Ancestry’s new African Breakdown: merely cosmetic changes?

Only a few weeks ago Ancestry went ahead with their ill-advised deletion of smaller DNA matches (6-8 cM). Resulting in a great loss of customer value. But already the next update is being rolled out. This time once again our Ethnicity Estimates are being reshuffled. It almost seems Ancestry is making it a yearly tradition to perform their ethnicity updates in the Autumn. Or should I say a yearly marketing ploy? Either way this is already the third time in a row! Starting with the first distastrous make-over of Ancestry’s African breakdown in 2018. Things did get better though in 2019. My verdict of last year was: “a step in the right direction but no substantial improvements for the most part. At least not when compared with the original African breakdown from the 2013-2018 version”. And really this assessment still stands also for this 2020 update. There have been a few positive changes, but nothing game changing

***

Map and full listing of all the 12 African regions available on Ancestry after its 2020 update. Despite the renaming of  a few regions the only new region is “Southern Bantu”. This region will however be minimal or absent for practically all Trans Atlantic Afro-descendants.  Because of the restored “Ivory Coast/Ghana” region the West African breakdown is now more coherent than in 2019. However it still has several shortcomings…All in all nothing substantial to make up for the recent purge of smaller DNA matches.

***

I find it disappointing that most of my suggestions for meaningful improvements (originally posted in 2018) have still not been taken up by Ancestry. Yet again this update seems to be centred mostly around providing greater regional resolution for Europe and Asia. The African breakdown seems to be merely a sideshow. The finer distinction to be made for Scottish ancestry is certainly striking but probably also overambitious.1 Such a focus is to be expected given that Ancestry’s customer database is overwhelmingly of European descent. While Asia probably represents a promising growth market. However the relative neglect of African & Afro-descended customer needs does go against Ancestry’s selfproclaimed goal to make their product experience inclusive for everyone. In my previous blog post I stated that Ancestry should seek to offer new tools geared to facilitate specialized research for Afro-descended customers. It should be clear that this update does NOT compensate for the loss of small African matches, earlier this month.

It is still my belief that each updated version as well as each separate DNA testing company should be judged on its own terms.2 For the sake of correct interpretation I have therefore yet again performed a comprehensive survey among 135 African Ancestry testers from all over the continent to evaluate the changes before and after this update. In addition I have also looked into a representative array of 50 updated results from across the Afro-Diaspora. These findings will be described in greater detail further below. Again for the most part no major changes. Which is why I will keep this discussion brief and only highlight the main outcomes:

  1. African breakdown for Africans before and after the 2020 update
  2. African breakdown for Afro-descendants before and after the 2020 update
  3. Is Ancestry getting sloppy?
    • When will we have genetic communities for West/Central Africa?
  4. Screenshots of African updated results

As always make an attempt to inform yourself properly without being overly dismissive. Despite shortcomings I do still think you can get valuable ancestral clues from Ancestry’s African breakdown. The macro-regional breakdown also to be taken into account to get a grasp on the greater picture. Instead of being preoccupied about the appearance of any surprising but minimal %’s. Which might very well disappear with the next update 😉 Such an approach to be combined with your remaining/salvaged African DNA matches, historical plausibility etc.. My previous discussion of the 2019 update may still offer helpful guidance. Ancestry’s FAQ is also useful:

1) African breakdown for African AncestryDNA testers

Table 1.1 (click to enlarge)

Based on the updated results for 135 African AncestryDNA testers from 34 countries, across the continent. Take notice that the predictive accuracy in most cases is quite solid. Although in a few cases it is still clearly in need of improvement. This goes especially for “Ivory Coast & Ghana” and “Eastern Bantu”. Although actually compared with 2019 there has already been some progress made also for these 2 regions. Follow this link for my spreadsheet containing all the individual results. Compare also with this overview from 2019.

***

In a nutshell the main features for Ancestry’s updated African breakdown are: 1) new region “Southern Bantu” 2) restoration of “Ivory Coast & Ghana”. Furthermore because of the addition of new samples into Ancestry’s Reference Panel the predictive accuracy of all 12 regions has also been on the increase for the most part. At least for native Africans themselves. For Afro-descendants with a more complex background the outcomes are more ambivalent. As was already the case after the 2019 update. Least significantly the region formerly called ““Southern & Eastern Africa Hunter-Gatherers” has now been renamed into “Khoisan, Aka & Mbuti people” (labeled “Khoisan/Pygmy” in my charts). For more insight do also compare with my previous findings from 2019 (based on the results of the same persons featured in my current survey):

“Southern Bantu”

Map 1.1 (click to enlarge)

You can consult this kind of useful information for each region mentioned in your results. Simply by clicking on “Learn more about this map and ethnicity”.  These maps are just meant to be seen as indicative however! This one for “Southern Bantu” is pretty much in agreement with my own survey findings. Do notice though that Madagascar is being left out. Unlike in Table 1.1 where it is mentioned that on average my Malagasy samples showed 58% “Southern Bantu” (within their scaled African breakdown).

***

The creation of the new “Southern Bantu” region is in itself a very welcome addition to Ancestry’s African breakdown. Something which I have been advocating already in 2018. Effectively the former “Southeastern Bantu” region has now returned. But at the same time it has been split up to indicate two separate portions for Bantu lineage from Southern and East Africa. “Eastern Bantu” already having been introduced in previous updates. While “Cameroon, Congo & Western Bantu” is now more properly equipped to detect Central African lineage. Naturally there is still a great deal of genetic overlap to account for. It will undoubtedly be a work in progress. But the current outcomes are hopeful already. As can be seen in Table 1.1 “Southern Bantu” peaks among South Africans. While for Angolans and Congolese it is barely showing up beyond trace level. Zambia, Zimbabwe, Mozambique and Madagascar being intermediate. A very similar outcome was already to be seen when the former “Southeastern Bantu” region was still in place. See also:

“Ivory Coast & Ghana”

Map 1.2 (click to enlarge)

Based on my own findings in Table 1.1 I strongly suspect these numbers are overly optimistic. “Ivory Coast & Ghana” scores of  greater than 75% being very rare, even for persons of Akan descent. Exception being the atypical cases of 100% scores whenever someone has been selected as a customer sample within Ancestry’s Reference Panel.  The inclusion of Liberia does make sense. But judging from my Ghanaian Ewe samples  I am very doubtful “Ivory Coast/Ghana” will reach such a substantial level of 50%+ in  eastern Ghana, let alone Togo…

***

I guess it is also to be applauded that Ancestry has recognized their huge “Ghana” mistake in 2019 by reinstalling the “Ivory Coast & Ghana” region. The effects are still far from optimal though. According to Ancestry’s new White Paper (p.9 ) “Ivory Coast & Ghana” still has the weakest predictive accuracy when measured by the very wide range of estimates among Ancestry’s own samples. When going by my survey findings (fewer in number but ethnically specified!) we can see that the predictive accuracy of  “Ivory Coast & Ghana” still has a long way to go. Especially among my Akan samples “Ivory Coast & Ghana” is only barely above 50%, on average. When formerly (2013-2018 version) it used to be over 90%! But at least the derailing trend has now been reversed.

Do also notice from table 1.1 that one further consequence is that Liberians are now pulling less to “Mali” and starting to move closer again to “Ivory Coast & Ghana”. Although to a much lesser degree than in the 2013-2018 version. Possibly because Ancestry has not yet fully restored all Ivorian Kru samples? This is explicitly being shown in the map above as well as in the new “Mali” map which no longer fully includes Liberia. Sierra Leone on the other hand is still firmly described by “Mali”, 100% even when going by Mende samples.

