West African Results (Upper Guinea)

Contents

  1. Introduction
  2. Summary of Findings
  3. Observations
    • “Senegal” & “Mali” predictive of DNA throughout Upper Guinea
    • “Mali” also predictive of DNA found in Burkina Faso and northern parts of Ivory Coast & Ghana
    • “Ivory Coast/Ghana” most notable among Sierra Leoneans
    • “Africa North” stand-out aspect of Fula genetics
  4. Screenshots & Youtube videos of West African AncestryDNA results (2013-2018)
  5. Is it possible to distinguish a genetic Upper Guinean component?
  6. Implications for Afro-Diasporans
    • Regional admixture DOES matter! Judge each case on its own merits. Combine insights from different fields to achieve complementarity!
    • How to make more sense of “Senegal” & “Mali” scores

1) Introduction

On this page I will be posting AncestryDNA results of West Africans with confirmed ethnic background(s)1. Unlike for Afro-Diasporans in the Americas these results can therefore be verified with known genealogy. This should be helpful to improve correct interpretation and also to determine how reliable/predictive the various African AncestryDNA regions might be.

This discussion of West African AncestryDNA results will be divided in two parts, each having its own page. Roughly corresponding with the categorization of Lower Guinea & Upper Guinea. And also in approximate alignment with a preponderance of either “Senegal” or “Mali” scores (see map 1.2) instead of “Ivory Coast/Ghana” and/or “Benin/Togo” scores among my West African survey participants. See also:

Please keep in mind that AncestryDNA’s ethnicity estimates have been updated several times now! On this page I am dealing exclusively with AncestryDNA version 2 which was current between September 2013 and September 2018. All matters being discussed on this page are therefore not pertaining to recently updated results (2018/2019). In my opinion especially version 3 (Sept 2018 – 2019) regrettably has been a downgrade rather than providing any meaningful improvement when wanting to learn more about one’s African origins. For more details see:

Senegal, Gambia, Guiné Bissau, Guinea, Sierra Leone, Mali & Mauritania

Map 1.1 (click to enlarge)

***

Map 1.2 (click to enlarge)

Senegal + Mali

“Senegal” and “Mali” regions. Source: Ancestry.com

***

Specifically on this page I will focus on results from Senegal, Gambia, Guiné Bissau, Guinea, Sierra Leone, Mali and Mauritania (shown in green in map 1.1). I will usually refer to these countries combined by using the term Upper Guinea (see this page for maps). Nigerian results have already been discussed on a separate page. Results from the remaining West African countries (shown in red in map 1.1): Liberia, Ivory Coast, Burkina Faso, Ghana, Togo and Benin have been discussed in the first part of this exploration of West African DNA results. If you want to skip the discussion just scroll to section 4 to see screenshots of individual West African AncestryDNA results.

As these results below show many West Africans are in fact of “mixed” background if you go along with the regional framework provided by AncestryDNA. As I will continue to say it is counter-productive to get distracted by the country name labeling. Rather consider the AncestryDNA regions to be proxies of ancestral components which have become more frequent in certain loosely defined areas but still show a wide dispersal in neighbouring areas as well due to ancient migrations and inter-ethnic unions occurring probably since the dawn of mankind! If you invest some time in learning how to truly “read” the results they can still be very informational in my opinion.

I like to thank all my West African survey participants for having tested on AncestryDNA and sharing their results so that it may benefit other people as well!See also:

_____________________________________________________________________________

2) Summary of Findings

 Selection of Upper Guinean AncestryDNA results (2013-2018)

***(click to enlarge)

COMPILUG

***

In my personal estimation AncestryDNA can report valuable information about your ancestral origins as long as you interpret the data correctly and as long as you are also aware of the relevant context and inherent limitations. Still these 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:

  • Don’t take the country name labeling of the regions too literally. The regional percentages firstmost signal close genetic similarity to the samples taken from the countries after which the regions have been named. Actual origins from neighbouring countries should not be ruled out.
  • Almost all African countries have been colonial creations with borders cutting right through the homelands of ethnic groups. Closely related ethnic groups can often be found on both sides of the border.
  • Most DNA is common in many populations, just at different frequencies. Due to either migrations or shared origins dating back from hundreds or even thousands of years. For this part of West Africa in particular the migrations associated with speakers of the Mandé and Atlantic language groups are a factor of paramount significance.
  • Sometimes (very) ancient admixture rather than any recent lineage is suggested by these AncestryDNA estimates. And in this way some of these regional scores might perhaps best be seen as genetic echoes from a (very) distant past. Beyond family tales, historical knowledge or even ethnic origin traditions.
  • Ethnicity is a construct which evolves across time due to ethnogenesis. Generally speaking therefore ethnic groups do not possess unique DNA markers. Especially in comparison with neighbouring ethnic groups or from within the same wider region. The most common scenario being a genetic gradient which causes ancestral components to gradually fan out. As can be verified from my survey.
  • However by closely studying the regional combinations being reported for West Africans we can still learn a great deal. In-spite of individual variation group averages do tend to provide more solid ground to make meaningful inferences when specifying someone’s ancestral origins. Also finding out where a specific region is most prominent or rather most subdued holds valuable lessons.

 West African group averages

Table 2.1 (click to enlarge) 

Stats, SEN, n=34

***

The statistical data displayed in Table 2.1 is based on a limited sample size. Although all countries within the Upper Guinea area are being covered now (incl. also Guiné Bissau!). While the sample size for Senegal & Gambia, Sierra Leone and especially my Fula (a.k.a. Fulani3) sample group is quite decent actually. These Fula results are from various countries across Upper Guinea. Which should contribute to a more robust data-set. All the more given that my combined survey group (n=122) contains a number of people which is almost three times greater than Ancestry’s Upper Guinean Reference Panel (n=44; 28 samples being used for “Senegal” + 16 samples for “Mali” in the 2013-2018 version).

All of the included countries themselves do harbour a multitude of ethnic groups. Right now however I can only make statistic calculations for a few relatively well represented ethnic groups within Ancestry’s customer database. It should be obvious therefore that a greater degree of genetic diversity and individual variation might be expected across West Africa and also within the listed countries.

This diversity is already clearly demonstrated by the Fula results representing almost 40% of my survey group (46/122). Although “Senegal” is primary for them their additional “Africa North” and “Middle East” regional scores are quite distinctive or even unique in this West African context! I will  discuss my Fula findings in greater detail in sections 3, 4 and 5. Within my Senegambian survey group the ethnic differentiation between my Wolof & Jola (from around the Casamance area) samples is more subtle. Possibly indicating a greater degree of shared ancient origins.

But among my Sierra Leonean samples again some noteworthy differentiation can be observed. The relative balance of “Senegal” versus “Ivory Coast/Ghana” possibly being a rough indicator of either northern (incl. Temne) or southern (incl. Mende) origins. While for my 3 Krio survey participants a much greater regional variety can be seen, which peaks in “Nigeria”. Despite very minimal sample size already mostly in line with their historically known origins! It is also noteworthy how so-called “South-Central Hunter-Gatherers” (labeled “Pygmy/San” in my survey) is so far consistently showing up for results from Sierra Leone, at a similar level to my Liberian findings (3%). For more discussion:

Eventhough only preliminary statements should be allowed at this stage I highly suspect that these group averages are already a reasonably good approximation of the main regional components to be found within the genepool of this part of West Africa. Atleast according to AncestryDNA’s reference panel used for its 2013-2018 version. The group averages I have calculated being more or less in line with how the “typical native” scores according to Ancestry’s information: respectively 100% for “Senegal” (n=28) and 39% for “Mali” (n=16).

However also other regions are frequently reported. Either as minor genetic components with lower confidence but often also with substantial amounts above trace level. Nearly always neighbouring and genetically overlapping West African regions are involved. These extra regions may be unexpected at first sight for people who according to their own knowledge are “100%” Senegalese, Gambian, Guinean, Malian, Sierra Leonean etc.. Taking these results at face value can therefore be misleading without correct interpretation. 

The disclaimers I already mentioned above as well as the next section below should provide sufficient clarification. As actually this outcome does overall still make sense. In short the advise would be to: don’t overfixate on the labeling of ancestral categories! In fact this does not only go for AncestryDNA but any kind of admixture analysis. In order to avoid jumping to premature conclusions I highly recommend that you atleast browse through some of the topics mentioned in the following links:

_____________________________________________________________________________

3) Observations

“Senegal” & “Mali” predictive of DNA across Upper Guinea

Map 3.1 (click to enlarge)

Senegal (Ancestry)

Source: ancestry.com (text in red added by myself)

Map 3.2 (click to enlarge)

Mali region

Source: ancestry.com (text in red added by myself)

***

As expected the highest “Senegal” scores (>80%) have been obtained for persons from either Senegal or Gambia. With the maximum score (97% “Senegal”) being reported for a possibly Wolof person. As can be seen in Table 2.1 the group average of “Senegal ” was also highest (84%) among 6 Wolof results. To be sure this does not mean that “Senegal” is an indicator of actual Wolof or even Senegambian descent! Other ethnic backgrounds might still be perfectly compatible with high “Senegal” scores. However this finding does clearly corroborate that AncestryDNA is indeed using samples taken from Senegambian populations to calculate their so-called “Senegal” estimates. Most likely Mandenka ones from the HGDP database. Although I have no certainty about this. And regrettably I was also not able to countercheck because as far as I know I have not included any Mandenka survey participants.

Ancestry mentioned that the predictive accuracy of “Senegal” was 100% for their own 28 reference samples (“typical native”). Quite extraordinary! But probably because these samples themselves were very genetically homogeneous and taken from one particular location. Naturally such a perfect predictive power was not attainable for other persons. Nonetheless in my overall AncestryDNA survey (both African & Afro-Diasporan) “Senegal” did prove to be one of the most reliable and distinctive regions. Despite the relatively small number of underlying samples.

But again “Senegal” is not to be seen as an indicator of specific ethnic lineage or a colonially derived nationality! Instead it is perhaps best to look at the presumable correlation with the spread of Atlantic speakers across Upper Guinea (see this map or also this one). And even beyond due to Fula migrations across the Sahel. To some extent there might also be a correlation with the spread of Mandé speakers. But to a lesser degree and possibly only with certain subgroups within the greater Mandé language family. Mostly because of overlap with “Mali” and “Ivory Coast/Ghana”. As can be verified from table 2.1 high “Senegal” group averages (>30%) can be found in several other countries aside from Senegal and Gambia. Going by high individual scores (>50%) this genetic cluster appears to be extending into western Mali (67% for a Soninké person) as well as beyond Guinea into Sierra Leone (55% for a possibly Temne person).

The relatively elevated “Senegal” level among the Temne (Atlantic speakers from northern Sierra Leone) especially being evocative. Also in light of the former trading zone called “Guiné de Cabo Verde” going from the Senegal river in the north to the Sierra Leone peninsula in the south (see this map). With the Temne being a significant part of Cape Verdean ethnogenesis (see “Sape” entry on this page). Very nicely corresponding therefore with the geographical distribution of the so-called “Senegal” region on Ancestry! The somewhat reduced level of “Senegal” among southern Sierra Leoneans (incl. Mende) possibly to be explained by a greater pull of “Ivory Coast/Ghana”. Due to them being southwest Mandé speakers who quite likely also absorbed a great deal of native populations while migrating through Liberia before settling into Sierra Leone. 

Generally speaking high “Senegal” scores were still always combined with secondary “Mali” scores as well.  In a way reinforcing their mutual predictive accuracy when taking a so-called macro-regional approach. Basically combining interrelated & neighbouring genetic regions in order to allow certain regional patterns to show up more clearly. Simply put “Senegal” +”Mali” = indicative of generic Upper Guinean lineage. Such a  perspective can be very valuable even when it lacks finer resolution. Especially when combined with other insights. For continued discussion see section 5 & 6.

“Mali” also found in Burkina Faso and northern parts of Ivory CoastGhana

Figure 3.1 (click to enlarge)

Malicompil

These results show some of the highest “Mali” scores in my survey (53% being the maximum). Not really being extra pronounced anywhere but rather reaching a higher level for southern Mali and areas directly across borders in Guinea but also in Burkina Faso, northern Ivory Coast and Ghana.

***

Going by Ancestry’s own info “Mali” was its least robust region in their 2013-2018 African breakdown. Only 16 samples being used and the “typical native” only scoring 39% for “Mali”. In sharp contrast to the much more well-defined “Senegal” region. Ironically among my also quite minimal Malian sample group (n=6) I calculated a nearly exact same group average of 38% “Mali”! The additional main regional components being “Senegal” and also “Ivory Coast/Ghana” for these six Malians.

Although quite exceptional throughout my survey I still did observe a few primary “Mali” scores among Guineans and Sierra Leoneans as well. But otherwise as already mentioned “Mali” was actually quite widespread across Upper Guinea. Usually in subdued but still clearly detectable amounts. At the beginning of my survey I speculated that “Mali” might attain greater prevalence among Sierra Leoneans. However this expectation has not really materialised. Although perhaps also due to a lack of samples with a relevant ethnic background (perhaps from northeastern Sierra Leone?).

Aside from this relative peaking in Mali, Guinea and Sierra Leone the so-called “Mali” region was also reported in high amounts for my samples from Burkina Faso and northern parts of Ivory Coast & Ghana. Again only supported by minimal sample size. But as I have always suspected the “Mali” region is also predictive of Gur and closely related Senoufo lineage. Which is not that surprising as actually such populations also reside within Mali’s borders, aside from Burkina Faso and the northern parts of Ivory Coast and Ghana. As much is also confirmed by Ancestry’s own information as shown in map 3.2

Quite likely the 16 Malian samples used by Ancestry were Bambara as well as Dogon. Bambara samples from south-central Mali providing a link for Mandé speaking populations across Upper Guinea. While additional Dogon samples from southeast Mali would probably be causing more genetic similarity for Gur and Senoufo speaking populations within & surrounding Burkina Faso. Aside from the minimal number of samples (n=16) such a possibly double ethnic basis might also explain the relative lack of focus of “Mali” (2013-2018 version). See also these links:

“Ivory Coast/Ghana” most notable for Sierra Leoneans

Map 3.3 (click to enlarge)

IvcGhana region

Source: ancestry.com (text in red added by myself)

***

The “Ivory Coast/Ghana” region (2013-2018 version) has been most predictive for Ghanaians, Ivorians and Liberians. However as my survey has demonstrated actually so-called “Ivory Coast/Ghana” is also being reported with substantial amounts in Sierra Leone. Much for the same reasons as for Liberia, its neighbouring country. After all Sierra Leone is also not appearing with its own region on AncestryDNA. The Sierra Leone context is however more complex because it shows a greater Upper Guinean influence (“Senegal” and “Mali”) compared with Liberia.

As mentioned already the relative balance of “Senegal” versus “Ivory Coast/Ghana” may possibly be seen as a rough indicator of either northern (incl. Temne) or southern (incl. Mende) background (for Sierra Leoneans themselves!). My survey participants of presumably Mende or otherwise southern Sierra Leonean background (n=4) having a more pronounced “Ivory Coast/Ghana” group average of around 50%. While for people of a Temne or otherwise northern Sierra Leonean background (n=12) a more reduced level of around 35% was obtained (see table 2.1).

Given known differences in migration history and ethno-linguistic background within Sierra Leone this could be a quite insightful finding. Especially the origins of the Mende have been debated extensively by historians. In light of their migration route by way of Liberia and also the probably high degree of mixing with native populations across the generations the increased level of “Ivory Coast/Ghana” for them does seem appropriate. It is worth repeating however that these are only sketchy research outcomes. Keeping in mind also that a plausible ethnic background was mostly established on an educated guess basis (surname & other profile details).

Such outcomes might perhaps not be replicated with greater sample size. Or when using a different constellation of reference samples and/or using a different algorithm to establish regional admixture. Still it is very interesting to compare with the new “Ghanaian, Liberian and Sierra Leonean” category on 23andme. Although based on a minimal number of samples again northern Sierra Leoneans appear to have a more noteworthy Upper Guinean component as measured by “Senegambian & Guinean” (see this page). For more detailed background info:

The “Ivory Coast/Ghana” region is also frequently showing up (above trace level) in the results of other West African countries. As far north as Senegal and as far east as Nigeria (see map 3.3). However much more subdued and practically never as primary region (except for one possibly Kissi person from Guinea). Unlike what I have observed for Sierra Leone therefore (see my spreadsheet). This goes especially for my survey participants from Guinea, Guiné Bissau as well as Mali (group averages all being >10%).

This wider distribution is possibly to be explained by the Ivorian reference samples being used by Ancestry. Although I have no certainty about it there might be Yacouba (south Mandé speaking) samples among them. Which would account for the genetic affinity being found also across the wider Upper Guinea area. And in particular it would explain the relative peaking among Mende individuals. Mende being a related southwestern Mandé language after all. See also page below for more discussion:

 

“Africa North” stand-out aspect of Fula genetics 

Figure 3.1 (click to enlarge)

Fulacompil

These Fula results are hailing from several countries: Senegal, Gambia and Guinea. In spite of some proportional variation it is very noteworthy how “Africa North” is consistently showing up as a main region (above trace level). “Middle East” is actually also quite consistent but usually with lower amounts. And sometimes also being reported as Low Confidence region or not at all.

***

My Fula (a.k.a. Fulani3) survey group has been quite outstanding in several aspects. First of all because I have been very lucky to gather a relatively big number (n=46) of Fula results.4 They represent the largest ethnically specified subgroup within my entire African survey! Furthermore they are from various countries, all over the wider Upper Guinean area (Senegal, Gambia, Guiné Bissau, Guinea, Sierra Leone and Mauritania).  See also my online spreadsheet for viewing all the individual results:

All results show a predominant “Senegal” and “Mali” combination, underlining the principal Upper Guinean origins of the Fula. This genetic component is also clearly detectable among people of (partial) Fula descent, such as the Hausa-Fulani. Nowadays living in other countries such as Nigeria or even further away like Sudan or even Saudi Arabia! A remarkable testimony of the Fula migrations all across the Sahel. But also indicative of the predictive power of especially “Senegal” in Ancestry’s 2013-2018 version. See section 5 for continued discussion.

Equally remarkable is how additional “Africa North” scores serve as a distinguishing marker for my Fula survey group. This is very obvious when comparing them with other non-Fula results from West Africa. “Africa North” amounts being practically absent among other West Africans in my survey. Unless they happen to be of mixed Fula descent or also of recent partial Moroccan descent.5

The Fula group average for “Africa North” was about 13%, which is minor but still quite substantial. The maximum score of “Africa North” among my 46 Fula results being 24% while the lowest amount was 5% (see this table). In other words this genetic component was consistently reported for all my Fula survey participants. Again a rather unique finding in my West African survey. In line with earlier preliminary outcomes based on admixture analysis. See for example:

In fact also minor “Middle East” scores were very common. But always with smaller amounts, often below trace level. And a few times even absent. So less consistent over all. Furthermore also quite common but with even more minimal trace amounts so-called “Iberian Peninsula”, “Europe South” and even “European Jewish” scores at times appeared among my Fula results. Generally speaking such regional scores are not indicative of recent genealogical ties with either Europe or the Middle East!Rather such scores are probably best to be seen in combination with “Africa North” as representing a (pre-) historical genetic link with North Africa and/or North African-like populations from a more distant past (“Green Sahara” era). “Africa North” not being a fully accurate region to describe North African lineage when going by actual North African results. See also:

The origins of the Fula people have been widely debated. And this topic tends to fascinate people for several reasons. Scientific research continues to reveal the distinctiveness of Fula genetics.7 Nonetheless the primary “Senegal” and “Mali” amounts observed during my survey (combined >70%) seem to be a quite solid indication of their principal Upper Guinean lineage. As also corroborated by historical record and their language being most closely related to other Senegambian (a.k.a. Atlantic) languages such as Wolof and Serer.

But their additional and consistent “Africa North” component demonstrates that there is more to their story! Despite recent advances being made there is still much left to uncover about this particular aspect of Fula ethnogenesis. The minor but still noteworthy “Southeastern Bantu” amounts among my Fula survey participants then also to be taken into consideration. With a group average of almost 3% and max. score of 11% these scores do not seem negligible (see this table). Possibly indicative of ancient Northeast African connections or else backward migrations from central and/or eastern Sahel back to Upper Guinea during historical periods? Follow these links for further reading:

_______________________________________________________________________

4) West African AncestryDNA results

The following screenshots of West African AncestryDNA results form the basis of my survey findings together with results which I obtained by way of the Compare Ethnicity tool. They are grouped by country and when possible also by ethnicity. The backgrounds behind these West African results were verified by me in the best way I was able to. Often based on statements made by the persons who did the AncestryDNA test themselves. But in many cases I also distilled a plausible ethnic/national identity by way of other clues (usually family names and family locations). In these latter cases I have added a question mark in the header. Therefore a 100% accurate depiction of self-identification is not intended.8

Disclaimer:

I am aware that ethnicity can be a sensitive topic. Aside from wanting to learn more about my own West African roots my motivation to research these so-called Ethnicity Estimates is purely scholarly. It stems from my deep fascination with West Africa’s ancient population migrations, its many connections with the Afro-Diaspora as well as a profound love for its vibrant & diverse culture. I do not condone the misuse of my research for identity politics! I am in full support of democratizing knowledge. However when parts of my work are being copied I do expect that a proper referral to my blog and the additional context it provides will be made. Please also keep in mind that DNA testing can be very educational and may have many positive effects. However in some cases it may also be abused by people with bad intentions.

Results from Senegal & Gambia

Table 4.1 (click to enlarge)

Stats, SG, n=34

***

Map 4.1 (click to enlarge)

senegal_gambia_pop_1972

See this page for more detailed ethno-linguistic maps of Senegambia.

***

In table 4.1 more detailed statistics can be seen for my Senegambian sample group (from both Senegal and Gambia). While reviewing the data do keep in mind that averages tend to hide underlying variation. It is always advisable to take into account other measures as well. Such as the median and especially the minimum & maximum values to get a sense of the range of the scores. Naturally also the sample size (mentioned in the row labeled “Number”) is essential to place this data in better perspective.

We can verify that for all Senegalese and Gambian results “Senegal” appears as primary region. In other words showing up with the biggest amount in a person’s breakdown. And in almost all cases “Mali” shows up as secondary region. There is some slight variation between my Wolof (n=6) and my Jola (a.k.a. Diola or Dyola) samples (n=3) (see table 2.1). The Wolof being a major ethnic group from the north of Senegambia and the Jola being located in the south (Casamance). But really given the minimal sample size this differentiation does not seem significant. Instead my overall survey findings for Senegal and Gambia seem to suggest a great deal of shared ancient origins and relative homogeneity. Notable exception being my Fula results from Senegambia which have not been included in table 4.1 and will be discussed further below.

I will be posting these Senegambian results grouped according to known ethnic background and ranked according to main regional score. Despite the obviously limited number of samples an insightful pattern seems to arise already. As perhaps expected “Senegal” peaks among my Wolof samples. But this is not an actual confirmation of Wolof or even Senegalese lineage! As shown in table 2.1 so-called “Senegal” is pervasive throughout the wider Upper Guinea area, and by no means exclusive to any country or ethnic group. As always it pays to be cautious and resist jumping to conclusions. See also sections 5 & 6. For further reading on Senegambian ethnicity:

***

SENEGAL (Wolof)

***

SG7

***

Very convincing Upper Guinean profile. The African breakdown practically only consisting of “Senegal” and “Mali”. The highest “Senegal” score I observed was actually 97% for a quite likely Wolof person (judging by his surname). A DNA match for one of my Cape Verdean survey participants.

***

GAMBIA (Wolof)

***

***

Again clearly a Upper Guinean composition (“Senegal” + “Mali” = 96%) with minimal trace regions. Wolof people are also to be found in Gambia. This small country is usually associated with Mandinka people. But Gambia is very much a multi-ethnic country with nearly all its ethnic groups also to be found in Senegal. See page 33 of this paper for a ethnic distribution of Gambia.

***

SENEGAL (Wolof)

***

SG6

***

Despite being of confirmed Wolof decent these results might appear to be a bit atypical due to the 15% “Ivory Coast/Ghana” score. The highest such score among my Senegambian survey group (n=34). Most people actually scoring 0% for this region (median: 0.3%, see table 4.1). Several explanations might be valid. Also involving inter-ethnic unions from perhaps several generations ago, beyond family recollection. Either way in itself the reporting of “Ivory Coast/Ghana” among people from the wider Upper Guinean area was not uncommon in my survey. Possibly caused by a subset of Ivorian samples used by Ancestry which might have been of Mandé background. See also previous discussion in section 3.

***

GAMBIA (Jola?)

***

SG1 (Jola)

***

Although not confirmed it seems quite likely that these results belong to a person of Jola background. Intriguingly the Jola people from southern Senegambia (Casamance) are sometimes said to be the oldest inhabitants of that area. Also fascinating in that regard is the minor but still detectable 2% “South-Central Hunter-Gatherer” score. Such scores being practically absent among my Senegambian survey group, except for other possibly Jola results. Further south (Guiné Bissau-Sierra Leone) I did however observe such scores on a regular basis. See also:

***

SENEGAL (Wolof, Serer, Fula)

***

SG5

***

This person is of mixed ethnic background. Which nowadays is pretty common for many Senegalese families. Especially when living in cities. His partial Fula background is very easily detectable by the 8% “Africa North”. As well as the 2% “Middle East” and <1% “European Jewish” actually. Combined reflecting a distinctive feature of Fula genetics. Which is otherwise not seen among my non-Fula Senegambian survey participants.

***

SENEGAL (?)

***

SG4

***

I have no confirmation about this person’s ethnic background. But it seems quite likely that she might also be of partial Fula descent, similar to the previous result.

***

GAMBIA (Wolof, Fula and Aku?)

***

SG3 (Wolof, Aku)

***

Again a person of mixed ethnic background. It seems she might be primarily of Wolof descent though. The Fula family line is from a few generations ago. Which would be in line with the “Africa North” amount being somewhat diluted. It is still clearly detectable though. Not shown in the screenshot but she actually also scored 1% “Middle East”.

Intriguingly this person is also aware of ancestors with English surnames. Which seems odd for a country like Gambia where almost all people carry native African surnames. However not so the so-called Aku, a sub-group of the Sierra Leonean Krio! They are only a small minority in Gambia (<2%). But have been quite influential around the capital of Banjul especially. I imagine many of them intermarried with native Gambians across the generations. Possibly this partial Aku lineage might be accounting for the otherwise atypical 12% “Benin/Togo” score. It being known that many Aku are actually of Yoruba descent this outcome would then make perfect sense. As after all my Nigerian survey has demonstrated that this region is very commonly reported for Yoruba, as primary region even. A thorough search for Nigerian matches might provide more clarification. More info on the Aku:

***

GAMBIA (Wolof) & SIERRA LEONE (Krio, Fula)

***

SGSL

***

Quite similar background to the previous result. But this time I actually received confirmation of a partial Krio family line from Sierra Leone. Which is why I have not included these results in my Senegambian group averages. Still very interesting to see that yet again the partial Fula lineage can be recognized by the 4% “Africa North” score as well as the 1% “Middle East”. The 41% “Senegal” + 13% “Mali” amounts are likely to be associated with this person’s Wolof side firstmost. But also partially inherited by way of her Fula grandfather. The quite elevated 27% “Ivory Coast/Ghana” possibly mostly reflecting her Krio side. While the 10% “Southeastern Bantu” score is more puzzling as this is a very wide ranging region. Intriguingly according to this person’s family traditions some East African lineage was expected though.

***

_______________________________________________________________________

Fula results 

Table 4.2 (click to enlarge)

Stats Fula (n=46)***

Table 4.3 (click to enlarge)

Stats Fula

***

Map 4.2 (click to enlarge)

Fula spread

 The Fulfulde/Pulaar language is closely related to the Wolof and Sereer languages spoken in Senegal. Historically it is known that the Fula originated from Futa Toro along the Senegal river valley. However due to ongoing migrations across the centuries the Fula people have expanded all the way east into Sudan. Often acquiring new local names such as Fulani (which is originally a Hausa term) and Fellata (a Kanuri term used in Chad and Sudan)

***

In table 4.2 more detailed statistics can be seen for my Fula (a.k.a. Fulani3) sample group hailing from various countries across Upper Guinea. While reviewing the data do keep in mind that averages tend to hide underlying variation. It is always advisable to take into account other measures as well. Such as the median and especially the minimum & maximum values to get a sense of the range of the scores. Naturally also the sample size (mentioned in the row labeled “Number”) is essential to place this data in better perspective.

I have been able to collect a wide array of Fula results from practically each country within the Upper Guinea area and even beyond (Nigeria and Saudi Arabia). This has allowed me to compile table 4.3 which shows group averages split up per country9. Which is quite extra-ordinary because as far as I am aware academic studies dealing with Fula genetics rarely attain such a variety of samples. Often focusing on Fulani results from the Central/Eastern Sahel actually (Burkina Faso, Nigeria, Chad). Even when Fula results from Upper Guinea might be assumed to be most relevant to understand the earlier ethno-genesis of Fula populations. Given minimal sample size and incomplete knowledge about my survey participants full background this overview is naturally very much preliminary and merely indented as an approximation! But it might still already hint at some insightful patterns.

The consistency of predominant “Senegal” + “Mali” scores (>70%) combined with additional “Africa North” scores (usually >10%) is pretty remarkable to start with. But still some proportional variation seems meaningful already. Possibly indicative of the degree of intermixing between the Fula and neighbouring ethnic groups. With my samples from Guiné Bissau seemingly being most influenced by non-Fula geneflow. Going by their lower “Africa North” group averages and more elevated “Senegal” and also “Ivory Coast/Ghana” scores. I have actually excluded a quite likely Fulakunda result from this Fula survey because his “genuine” Fula lineage seems quite diluted (his screenshot is posted further below under Guiné Bissau).

In table 4.3 I have also made a comparison between my western Fula survey participants and persons with a most likely Hausa-Fulani background. Contrasting their group averages might be instructive to illustrate the genetic impact of eastward bound Fula migrations into Nigeria and also their intermingling with local populations. My Hausa-Fulani survey group does indeed mirror both the Upper Guinean and North African regional scores of my western Fula survey group quite well.

Given their (presumed) partial Hausa background it seems appropriate that their combined Upper Guinean component (“Senegal” + “Mali”) seems to be about half of my western Fula survey group (30% vs. 71%). Interestingly much less distance between their “Africa North” group averages (11% vs. 13%). Perhaps suggesting an additional (Tuareg?) source for the Hausa-Fulani. But the striking difference is the additional and primary “Nigeria” component (45%) for the Hausa-Fulani. Which is quite minimal among my Fula survey participants from the wider Upper Guinea area. Highlighting how Ancestry’s 2013-2018 version, despite its shortcomings, was still quite useful already to make even these finer ethnic distinctions. For more discussion of Hausa-Fulani AncestryDNA results see sections 2 & 4 of this page:

I have already discussed in section 3 how especially Africa North” has been a distinguishing marker of Fula results within my West African survey. In combination also with lesser “Middle East” and minimal but still non-trivial “South European” scores. Follow links below to check individual results and see an extended version of table 4.2:

It has also been mentioned already that the results of my Fula survey group still consist of a predominant combined share of “Senegal” & “Mali”. Indicating their principal Upper Guinean origins. This is something I will discuss in greater detail also in following section 5. Because this finding seems to corroborate the predictive accuracy of these regions.

My Fula survey findings would not have been possible without the tremendous help of my Fula friend. He not only shared his own insightful DNA results with me but also of his closest DNA cousins. These matches could be identified as very likely to be Fula because of shared DNA amounts, plausible regional admixture, surnames and other relevant profile details. In addition he also shared his profound thoughts with me about DNA testing and the ethnogenesis of the Fula people. I have had many stimulating discussions with him which greatly inspired me. Merci mille fois!

There are many online sources to be found about the Fula people. Regrettably not all of them are always reliable though. Just to provide a small selection:

***

FULA  (Senegal/Mauritania)

***

FULASEN2

***

This person’s family is from the Senegal river valley which forms the border of northern Senegal with southern Mauritania. This region is also known as Futa Tooro, often claimed to be the original homeland of the Fula people. Appropriately this breakdown features the highest “Senegal” score I observed during my Fula survey. Still also a noteworthy 9% “Africa North” score and 2% “Europe South” (most likely indicative of very ancient Trans-Gibraltar geneflow, see section 3) but “Middle East” being absent.

***

FULA  (Senegal)

***

FULASEN

***

FULA  (Gambia?)

***

FULA2

***

No confirmation of this person’s exact background. The rather high “Benin/Togo score seems quite atypical and was the second-highest in my Fula survey. Possibly indicative of some inter-ethnic union from a few generations ago.

***

FULA  (Guinea?)

***

FULA1

***

FULA  (Guinea)

***

FULAGC1

***

Might just be individual variation but still interesting to see the somewhat elevated level of “Mali” appearing in these three Guinean Fula results directly above and below. According to Ancestry’s own info the “Mali” region also being prevalent in Guinea. Although actually the highest “Mali” average was obtained for my 3 Fula results from Sierra Leone (22%, see table 4.3). Also noteworthy “Southeastern Bantu” score above tracelevel.

***

FULA  (Guinea)

***

FULAGC3

***

FULA  (Senegal & Gambia)

***

FULASG

***

One of the highest “Africa North” scores in my Fula survey. Although actually also several people scored higher than 20% and the maximum score was 24% (see table 4.2). Combined with a considerable “Middle East” score. As these two regions tend to be interrelated.

***

FULA  (Guinea)

***

FULAGC2

***

Intriguing “Southeastern Bantu” score. The highest within my survey but in fact similar amounts (above trace level) were reported also for some of my other Fula survey participants. It might just be a fluke on Ancestry’s part. But with a group average of almost 3% and max. score 11% these scores do not seem negligible (see table 4.2). Possibly indicative of ancient Northeast African connections or else backward migrations from central and/or eastern Sahel back to Upper Guinea during historical periods?

***

FULA  (Saudi Arabia)

***

FULASA

***

Very special breakdown for several reasons. This was the very first Fula result shared with me, already four years ago! It took me quite a while to gather additional results. Even when it seems that DNA testing has become particularly popular with Fula people when compared with other West Africans (see this page).

This person is actually not West African-born but Saudi Arabian rather and her family was born there as well. However she is of Fula descent on both sides. Most likely from somewhere in the wider Upper Guinea area with migration possibly taking place in the late 1800’s. Fula communities can be found all across the interior of West Africa, a.k.a. the Sahel region but also beyond! (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 also:

And the amazing thing is that despite centuries of migration all across the Sahel this person’s origins from Upper Guinea 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” + “Africa North” %’s. It is an (extra detailed) corroboration of what was obtained in a DNA study about Sudanese ethnic groups. In which the authors found that their Sudanese Fula samples were clearly distinct from the others. See Dobon et al. (2015, fig.3) or also this blogpost for more in depth analysis.

One give-away aspect though of this person’s eastern-bound migration and possibly mixed family history appears to be the relatively high 17% “Middle East” score. I have actually not included this person’s results in my western Fula survey (because of her Saudi background). But otherwise this score would have been the highest (11% being the maximum, see this table). Unlike what was reported for all my western Fula results in this case “Middle East” is also (slightly) greater than “Africa North”. The 4% “Iberian Peninsula” is noteworthy too but most likely indicative of very ancient Trans-Gibraltar geneflow (see section 3). The minuscule amount of “Asia South” might seem like noise but I suppose given some partial Gulf Arab admixture might actually turn out to still be genuine. A thorough scan for South Asian DNA matches might provide more clarity.

***

FULA  (Senegal, Guinea and Mali)

***

FULAGCSEN2

***

Another illustrative breakdown. It features one of the highest “Mali” scores in my Fula survey. The maximum score being 39% for the only Fula result which also featured “Mali” as primary region (see table 4.2). Interestingly this person is aware of actual Malian lineage! Although his parents are from Senegal and Guinea. His grandfather is actually from the Macina (a.k.a. Maasina) region which has a numerous Fula population (see this link). Also one of his 2nd great grand mothers was Bambara from Mali. As discussed in section 3 the so-called “Mali” region did not have a robust foundation even when generally speaking it has been indicative of Upper Guinean/western Sahelian origins. But in this particular case some correlation with actual Malian lineage seems quite likely.

Another striking aspect about this breakdown is that aside from the above average “Africa North” and the relatively low “Senegal” amounts no other African regions are being reported (also not as trace region). Grouping “Senegal” and “Mali” together (76%) this seemingly makes for a “perfect” Upper Guinean component (at least when going by AncestryDNA’s imperfect format 😉 ). This stands in some contrast with the other Fula results in my survey for whom additional African scores were reported. Especially so-called “Ivory Coast/Ghana” and intriguingly also the (mislabeled) “Southeastern Bantu” region.

Naturally Ancestry’s 2013-2018 version had several shortcomings caused by misleading labeling and inadequate sampling among other things. Still through comparison with other results, bench marking if you like, you can determine your relative position. Which combined with additional context can potentially be highly informational and in accordance with documented history and/or genealogy. Even when ancestral categories/clusters and their labeling will usually tend to be fuzzy indeed. As said by this person himself:

You know, Fulanis themselves have different theories when it comes to their ethnic history, while there is a common old lineage, generally they are considered to have moved around quite a bit and mixed with other groups so the possibilities can be endless, depending on the region, each individual and its family history.”

My genetic reflects the many movements, migrations and alliances which have been formed across the Sahel.”

Finally also remarkable how this person is assigned to no less than 3 genetic communities, a.k.a. “migrations”. Based on his matching strength with various parts of the Afro-Diaspora. Many Africans actually tend to get grouped with at least one “migration” (usually “African Caribbeans”). However from what I have seen the Fula people are particularly “matchy” with Afro-descendants, as measured by the total number of DNA matches they receive. See also:

 

_______________________________________________________________________

Results from Guiné Bissau

Map 4.3 (click to enlarge)

ethnies_guinee_bissau

The ethnic groups of Guiné Bissau are being shown in this map with their French names. The Diolas being the equivalent of the Jola from Senegambia. While the Fula are described as “Peulhs”.

***

Because of a relatively small migrant presence in the USA there are very few DNA testers from Guiné Bissau to be found within Ancestry’s customer database. However I did eventually manage to include several Guinean Bissau results! Which is especially relevant for me personally as I am of Cape Verdean descent. And the African roots of Cape Verdeans will usually trace back to Guiné Bissau (in addition to other parts of Upper Guinea, see this page).

In fact almost all of my 8 Guinean Bissau results (incl. 4 Fula for whom I calculated a separate group average, see table 4.3) are DNA cousins of my Cape Verdean survey participants! Despite the minimal sample size it is still also remarkable how within my survey the group average for “Senegal” is practically equal among Guineans (n=4) and Cape Verdeans (n=100). Resp. 58% & 59% (see table 2.1). Also the remaining part of their African breakdown is quite similar (when scaled to 100% for Cape Verdeans). See also:

Despite being a small country Guiné Bissau does have a multi-ethnic make-up, similar to southern Senegambia. As can be seen in map 4.3. Most of the coastal groups are likely to be closely related. While the Fula and perhaps also the Mandinga will be showing greater differentiation. As far as I am aware only haplogroup studies have been performed so far on Guinean Bissau genetics. Further reading:

 

GUINÉ BISSAU (Fulakunda?)

***

GB1

***

Highly fascinating results! I do not have full certainty but a Fulakunda background seems very likely. The Fulakunda, a.k.a. Fula Preto (= “Black Fula” in Portuguese) are a Fula speaking population in Guiné Bissau and also in Senegambia who are known to have intermarried with local populations to a great degree. The 2% “Africa North” score definitely seems to be in line with partial Fula lineage, albeit in a diluted form (see also discussion under Fula results).

The exact origins of the ethnically mixed Fulakunda are not always traceable. Also due to the practice of indigenous slavery within Fula society. This can be a sensitive topic. And I naturally do not intend to be offensive in my following speculations. Nor do I mean to exclude other possibilities or simplify complex family histories. This person’s combined “Senegal” & “Mali” score (76%) certainly makes for a convincing Upper Guinean profile. However the 19% “Nigeria” score appears to be atypical. Within my Upper Guinean survey I did not come across such high “Nigeria” amounts except among Krio results from Sierra Leone. Although also 2 Senegalese results showed considerable “Nigeria” amounts >10%.

The person who took the test is not aware of any Nigerian lineage. And I suppose it might also be a misreading on Ancestry’s part. Because after the update of September 2018 this person’s breakdown reads as follows:

  • “Mali”: 67%
  • “Senegal”: 31%
  • “Ivory Coast/Ghana”: 2%

But actually also for Nigerians themselves “Nigeria” amounts decreased drastically or even disappeared completely after the 2018 update. While for northern Nigerians often “Mali” has suddenly become very high (see this blogpost). So possibly partial Hausa or otherwise central Sahelian lineage might still be hinted at? Given continuous migrations and trading by both Fula and Hausa populations all along the Sahelian corridor this does not seem far-fetched. This persons’ African DNA matches might very well provide more clarification. Further reading:

***

GUINÉ BISSAU (Mestiço?)

***

GB2

***

These results were obtained from a very fascinating Youtube video (also posted further below). Although not explicitly mentioned a Mestiço (=mixed-race in Portuguese) background seems very likely. Aside from a small Cape Verdean minority there has also always been a small, locally born &  multi-generational Mestiço/Crioulo community within Guiné Bissau as well. They have a long historically documented presence as traders in Guinean coastal cities. Their African origins can be assumed to be mostly from surrounding ethnic groups. However due to their wider trading networks across the Upper Guinea area and possibly also with other ex-Portuguese colonies I suppose also origins from further away are possible. See also “Kriston” references in:

Starting with the African breakdown the primary “Senegal” score is of course in line with a generic Upper Guinean profile. A secondary score for “Ivory Coast/Ghana” makes sense because such substantial scores (>20%) are also common for neighbouring Guineans (Conakry) and Sierra Leoneans. However interestingly it does somewhat look distinctive when compared with an average Cape Verdean profile. Because for them usually “Mali” came in second place and only rarely “Ivory Coast/Ghana” (2013-2018 version). Even more so when also taking into account the notable “Southeastern Bantu” score! Then again 1 result hailing from quite likely Santiago island did show a similar breakdown with 20% “Ivory Coast/Ghana” as well as actually “European Jewish” being primary in the European breakdown! See also:

Although at times European admixture among Guinean Mestiços might have been reinforced by recent intermarriage during the late Portuguese colonial era (1800’s-1900’s). It may also be traced back further back (1500’s-1700’s). Intriguingly this person does not seem to be aware of any recent European ancestry. It turns out that this lineage is specifically Jewish, highlighting the significance of regional admixture! While any Portuguese admixture does not seem to be suggested. Pretty astonishing. And seemingly in agreement with the well-documented historical presence of Jewish traders in Guiné Bissau. Although they were rather Sephardi and not Ashkenazi as suggested by “European Jewish”. And also mostly to be traced back to the 1500’s/1600’s which would suggest a greater degree of dilution. Again a thorough scan of this persons DNA matches might very well provide more clarification. As she is bound to have a high number of Jewish matches (as they are an endogamous population and well represented in Ancestry’s customer database). See also:

***

_______________________________________________________________________

Results from Sierra Leone

Table 4.4 (click to enlarge)

Stats SAlone

***

Table 4.5 (click to enlarge)

Stats Salone

***

Map 4.4 (click to enlarge)

sierra_leone_ethnic_1969***

In table 4.4 more detailed statistics can be seen for my Sierra Leonean sample group. While reviewing the data do keep in mind that averages tend to hide underlying variation. It is always advisable to take into account other measures as well. Such as the median and especially the minimum & maximum values to get a sense of the range of the scores. Naturally also the sample size (mentioned in the row labeled “Number”) is essential to place this data in better perspective.

In section 3 I have already discussed how my Sierra Leonean findings might be interpreted. Especially in regards to their notable “Ivory Coast/Ghana” as well as “Senegal” scores. In table 4.5 one can see how the relative balance of “Senegal” versus “Ivory Coast/Ghana” might possibly be related with either northern (incl. Temne) or southern (incl. Mende) origins. Although again really my sample size is quite limited and I often did not have full certainty about my survey participants background. Most of their results were actually obtained as DNA matches by way of the Compare Ethnicity tool.9

Still it is most likely no coincidence that all 5 primary “Senegal” scores were reported for people with a (presumed) northern background. Also the three primary “Mali” scores in my survey seem to belong to people with a northern background. Not only Temne but also Limba, Koranko, Susu etc. backgrounds being possibilities, see map 4.4. Among my 4 usually confirmed Mende survey participants however only primary “Ivory Coast/Ghana” scores were reported. Often also in predominant amounts (>50%). Then again 4 primary “Ivory Coast/Ghana” scores did also show up among my presumably northern Sierra Leonean subgroup. Furthermore practically all my Sierra Leonean survey participants showed both “Senegal” and “Ivory Coast/Ghana” as main region and often “Mali” too for that matter.

Obviously there will always be individual variation and nothing is set in stone! Both recent and more ancient inter-ethnic unions throughout Sierra Leone can be assumed to have occurred frequently. It seems inevitable that most ethnic groups within Sierra Leone will display close genetic affinity as a testimony of a great degree of shared origins. See for example the below referenced haplogroup study (Jackson et al., 2005). The authors did not find any significant differentiation between their Mende and Temne samples, based on mt-DNA frequencies. But intriguingly they did establish that their Limba samples stood out somewhat. The Limba being from the north and their language used to be classified as Atlantic.10

Therefore my overview in table 4.5 is naturally only intended as an exploratory exercise. But it might still already hint at some insightful patterns. Helpful for determining how to interpret AncestryDNA’s African breakdown (2013-2018). But also useful to shed more light on Sierra Leone’s highly fascinating (pre)historical migrations. For more detailed background info:

Further reading on genetics:

My 3 Krio results have been omitted from table 4.4 merely for the purposes of my research. On the one hand wanting to learn how AncestryDNA describes the DNA of Sierra Leone’s ethnic groups which were established already before the colonial era (1787). But obviously the Krio themselves are also a very fascinating research topic! They have been living in Sierra Leone for many generations already and were crucial in their country’s formation. Because of their unique history they have a quite mixed (within Africa) background. Intriguingly making them very similar to Afro-Diasporans from across the Atlantic!

Given their expected more diverse and distinctive African breakdown they also make for a perfect control group to see how well AncestryDNA predicts African lineage. Despite minimal sample size one of my major research findings has indeed been that AncestryDNA can be useful to identify the diverse African heritage of the Krio, which often goes beyond Upper Guinea. The same actually also goes for related Aku lineage among Gambians and Americo-Liberians. The absorption of Recaptive Africans from especially southern Nigeria but also other places (such as the Congo) among already mixed (within Africa) Afro-Diasporans from the USA & the West Indies leaving a distinctive genetic imprint. Quite impressive that AncestryDNA (2013-2018) has been able to single it out!

We can easily verify this from table 4.5 which shows that “Nigeria” has been the most significant region for my 3 Krio survey participants. Group average being 33%. While also “Benin/Togo” and “Cameroon/Congo” show substantial group averages. In sharp contrast with my other Sierra Leonean survey participants for whom these regions (aside from some outlier scores) tended to be minimal. See not only group averages but also the median scores in table 4.4. Furthermore “Ivory Coast/Ghana” and even more so “Senegal” show up quite subdued for my Krio sample group as a whole. Even when these two regions have been the characteristic main regions for my other Sierra Leonean survey participants. Follow these links for further reading:

***

SIERRA LEONE (Northern/Temne?)

***

SL5

***

I have no certainty about this person’s possibly northern background. Let alone her actual ethnicity. However from her profile details as well as her actual results this seems quite plausible. Notice the predominant “Senegal” and “Ivory Coast/Ghana” combination. The 10% “Southeastern Bantu” score is a bit peculiar. But it was actually not completely exceptional during my Upper Guinean survey to see similar amounts. It might just be a fluke on Ancestry’s part. And given the Sierra Leone context it might be indicative of various ancestral scenario’s. Also involving African Recaptives I suppose.

***

SIERRA LEONE (Mende)

***

SL4 (Mende)

***

This person is of confirmed Mende descent on both sides. His results show one of the highest “Ivory Coast/Ghana” scores in my Sierra Leone survey.  As discussed in section 3 this seemingly higher occurrence among Mende results is possibly to be explained by the Ivorian reference samples being used by Ancestry. Although I have no certainty about it there might be Yacouba (south Mandé speaking) samples among them. The Mende speak a related southwestern Mandé language which makes it likely that a greater genetic affinity would be picked up.

The remaining part of the breakdown is also very much in line with a broader Upper Guinean profile. Only 3% “Nigeria” mentioned as a trace region. While the 3% “South-Central Hunter Gatherer” score has been a very consistent feature of both Sierra Leonean and Liberian profiles during my survey. And I would venture to say it can be considered to be “indigenous” for several millennia even. See also:

***

SIERRA LEONE (Mende)

***

SL6 (Mende)

***

Again a person of confirmed Mende background. And with a quite similar breakdown as the previous one. Especially in regards to the main combination of “Ivory Coast/Ghana” and “Senegal” as well as the minor but still clearly detectable 4% “South-Central Hunter Gatherer” score. 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 again 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.

Then again I suppose it might also be a misreading on Ancestry’s part. Because after the update of September 2018 this person’s breakdown reads as follows:

  • “Mali”: 59%
  • “Ivory Coast/Ghana”: 37%
  • “Senegal”: 3%
  • “Benin/Togo”: 1%

But generally speaking I find Ancestry’s update to have resulted in less informative results (see this blogseries). In particular I greatly suspect that “Mali” amounts have been inflated after the update. And possibly therefore the striking “Benin/Togo” amount might have disappeared. An independent method of verifying would be to analyze African DNA matching patterns.

Which I have actually done for this person as his profile was kindly shared with me. Using my scanning & filtering method I was able to find around 20 most likely African profiles. Overwhelmingly from Sierra Leone as expected. But also a few from Liberia, Gambia and Ivory Coast. Not that surprising in the greater scheme of things. As this is all in the wider Upper Guinea area. However I did also find 1 quite likely Nigerian match which might possibly connect with the 20% “Benin/Togo” score! As more Africans test with AncestryDNA hopefully also more illuminating DNA matches might be forthcoming.

***

SIERRA LEONE (?)

***

SL1

***

I only received confirmation about this person’s Sierra Leonean background. But nothing more specific. Still clearly recognizably as such due to the predominant combination of “Ivory Coast/Ghana” and “Senegal” as well as the minor but still clearly detectable 4% “South-Central Hunter Gatherer” score. Similar to the previous result still also an atypical component. 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 in the 1800’s 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. Some of these unions involving either Krio or Recaptives might have taken place many generations ago already. And therefore some people might not even be aware and just self-identify as being from one single ethnic group.

***

SIERRA LEONE (Krio) & GUINEA (1/4)

***

SL7 (Krio)

***

This person is of confirmed Krio descent on one side and of confirmed Guinean & Sierra Leonean on the other. The breakdown does look extra diverse indeed. Featuring no less than 5 main regions! Tempting to see the “Ivory Coast/Ghana” and “Senegal” scores mainly to be associated with her Guinean side. And the “Benin/Togo”, “Nigeria” and “Cameroon/Congo” scores with her Krio side. I suppose testing family members might bring greater clarity about how these regional scores were inherited. Aside from also investigating African DNA matching patterns.

***

SIERRA LEONE (Krio?)

***

SL2 (Krio)

***

These results belong to a person of very likely Krio background. But probably her family tree still also includes recent intermarriage with other ethnic groups from Sierra Leone as well. Judging from her surname as well as the quite elevated “Senegal”, “Mali” as well as “Ivory Coast/Ghana” scores. And actually also her African DNA matches which I analyzed in 2018 by way of my scanning & filtering method. Mostly being Fula and Liberian matches. Then again her primary “Benin/Togo” score is clearly outstanding. Also otherwise her African breakdown is very diverse. As to be expected for a person of Krio descent.

Very interestingly also her minor non-African admixture is not European but South Asian! During my survey I have come across many Krio profiles in preview (when the Compare Ethnicity tool was not yet available). And I found that these Krio profiles usually tended to be 100% African. Whenever European admixture did show up it was usually residual. It will be very interesting to see large scaled research into the origins of the Sierra Leone Krio. Despite individual variation and documented evidence of mixed-race Krio individuals I do suspect that the degree of non-African admixture among them might have been over-estimated by some people.

***

1/2 SIERRA LEONE (Krio) & 1/2 UK

***

SL3 (Krio)

***

This person has one parent who is of confirmed British/Irish background and one parent who is of confirmed Sierra Leonean Krio background. The practically 50/50 continental breakdown correctly reflects this main mixture. But even more amazingly it indicates that her Krio side seems to be mostly hailing from beyond Upper Guinea. As was to be expected it shows regions far removed from Sierra Leone itself. It is known that among the socalled Recaptives, people from Nigeria (especialy Yoruba) were predominant and also many people from the Congo were among them. So in that way her top 2 main regions makes a lot of sense. See also another confirmed Krio result in the Youtube section who received a primary “Nigeria” score as well.

Interestingly this person also tested on 23andme. And her predominantly Nigerian lineage seems to be confirmed there even more so! Another fascinating aspect about this person’s family history is that she is aware of connections to Nova Scotia /Canada. Tracing back to the Black Loyalists, some of the earliest documented Krio origins! Very amazing to see these ocean spanning historical connections being confirmed! See also:

_______________________________________________________________________

Results from Mali & Mauritania

Map 4.5 (click to enlarge)

Mali ethnique

Map 4.6 (click to enlarge)

Mauritania

Notice how the ethnic groups along the northern bank of the Senegal river are basically a continuation of Senegal’s ethnic groups. The Pulaar being a reference to the Fula.

***

Because of a relatively small migrant presence in the USA there are very few DNA testers from Mali or Mauritania to be found within Ancestry’s customer database. However luckily I did eventually manage to add a few such results in my survey! Many of them are however of uncertain or mixed background. Which is why I have only included results into my main overview (table 2.1) for whom I had solid indications of a plausible Malian or Mauritanian background.

Both Mali and Mauritania are among the biggest African countries with a great diversity of ethnic groups. Because of their geographical location in the Sahel/Sahara region also showing overlap with North Africa! Therefore a much greater variation might be expected had I been able to collect more results.

In section 3 I have already discussed in more detail how the “Mali” region was not particularly predictive because of its weak sample foundation. Quite likely the 16 Malian samples used by Ancestry were possibly Bambara as well as Dogon. Bambara samples from south-central Mali providing a link for Mandé speaking populations across Upper Guinea. While additional Dogon samples from southeast Mali would probably be causing more genetic similarity for Gur and Senoufo speaking populations within & surrounding Burkina Faso. Aside from the minimal number of samples (n=16) such a possibly double ethnic basis might also explain the relative lack of focus of “Mali” (2013-2018 version).

As confirmed by Ancestry’s own information “Ivory Coast/Ghana” and “Senegal” were also very likely to appear as main regions for native Malians. And this indeed occurred during my survey as well. Ironically I calculated a nearly exact same group average of 38% “Mali” for my 6 Malian survey participants as Ancestry’s median of 39% for their 16 Malian reference samples. For my Mauritanian survey participants a relatively high “Africa North” contribution appeared (32.7%, see table 2.1). As expected given that country’s demographics. Although really much depends on ethnic background. Southernmost Mauritania (right across the Senegal river) basically being an extension of Senegambia (see map 4.6).

Further reading:

***

MALI or IVORY COAST (Bambara)

***

ML1 (Bambara)

***

This person might possibly also be from right across the border with Ivory Coast. But he is still a confirmed Bambara. Hailing from the largest ethnic group in Mali (see map 4.5). And quite likely fellow Bambara samples have been used by Ancestry to determine his “Mali” score. Perhaps a testimony of greater genetic variation to be found among the Bambara but the genetic similarity does not seem very convincing. Although 37% “Mali” is very close to what the “typical native” would score according to Ancestry (39% based on their 16 samples).

Despite the substantial “Senegal” and “Ivory Coast/Ghana” amounts still “Mali” did get reported as primary region. This also occurred for 5 other Malian survey participants. With the highest “Mali” amount being 53%, reported for a person of partial Bambara & Gambian descent. Within my entire African survey I did also observe primary “Mali” scores among other nationalities, but not often. Three times for survey participants from Sierra Leone, one time for a Guinean, Moroccan, Fula (country unknown), Ivorian (Senoufo) and a Ghanaian (northern/Mossi). The minor but still clearly detectable “Southeastern Bantu” score is also noteworthy. Perhaps due to some distant cross-Sahelian ancestry from Chad or beyond?

***

MALI (Soninke)

***

ML2

***

Very interesting to see these results for a confirmed Soninke person. The Soninke are a Mandé speaking population but they are located more so to the northwest of Mali. And they also live in neighbouring countries. To be found along the Senegal river valley as well (see maps 4.5 & 4.6). All in all it is clearly a Upper Guinean profile by way of the “Senegal” and “Mali” combination (93%). However “Senegal” is predominant. In fact for a few other possibly Malian profiles (which I did not include in the group averages due to missing details or for being mixed) “Senegal” was also reported with the biggest amount.

As to be expected perhaps because of “Mali”‘s weak predictive accuracy. But still useful to know that “Senegal” can also be prevailing in western Mali. Actually after the update in September 2018 this person’s breakdown changed into 70% “Mali” + 30% “Senegal”. Highlighting how such results are not set in stone but rather subject to changes in reference samples and also algorithm. Eventhough intuitively this change may be seen as an improvement (for Malians themselves). I still find that generally speaking Ancestry’s 2018 update has resulted in less informative results (see this blogseries). In particular I greatly suspect that “Mali” amounts have been inflated after the update.

***

3/4 MALI (Bambara) & 1/4 MOROCCO 

***

MLMO2

***

Very nice to see this person’s breakdown which clearly reveals his partial Moroccan heritage. His mother being half Moroccan and her results are shown directly below. The genetic inheritance from his 1 Moroccan grandparent would consist out of 14% “Africa North” + 7% “Middle East” and 3% “European” traceregions. Combined around 25% as expected! Moroccans being described by AncestryDNA (2013-2018) as a composite of genetically overlapping regions rather than just simply “Africa North”. See also:

This person’s otherwise Malian/Bambara side is being described as a combination of “Senegal”, “Ivory Coast/Ghana” and “Mali”. As was to be expected. Although again “Senegal” seems to be clearly predominant. Confirming how both “Senegal” and “Mali” scores are basically pinpointing generic DNA found all over the Upper Guinea area.

***

1/2 MALI (?) & 1/2 MOROCCO 

***

MLMO1

***

Mother of the person whose results are shown directly above. She is half Moroccan and half Malian. But I have no details about any specific ethnic background. Again the partial Moroccan component can be clearly distinguished: 20% “Africa North” + 17% “Middle East” + 7% “European” traceregions. The remaining part of her breakdown looks less recognizable as being from Mali. Due to the high “Senegal” and also considerable “Benin/Togo” score. Within the trace regions are mentioned as well however: 5% “Mali” and 4% “Ivory Coast/Ghana.

***

3/4 MAURITANIA (Fula & Moroccan) & 1/4 LIBERIA (Americo-Liberian)

***

MAUR1***

Although mostly Mauritanian this person still is very mixed. Three of her Mauritanian grandparents are themselves mixed between Fula and Moroccan heritage. While extra-ordinarily her fourth grandparent is an Americo-Liberian. Similar to the Krio from Sierra Leone bound to have a greater diversity within their African breakdown. Usually going beyond Upper Guinea. Amazingly this person is also aware of a direct Nigerian great-grandparent by way of her Liberian grandparent! It all makes for quite a mix to entangle. But Ancestry still does quite a good job at it! Notice for example the striking 10% “Nigeria”!

Her principal 75% Mauritanian/Moroccan/Fula origins seem to be described by Ancestry as follows: 39% “Africa North” + 25% “Senegal” + 13% “Middle East”. Quite likely the 6% so-called “Iberian Peninsula” is also to be included. Even when then an approximate 75% total is being surpassed. Still such (mislabeled) “European” scores are also very common for North Africans and also Fula people in fact (see previous results). Then again I suppose theoretically some of it might might also be due to her Liberian side.

***

How representative are African DNA testers on Ancestry?

Somewhat going off on a tangent. But one thing I have always found very fascinating is to contemplate the family backgrounds of the various African DNA testers I have been in contact with across the years due to my ongoing surveys. It appears to me that they are often more cosmopolitan in origin than strictly rural Africans might be. Possibly due to their recent migrant history to the USA or other parts of the West. And perhaps also because they might often be hailing from multi-generational urbanized families (already within Africa) or were part of trading diasporas.

What I found especially striking is the number of Krio related descendants I have come across during my survey. Not only from Sierra Leone and/or Liberia, but also from Gambia (where they are known as Aku), Nigeria (where they are also known as Saro), the Ivory Coast and in this last case even Mauritania! Although initially being disadvantaged because of their former enslaved status these Krio and/or Americo-Liberians quickly managed to improve their social standing by profiting from their relatively greater exposure to English language/culture etc.

For practical reasons this has raised some concern with me of how representative these samples may be. Because it may be assumed that people with such backgrounds may have intermingled with other ethnic groups to a greater extent than people from more isolated rural settings. In fact this is often also expressed in greater regional diversity in their admixture results. Also for their Afro-Diasporan DNA matches it pays to be very cautious when wanting to zoom into a mutual ancestor! But from a socio-political and also just a human-interest perspective I find this group and their family histories very fascinating and inspirational!

_______________________________________________________________________

Youtube Videos

***

CABO VERDE (Santiago)

***

***

GUINÉ BISSAU (Mestiço)

***

***

GUINEA CONAKRY (Fula)

***

***

SIERRA LEONE (Krio)

***

 

_______________________________________________________________________

5) Is it possible to distinguish a genetic Upper Guinean component?

 

“Senegal” + “Mali” = Upper Guinean DNA

Table 5.1 (click to enlarge)

Stats, UG component

“Senegal” and “Mali” scores have been added together to arrive at an approximation of a Upper Guinean genetic component. Obviously within the constraints of AncestryDNA’s imperfect regional format.

***

Map 5.1 (click to enlarge)

Senegal freq. 2013-2018

This map shows the distribution of “Senegal” group averages among my West African survey participants. See also table 2.1. As expected it peaks in Senegambia and fans out in surrounding countries. But actually high individual scores (>50%) were reported for persons from western Mali and northern Sierra Leone as well. Because of a higher predictive accuracy (75% for 34 Senegambians) and greater consistency this region has proven to be the most useful indicator of Upper Guinean lineage in Ancestry’s 2013-2018 version. Although usually reinforced by “Mali”.

***

According to many pundits only continental admixture is to be taken seriously in DNA testing. Sub-continental, a.k.a. ethnicity estimates, a.k.a. regional admixture only being fit for entertainment purposes.11 I myself have never taken this stance. Preferring to judge each case on its own merits. Attempting to maximize informational value despite imperfections and avoiding source snobbery. Which is why I have conducted my AncestryDNA surveys among both Africans and Afro-descendants in the past. Applying an additional macro-regional framework for extra insight (Upper GuineaLower GuineaCentral/Southeast Africa).

Basically combining interrelated & neighbouring genetic regions in order to allow certain regional patterns to show up more clearly. This turned out to be particularly helpful when wanting to explore any rough correlation between aggregated AncestryDNA results for Afro-Diasporans and slave trade patterns. Which I indeed established for the most part. For more detailed discussion see also:

Also for Africans themselves such an approach can be useful. As illustrated by table 5.1 & map 5.1 above my West African survey has demonstrated that:

  • when combining scores for “Senegal” & “Mali” a prevalent “Upper Guinean” genetic component can indeed be observed. As expected it is (nearly) predominant for all my survey groups from the wider Upper Guinea area. Additional double-digit scores for “Ivory Coast/Ghana” and “Africa North” to be explained by geography and/or distinctive ethno-genesis (see section 3 for more discussion).
  • Because of a higher predictive accuracy and greater consistency the “Senegal” region has proven to be the most useful indicator of Upper Guinean lineage in Ancestry’s 2013-2018 version. Actually “Senegal” turned out to be one of the most reliable regions. When judged by its group average of 75% among 34 actual Senegambians. Within my African survey only “Cameroon/Congo” reached a higher predictive accuracy of >80%  among my Cameroonian & Congolese survey participants (see this blog post).
  • Given geographical proximity some (inevitable) genetic overlap with especially “Ivory Coast/Ghana” and Lower Guinean DNA did occur as well. So this Upper Guinean genetic component is obviously only intended as an approximation! Then again to a great extent Central African DNA & Upper Guinean DNA were mutually exclusive, during my survey. As measured especially by “Senegal” scores being either absent or at minimal trace level for my Central African survey participants (see this table). And also vice versa for “Cameroon/Congo” scores among my Upper Guinean survey participants (see table 2.1). Which only occurred in significant amounts for Krio Sierra Leoneans because of plausible historical reasons.

Upper Guinean DNA correlating with Atlantic language group?

Map 5.2 (click to enlarge)

Atlantic

The Atlantic language group is shown as being mainly located in Upper Guinea and even extending to northern Liberia (see also this map or this one). However due to cross-Sahelian migrations of the Fula people their Atlantic/Senegambian language is nowadays also spoken in Mali, Burkina Faso, Niger, northern Nigeria and beyond. Source: The Languages and Linguistics of Africa (Gueldemann, 2018, p.182)

***

In the first part of my West African AncestryDNA series I already discussed how learning about ancient population migrations and the distribution of African language groups can be very helpful to make more sense of your AncestryDNA results. In particular I mentioned how “Senegal” seems to correlate mainly with the distribution of Atlantic and to a lesser degree also with Mandé languages (northern & central subgroups). Both language groups concentrated within the area which is also known as Upper Guinea. The “Mali” region being less specific due to also partially describing genetic links with Gur/Senoufo speakers. While southern Mandé speakers from Ivory Coast, Liberia and Sierra Leone tend to be also covered by “Ivory Coast/Ghana”. These tendencies indeed also manifested themselves during this second part of my West African survey. See also:

The distribution of language groups is often a good indicator of broader genetic patterns/clusters to be taken into consideration. As generally speaking the initial expansion of distinctive populations was accompanied by the spread of their language and often also their particular mode of livelihood. Of course language replacement, inter-ethnic unions etc. will result in a less clear cut situation. But various scientific papers have already demonstrated that broad-scale population structure within Africa indeed reflects both geography, (pre-)history and language. Some of these studies will be referenced further below.

In the context of my Upper Guinean survey I would just like to point out the following rather evocative findings. As shown in map 5.2 the broader Atlantic language group includes mainly Senegambian languages but also languages spoken in Guinea and Sierra Leone (such as Temne). Largely corresponding with my survey findings for “Senegal” extending into northern Sierra Leone! Even when these Atlantic subgroups (indicated by Mel, Gola, Limba etc.) often appear to be isolates. And according to most recent research may not not be closely related with northern Senegambian languages after all.

Furthermore due to the eastbound migrations of the nomadic Fula people the Atlantic language group has expanded across the Sahel into northern Nigeria and beyond (see also map 4.2). The language of the Fula people (Pulaar or Fulfulde) after all is classified as Atlantic/Senegambian by linguists. Although apparently many urbanized Hausa-Fulani nowadays have switched to speaking the Hausa language rather. It is still very striking how through my Nigerian AncestryDNA survey findings I was able to pinpoint a significant Upper Guinean genetic component for the Hausa-Fulani (see table 4.3). And very fittingly it seems that increased “Senegal” scores beyond Upper Guinea can indeed be correlated with the Atlantic language group. Or at least the descendants of Fula nomads who initially spoke an Atlantic language (Pulaar/Fulfulde). Again it is important to note though that the linguistic classification of African languages is very much a work under progress.12 See also:

Academic studies based on autosomal DNA analysis

Chart 5.1 (click to enlarge)

Triska 2015 Atlantic WA component

This is a zoomed-in portion which I took from a bigger chart “Population Structure in the Sahel” (Triska et al., 2015, see this link for complete chart). Performing admixture analysis for K = 7 two clusters are revealed. The orange cluster, labeled “Atlantic West Africa” clearly distinguishes Senegambian samples from Nigerian ones. It seems to be a close equivalent of the Upper Guinean genetic component from my survey findings. Also take notice of how the Fulani samples to the right are more complex but still with the orange cluster being predominant. In line again with my own findings.

***

When I talk about an “Upper Guinean genetic component” this is strictly speaking from a macro-regional perspective. Relating to generic DNA to be found in higher frequency within the wider Upper Guinea area. And therefore not unique to any given ethnic group or country! However this approach will allow for a distinction to be made with the remaining part of West Africa (Ghana-Nigeria etc.), Central & Southeast Africa as well as North & East Africa. Such a perspective can be very valuable even when it lacks finer resolution (it is still sub-continental though!). Especially when combined with other insights.

Just to clarify I am not advocating that a reliable distinction can be made between ethnic groups within Upper Guinea. This is a much more complex research question than many people may realize or (unrealistically) hope for. At least generally speaking. Then again within my survey a clear ethnic distinction did occur for the Fula and also the Sierra Leonean Krio (see table 2.1). Also the Gambian findings by Jallow et al. (2009) described below do seem promising!

It is important to stress that generally speaking the basis for this distinction is not in some unique DNA markers which can only be found among one particular ethnicity. But rather because of a distinctive proportional mix of ancestral componentsIn fact this is also the key to correct interpretation of Ancestry’s socalled Ethnicity Estimates!  See also:

Below a summary of the references to be found in major scientific papers which also identify a distinctive Upper Guinean genetic component. Not meant to be exhaustive. All of these studies based on autosomal admixture analysis. Therefore comparable to my AncestryDNA findings. Albeit that AncestryDNA (2013-2018) usually provides a finer regional resolution for its African breakdown (K=9). I have reviewed some of these studies already in the past. While the more recent ones might eventually get reviewed in greater detail in the near future. See this section of my blog for an overview:

***

Jallow et al., 2009

Chart 5.1 (click to enlarge)

Jallo STRUCTURE

Chart taken from the supplement of Jallow et al. 2009. Quite an ambitious exercise to focus on the genetic basis of ethnic differentiation for such a small country as Gambia! Notice that the Mandinka samples seem to be least well-defined. Which would make sense given their known absorption of local populations.

***

A pioneering study which makes use of a HUGE data-set of Gambian samples, incl. Jola, Wolof and Fula aside from the usual Mandinka. Possibly more than 2000 individuals were genotyped! Actually ever since I learnt about the existence of this study I have been wishing for these samples to also be utilized by either 23andme or Ancestry 🙂 As an addition of these samples would surely contribute to a more solid identification of an Upper Guinean genetic component. And possibly even enable finer resolution beyond that. The samples seem to be available through the MalariaGEN database.

Their research outcomes are still highly relevant. On the one hand they establish that:

____________________

“The Gambian sample can be clearly distinguished by PCA from the Yoruba people of Ibadan, Nigeria (a different part of West Africa)

____________________

Focusing on their Gambian samples to possibly uncover population substructure within Gambia revealed the following perhaps more ambiguous outcome:

____________________

Some individuals could be confidently assigned to a specific ethnic group, whereas others seemed to have a more complex ancestry.”

Self reported ethnicity correlates with genetically defined subpopulations”

____________________

 

Zakharia et al., 2009

One of the first studies to clearly observe how Mandenka samples can be genetically separated from Yoruba and Bantu samples. Going beyond the mere ethnic labels. And already implying therefore how a macro-regional approach is feasible. Distinguishing Upper Guinea from Lower Guinea and Central Africa. Something which they actually applied on their 136 African American samples! Again quite pioneering and akin to my own African American research findings from 2015.

____________________

despite their genetic similarity, PCA shows clear separation among the Yoruba, Mandenka, and Bantu populations

____________________

 

Bhatia et al., 2011

Similar findings as Jallow et al. (2009) also using the very same Gambian data-set apparently. Based on 500 Gambian and 500 Nigerian samples they find that:

____________________

the Gambian samples are separated from the Nigerians […], form separate but overlapping clusters, […]

” We label each of the Gambian individuals with their subpopulation label (Mandinka, Jola, Fula, Wolloff) and note the existence of cryptic population structure within the Gambia.

____________________

 

Triska et al. (2015)

Very insightful study which includes Fulani samples from Burkina Faso, Niger and Chad. They are compared with other samples from along the Sahel belt but also from Senegambia and Sierra Leone! As can be seen also in chart 5.1 above an orange cluster labeled “Atlantic West Africa” clearly distinguishes Senegambian samples from Nigerian ones. It seems to be a close equivalent of the Upper Guinean genetic component from my survey findings.

Also take notice of how the Fulani samples are more complex but still with the orange cluster being predominant. In line again with my own finding. The term “Atlantic West Africa” is strictly intended to be geographical though. As far as I can tell. In the words of the authors:

____________________

Western African populations present varying proportions of two clusters, one being more frequent in Atlantic Western populations (orange in fig. 1C; 90% in Mandenka), whereas the other is more frequent in Western/Central populations, especially in Esan and Yoruba of Nigeria (dark blue; reaching 74–81% frequency).

“The nomadic Fulani present the reverse pattern for the Atlantic Western versus Eastern African components (55% and 11%, respectively), […]. These results support the hypothesis of […] a Western to Central Africa past migration for Fulani.” (Triska et al. , 2015, p.3488)

____________________

 

Dobon et al. (2015)

Focusing on Sudanese genetics actually. But still relevant because one of their sample groups happened to be Sudanese Fulani (a.k.a. Fellata). Using both PCA and admixture analysis they report that as expected Sudanese Fulani can be clearly distinguished from other Sudanese groups. In fact the Fulani samples are showing up with their own component which also differentiates them from like-wise West African Yoruba. Still in regards to the additional East African admixture they found they also caution that:

____________________

This finding does not agree with studies of Fulani people in the Lake Chad Basin which reported that Fulani from West Africa’s Sahel usually have consanguineous marriages and do not seem to have admixed with local farmers. These data together suggest differentiated genetic legacy in different Fulani populations from various geographic regions of the continent.” 

____________________

 

Gurdasani et al., 2015 & Fortes-Lima et al., 2017 

Both studies demonstrate clear distinction between Senegambian and Igbo/Yoruba samples. Yet again including the Wolof, Jola and Fula samples from Gambia! See for example this chart and this one.

 

Vicente et al. (2019)

Most recent study and again very useful! This time Fula samples from not only the central Sahel (Burkina Faso, Chad, Niger) are included but also Gambian Fula samples. Actually in regards to Fula genetics their lactase persistence finding is most salient to understand the “Africa North” and also “European” scores in my survey (see previous discussion in section 3).

Their admixture analysis however does not seem to allow for a clear identification of a Upper Guinean genetic component among eastern Fulani (as was done in Triska et al., 2015). Instead a Fula-specific component seems to arise (as was also the case already in Tishkoff et al., 2009). Then again their main admixture findings do clearly underline how Senegambian samples can reliably be distinguished from both Nigerian and Ghanaian samples. And to a lesser degree also samples from Burkina Faso (Mossi etc.) which were also included!

 

_______________________________________________________________________

 

6) Implications for Afro-Diasporans

To conclude this page I will now describe some of the most important implications for Afro-Diasporans in an attempt to improve proper interpretation of their West African regional scores. In particular for “Senegal” and “Mali”.13 This section firstmost reflects my own reasoning and naturally other interpretations remain possible as well. As always context is everything and solid genealogical research combined with historical plausibility should be leading instead of wishful thinking.

1) “Senegal” + “Mali” combined is a solid indication of lineage across Upper Guinea 

  • Neither “Senegal” or “Mali” is tied to any particular country or ethnic group. As my survey findings have shown these regions are reported for people throughout Upper Guinea: Senegambia, Guinea Bissau/Conakry, Mali and Sierra Leone. Therefore these regions combined are firstmost an indication of generic Upper Guinean DNA.

2) “Mali” can also be predictive of DNA found in Burkina Faso, northern parts of Ivory Coast & Ghana.

  • Due to quite likely inclusion of Dogon samples the “Mali” region may also point towards genetic similarity with DNA found among Gur and/or Senoufo speakers in the southeast of Mali. And across borders into Burkina Faso, Ivory Coast and Ghana.

3) “Ivory Coast/Ghana” can also be predictive of Sierra Leonean DNA

  • The so-called “Ivory Coast/Ghana” region is indeed quite predictive of both Ghanaian and Ivorian and in fact also Liberian origins (see part 1). However in addition ancestry from Sierra Leone and to a lesser degree Guinea might also be described by this region. You will need to perform your own follow-up research in order to find out more specifics.

4) “Africa North” might also be inherited by way of Fula ancestors

5) “South-Central Hunter-Gathers” can also be predictive of West African ancestors

  • In particular Sierra Leonean and Liberian ancestors might have passed on so-called “Africa South-Central Hunter-Gatherers” DNA markers. As these are also present in their own genome at around 3%. The “Africa South-Central” labeling by Ancestry is therefore not to be taken too literally. Despite usually appearing as “low confidence” trace region this still represents a very distinctive type of DNA. For more discussion see:

 

Regional admixture DOES matter!

Chart 6.1 (click to enlarge)

***

A very insightful selection of West African AncestryDNA results has been shown on this page. Even if still quite limited in number and scope. Keeping in mind the relevant context and limitations already discussed in the previous sections it seems that “Senegal” is quite predictive but “Mali” less so. This is going both by my own survey findings (see table 2.1) as well as by Ancestry’s information in regards to how the “typical native” scores: respectively 100% for “Senegal” (n=28) and 39% for “Mali” (n=16).

See also chart 6.1 above, which shows the entire range of estimates for Ancestry’s own reference panel. The black line in the middle representing the median. And the coloured bar would be the most common range. Even when the “Mali” region has arguably been Ancestry’s weakest defined region. It usually did reinforce the predictive accuracy of the “Senegal” region. Combined acting to describe a generic Upper Guinean component. While “Senegal” itself has proven to be the most reliable region, along with “Cameroon/Congo”. That is when judged by how my survey participants from Senegal & Gambia score for it (75% on average, n=34).

After going through this whole page it should be abundantly clear that the regional labeling by AncestryDNA is not intended to be taken too literally! You may have noticed for example that I have consistently used “..” when mentioning any of AncestryDNA’s regions 😉. My survey findings seem to indicate that ethno-linguistic background will be a more relevant determinant for West African AncestryDNA results rather than nationality per se. Correlating with the selection of West African samples by AncestryDNA in their reference panel and how well these samples happen to match with someone’s DNA.

The regional percentages reported by AncestryDNA firstmost signal close genetic similarity to the samples taken from the countries or areas after which the regions have been named. And not actual descent or some kind of blood quantum as is too often assumed. The regional scores are also often reflective of ancient migrations & overlapping genetics/geography rather than any recent lineage (see section 3 for more detailed discussion).

Admixture analysis such as provided by AncestryDNA is often drawing criticism for not being in line with unrealistic expectations. Specifically in regards to how ancestral categories should conform exactly to a person’s family tree and all the known ethnic lineage it may contain. Disregarding how such over-specified information is simply not to be found in our DNA. Atleast not given the current state of knowledge. However I myself do strongly believe that AncestryDNA’s Ethnicity Estimates can still be of great informational value as long as you:

  1. focus on correct interpretation of your results
  2. educate yourself about inherent restrictions
  3. combine with other research findings.

One factor many critics fail to take into account is that Afro-descendants usually do not have the privilege of being able to fall back on paper trails when wanting to learn more about their African origins. A strategy based strictly on genealogy will only very rarely lead to identifiable African ancestors. And even if valuable success stories do exist.14 Such discoveries still usually only pinpoint one single African family line out of potentially hundreds of others!

Because of this lack of viable alternative Afro-Diasporans cannot afford to be snobbish about the imperfections of admixture analysis. Basically we will want to maximize all the informational value we can obtain. Taking any promising lead we can get and combine with other clues. Seeing the glass as half full and not half empty. Rather than being overcritical and resolute in (overhasty) dismissal. Because then you risk loosing out on helpful information. Even when only approximate or incomplete. Your admixture results will always be relevant to put things in perspective. If only to be able to (roughly) distinguish between major sources of ancestral origins versus minor lineage.

Instead of taking a generalizing and dismissive stance each aspect of admixture analysis should be judged on its own strengths and weaknesses. Precisely because of its still unrivaled West African regional specificity I find AncestryDNA (2013-2018 version) to have been more insightful than anything on offer by other commercial DNA testing companies. As well as any thirdparty analysis such as available on Ged-Match, DNA Land etc.. At least up till the latest update in 2018/2019 by 23andme.15

My assessment is based on the more than 500 AncestryDNA results of native Africans I have seen by now. Which were usually in alignment (broadly) with their verifiable background (see this overview). Also my survey of Afro-Diaspora results was largely a confirmation of historically documented African origins for each nationality. Again such potentially profound information is not something to carelessly brush aside when wanting to Trace African Roots!

Combining various levels of regional resolution

Map 6.1 (click to enlarge)

reg. resolution

***

The three maps shown above illustrate three different levels of regional resolution which may all be useful when wanting to explore your West African origins. On the left the various bigger language groups of western Africa are being displayed. As discussed in part 1 these language groups tend to overlap to a great extent with AncestryDNA’s West African regions which are displayed in the right bottom corner. For a historical perspective I have also included a map which lays out all the relevant slave trade regions for West Africa. Again seemingly corresponding with AncestryDNA’s regions. The two-way distinction being made between Upper Guinea & Lower Guinea will not only be helpful to connect the dots when it comes to Afro-Diasporan history. In fact it may also lead to a greater understanding of genetic overlap patterns within West Africa and identifying West Africa’s main (pre)historical driving forces before the advent of Europeans.

Combining insights from various disciplines is central to my personal research strategy. Always looking for complementarity, correct interpretation & plausibility. Critical but also staying open-minded and careful not to be dismissive when informational value can still be obtained. From my experience it can be very beneficial to acquire a basic grasp of this interplay between genetics, ethno-linguistics and (pre)history. It has definitely provided a deeper understanding of my own personal African origins as a Cape Verdean descendant.16 And in addition it has also clarified many of the questions which came up during my AncestryDNA survey. Helping me to see the origins of not only the Afro-Diaspora but also Africans themselves in much sharper focus.

Obviously we’re not dealing here with exact science. The causal relationships will be complex, tentative and multi-faceted. Categories will tend to be blurry, overlapping and not just strictly “black or white”. Just like in real life 😉 Understandably many people desire to have the most specific degree of resolution when searching for their African roots. They want to be able to pinpoint their exact ethnic origins and preferably also know the exact location of their ancestral village. In a way following in the footsteps of the still very influential ROOTS author Alex Haley. See also:

Unfortunately these are rather unrealistic expectations to have in regards to DNA testing. Not only given current scientific possibilities. But also because such expectations rest on widely spread misconceptions about ethnicity, genetics, genealogy as well as Afro-Diasporan history. Too often people ignore how the melting pot concept is really nothing new but has always existed. Also in Africa where inter-ethnic mixing has usually been frequent! Throughout (pre) history and maybe even more so in the last 50 years or so. Generally speaking ethnicity is a fluid concept which is constantly being redefined across time and place.

Too often people fail to take into consideration how due to genetic recombination our DNA will never be a perfect reflection of our family tree but might actually also at times suggest very ancient migrations. Too often people underestimate the actual number of relocated African-born ancestors they might have (dozens or even hundreds!). As well as the inevitable ethnic blending which must have taken place across the generations. Too often people are still not informing themselves properly about Africa itself and the documented origins of the Afro-Diaspora. Many specific details might have been lost forever but there is a wealth of solid and unbiased sources available which can help you see both the greater picture as well as zoom in more closely to your own relevant context. See also:

AncestryDNA might not provide the “100% accurate” ethnic granularity many people unrealistically wish for. But it does offer regional resolution which goes one big step beyond just describing your DNA as “West African”. Unlike any of its commercial competitors AncestryDNA is able to distinguish no less than 5 different West African regions. In addition its 2013-2018 version also did a reasonable job at separating West African DNA from Central/Southern African DNA (“Cameroon/Congo” and “Southeastern Bantu”). Which is essential for getting a full scope perspective on the origins of the Afro-Diaspora.

The analysis performed by AncestryDNA, even when pioneering and already very valuable in my estimation, is still only a first rudimentary step in deciphering one’s origins within Africa. The nine African regions in Ancestry’s 2013-2018 version might very well have different ethnic implications for each separate nationality within the Afro-Diaspora but also in individual cases. Follow-up research may enable you to still zoom in closer to more specific West African lineage.

In this regard group averages of AncestryDNA results for West African ethnic groups or nationalities can be very helpful in order to know what to expect more or less. And also in order to make a proper inference of how you yourself fit in the scheme given any other ancestral clues you might have. Even when entangling the intricate origins of an Afro-Diasporan will by default almost always be very tricky. Being aware of the greater inter-relatedness of African genetics, history and ethnicity as shown in map 6.1 may still provide added value to your initial results and produce new research findings.

 

How to make more sense of “Senegal” & “Mali” scores

Map 6.2 (click to enlarge)

Senegal &amp; Mali2

Source: Ancestry.  The “Senegal” and “Mali” regions combined are first most indicative of generic Upper Guinean DNA! Take note how “Senegal” expands all the way into Sierra Leone and also into western Mali. While according to my survey findings “Mali” actually also is frequently reported for people in Senegambia, Guinea Bissau/Conakry and again Sierra Leone. And aside from that also Burkina Faso and northern parts of Ivory Coast & Ghana!

***

Performing follow-up research based on your AncestryDNA score for the so-called “Senegal” & “Mali” regions might very well enhance your understanding of the actual ancestors who passed on this part of your DNA. In some cases it might even help you pinpoint specific ethnic lineage! However as always it pays to be careful and not jump to conclusions. Lest you may misidentify any of your ancestors. Below I will describe a schematic approach. Intended in the first place for people with substantial amounts of “Senegal” and/or “Mali”. Naturally other approaches can be fruitful as well.

Some considerations

  • Realize that any so-called “Senegal” or “Mali” % is generally speaking indicative of generic Upper Guinea DNA. And not tied to any particular country or ethnic group! However in addition “Mali” might also be suggestive of DNA inherited from Gur/Senoufo speaking ancestors from Burkina Faso and surrounding areas in northern Ghana & Ivory Coast.17
  • Realize that your “Senegal” and/or “Mali” amount is likely to be traced back to numerous family lines and not just one (unless you happen to have relatively recent West African ancestry). Just as an example: a 25% “Senegal” score for a typical Cape Verdean might be due to the genetic contributions of at least 250 different ancestors born throughout Upper Guinea! On average the DNA contribution of an ancestor living in the mid 1600’s could be a mere 0.1% (leaving aside pedigree collapse). And in fact for Cape Verdeans and also many Latin Americans it might even be that much of their Upper Guinean ancestry traces back further even. Into the 1500’s! See also:
  • Realize that therefore your “Senegal” and/or “Mali” scores could include ancestors from various countries, all at the same time. For example a 25% “Mali” score for a typical African American might be traced back to 10 ancestors from Guinea Conakry, 5 ancestors from Mali, 5 ancestors from Sierra Leone, 3 ancestors from eastern Senegambia and 2 ancestors from northern Ghana.

Action plan:

  • Build up your family tree and attempt to trace back to your earliest known ancestral location in the Americas for all lines. The so-called “migrations” (a.k.a. genetic communities) you have been assigned to by AncestryDNA will usually be quite indicative in this regard! For example for African Americans it might be truly worthwhile to know if most of their ancestry is to be traced back to either Virginia or South Carolina. For West Indians it will be useful to know if there has been any inter-island migration within their family tree.
  • Research the documented slave trade patterns for your earliest known ancestral locations. This will give you an approximate idea of the odds involved with having for example either Senegambian or Sierra Leonean ancestry. It is advisable to also study all other relevant aspects of local history (incl. possibly African retention) associated with these earliest known ancestral locations in the Americas. See also:
  • Perform a full scan of your DNA matches and filter them for 100% African profiles. Determine if there are any matches among them from the wider Upper Guinea area: Senegal, Gambia, Guiné Bissau, Guinea, Sierra Leone, Mali and Mauritania. Be sure to also check if their own regional admixture is in line with an Upper Guinean profile.18 Provided they are genuine IBD matches this could directly inform you if you happen to have any lineage from anywhere in the Upper Guinea area. However many reservations are to be taken into consideration! Firstly your MRCA might not be of the same background as your match. Also be aware that these matches will give you an indication of only one single family line.
  • Realize that most Upper Guinean nationalities are underrepresented within Ancestry’s customer database (esp. Guiné Bissau and Mali). While a few might be fairly well represented (esp. Sierra Leone). Furthermore also Fula people (from various countries) seem to be overrepresented as they have been taking DNA tests in greater numbers. Stacking the odds in favour of receiving matches from people with a latter background. Which however may not per se be in line with their overall proportional contribution to your DNA. Moreover, absence of evidence is not evidence of absence!
  • Try fitting your Upper Guinean matches into your family tree. This will be very tricky given scanty information and will also require a lot of patience. But it might still be worthwhile for eventually zooming in closer to your West African origins along a certain family line or even actually identifying a West African ancestor! Breaking down those brick walls based on paper trails! Having either your parents or other close family members also DNA tested will of course be greatly helpful. Especially if they share the same West African matches with you. However any shared matches for your West African match might do the trick I suppose. As long as you can figure out how these shared matches are to be placed in your family tree. More advanced techniques such as triangulation and DNA Painting might also open up promising avenues.

“Senegal” + “Mali” across the Diaspora (sorted) 

Table 6.1 (click to enlarge)

Stats Upper Guinea (diasp)

This table features an approximation of an Upper Guinean component by combining “Senegal” and “Mali” group averages (see also section 5). The ranking among Afro-Diasporans is more or less in line with historical sources. Illustrating how a Upper Guinean founding effect among Hispanic Americans may have been very significant!

***

Table 6.2 (click to enlarge)

Source: TAST Database (2019). Follow this link for underlying numbers. Keep in mind that Inter-Colonial slave voyages are NOT included. The data is based on documented Trans-Atlantic Slave Trade voyages.  Amazonia is referring to northern Brazil; Hispaniola is referring to the Dominican Republic; Spanish Circum-Caribbean is referring mainly to Colombia & Mexico; Saint-Domingue is referring to Haiti.

***

Table 6.1 is displaying my survey findings of AncestryDNA results for people across the Afro-Diaspora. And it also includes Senegambians, Guineans and Malians as a control group (see this page for a full overview). The sample size is limited for a few survey groups but otherwise already quite decent. It is also cross-sectional because it was collected at random from various parts of the Diaspora and also within West Africa. Overall contributing to the robustness of the data. Table 6.2 is taken from the Trans Atlantic Slave Voyages Database and features the proportional shares of relevant African slave trade regions for selected parts of the Diaspora.19 The Upper Guinean coast line generally being divided in “Senegambia” (which includes Cape Verde & Guiné Bissau!); “Sierra Leone” (which includes Guinea Conakry!) and the “Windward Coast” (which includes Liberia and the Ivory Coast). See also:

Cape Verdeans & Hispanic Americans

When measured by “Senegal” + “Mali” we can establish that the predominant Upper Guinean component among Cape Verdeans is in line with Senegambians & Guineans as well as Malians actually and my other Upper Guinean survey groups (see also table 5.1). Given the discussion in the previous sections this outcome can be considered as corroborating the prediction accuracy of especially the “Senegal” region among West Africans themselves.

Not shown in table 6.1 but actually aside from standing out for their overall Upper Guinean component the relative balance of “Senegal” versus “Mali” was also much more pronounced among Cape Verdeans. Clearly tending towards “Senegal” in the case of Cape Verdeans. This was also somewhat true among Latin Americans, but to a much lesser degree. While for my other survey groups in the Americas it was usually either tending towards “Mali” or just evenly balanced (see this chart). For more detailed discussion see:

In regards to the (Trans-Atlantic) Afro-Diaspora we can observe how Upper Guinean DNA is most prevalent among Mexicans, Puerto Ricans and Dominicans. While it is most subdued for Jamaicans and Haitians. This conforms with what we know about slave trade patterns. The data shown in table 6.2 representing the most recent update of the Slave Voyages database (TAST) which is widely considered to be authoritative. Do keep in mind however that this overview does not include Inter-Colonial slave voyages!

A greater part of the earliest slave trade to Latin America (1500’s) was conducted from Upper Guinea. And actually also routed by way of Cape Verde! Naturally tables 6.1 & 6.2 are only to be considered as approximations. And several other factors are to be taken into account. Especially for Puerto Rico and the Dominican Republic (“Hispaniola”) clandestine and therefore undocumented slave trade was often standard practice. Also slave trade from Sierra Leone is bound to be underestimated.20

Still it is striking how much the estimated proportional shares of Upper Guinean DNA mentioned in table 6.1 conform with slave trade originating within that area. On both counts around 30-40%. Mexican slave trade data being included in “Spanish Circum-Caribbean”. Seemingly reflecting a major Upper Guinean founding effect among Hispanic Americans. I have blogged about this topic many times already. And I intend to do so again eventually as my updated 23andme surveyfindings are also in support of this remarkable phenomenon! See also:

Haitians & Jamaicans

The subdued “Senegal” + “Mali” scores for my Jamaican and Haitian survey participants are in line with what we know about slave trade from Lower Guinea (Ghana, Togo, Benin and Nigeria) as well as Central Africa (esp. for Haitians) being much more important for them. But it is again quite striking how closely their estimated Upper Guinean components of 10% correspond with the combined shares of slave trade from Senegambia, Sierra Leone and the Windward Coast. Around 11% for Haitians (“Saint Domingue”, 6.9% + 3.4% + 1.2%) and around 10% for Jamaicans (2.3% + 3.7% +4.2%).

The Windward Coast (4.2%) actually being most significant for Jamaicans. Both Liberia and the Ivory Coast can be said to be intermediate between Upper & Lower Guinea, because of their ethno-linguistic composition also including Atlantic and Mandé speakers (see this page). However during my survey it turned out that both countries are genetically speaking mostly characterized by “Ivory Coast/Ghana” (2013-2018). And in that sense they largely  fall outside of the Upper Guinean genetic component described in table 6.1.

Jamaica’s strictly Senegambian slave trade proportion is minimal (2.3%). And amazingly it corresponds very closely with the group average of “Senegal” I found for 100 Jamaicans: 2.4% (see this sheet). “Mali” attaining a higher group average of 7.4%. And at times even reported with double digit amounts or as primary region (2x). I suppose this might imply that in particular for Jamaicans the alternative interpretation of “Mali” might be quite relevant. That is “Mali” being indicative also of Gur speaking lineage from northern Ghana and beyond into Burkina Faso and surrounding areas. As indicated also by a high frequency of so-called “Chamba” captives in Jamaican runaway advertisements.

Similar to Jamaicans also for Haitians I obtained a higher “Mali” group average (6.6%) than for “Senegal” (3.7%). With higher outliers for “Mali” as well. Although among my 97 Haitian survey participants I only observed 1 primary “Mali” score. Generally speaking the Upper Guinean lineage of Haitians appears to be diluted to a large extent. Still in some contrast with other parts of the Diaspora it may be assumed that Upper Guinean roots of Haiti hail mostly from northern and interior parts of Senegambia as well as Mali. Corroborated by principal slave trade from actual Senegalese slave trade ports and a high incidence of documented “Bambara” captives in plantation inventories. See also:

Brazilians

The top-ranking position of Amazonia (=northern Brazil, esp. Maranhão) in table 6.2 might be surprising to many people. Due to favourable navigation conditions unique to northern Brazil it received a great number of captives which were shipped from Guiné Bissau. However generally speaking Upper Guinean lineage is of much less importance in southeast Brazil (Rio etc.) as well as Northeast Brazil (Bahia etc.). And these are the areas where the greater majority of Afro-descended Brazilians live. This is also reflected in the lower “Senegal” + “Mali” group averages for my Brazilian survey group (14% versus 58% share of Senegambia in slave trade to northern Brazil).

Northern Brazil/Amazonia continues to be relatively underpopulated. It is still noteworthy that Brazil does not have a uniform African heritage. After all Brazil is a huge country and covers about half of South America! Regionally speaking things will vary according to differentiated slave trade patterns. Central/Southeast African and Bight of Benin lineage being predominant in the southeast and northeast, generally speaking. But with a more pronounced Upper Guinean influence in the north. Although domestic slave trade and post slavery migrations within Brazil might also have had substantial impact. See:

African Americans

The data for the USA (as measured by Virginia and South Carolina) is somewhat intermediate in all of this. As shown in table 6.1 the estimated Upper Guinean component being around 17%. Which would correspond with the slave trade records for only Senegambia (18-19%). But additional slave trade with Sierra Leone is then not covered. This being especially significant for South Carolina (14.4%). And even though more ambiguous some historians still also include the Windward Coast when discussing Upper Guinean slave trade in its broadest sense for the USA. But this portion of around 4-7% again seems to be not covered by the combined “Senegal” & “Mali” scores for my African American survey group.

Therefore also in light of my actual West African survey findings it seems logical to assume that additional “Ivory Coast/Ghana” scores are needed to do justice to the complete Upper Guinean inheritance for African Americans. As this region was after all very prevalent among Liberians as well as Sierra Leoneans. And to a minor degree even to be found in double digit amounts among Guineans and Malians (see part 1 and table 2.1). Obviously the actual correlation between genetic inheritance and known slave trade patterns is far more complex and might also involve other variables! Therefore tables 6.1 & 6.2 do not need to correspond per se. Then again exploring the discrepancies may lead to more insight.

As I have already speculated in 2015 it seems that strictly Senegambian origins for African Americans might have been overestimated by many people. Naturally genuine Senegambian lineage (from either Senegal or Gambia) does exist among African Americans. However not to the expected higher degree when going by perhaps somewhat misguided historians. The group average of “Senegal” being around 8% among my 350 African American survey participants. And “Mali” attaining a slightly higher group average of 9%.

A striking outcome of my African American survey in 2015 was a greater frequency of primary “Mali” scores when compared with primary “Senegal” scores (20/350 versus 5/350, see this chart). Possibly implying that the Upper Guinean roots of African Americans are more so to be found to the east and south of Senegambia proper. That is Guinea Conakry, western Mali and Sierra Leone. In line with principal slave trade ports for the USA being located in Gambia, Sierra Leone and Guinea Conakry. Mostly receiving captives from their hinterland which of course was not bounded by modernday borders.

Obviously more research is needed and and also some variation along USA state origins is to be taken into consideration. On the one hand it seems likely that “Senegal” and “Mali” scores might also have been partially inherited by way of (northern) Sierra Leonean and Guinean (Conakry) ancestors. But we can also be practically sure that the genetic contributions by way of Sierra Leone will be more differentiated. And similar to the Windward Coast it will also reflect in so-called “Ivory Coast/Ghana” scores for African Americans. Especially for those with strong South Carolina connections. I will explore this in greater detail in the next paragraph. See also:

 

Comparing slave trade from Senegambia & Sierra Leone: relative proportions

Table 6.3 (click to enlarge)

TAST, Diaspora comp. 2 coasts

Source: TAST Database (2019). Keep in mind that Inter-Colonial slave voyages are NOT included. The data is only based on documented Trans-Atlantic Slave Trade voyages. Amazonia is referring to northern Brazil; Hispaniola is referring to the Dominican Republic; Spanish Circum-Caribbean is referring mainly to Colombia & Mexico; Saint-Domingue is referring to Haiti.

***

Seeking a proper context is always essential when trying to interpret your own personal “Senegal” and/or “Mali” test results. Not only your unique family history will matter in this regard but also your nationality or even your provincial/state origins within your home country. Especially your earliest known ancestral locations in the Americas might be very indicative as discussed above. In order to provide more solid ground when seeking to determine the plausibility of various ancestral scenarios I have performed a customized search in the Trans Atlantic Slave Trade (TAST) Database.

Table 6.3 features the relative shares of slave trade originating from either the “Senegambia” coast line (incl. Guiné Bissau & Cape Verde) or the “Sierra Leone” coastline (also includes Guinea Conakry). Take note that it adds up to 100% along the row! In order to focus solely on the slave trade within this restricted Upper Guinean area I have left out all other slave trade regions from the equation. In other words it is purely a comparison meant to establish the relative mutual balance. The %’s are NOT referring to proportional shares of overall slave trade.

Table 6.3 is quite insightful for particularly African Americans and Jamaicans. But due to the exclusion of inter-colonial slave trade voyages this overview is regrettably least useful for Hispanic Americans. As table 6.3 is bound to underestimate the true proportion of slave trade originating along the Sierra Leone/Guinea coastline. Early slave trade to Latin America was often routed by way of Cape Verde. However the Upper Guinean captives which were transferred on Cape Verde actually hailed from various parts of Upper Guinea, incl. also Sierra Leone (see this map).

Furthermore the Hispanic Caribbean might later on also have received many captives from Sierra Leone by way of English/American contraband traders. In agreement with the frequently mentioned so-called Ganga or Canga, a slave ethnonym for captives brought in from Sierra Leone and northern Liberia. This data is however mostly not included in the TAST database. Although it can be gauged somewhat by the number of documented slave voyages departing from Sierra Leone to Puerto Rico. As well as for Cuba (not shown in this overview though).

In previous discussion of my Hispanic American findings (2015) I have pointed out the remarkably higher level of “Mali” scores for especially Mexicans. I speculated this outcome might be suggestive of specific ethnic lineages from Upper Guinea being more prominent for them than for Cape Verdeans. Given historical evidence these lineages are more likely to hail from Sierra Leone rather than Mali itself. In particular the so-called Zape or Sape captives from Sierra Leone being a likely candidate. An umbrella term referring to esp. the Temne and related ethnic groups. It was especially current in the 1500’s. Pending further research the validity of this hypothesis remains to be seen though. This outcome might also just be caused by an imprecise assignment of either “Senegal” or “Mali” scores for people with only minor total African DNA (<10%).  See also:

Table 6.3 shows that the relative importance of either Senegambia (incl. Guiné Bissau & Cape Verde) or Sierra Leone (incl. Guinea Conakry) varies for each particular destination within the Americas. But for most destinations in the Americas Upper Guinean slave trade by way of Senegambia would have been prevalent. Even when correcting for inter-colonial or clandestine slave voyages. There are however some notable exceptions. Jamaica and South Carolina clearly show a higher proportion of slave trade with Sierra Leone. This could have important implications on how to interpret their “Senegal” and “Mali” scores. As well as “Ivory Coast Coast/Ghana” actually. This latter aspect has already been discussed though. See also:

 

Upper Guinean slave trade by way of Gambia & Sierra Leone most prevalent for the USA

***

Map 6.3 (click to enlarge)

TAST - USA - countries of origin (percentages)

Source: TAST Database. Follow this link for underlying numbers. Based on estimated slave trade rather than documented slave voyages (as shown in tables 6.2 & 6.3). See also this section of the Slave Voyages website.

***

Map 6.3 is also based on the TAST Database. But this time not on documented slave voyages (as used for tables 6.2 & 6.3). But rather its estimates of 388,747 captives being disembarked in North America. Very useful because it is providing more detail beyond the traditional slave trade regions. Take note for example how the so-called “Sierra Leone” slave trade region is now split up between the modernday countries of Guinea Conakry & Sierra Leone (based on embarkation port details, see this overview). The same goes for the Senegambia information which is now split up between Senegal, Gambia, Guiné Bissau and Cape Verde. Just like the AncestryDNA regions these modernday country origins are not to be taken too literally though! Inland slave trade routes could very well cross borders after all.

Taking a cue from map 6.3 the ranking for the USA as a whole would be as follows (take note this time the estimated proportional shares are reflecting the entire slave trade from Africa):

  • Gambia: 17.01% (66,100)
  • Sierra Leone: 7.10% (27,600)
  • Senegal: 6.41% (24,900)
  • Guinea Conakry: 4.45% (17,300)
  • Guiné Bissau: 0.15% (600)
  • Cape Verde: 0.10% (400)

This overview can be quite useful when interpreted correctly. Direct slave trade from either Guiné Bissau or Cape Verde to the USA was negligible. While slave trade routed by way of Gambia and Sierra Leone was most prevalent. Again due to heavy movement of people across modernday borders the exact implications might be tricky to establish. However also taking into account other historical clues it seems that especially Mandé speakers from eastern Senegambia, Guinea Conakry, western Mali as well as both Atlantic and southwest Mandé speakers from Sierra Leone would then be involved.

Combined these more specific Upper Guinean origins for African Americans should be correlating with their “Mali”, “Senegal” and additionally also “Ivory Coast/Ghana” scores. As can be deduced from the actual AncestryDNA results of Upper Guineans (see table 2.1). Only by proxy of course. Although again it seems telling that “Mali” showed up somewhat more pronounced than “Senegal”, over all speaking for my African American survey group.  See also:

 

Looking for corroboration? Search for Upper Guinean matches!

As already mentioned for follow-up research it is highly advised to search for Upper Guinean DNA matches on Ancestry in the first place. 

Do keep in mind then that Fula matches might be over-represented due to a skewed customer databaseFrom my Cape Verdean survey findings it turned out that DNA testing has become especially popular among Fula people, compared to other West Africans. Stacking the odds of being matched with them. However this may not per se be in line with their overall proportional contribution to your DNA. Also remember that the Fula people did not only intermingle with Hausa people in Nigeria.

Also throughout Upper Guinea they have mixed with other ethnic groups. Such as the Wolof, Sereer, Mandinga, Bambara etc.. Usually by way of their maternal line. Despite some degree of endogamy among more traditional and still nomadic subgroups this is especially true for sedentary subgroups such as the Halpulaar or Toucouleur from Senegal and the Fulakunda or Fula Preto from Guiné Bissau. In this way receiving a Fula match does not automatically imply that your common ancestor was also Fula! Also Mandinga, Wolof or Sereer etc. MRCA’s can very well be possible.

Also when receiving a Sierra Leonean match you will still have to be very careful and check whether they do not have any partial Krio ancestry. Because in that case the ancestral connection might actually be reversed! And your MRCA could even be African American! Both Liberia and Sierra Leone having received freed ex-slaves from the US and the West Indies as settlers. In fact as testified by my own African survey group inter-ethnic unions are quite frequent across West Africa and not only among the Fula or in Sierra Leone!

Just as a final reminder it should be realized that the actual amounts being reported for the so-called “Senegal” & “Mali” regional components (2013-2018) are first-most a measurement of genetic similarity to most likely Mandenka, Bambara & Dogon samples. However this will by default only be a rough estimate as more fitting sample groups to compare with may very well have been missing within Ancestry’s 2013-2018 Reference Panel. Then again Ancestry’s arguably flawed update in 2018 has illustrated how an addition of mostly Malian samples was not per se beneficial. Because it resulted in a dramatic decrease of “Senegal” scores and inflated “Mali” scores all over West Africa (see this blogseries).

This loss of informational value will hopefully be corrected by Ancestry in its upcoming update (scheduled for 2019). However as an alternative you might also want to test with 23andme to get a second opinion on your Upper Guinean lineage. Going by actual West African 23andme results the predictive accuracy of their 3 new West African regions (“Senegambian & Guinean”, “Ghanainan, Liberian and Sierra Leonean” and “Nigerian”) is quite impressive and compares quite favourably with AncestryDNA’s 2013-2018 version. While it is arguably (much) more reliable than Ancestry’s current version (September 2018 update). See also:

 

Map 6.4 (click to enlarge)

23andme vs Ancestry

The map on the left shows the distribution of “Senegambian & Guinean” scores among my African 23andme survey participants. The map on the right shows the group averages for “Senegal”. Both regions appear to be quite similar in scope. The predictive accuracy being somewhat greater on 23andme though and also extending into Guinea Conakry. Going by preliminary group averages. The “Mali” region on Ancestry however does not (yet) have an equivalent on 23andme. Although ambivalent “Mali” (2013-2018) could still be of added value when interpreted correctly. Then again through its “Ghanaian, Liberian & Sierra Leonean” region 23andme can now provide a much better coverage of Sierra Leonean DNA.

_______________________________________________________________________

Notes

1. I am aware that ethnicity can be a sensitive topic for West Africans. Aside from wanting to learn more about my own West African roots my motivation to research these so-called Ethnicity Estimates is purely scholarly. It stems from a deep fascination with West Africa’s ancient population migrations, its many connections with the Afro-Diaspora as well as a profound love for its vibrant & diverse culture. I do not condone the misuse of my research for identity politics! I am in full support of democratizing knowledge. However when parts of my work are being copied I do expect that a proper referral to my blog and the additional context it provides will be made.

2. I like to express my sincere gratitude to all the persons who made this survey possible! Almost all results featured with a screenshot on this page have been shared with me directly by invitation. Furthermore I received tremendous help of friends who shared the results of their Upper Guinean DNA matches with me. Also my survey of African DNA matches being reported for Cape Verdeans provided me with many Upper Guinean results (by way of Ancestry’s Compare Ethnicity tool, see this link for an overview).

A few results were also collected by me from public websites or social media. As I found them to be of potentially great educational benefit for others. I have asked for prior consent whenever I could but regrettably was not able to do so in all cases. I have naturally taken great care to cut away any name details in order to safeguard everyone’s privacy. Apologies in advance to anyone who recognizes their results and is not comfortable with this blog page featuring them. Please send me a PM and I will remove them right away.

3The Fula people go by several names, usually borrowed from other languages. Illustrating their extensive geographic distribution across the Sahel region. From Upper Guinea into Sudan. The ethnonym used by Fula people themselves would be Fulbe or Pullo (singular). I am choosing however to use Fula, which I will use throughout this pageBecause historically speaking this was the main term they have been known by in Cape Verde. This name originally being derived from the Mandinga language, which has been greatly influential on Cape Verdean Crioulo (see this link).

In other languages other ethnonyms are more current. For example Peul in French and Fulani in English. These names are likewise historically derived from neighbouring ethnic groups (resp. the Wolof in Senegal and the Hausa in Nigeria). Whenever I am using the word Fulani instead of Fula I am therefore specifically referring to the eastern offshoot of the Fula people. Nowadays residing in places such as Niger and northern parts of Nigeria and Cameroon. Often intermingling with local Hausa people etc. but originally arriving from Upper Guinea. And more specifically the Senegal river area being their presumed homeland. For more details see also:

4Many of the Fula results (10) within my survey were shared directly with me by the persons themselves. Also their ethnic background was then usually confirmed to me by PM. However the greater part of my Fula results were obtained by the very gracious help of a friend of mine who happens to be Fula himself. Analyzing his closest DNA matches by way of surname and other relevant profile details we were able to single out persons with a very high plausibility of being of Fula descent. Ancestry’s Compare Ethnicity tool (available since 2018) enabled me to have access to the complete ethnicity estimates of these matches. His Fula DNA cousins tend to share several segments with him. Illustrative of Fula endogamy I suppose. These matches were scattered across many countries, reflecting wideranging Fula migrations across the centuries. See this spreadsheet for more details:

5. The substantial “Africa North” group average (~13%) for my Fula survey participants is very distinctive. However this component is not entirely  unique to them. Within a strictly (mainland) West African context “Africa North” is indeed very rare among persons without any (distant) Fula lineage. However three of my West African survey participants confirmed to me that they are of partial Moroccan descent. Because they are therefore of mixed background I did not include their results in any of the group averages I calculated for this page. This involves one mother and son from Mali and also one person from Mauritania/Sierra Leone. See section 4 for screenshots of their results

For different reasons and therefore requiring very careful interpretation “Africa North” and also “Middle East” was also reported in considerable amounts among my Northeast African survey participants. Furthermore also among Cape Verdeans “Africa North” scores were very common in the 2013-2018 version. In their case again careful reasoning needs to be applied. Also taking into consideration geneflow by way of Portuguese ancestors. See also:

or see also this chart (based on unscaled amounts!):

6Obviously in selected cases Fula people might have genuine ancestral ties with the Middle East and even Europe. As indicated by minor “Middle East” or also “Iberian Peninsula” and “Europe South” scores.  However unless substantiated by more solid evidence (genealogy!) it would be a mistake to take regional admixture results too literally. As always proper interpretation and awareness of the relevant context being essential. Even when not directly leading towards recent ancestors (last 500 years or so) from either Europe or the Middle East such results are however pointing out the distinctiveness of Fula genetics!

I have had an intriguing discussion about this topic with a very knowledgeable Fula friend of mine who tested with Ancestry. He scored 3% “Iberian Peninsula” and <1% “Europe South”, similar to several other Fula results within my survey.  This is most likely just North African (like) DNA which is being mislabeled. Because actual Moroccans and Algerians do not tend to score 100% “Africa North” on Ancestry but rather are described as a composite of “Africa North”, “Middle East” and South European regions.

I have made a thorough analysis of my Fula friend’s DNA matches in March 2018 (by way of my scanning & filtering method). Out of his 6,305 DNA matches in total (see this overview), I was able to filter out only 25 matches without any African (Sub-Saharan) regions in their profile. While only a handful also did not show any “Africa North” or “Middle East”. Quite astonishing in itself but some of these matches may yet turn out to be false positives or glitches due to incomplete display or perhaps they still do have diluted African admixture but Ancestry wasn’t able to detect it. I believe if his European scores had been legit he would have gotten a far greater number of European American matches. For a Liberian with some Krio lineage and only <1% European admix on Ancestry I found over 500 European-American matches with no African regions at all, out of a total of 3,372 DNA matches (again based on a scan performed in 2018)!

Still in selected cases I do think that some Fula people could have genuine European lineage. I have actually asked my Fula friend about this as well. I first said:

____________________

“I am wondering if perhaps in their dealings with mixed-race (Euro-African) trader communities living along the coasts of Guinea and Sierra Leone the Fula might not also have been “given” any of these mixed-race traders daughters as wives to cement their mutual trading relations if you like. These mixed-race traders have been documented from early on. In regards with the Portuguese and Cape Verde they were often known as Lançados or Tangamaos. But in the 1700’s they were also joined by the Euro-African offspring of American and British traders, along the Rio Pongo and Rio Nunez. These mixed-race Anglo families are well documented and are described especially in the work of George E. Brooks (1993, 2003, 2010).

____________________

Upon which he replied:

____________________

” You are absolutely right, when you describe how marriages/alliances used to cement relationships between different communities. This is precisely how my father explained it to me. If we consider the fact that more importance was given to the paternal lineage than to the one of the maternal side, the circumstances were sort of reunited for these alliances to become realities. I had heard of the Creole from Sierra Leone and communities of returnees from the Afro-Diaspora in West Africa. It seems that they also had privileged relationships as traders and intermediate business “partners” with Europeans at some point. I have wondered if in one of my ancestral lines, individuals may have been integrated”

____________________

7. A pioneering study of Fula genetics using autosomal DNA has been Tishkoff et al. (2009) which identified a separate Fulani cluster. And also mentioned (p.16) that:

____________________

“The origin of European (possibly via the Iberian peninsula) and/or Middle Eastern ancestry in the Fulani requires further exploration with additional genetic markers”

____________________

Most recently a very intriguing paper (Vicente, 2019) found that the lactase persistence trait found among Fula nomads might possibly be inherited by way of ancient Trans-Gibralta geneflow from a neolithic Iberian population. Which also was absorbed within the North African genepool. This finding could be used to explain the minor South European scores being reported for Fula people on Ancestry and also on 23andme I suppose. Follow these links for full articles:

8. The ethnic headers are merely meant to provide additional perspective. Even if at times only approximate info was available to me. Naturally I respect everyone’s right to self-identify as they please. My accompanying comments should be taken as informed speculation on my part. They are not meant to exclude other possibilities or simplify complex family histories.

9. I usually did not have the birth locations of all 4 grandparents behind each result. Aside from PM’s and public family trees I also tried to verify the background of my survey participants in other ways. Naturally to the best of my ability. Going by any clues given but in particular: plausible surnames and plausible regional combinations as well as the “migrations” mentioned in AncestryDNA’s Ethnicity Estimates. Taking a cautious approach and preferring to leave out possible survey participants when in doubt. For each possibly African surname, first name or even nickname I have performed a search on not only Google but also Facebook to find out more details. The forebears.io website has been a particularly helpful resource in these efforts. See also section 5 (Methodology) of this page:

10. It would have been very interesting to see how high the “Senegal” group average might get for a sizable Limba survey subgroup! The Limba people being from the north and apparently assumed to be the oldest inhabitants of Sierra Leone. While their language used to be classified as Atlantic.

____________________

Although there are numerous ethnic groups in Sierra Leone, the Mende and Temne together account for approximately 60% of the total population. To see if genetic differences could be observed among ethnic groups in Sierra Leone, the nucleotide sequence of the hypervariable 1 (HV1) region of mitochondrial DNA (mtDNA) was determined from samples of the two major ethnic groups, the Mende (n=59) and Temne (n=121), and of two minor ethnic groups, the Loko (n=29) and Limba (n=67).

Among these 276 HV1 sequences, 164 individual haplotypes were observed. An analysis of molecular variance indicated that the distribution of these haplotypes within the Limba sample was significantly different from that of the other ethnic groups. No significant genetic variation was seen between the Mende, Temne, and Loko. These results indicate that distinguishing genetic differences can be observed among ethnic groups residing in historically close proximity to one another.” (Jackson et al., 2005, p.156)

____________________

The study mentioned above is of course restricted as it is based on haplogroup testing. And not on genome-wide autosomal testing, as performed by AncestryDNA. Still this particular study resonates with me personally as well. Because my maternal haplogroup happens to be L3e4 (as tested by 23andme). And it is mentioned explicitly by the authors:

____________________

“For instance, the five members of haplogroup L3e4 in the Limba sample were the only ones found among the Sierra Leone samples” (Jackson et al., 2005, p.162)

____________________

I do not believe my maternal haplogroup is exclusive to any ethnic group or can be used to pinpoint a likely ethnic background. This is sadly still a very common misconception. And in fact the authors also mention L3e4 being found among Mandenka and Wolof samples (see this table). However from a regional perspective (wider area of Upper Guinea) and also taking into account a possibly more ancient time frame of Atlantic speaking migrations this finding does seem quite relevant for my direct maternal line.

11. The belief that especially subcontinental predictions in admixture analysis are bound to be very inaccurate has been repeated almost like a mantra by some people. Even when from my AncestryDNA findings it may be deduced that great informational value can be obtained (given correct interpretation). For an overview of blog articles discussing the usefulness of admixture analysis, see heading “Blog Posts” in this ISOGG entry: Admixture Analysis. To be sure each particular DNA testing company (and also any update of their results) will indeed have its own flaws and strengths. Also the perceived accuracy of ethnicity results may vary according to a person’s own main background. As well as their expectations of what admixture analysis should deliver.

Without wanting to be divisive I do find it noteworthy how the ethnic backgrounds of  bloggers within the genetic genealogy “world” seem to often determine their outlook on admixture analysis. Perhaps not that surprising given that their evaluation will be based mostly on how their own DNA is being described. Nothing wrong with personal reviews. But I do think that people of African descent should realize that Afro-Diasporan predicaments are not the same as those of bloggers of fully European descent. Still because these bloggers tend to have a great following their stance is often also copied by African Americans or other Afro-descendants.

White Americans/Canadians and Europeans who take DNA tests tend to have well researched family trees and detailed knowledge of their recent ethnic origins. This allows them to be “picky” about their ethnicity estimates. As after all they will have the luxury to verify their results. Which often tends to make them obsess on details such as the labeling of ancestral categories rather than trying to grasp the overall added value it may have for other people.

For example they might insist on seeing their Irish ancestry clearly distinguished from their Scottish one. And will get upset when this is not accomplished. Even when naturally Irish & Scottish DNA will be very tricky to separate due to a great degree of genetic overlap. But with an overly dismissive stance you risk loosing out on valuable insights to be obtained! For example when taking a macro-regional approach you might still want to appreciate the ability of DNA testing companies such as Ancestry & 23andme to distinguish quite reliably between Northwest European DNA and East European DNA. Or to quite reliably indicate Jewish lineage.

I find it interesting to contrast this with the more constructive attitude among bloggers of mixed or non-European descent. Often seeking ways to already use admixture analysis pragmatically despite obvious shortcomings. See for example:

Generally speaking I suspect some degree of ethno-centric bias is quite likely to influence how people will judge the usefulness of admixture analysis. Again not trying to be accusatory or anything. Because I believe I could very well be guilty of such bias myself as well! I am of both Cape Verdean and Dutch descent. My blogging interests are more wideranging. After all I am focused on covering the entire Afro-Diaspora and also regularly discuss African topics. Still my own specific background may (consciously or unconsciously) direct me in certain ways. I guess what I’m trying to say is when deciding on the usefulness of admixture analysis don’t let other people, social media, genetic genealogists or even bloggers (myself included 😉 ) dictate your choices! Do inform yourself properly but make up your own mind based on facts relevant to your own situation rather than on sometimes subjective opinions.

12. Actually it seems that quite recently the broader Atlantic language group has been declared obsolete because of lack of coherence between the subgroups. At least according to Wikipedia. I have nonetheless maintained it on this page. Partially because it is still a convenient geographic umbrella term. But also because regardless of whether formerly Atlantic subgroups from the south (Mel etc.) may be deemed too divergent from northern Senegambian languages (such as Wolof, Sereer & Fula). The fact remains that these languages combined still represent populations groups which inhabited Upper Guinea before the expansion of Mandé speakers. See also:

13. I will be working under the assumption that these “Senegal” and “Mali” scores are reported as main region and also with a considerable relative share within the complete African breakdown. Arguably the main regions appearing in your results might be more deserving of your research efforts. Your research results might then be more fruitful and covering a wider span of your ancestral make-up. As afterall the regions with the biggest amounts can be deemed to be more solidly based on your most important regional origins (even when taken as mere proxies).

Therefore any reporting of “Senegal” and “Mali” as mere trace regions with low confidence will be left out of consideration. As the labeling already implies in such cases you are dealing with an increased possibility of a false positive or misreading of your DNA.  Depending on the particulars actually such scores might still have informational value. But the array of ancestral possibilities simply becomes too broad for any meaningful discussion on this page.

14The very low odds of tracing back to Africa for African Americans were highlighted during an episode of Finding Your Roots by Henry Louis Gates Jr.. The guest of his show, musician and producer Questlove, was found to be descended from one of the last enslaved Africans to arrive in the US on a slave ship.

____________________

“The discovery of ancestors on the Clotilda isn’t just an interesting genealogical fact. As Gates says, it means that Questlove is the only African-American he knows who can answer a question that many have asked: not only where in Africa his ancestors came from, but how exactly they got to the U.S. in the first place.”

____________________

This certainly was a very remarkable and precious finding! But such a verifiable paper trail leading back to Africa must be extremely difficult to reproduce for ordinary African Americans. For one thing they will not be assisted by professional teams of historians and genealogists. Furthermore this finding only concerns one particular family line (relatively recent) among possibly hundreds of others. All individually to be traced back to several parts of Africa! How is a layman expected to ever uncover a majority of these lines, let alone one single one?  This (near) impossibility of the genealogical route for me underlines once more how there is a lot more at stake for Afro-Diasporans when taking a DNA test than for people with plentiful documented knowledge about their ancestral origins.

15. Currently you are much more likely to get a realistic estimate of your West African lineage on 23andme. Even if of course not 100% accurate 😉 Still going by actual West African 23andme results the predictive accuracy of their 3 new West African regions (“Senegambian & Guinean”, “Ghanainan, Liberian and Sierra Leonean” and “Nigerian”) is quite impressive. To be fair Ancestry’s African breakdown during 2013-2018 was far better in my opinion than 23andme’s African breakdown during that same period (before their update in 2019). Unlike any of its commercial competitors AncestryDNA was able to distinguish no less than 5 different West African regions. In addition it also did a reasonable job at separating West African DNA from Central/Southern African DNA (“Cameroon/Congo” and “Southeastern Bantu”). Which is essential for getting a full scope perspective on the origins of the Afro-Diaspora. See also:

16In 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 have surveyed (see this sheet) and “Mali” usually coming in as second main region.

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 and surrounding countries. Usually also a variety of sometimes surprising Trace Regions were reported for Cape Verdeans but with reduced confidence level. For more analysis of Cape Verdean AncestryDNA results see this page:

17. The mere fact that within my survey “Mali” was also reported in high amounts for Gur/Senoufo speaking individuals could be highly relevant for Afro-Diasporans. Many might be too quick to link “Mali” solely with Mandé lineage but Mali is very much a multi-ethnic country! And so is Ghana which is not only home to Akan or Gbe speakers but also Gur speakers in the north (see this map)! It is very useful to be aware that throughout the West Indies “Chamba” was once a frequently used ethnonym for enslaved captives hailing from northern parts of Ghana, Togo and Burkina Faso. According to some historians it is a reference to Gur speakers in particular (such as the currentday Chamba). Again I like to underline that so-called “Mali” is not an exclusive marker for any given ethnic group or nationality. But it is also not to be exclusively associated with the wider Upper Guinea area. For more info:

18. Ideally you would want to verify if the shared DNA segment with your Upper Guinean matches is also showing up as either “Senegal” or “Mali”. Or even “Africa North” and “Ivory Coast/Ghana” to verify Sierra Leonean or Fula matches in particular. Because that way you could have more certainty that these matches will indeed relate to your own Upper Guinean admixture. Regrettably this potentially very useful information is not available because Ancestry so far has not implemented a chromosome browser. See also:

19. It should be pointed out that only Trans Atlantic slave voyages are being included in tables 6.2 & 6.3 as well as map 6.3. And not Inter-Colonial slave voyages! So this data is not intended to reflect the full picture. For example largely undocumented English contraband slave trading was very significant for the Hispanic Americas and to a lesser degree also Haiti (“Saint Domingue”). While for the USA especially Domestic Slave Trade from the Upper South has made a great impact. For all places mentioned obviously also post-Slavery migrations should be taken into consideration (for more disclaimers see this page).

20. Early slave trade to Latin America was often routed by way of Cape Verde. However the Upper Guinean captives who were transferred on Cape Verde actually hailed from various parts of Upper Guinea, incl. also Sierra Leone. Furthermore especially for the Hispanic Caribbean the factor of largely undocumented slave trade looms large. It is known that Puerto Rico and the Dominican Republic also received many captives by way of English, French, Dutch and Danish contraband traders. Incl. again from Sierra Leone and actually also the Windward Coast (often named “Canga”). This data is mostly not included in the TAST database or table 6.2. Although it can be gauged somewhat by the number of documented slave voyages for Puerto Rico.

In addition we should probably also take into account post-colonial migrations by English speaking West Indians into Puerto Rico and the Dominican Republic. Making for a rather convoluted situation which thanks to genetics is now however becoming more clearer. In my previous discussion of Hispanic AncestryDNA results I have already established that any Upper Guinean founding effect will be most noticeable among Hispanic Caribbeans with a relatively smaller total amount of African DNA (>25%). In line perhaps with more diluted and therefore older African origins (1500’s/1600’s versus 1700’s/1800’s). I will elaborate on this so-called sub-structure eventually.  See also: