out either of course. It’s noteworthy therefore that not only the predominant Senegal+Mali combination is mirrored (around 60% for both) but also to lesser degree North Africa is showing up (2% versus 16%) and even the Middle East (6% versus 17%).
- The minimal amounts of supposedly European (all Iberian), Asian (South Asia) and even Pacific Islander ancestry for this fully Fula descended person (afaik) are rather peculiar to say the least :-D. It goes to show that Trace Regions should always be taken with an EXTRA grain of salt, not only the geographical labeling might be off but also the estimates themselves usually include zero %! All sorts of speculative scenarios might apply but it could very well be that these trace %’s merely represent DNA segments which AncestryDNA finds hard to classify given it’s current reference panel. The 2% Southeast Bantu is interesting though as it might be indicative of minor East African connections.