 

Ancestry’s new Reference Panel

Table 1.2 (click to enlarge)

Aside from the brandnew “Southern Bantu” region (+162) the greatest increase in samples took place for “Benin & Togo” (+179) and “Ivory Coast & Ghana” (+127). Otherwise no addition of West or Central African samples. However the 3 East African regions did receive some extra samples, which also contributed to an increased predictive accuracy for these regions (see table 1.1).

***

For a greater understanding of your Ethnicity Estimates it is always advised to learn more about the methodology used by Ancestry to produce their results. In particular Ancestry’s Reference Panel and Ancestry’s customized algorithm are key aspects. For previous updates in 2018 & 2019 I dealt with this topic in great detail already. And much of what I discussed then is still relevant also for this 2020 update. The composition and number of African samples contained in Ancestry’s Reference Panel has once again changed somewhat. But not as much as in 2019. As can be seen in the overview above in fact most regions did not see any change at all. Recommended reading:

Ancestry’s algorithm apparently has been “freshly updated” in the way it deals with phasing. But because of the way it has been designed this algorithm continues to have a homogenizing or over-smoothing effect. We can observe this for example by how 3 of my South African survey participants (Swazi & Xhosa) are described as being practically 100% “Southern Bantu”. Even when in previous results as well as when tested by 23andme they used to receive minor but still very substantial “Hunter-Gatherer” scores within the range of 10-20%. Renamed now into “Khoisan, Aka & Mbuti”. Khoisan lineage is however a confirmed ancestral component for Bantu speaking South Africans! Going both by oral traditions as well as several DNA studies. This is just one example because the same thing (concealing of minor admixture which is consistently present on a population level) also happens for Central Africans, Moroccans, Ethiopians or any other groups which happen to be a perfect match for the samples contained within Ancestry’s Reference Panel (see table 1.1).

In regards to Ancestry’s updated Reference Panel (see table 1.2) I would just like to make the following observations. Otherwise please refer to my 2018 & 2019 discussion.

  •  The proportion of African samples in Ancestry’s current Reference Panel is only about 6% right now (2856/44703). This does mean a slight increase when compared with 2019. When this relative share of total African samples was around 5% (2310/40017). However it is clearly a reflection of the minor importance of African samples within Ancestry’s evolving Reference Panel.
  • Still the new “Southern Bantu” region is starting off with a robust number of samples (162). I have not yet come across any reference on the specific background of these samples. However it seems fair to assume these samples were taken among South African Bantu groups. As the highest (nearly 100%) scores for the new “Southern Bantu” region were obtained for 3 of my South African survey participants (Swazi & Xhosa).
  • Otherwise most of the increase in African samples went to just two regions: “Benin & Togo” (+179) and “Ivory Coast & Ghana” (+127). This seems to have been beneficial especially for improving on the predictive accuracy of “Benin & Togo”. Probably because numberwise it is now on a more equal standing with “Nigeria” (466 vs. 521). Notice for example in table 1.1 how “Benin & Togo” is around 90% for 3 Beninese DNA testers and around 80% on average for 3 Ghanaian Ewe DNA testers. This is a higher level than achieved in 2019 (around 65%) Also noteworthy how among Yoruba Nigerians “Benin & Togo” is again forming a considerable part of their breakdown. Even when “Nigeria” will usually be clearly predominant. For Nigerian Igbo’s I have not observed any meaningful “Benin & Togo scores though, beyond trace level.
  • The restoration of the “Ivory Coast & Ghana” region probably implies Ancestry has once again added Ivorian samples into their Reference Panel. However I have not come across any details about this. One wonders if aside from Ivorian Akan samples this time also Ivorian Kru and South Mandé samples have been added again? The 2019 “Ghana” region was obviously a flop. Despite some easy gains in predictive accuracy “Ivory Coast & Ghana” still needs plenty of work to be a meaningful aspect of Ancestry’s African breakdown.
  • The former sample imbalance within West/Central Africa has been resolved for the most part. However relatively speaking “Senegal” looks undersampled now with 114 samples when compared with “Mali”, “Benin & Togo”, “Nigeria” and “Cameroon, Congo & West Bantu”. These latter regions all having 400+ samples. Actually “Senegal” is working great now for people with relatively homogenous African backgrounds (incl. Cape Verdeans!). However for the Afro-Diaspora the outcomes are more mixed. Especially among non-Hispanics “Senegal” seems to be structurally underestimated right now.
  • The moderate increase in samples for the 3 East African regions has also resulted in an improvement of predictive accuracy for “Ethiopia & Eritrea”, “Somalia” and to a lesser degree “Eastern Bantu”. This only goes for people native to these regions though. Because of missing reference samples from Nilotic and Chadic groups (Sudan & Chad) Ancestry’s African breakdown is still not fully covering the whole African continent in its main aspects. This shortcoming actually may also explain the surprising but historically implausible occurence of East African trace region scores being reported for Atlantic Afro-descendants. A Sahellian proxy explanation being more likely.

 

2) African breakdown for AncestryDNA testers across the Afro-Diaspora 

Table 2.1 (click to enlarge)

Based on the updated results for 50 AncestryDNA testers from across the Afro-Diaspora. Much has remained the same more or less. Hardly any major differences when compared with their pre-update results. The primary regional scores for each nationality still make sense, historically speaking. Arguably a bit more so compared to 2019 when “Nigeria” and also “Cameroon, Congo, Southern Bantu” were overstated for many people. Despite some improvement  “Ivory Coast & Ghana” is still the weakest link. The individual results can be seen by way  of my spreadsheet.

***

In order to judge how Ancestry’s 2020 update has impacted the results of people across the Afro-Diaspora I have also performed a small but still comprehensive survey for various Afro-descended nationalities (n=50). As already argued above there were hardly any major changes. Much remained the same, give or take a few %’s. As can be verified from table 2.1. above.The main outcomes are still more or less consistent with historical plausibility. Especially when applying a macro-regional framework. A 3-way division into: Upper Guinea, Lower Guinea ,Central & Southeast Africa. Much of what I discussed for the 2019 update still stands. So in case you like to have more details just reread section 3 of the blogpost below. It might also be insightful to see how the individual results of my survey participants have changed during all 3 updates till now. 

 

General trends

Luckily Ancestry’s West African regional framework seems to be gradually recuperating after the blows it had to deal with in the 2018 & 2019 updates. However there are still some serious shortcomings. Below a quick summary of my thoughts when comparing the updated results with the group averages of 2019.

  • “Benin & Togo” is showing up more strongly for most of my survey groups. In particular Barbadians. It seems that the overcorrection of “Nigeria” scores that took place in the 2019 update has been brought back to a more realistic level. Due to inevitable genetic overlap the distinction between these two regions will always be tricky though. Furthermore in many cases “Benin & Togo” will also be hiding actual Ghanaian lineage as well!
  • Predictive accuracy of “Ivory Coast & Ghana” is still under par. Although some progress has been made.  Especially for Jamaicans and Barbadians it will continue to greatly underestimate genuine Ghanaian DNA. For a more realistic estimate of such lineage you are still better of with 23andme’s “Ghanaian, Liberian & Sierra Leonean” category.
  • The Upper Guinean regions of “Senegal” and “Mali” are pretty much on point for Cape Verdeans and Latin Americans. However especially “Senegal” seems to be structurally underestimated for African Americans and other survey groups. For many African Americans and West Indians the decline in “Mali” scores is probably caused by how their Liberian DNA is now more so being described by “Ivory Coast/Ghana” (as was also the case in the 2013-2018 version).
  • “Northern Africa” scores have been increasing for Cape Verdeans and Latin Americans. Probably in most cases at the expense of either “Portugal” or “Spain” scores. This brings back the levels of minor but still distinctive North African admixture seen already in the 2013-2018 version. This is not just some fluke outcome but in line with historical expectations. And also actually to be backed up by associated North African DNA matches.
  • For many people “Cameroon, Congo & Western Bantu” scores will be practically the same as their previous “Cameroon, Congo & Southern Bantu” scores. Underlining how Central African lineage is much more prevalent than Southeast African lineage for Atlantic-Afro-descendants. Only for Puerto Ricans and Dominicans this region declined without any counterveiling increase in “Southern Bantu”. Most likely because their “Cameroon, Congo etc.” scores were overstated after the 2018 & 2019 updates. Which actually still seems to be the case somewhat also for Brazilians and Haitians.

 

How to interpret “Southern Bantu”

*** (click to enlarge)

These are the updated results for my DNA cousin whose father is from Mozambique. We share a mutual Cape Verdean ancestor through his mother who is from São Tomé & Principe. His Mozambican side is convincingly described by “Southern Bantu”. His minor Cape Verdean connection is also still clearly distinguishable by way of  “Senegal” and most likely also the “Portugal” score. For his previous results in 2019 see this screenshot.

***

The above results belong to my DNA cousin whose father is from Mozambique, while his mother is from São Tomé & Principe. It provides an excellent illustration of how the “Southern Bantu” region can also be useful for people with a mixed background. However generally speaking such a high degree of “Southern Bantu” will be extremely rare for Atlantic Afro-descendants (but not so for Afro-descendants in the Indian Ocean area!). Simply because that type of lineage is either absent or heavily diluted. The “Southern Bantu” scores were uncommon and minimal in my survey. Only showing up for 7 persons out of 50 in total. As expected this new region is peaking among Mexicans (4.2%) and Brazilians (2.8%). Because previously the “Southeastern Bantu” region was also peaking for them (see this chart). Due to my scaling correction the Mexican group average might be a bit distorted though. The highest absolute “Southern Bantu” score in my survey being a mere 2%, which was only reported for Brazilians (2x). Otherwise among my 50 survey participants only scores of 1% “Southern Bantu” were obtained for 1 Mexican, 1 Brazilian,  2 African Americans and 1 Haitian.

Although this new “Southern Bantu” region has made a minimal impact it can still provide informational value. Even when not reported at all! Because then it might clarify how your previous “Southeastern Bantu” score from the 2013-2018 version was really indicating Southwestern Bantu lineage from perhaps Angola or even still DRC Congo. In line with historical plausibility for the Trans-Atlantic Diaspora. For example my African American survey participant AA08 used to have a very striking “Southeastern Bantu” score of 24% in the 2013-2018 version. However after this update she did not get any “Southern Bantu” at all. Instead her renamed “Cameroon, Congo & Western Bantu” score is now 26%. See also:

Whenever trace amounts of “Southern Bantu” are being reported the odds will be high I suppose it will indeed be indicative of Bantu lineage from beyond Central Africa. Although actually already among Zambians this region is showing up in substantial amounts (>30%, see table 1.1). It will depend on your specific Diasporan background which ancestral scenario will be more plausible. However generally speaking it might be either Mozambican or Malagasy lineage. To be corroborated whenever possible of course by additional clues such as corresponding DNA matches as well as Southeast Asian admixture or haplogroups in case of a possible Madagascar connection.

It seems that Ancestry has most likely been adding South African samples instead of any Angolan, Mozambican or Malagasy samples into their Reference Panel. Which is too bad because such samples would be historically most relevant for the greater part of Ancestry’s Afro-descended customers. But either way this new “Southern Bantu” region does enable a very useful distinction to be made between western and southeastern Bantu origins! For a second opinion it is recommended to also test with 23andme. Their corresponding “Southern East Africa” region is actually more so based on samples from Swahili speaking countries (Kenya, Tanzania etc.). But still already greatly indicative of any possible Southeast African lineage. And probably with greater accuracy because from my experience 23andme’s algorithm is better equipped to deal with trace amounts of admixture than Ancestry’s homogenizing algorithm. See also:

 

East African trace amounts

As always the country name labeling is not to be taken too literally. Notice that according to Ancestry’s own information so-called “Ethiopia & Eritrea” is also prevalent in neighbouring Sudan. Based on my survey findings for Fula and Hausa-Fulani persons I greatly suspect so-called “Ethiopia & Eritrea” in fact is also to be found further west, all along the Sahel. Albeit probaby only in small amounts. Indicative perhaps of very ancient back & forth migrations between West and East Africa along this corridor.

***

The 3 East African regions “Ethiopia & Eritrea”, “Somalia” and “Eastern Bantu” were already introduced during previous updates in 2018 & 2019. In my 2019 survey the first 2 strictly Northeast African regions did not show up at all for any of my survey participants. Not even in trace amounts. This only happened for “Eastern Bantu”, but then still only at most 1%. Such an outcome is of course in accordance with historical plausibility for Atlantic Afro-descendants. However this time the frequency of minuscule East African trace amounts has increased somewhat, incl. also scores for “Ethiopia & Eritrea”. Even when last year this was not showing up for them. This probably has to do with the addition of new Ethiopian and Somali samples by Ancestry into their Reference Panel (see table 1.2). Or perhaps also because of Ancestry’s somewhat changed algorithm.

Understandably this might be intriguing for people who receive such scores. However to keep things in perspective the group averages for these regions are still very minimal (see table 2.1). The highest individual score observed during this survey was 3% “Ethiopia & Eritrea” for a Mexican. But otherwise it was usually only scores of 1% being reported. Among 50 people this happened 5 times for “Eastern Bantu” and 7 times for “Ethiopia & Eritrea”. Of course one always has to be very careful with trace amounts of unexpected admixture. Afterall admixture analysis is not 100% accurate and especially with smaller amounts it might produce so-called “noise” results. Misleading estimates which fall within the inherent margins of error. It might very well be that these trace amounts will disappear again with the next update 😉

Given that I have seen minor “Ethiopia & Eritrea” scores being reported also for Fula and Hausa-Fulani people (see for example this screenshot) I am inclined to think this coutcome is rather a proxy of Sahelian ancestry for Atlantic Afro-descendants. Although of course other explanations might be valid too in individual cases. The “Eastern Bantu” scores might actually be more so indicative of Southeast African ancestry, similar to “Southern Bantu”. To get a second opinion it might be worthwhile to also test with 23andme. As from my experience they have a better trackrecord when dealing with trace amounts of admixture. Plus their Northeast African categories are better defined than on Ancestry. Also incl. a separate “Sudanese” category which is right now missing on Ancestry. Because of missing reference samples from Nilotic and Chadic groups (Sudan & Chad). This “Sudanese” category on 23andme is actually also most likely a proxy of DNA markers inherited from Sahelian ancestors (when reported for Atlantic Afro-descendants).

As I always mention on my blog I tend to be extra careful when dealing with DNA results suggesting Northeast African ancestry for Atlantic Afro-descendants.3 Simply because it is both historically unsupported (on a group level) and also the actual science behind DNA testing and especially ethnicity estimates has proven to be far from exact! Of course this is not to deny the potential informational value to be gained from smaller regional scores. But close scrutiny and independent verification are always called for. Each case to be judged on it own merits.  Unless you have additional clues and corroborating evidence it will usually not be worth your time to investigate any further. As I once said it is best to keep your eye on the main prize and don’t get distracted by the small change, no matter how shiny they appear to be! 

 

3) Is Ancestry getting sloppy?

*** (click to enlarge)

This map is showing where the so-called “Khoisan, Aka & Mbuti Peoples” region is to be found. Only South Sudan and Tanzania are being highlighted. Even when this region should be most prevalent in South Africa and Central Africa. Afterall the Khoisan people and the Pygmy peoples  (Aka & Mbuti) are located in South Africa and Central Africa! Needlessly misleading and disorientating for anyone who’s not aware of the underlying context. With a minimal effort at some decent quality control such a blooper could easily have been avoided…

____________________

“To be sure there are still plenty of helpful sections/pages offering guidance and context […]. But compared with the 2013-2018 version it is clearly getting worse. […] Also the regional descriptions have now become very minimal and bland. Even containing obvious and sloppy mistakes at times” (Fonte Felipe, 2019)

____________________

The regional maps integrated within your Ethnicity Estimates on Ancestry have always been a useful tool to clarify that Ancestry’s regional labeling is not to be taken too literally. Some people might think these maps are custom-made for their own personal results. But that’s not true. These maps will be displayed in exactly the same way for everyone. Also important to keep in mind is that these regional maps are only indications and approximations. Based on inherently limited information available to Ancestry. Unfortunately it seems that which each next update Ancestry has become more sloppy in providing this type of essential contextual information.

Regrettably the quotation above from 2019 still applies. Perhaps even more so. The map shown above for the renamed “Khoisan, Aka & Mbuti Peoples” region is glaringly wrong and misleading. And what makes it even worse is that this was already the case in 2019 (see this map). So no lessons learnt apparently… Also the regional maps for “Nigeria” (featuring Central African Republic!) as well as “Mali” (highlighting Chad!) are needlessly misleading and disorientating for people who are not ware of the underlying context. And this is just mentioning a few of the mistakes I have seen in Ancestry’s new regional descriptions.4

I have been noticing this sloppiness trend ever since the first update in 2018. Also the general level of transparency on Ancestry seems to be decreasing with each update. A sad affair because Ancestry’s original 2013-2018 version really stood out to me because of its genuine attempt to inform people about all aspects of their Ethnicity Estimates. Unlike other DNA testing companies who often appear to be rather careless in this regard. I am not quite sure what the underlying reason could be. Possibly the original team behind AncestryDNA’s Ethnicity Estimates has been replaced by now. Or also higher management may have set new priorities… I can imagine this update may very well have been a rush job.5 Either way Ancestry is really doing their customers a great disservice in this regard! A minimal effort at some decent quality control could already have prevented most of the mistakes mentioned above.

 

Take it to the next level!

What is adding to this failure to provide useful information & context is Ancestry’s slow response to their customers’ feedback. In 2018 I blogged about several suggestions for improvement of Ancestry’s African Breakdown. Many of these ideas will provide quick-win solutions and will be relatively easy to implement. Just to repeat a few relevant ones:

____________________

Bring back the continental breakdown display (subtotals specified for each continent)

Avoid misleading labeling of ancestral regions. Providing a false sense of accuracy.” 

Knowledgeable scholars of African & Afro-Diasporan history should be involved in a re-writing of the regional descriptions

____________________

To be fair Ancestry has actually followed up on some of my suggestions as well in the meanwhile. Within Ancestry’s regional maps you can now find a feature that is reminiscent of the former “genetic diversity” tabs in the 2013-2018 version. Providing detailed statistical information about the predictive accuracy of each single region. Simply by clicking on “Learn more about this map and ethnicity“. Regrettably Ancestry has been rather sloppy with processing their data. The ethnicity averages estimated by Ancestry for “the typical native” often seem incomplete, contradictory or even downright wrong when contrasted with my own surveyfindings in table 1.1. See for example map 1.2 for “Ivory Coast & Ghana”, discussed above. Then again the “also found in” descriptions do tend to be more accurate and complete. You can access this info by first clicking on your regional score and then clicking on “Read Full History“. 

The addition of the new “Southern Bantu” region and the steady overall expansion of African samples is also following up on my 2018 suggestions. However many historically relevant populations are still missing in Ancestry’s Reference Panel. Historically relevant to African Americans and other Atlantic Afro-descendants that is! With a primary focus on finer resolution for West & Central Africa. Where almost all their African roots are to be traced, going by historical plausibility and cultural retention. Instead of Southern & East Africa, even when this is of course also welcome in itself for the sake of completeness. As well as to service the (much smaller) group of Ancestry’s Afro-descended customers who happen to hail from these parts of Africa. Naturally all of this is a work in progress. And despite some mishaps in this current update I do feel that Ancestry is slowly moving in the right direction.6 But really I believe that it should be Ancestry’s ambition to take its African Breakdown to the next level! And move beyond what was already achieved during the 2013-2018 version. Keeping in mind their own selfproclaimed goal to enable everyone’s story to be found!

 

When will we have Genetic Communities for West/Central Africa?

Map 3.1 (click to enlarge)

Ancestry has more than 1100 regions on offer, incl. also potentially very useful genetic communities. European regions form the overwhelming majority of these regions: 72% (838/1161)! A great contrast with the rest  of the world. The number of strictly African regions is only 12 . To be augmented with 4 genetic communities located within Africa. Combined this implies that African regions represent a share of less than 1% (16/1161). Ancestry mentions a number of 110 African regions but actually 96 of these regions are genetic communities meant for African Americans and African Caribbeans. Very useful for pinpointing relatively recent origins within the Americas. But not helpful for tracing back to Africa!

***

I have already mentioned how currently Ancestry’s African Breakdown seems to be merely a sideshow. We can measure this for example by the very low share of African samples within Ancestry’s Reference Panel: only 6% (2856/44703). But also the number of regions relevant for Tracing African Roots for African Americans and other Atlantic Afro-descendants continues to be very limited. In fact because “Southern Bantu” is so minimal for the Atlantic Afro-Diaspora arguably it is still only 6 African regions which really matter, for them. At least when going by double digit scores (“Senegal”, “Mali”, “Ivory Coast & Ghana”, “Benin & Togo”, “Nigeria” and “Cameroon, Congo & Western Bantu” ). “Northern Africa” also being significant for Cape Verdeans and Latin Americans, but in most cases likely to be related to their Iberian ancestry and not their otherwise African roots (except for Fula lineage). See also table 2.1.

If we have a look into the complete overview of regions currently on offer by Ancestry it certainly looks very impressive. Among its 1100+ regions (incl. genetic communities/migrations) are also mentioned 110 African regions (see map above). Which would mean a share of about 10%. However if you look closely into these 110 African regions it is actually mostly genetic communities set up for African Americans and African Caribbeans but not within Africa! Only 12 main regions are currently being located in Africa. Which forms a great contrast with the much greater specificity being provided for Europe, Asia and the Americas. The effective share of regions located in Africa would then only be around 1%! This glaring regional imbalance is naturally firstmost a reflection of Ancestry’s customer database and the difficulty of obtaining African samples. However I do believe that with some creativity and most importantly some true committment Ancestry should be able to provide greater African resolution than right now.

Including also African Genetic Communities!7 This tool based on matching strength might not be 100% accurate. But it has already proven to be greatly useful for zooming into the known micro-level origins of especially Europeans and Americans. It therefore potentially holds immense importance for people who are not aware of their specific origins beforehand! Actually right now there already seem to be 4 genetic communities located within Africa. But most of them are not picking up on strictly African DNA:

  • Morocco & Algeria“: a very useful specification of North African ancestry one whishes would also be applied for other parts of Africa!
  • Asians in South Africa“: will often also include Malagasy and Indian Ocean islanders aside from South Africans.
  • South Africa, European Settlers“: also includes South African Coloureds, besides Afrikaaners.
  • Portuguese Islanders in the Eastern U.S.”: despite the labeling actually referring to Cape Verdeans! Cape Verde obviously being an independent state in West Africa! A gross oversight which needs to be corrected by Ancestry: see also this comment of mine on Ancestry’s own blog.

As I have said several times now I really think Ancestry should finally begin expanding their Genetic Communities tool into West & Central Africa. This is needed now more than ever. Given that Ancestry’s deletion of smaller DNA matches (6-8 cM) has been harmful especially for Afro-descendants! Especially Nigeria and Ghana being promising first candidates, based on greater availability of customer samples. But when enabling DNA matching with samples from Ancestry’s Reference Panel possibly also additional genetic communities could be realized for Upper Guinea, Sierra Leone/Liberia and Cameroon/Congo? Such an expansion would be truly beneficial for Ancestry’s customers who aim to Trace African Roots. Ancestry would then truly be stepping up their game. In line with their self proclaimed goal to make Ancestry’s experience truly inclusive for everyone! 

 

4) Screenshots of African updated results

I will only be posting a limited selection of screenshots. But this array does cover all main parts of Africa. Obviously there might be greater individual variation within each country or even within a particular ethnic group. Still this should already be quite illustrative of the main patterns I have described above. As far as I was able to verify all of the following screenshots below are from persons with four grandparents from said nationality/ethnicity, unless specified otherwise. But naturally I did not have absolute certainty in all cases. Practically all results have been shared with me by the DNA testers themselves.  I like to thank all my survey participants for having tested on AncestryDNA and sharing their results online so that it may benefit other people as well!

 

GAMBIA

***

 

***

SENEGAL (Wolof, Fula, Serer)

***

***

SENEGAL (Wolof)

***

***

SENEGAL (Fula)

***

***

SENEGAL (Fula)

***

***

SENEGAL & GAMBIA (Fula)

***

***

FULA? (?)

***

***

SENEGAL & GUINEA (Fula)

***

***

 GUINEA (Fula)

***

***

GUINÉ BISSAU  (Fulacunda)

***

***

MALI (Bambara & Fula)

***

***

MALI (Soninke)

***

***

MALI (Bambara & 1/4 Moroccan)

***

***

MALI & MOROCCO 

***

***

SIERRA LEONE (Mende)

***

***

SIERRA LEONE (Mende)

***

***

LIBERIA (Northern?)

***

***

LIBERIA (Vai, Bassa, Gbandi & Americo-Liberian)

***

***

LIBERIA (Bassa & Americo-Liberian)

***

***

LIBERIA (Grebo & Vai)

***

***

LIBERIA (Kru/Kpelle)

***

***

LIBERIA (Kru/north)

***

***

LIBERIA (Kpelle, Grebo, Gola)

***

***

LIBERIA (?)

***

***

LIBERIA (?)

***

***

LIBERIA (?)

***

***

LIBERIA (Grebo & Lofa county)

***

***

IVORY COAST (7/8 Akan & 1/8 Krio (Sierra Leone))

***

***

IVORY COAST (3/4 Akan & 1/4 Krio (Sierra Leone))

***

***

IVORY COAST (North, part Fula)

***

***

GHANA (Akan:Ashanti & Sefwe)

***

***

GHANA (Akan)

***

 

***

GHANA (Akan)

***

***

GHANA (Akan & Ga)

***

***

GHANA (?)

***

***

GHANA (Akuapem & Ewe)

***

***

GHANA (Northern)

***

***

GHANA (Ewe)

***

***

GHANA (Ewe)

***

***

GHANA (Ewe?)

***

***

NIGERIA (Igbo)

***


***

NIGERIA (Igbo)

***

***

NIGERIA (Igbo)

***

***

NIGERIA (Igbo)

***

 

***

NIGERIA (Yoruba)

***

 

***

NIGERIA (Yoruba)

***

 

***

NIGERIA (Yoruba)

***

***

NIGERIA (Yoruba)

***

***

NIGERIA (Yoruba)

***


***

NIGERIA (Yoruba)

***

***

NIGERIA (Yoruba)

***

***

NIGERIA (Edo?)

***

***

NIGERIA (Igbo, Edo & Yoruba)

***

***

NIGERIA (Bini, Itsekiri, Urhobo & Isoko)

***

***

NIGERIA (Hausa-Fulani)

***

***

NIGERIA (Hausa-Fulani)

***

***

CAMEROON (Oroko/Balundu)

***

***

CAMEROON (Bulu)

***

***

CAMEROON (Bulu & Bandem)

***


***

EQUATORIAL GUINEA (Fang)

***

***

GABON (Fang & Bateke)

***


***

CONGO BRAZZAVILLE (Bakongo)

***

***

CONGO (DRC) (?)

***

***

CONGO (DRC) (?)

***

***

ANGOLA (Bakongo)

***

***

ANGOLA (?)

***


***

ANGOLA (?)

***

***

ZAMBIA (?)

***

***

ZAMBIA (?)

***

***

ZAMBIA (?)

***

***

ZIMBABWE (Shona?)

***


***

ZIMBABWE (Shona?)

***

***

ZIMBABWE (Shona?)

***

***

ZIMBABWE (Shona?)

***

***

MALAWI 

***

***

MALAWI 

***

***

MOZAMBIQUE & SÃO TOMÉ & PRINCIPE (+Cape Verde)

***

***

SOUTH AFRICA (Swazi?)

***

***

SOUTH AFRICA (Swazi?)

***

***

SOUTH AFRICA (Xhosa)

***

***

SOUTH AFRICA (Coloured)

***

***

SOUTH AFRICA (Coloured)

***

 

***

SOUTH AFRICA (Coloured)

***

***

SOUTH AFRICA (Coloured)

***

***

SOUTH AFRICA (Coloured)

***

 

***

SOUTH AFRICA (Coloured)

***

***

MADAGASCAR (Northeast)

***

***

MADAGASCAR (Southwest)

***

***

MADAGASCAR (Merina?)

***

 

***

COMOROS 

***

***

RUANDA (Tutsi)

***

***

BURUNDI (Tutsi)

***

***

TANZANIA (western (Ha, Hangaza))

***

***

TANZANIA (?)

***

***

TANZANIA (Jita & Kuria)

***

***

TANZANIA (Kuria)

***


***

UGANDA (northern)

***

***

UGANDA (?)

***

***

UGANDA (Nilo-Saharan: Aringa & Kakwa)

***

***

KENYA (Kikuyu)

***

***

KENYA (Kikuyu)

***

***

KENYA (Kikuyu)

***

***

KENYA (1/2 Taita, 1/4 Kikiyu, 1/4 Kisii)

***

***

KENYA (Taita)

***

***

KENYA (?)

***

***

KENYA (?)

***

***

KENYA (?)

***

***

KENYA (Swahili)

***

***

SUDAN

***

***

***

ETHIOPIAN  & AFRICAN AMERICAN

***

***

SOMALIA

***

***

ALGERIA

***

***

ALGERIA

***

***

MOROCCO 

 

***

***

 

 

___________________________________________________________________________

Notes

1) This current update on Ancestry really seems to be all about Scotland! Dividing British ancestry into 4 separate regions for the first time. I am a bit ambivalent about this attempt to distinguish genetically very closely related populations. As I really prefer for DNA testing companies to go by the motto: don’t be more specific than your underlying data allows for. Generally speaking when it comes to regional admixture I would say just provide more generic but still informative (macro)-regions (in this case just simply “British” or even “Northwest European”) and leave it up to the DNA matches/genetic communities for further specification. In order to avoid confusion and people being mislead about their genealogical lineage, which will often not fully allign with their genetic inheritance. Regrettably this also happened with LivingDNA’s overdetailed African breakdown

On the other hand I am also curious how far Ancestry can eventually get with this project. Obviously this is only a first step because they seem to be quite committed to take this further to the next level. Unlike Ancestry’s African breakdown… This is what Ancestry had to say about it:

People from England or the northern counties of Ireland may be getting more Scotland than they might expect given their family history—and sometimes vice versa. This is a natural consequence of trying to distinguish several closely related peoples apart at just the DNA level. This update is the first time we have been able to identify these four regions separately, so don’t be surprised if we are able to make more refinements to these regions in the future.

See also:

2) According to some people only continental admixture is to be taken seriously in DNA testing. Sub-continental admixture, a.k.a. ethnicity estimates, a.k.a. regional admixture only being fit for entertainment purposes. 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 Africans and Afro-descendants in the past. Applying an additional macro-regional framework for extra insight (Upper GuineaLower GuineaCentral/Southeast Africa).

From these ongoing research efforts I have learnt that regional admixture DOES matter and Ancestry’s Ethnicity Estimates are of course NOT randomly determined. Ancestry’s predictions may not be 100% accurate but still in most cases they are reasonably well-aligned with the known or historical plausible backgrounds of my African & Afro-Diasporan survey participants. To be combined with any other ancestral clues you may have, especially DNA matches. Which is why it is such a shame that Ancestry has deleted so many potentially very informative smaller African matches earlier this month. See also:

3)  Generally speaking I believe that any DNA test result indicating East African ancestry should always be critically scrutinized. I tend to be very sceptical about the degree or frequency of “East African” DNA results reported for  Atlantic Afro-descendants. Because this does not fit well with historical plausibility nor cultural retention. Of course one must remain open minded and within itself this topic of any possible East African connection for the Atlantic Afro-Diaspora is quite intriguing. Hypothetically speaking in rare and individual cases I suppose it might be possible to have one single East African ancestor. However I am assuming it would be greatly diluted then in most cases.

But certainly such atypical cases do not justify the currently much inflated level of so-called East African DNA results being reported by various DNA testing companies, across the board. Either haplogroups, admixture scores or DNA matches suggesting such connections. The tricky thing is that DNA testing of course is no exact science. Due to faulty algorithms, lack of reference populations etc.. On all fronts, also including DNA matches! Which can very well be IBS or false positives. In particular smaller matches (see this chart). Especially the implied time framing is often unclear. I highly suspect DNA test results suggesting East African ancestry are often merely a consequence of VERY ancient population migrations across the continent (going back millennia instead of centuries). Something which would also be detected among actual West or Central Africans.  Irrelevant therefore from a genealogical perspective (last 500 years or so).

I think it helps to be as specific as possible when outlining such results. Making a distinction between various parts within East Africa. Many people may already be aware of the legitimate Southeast African connection (mostly Madagascar & Mozambique). Strictly speaking East Africa for me would be the Swahili speaking countries: Tanzania, Kenya, Uganda and possibly also eastern DRC, Ruanda/Burundi. While Northeast Africa would be Somalia, Ethiopia, Eritrea, Djibouti and Sudan.

Now when it comes to historically documented ancestral connections between these latter areas and the Atlantic Afro-Diaspora it is far less apparent than for Southeast Africa. Although it is known that some atypical Trans-Atlantic slave voyages did depart from the Swahili Coast: Mombasa (which used to be ruled by the Portuguese!) and also Zanzibar. But going by actual numbers as well proportionally this flow of people was quite minuscule. Based on the records available in the standard reference database of the Slave Voyages website this East African share in Trans-Atlantic Slave Trade might be less than 0.1% (=6324/6709327). Comparing the total number (6,324) of disembarked captives for Kilwa, Zanzibar and Mombasa, all Swahili ports to the north of Mozambique, with the total number of disembarked captives from Southeast Africa (308,775, overwhelmingly from Mozambican ports with main destination being Brazil) and all of Africa (6,709,327). Naturally all of this is according to what has been documented and excluding voyages with unknown itineraries, obviously the estimates will be higher.

***(click to enlarge)

TAST (Swahili ports, numbers, destinations specified)

 

Source: Trans-Atlantic Slave Trade Database 2016 (http://www.slavevoyages.org)

***

4) The  “Khoisan, Aka & Mbuti people” region was previously named “Southern & Eastern Africa Hunter-Gatherers” after the 2019 update. And originally (2013-2018 version) it was named “Africa South-Central Hunter-Gatherers”. It seems Ancestry cannot make up its mind what to do about this region… However the underlying samples (around 34-38) appear to have been the very same ones, right from the start! Derived from genetically very distinctive but also extremely marginalized hunter-gathering populations. Going by Ancestry’s current labeling and also their regional description these would be the Khoisan from South Africa and the Aka & Mbuti people from Central Africa (combined also often referred to as Pygmies).

In previous blog posts I have speculated that also Sandawe samples from Tanzania may have been used by Ancestry. Another much studied yet very marginalized hunter-gathering population. This might explain the eccentric highlighting of only Tanzania and South Sudan in the regional map for “Khoisan, Aka & Mbuti people”. And previously also for the “Southern & Eastern Africa Hunter-Gatherers” region (see this map). In accordance with these regional maps actual people from Northeast Africa (incl. Tanzania, South Sudan but also Kenya and Uganda!) do indeed show genetic affinity with these hunter-gatherer DNA samples. However this outcome is principally caused by lack of better alternative. Afterall Ancestry’s Reference Panel is still lacking Nilotic samples which would provide a much better fit!

As I have said before in 2018:  “I do not believe there is much added value in reporting these genetic affinities with marginalized hunter-gathering populations (no matter how distinctive and fascinating in itself) as they usually go back thousands of years. It only leads to confusion while also the labeling may be perceived as akward by some people. I would only keep in the Khoi-San samples as they are meaningful and relevant to describe the recent origins of especially South Africans.” See also:

See maps below for examples how Ancestry continues to be sloppy for other regions as well. Not meant to be exhaustive. But another illustration would be the way Ancestry has simply maintained their regional descriptions from the 2013-2018 version. Even when in the meanwhile things may have changed. This goes especially for “Southern Bantu” which is exactly the same now as for “Cameroon, Congo and Western Bantu” (except for the map and the “also found in” info). But also “Benin & Togo” which still mentions the obsolete genetic diversity statistics from the 2013-2018 version

“Mali” region is mysteriously highlighting Chad. Although this country is not mentioned at all in the “also found in” details! (although the Central African Republic is…) Because of a current lack of Chadic samples it would make sense actually that people from Chad are described by neighbouring regions, especially across the Sahel. As a second-best solution so to speak. In fact I have observed rather elevated levels of “Mali” among Hausa-Fulani from northern Nigeria in my survey (see table 1.1). But for some strange reason Nigeria as well as Niger are being left out of this map. Furthermore I have a strong hunch that even if actual Chadians might indeed receive “Mali” scores it will not be to the exaggarated degree (>75%) as suggested by this map. And more importantly for Trans-Atlantic Afro-descendants such seemingly random information will only be misleading and not contributing to a better understanding of their personal “Mali” results.

***

The display below for “Nigeria” is in itself very useful! Because it also contains average ethnicity scores for “locals”. Such benchmark information being crucial for correct interpretation. However because of Ancestry’s sloppiness it is also incomplete. Notice how Benin and Cameroon are not mentioned in this map. Even when from my survey findings I am certain that some locals from these countries will also score considerable scores of “Nigeria”. Being aware of such genetic overlap (especially with Benin) is much more relevant than being aware of a seemingly mysterious highlighting of the Central African Republic (CAR)! To be sure people from the CAR might very well score “Nigeria” amounts in theory. Because afterall Ancestry’s Reference Panel currently does not contain samples from the CAR itself. Which is why their DNA by necessity will be described by the next best thing. Which could be “Nigeria” indeed. But again I am doubtful that actual people from the CAR will indeed receive “Nigeria” scores to the elevated degree (25-50%) as suggested by this map. I imagine “Cameroon, Congo and Western Bantu” will be more significant for most. While also “Ethiopia & Eritrea” may be seen for some of them. Naturally all depending on actual ethnic background.

Again for the sake of correct interpretation of one’s “Nigeria” scores it would have been much more useful if Ancestry were to elaborate on the differences within Nigeria. My surveyfindings for Hausa-Fulani suggesting that actually “Nigeria” amounts are significantly lower for northern Nigerians than for southern Nigerians. While Yoruba Nigerians tend to receive considerable “Benin/Togo” scores. Even when their “Nigeria” scores will still be predominant. Igbo Nigerians receiving the highest “Nigeria” scores (90%+), usually with a trace amount of “Cameroon, Congo & Western Bantu”. A similar but somewhat different constellation than obtained during the 2013-2018 version. But basically a consequence of the same underlying genetic differences for Nigerians. See also:

***

5) The timing of this Ethnicity update as well as Ancestry’s recent deletion of small DNA matches does seem suspicious when being aware that Ancestry was taken over in August by an investment firm called Blackstone. See also:

6)  In my blogpost about Ancestry’s update in 2019 I also included a poll. Despite misgivings most people did indeed think this update was an improvement over the update in 2018. As did I myself. But it is becoming increasingly frustrating when you are aware that Ancestry is able to do so much more! As proven by the detailed regional resolution Ancestry is offering for especially their European and American customers.

7)  For more details about Ancestry’s Genetic Community tool follow these links:

Below my original suggestion from 2018:

____________________

“Create new African “migrations”, a.k.a. genetic communities.  […] I understand that this potentially very insightful “migration” feature is a work under progress. Naturally a certain minimum number of DNA tested Africans with a common background will be required to create new genetic communities. Then again from my ongoing survey of African AncestryDNA results I have learnt that there could very well be hundreds [thousands by now] of Africans (incl. 1st and 2nd generation migrants) already within Ancestry’s customer database. Especially for Nigerians and Ghanaians I would imagine something could already be set up. Even more so when appropriate academic samples can be added. Given the pressing need for more specification of African lineage I would argue for a loosening of certain thresholds and/or requirements provided that a minimum level of robustness for this “migration” tool can still be maintained.”

____________________

Again availability of African samples is of course an essential precondition but should not be used as a convenient excuse. As I have argued also in 2018: 

____________________

creative recruiting of African samples: through targeted marketing for example offering free kits among African migrant associations in either the USA or Europe. Or else also by crowd-sourcing: third parties/ individuals travelling to Africa in order to test local Africans. Afterall where there is a will there is a way 😉  When carried out effectively the costs involved may be quite minimal while the added value could be enormous!

____________________

An expansion of Ancestry’s Genetic Community tool into West/Central Africa is needed now more than ever. Given that Ancestry’s deletion of smaller DNA matches (6-8 cM) has been harmful especially for Afro-descendants! See also:

66 thoughts on “Ancestry’s new African Breakdown: merely cosmetic changes?

    • Yes “Ivory Coast & Ghana” is still greatly underestimating true Ghanaian lineage. You’re better off going by DNA matches indeed or also 23andme’s “Ghanaian, Liberian and Sierra Leonean” category.

      Because of the introduction of “Scotland” in this update it seems to me that the European breakdown of many African Americans and other Afro-descendants with only minor European admixture has become somewhat more random again. A bit like the situation in the original 2013-2018 version when European trace regions for AA’s were often all over the place. But usually just a misreading of actual British DNA. See also this old blogpost of mine:

      https://tracingafricanroots.wordpress.com/2018/06/16/update-afro-diaporan-ancestrydna-survey-part-2/

      Basically with such surprising trace amounts (which usually will include a 0% score in the range!) it is indeed wise to check if it’s consistent with other DNA testing companies. Also it will be useful to know if your personal score is distinctive / above average when compared with other African Americans. Afterall it could also merely be some fluke outcome which is commonly reported for everybody. In previous versions of 23andme as well as Ancestry I recall some minor South European scores were often reported for AA’s. But this in fact also happened for native British people. Simply because their genetics are too complex to be described as 100% “British”.

      A more throrough way to investigate the possibility of having actual Portuguese ancestry is by looking into your DNA matches. My scanning & filtering method also works for zooming into 100% European matches. Singling out Portuguese matches is in fact relatively easy because Ancestry’s “Portugal” region is quite predictive. And practically all native Portuguese persons will have this region show up as their main region. Usually combined with minor scores for “Spain” as well as “France” due to genetic overlap.

      Because there is a relatively great number of Portuguese Ancestry testers the odds of matching some of them are going to be pretty high. As long as your Portuguese ancestry is indeed legit. I have myself confirmed this by scanning for European matches among 50 Cape Verdeans. Also for Jamaicans and Guyanese with some known minor Portuguese ancestry it is very easy to find Portuguese matches. I have in fact also scanned for European matches among African Americans but I don’t recall ever coming across Portuguese matches for them yet.

      Liked by 1 person

      • Yesterday, I found an african dna match by accident. She’s half african and half french and english. Her mother’s side is the african side and they hail from Gabon, but she also has roots in Cabinda, Angola. She said she isn’t sure if its bakongo, neverthless we share 9cm across 1 segment.
        This partially explains my 13.4% Angolan & Congolese score on 23andme. Ancestry just blows it out of proportion with 32% Cameroon, Congo and Western Bantu. This is my first match from Central Africa so it’s pretty exciting.

        Liked by 1 person

        • Congratz! Ancestry DNA testers from Central Africa are still relatively rare so finding such a match is indeed awesome. I have been surveying African DNA matches being reported for people all across the Afro-Diaspora. And often they would not get any Central African DNA matches even when their “Cameroon, Congo etc” scores were clearly predominant. This also goes for Brazilians and Haitians. However small matches (<8 cM) from likewise Bantu speaking countries in Southern and East Africa did regularly show up. I suspect these are for the most part population matches though. Not genealogically meaningful (~ 500 years). But rather picking up on commonly shared DNA segments which may have been circulating within the greater genepool of Bantu speaking people for thousands of years already perhaps.

          Like

          • In 2019, I found my only African cousin on 23andme. He is from Ivory Coast, and he is Mahou(Malinke) on his dad’s side and tagbanna(which is a sub-group of Senoufo on his mother’s side. 23andme lists him as my fourth cousin, and we share 17cm. On Ancestry my highest cm match is an Igbo woman and we share 12cm.
            I thought this was weird since I have 12 confirmed IBD Igbo matches on ancestry but they are all <10cm. Still, I know I have strong igbo roots, since I do have the virginia community and I’ve always maintained a Nigeria score of >40%.

            Is it odd that dna I share with a Malinke man is much more than my igbo matches?

            Like

            • You will connect to your DNA matches by way of many different shared ancestors from different parts of Africa. Some of these shared ancestors may have been from relatively early time periods. While others may be traced to relatively recent timeperiods. This will be expressed in the amount of shared DNA. This Ivorian match might possibly just indicate an ancestor who is from a closer branch than the ancestry implied by your Igbo matches. These Igbo matches may actually also imply multiple branches. Which add up to a greater overall degree of Igbo ancestry, even if perhaps from earlier timeperiods. Naturally other ancestral scenario’s might still be valid too:
              See also:

              https://tracingafricanroots.wordpress.com/2015/01/20/fictional-family-tree-incl-african-born-ancestors/

              Another thing to keep in mind is that 23andme currently limits the number of matches you can receive. This means all your DNA matches on 23andme are going to be quite close matches by default. And you will be missing out on smaller but still informative matches which are still being reported on Ancestry (eventhough they did recently raise their threshold as well to 8cM).

              On the other hand 23andme does have more tools available to do follow-up research on your DNA matches! Incl. their chromosome browser which lets you know the exact location of the shared DNA segment. As well as the overview of shared matches under the button “Find Relatives in common”. You should check it and see if any of these shared matches triangulate (shared DNA = Yes). If so, by comparing family trees with the other matches who also have the Ivorian match in common you might possibly be able to zoom into an earliest family line associated with this Ivorian match.

              Like

  1. FamilyTree updated their regions, and they have a very interesting breakdown for Africa. I have uploaded my data there since 2014 and my results look more plausible. I have nearly the same % of Nigeria as I had in the 1st Ancestry update and 23andMe, and my remaining % is a combination of Windward coast, Sierra Leone, and Ghana/TG. Their breakdown for WEST AFRICA is as: Ivory Coast/Liberia, Guinea/Sierra Leone, GH/TG/Benin, Nigeria, Senegal/Gambia/Guinea Bissau, and CENTRAL AFRICA as: Northern Congo Basin, Southern Congo Basin and Atlantic Equatorial Africa. Did west African you know has uploaded their raw data on their platform???

    Liked by 1 person

    • Nice, please send me a screenshot! Yeah I already saw a couple of African FTDNA results. It looks pretty good for most of them. I really like what they’ve done with the expansion of their African breakdown. However I do fear that because of FTDNA’s screwy algorithm results for people with more complex backgrounds will tend to show weird outcomes. Similar to MyHeritage I guess. I ‘ve seen this being reported by many European customers already.

      Like

  2. Ok, I think I understand now why it is less important to focus on specific tribes and more on groups of African people like mande people or Atlantic for example. I have noticed a trend with my upper and lower guinea DNA matches. Even though they all hail from different countries, all their tribes are related to the mand people. By mistake, I ended up finding another African match who is my first match from Liberia. Her father’s tribe is mano, (they are a mande speaking people of northern Liberia and southeast guinea). The mano tribe are closely knit to the neighboring tribe gio (they can be found in west and southwest ivory coast). This tribe coincides with my Ivorian match on 23andme whos father’s side is malinke/dan from west ivory coast. Even all my Igbo matches don’t necessarily mean I am of 100% Igbo lineage, since in the biafra slave trade there were ibibio/ikwerre,ijaw etc. Btw me and my liberian match share 19cm and we also have a shared match whom I share 23cm with.

    Liked by 1 person

    • Interesting! The Mahou are Malinke/Mandingo (Northern Mande) but they are also closely related to the Dan (Southern Mande), these two mixing gave birth to another population that we call Tura in Ivory Coast. The Dan, the Tura, the Mano (of LIberia and Guinea, in Guinea they are also called Mah) and the Mahou are all neighbors, so it makes sense that your Mahou match is also related to your Mano match from Liberia. What I wonder now is if all the Kpelle, Mende, and Gio/Dan are actually the same people, because I know one Dan from IVC who claimed that she could understand Mende from Sierra Leone, that it is the same language as Dan, with differences in accents and expressions. So you are right not to focus literally on specific tribes.

      Liked by 1 person

    • Yes and also samples from Chadic speakers would be beneficial. First of all of course for people who have genuine and recent ancestry from these places. That is to say actual Northeast Africans as well as people from the central/eastern Sahel.

      But it should also be helpful to avoid misleading reportings of Northeast African admixture for Atlantic Afro-descendants. In most cases for them trace amounts of unexpected East African admixture will be indicating Sahelian ancestry in fact. As I discuss under section “East African trace amounts”.

      The addition of such Northeast African samples will indeed fill in a major gap within Ancestry’s Refernce Panel. However there are plenty of other missing reference populations still as well! For West Africa I think a crucial one still missing are Gur speaking samples from Burkina Faso and northern parts of Ivory Coast, Ghana and Togo.

      As mentioned in my blog post I would personally prioritize the addition of more West African as well as Central & Southeast African samples. As these will have the greatest added value for Ancestry’s Afro-descended customers. Who are afterall overwhelmingly either African American, West Indian or Latin American and not Northeast Africans.

      However many historically relevant populations are still missing in Ancestry’s Reference Panel. Historically relevant to African Americans and other Atlantic Afro-descendants that is! With a primary focus on finer resolution for West & Central Africa. Where almost all their African roots are to be traced, going by historical plausibility and cultural retention. Instead of Southern & East Africa, even when this is of course also welcome in itself for the sake of completeness. As well as to service the (much smaller) group of Ancestry’s Afro-descended customers who happen to hail from these parts of Africa.”

      Like

  3. How close to Mali do you suspect Gur speakers might be? I’ve been wondering if maybe some of the high Mali we see in the Akan might not be either Gur or maybe Mande or even Hausa. Didn’t the Asante conquer the Dagomba?

    Like

    • It will vary, I have seen the results for a Senoufo person from Ivory Coast who scores 75% “Mali” after the update. If you look into table 1.1 you can see the results for a most likely Gur speaking person from Burkina Faso. His breakdown is almost evenly split up between 4 regions. “Mali” being more subdued with 27%. Indicating how genetically speaking many Gur speakers might be right in the middle of the current clusters available on Ancestry. And therefore for greater delineation a separate region based on Gur speaking samples would be very useful.

      Hopefully leading towards sharper definition of other regions as well. And in particular improving the prediction accuracy of “Ivory Coast/Ghana”!

      With this update the “Mali” scores for Akan people have been decreasing. It is only 7.6% right now (see table 1.1). I don’t think the occurence of these scores after the 2019/2019 updates was indicating any recent non-Akan lineage for them per se. It was first most a consequence of Ancestry messing up the “Ivory Coast/Ghana” region which used to work just fine in the 2013-2018 version!

      Within my survey are also the results of a person from northern Ghana. And for him I do think it is very likely he has some additional Hausa lineage. Because his family history he shared with me is actually in support. You can see his screenshot in the last section. He scored only 2% “Mali” however his “Nigeria” score was 26%! Keep in mind that the “Mali” scores appearing for Nigerian Hausa-Fulani are probably more so related to their Fula side than it is to their strictly Hausa side.

      Like

    • I have seen 2 additional northern Ghanaians results, one from Paga Zenga, near to the Burkinabe border, and another who I think might be from Chiana Kaleveo, not too far from Burkina Faso as well. See their % below:

      Paga Zenga
      -Benin: 45%
      -Mali: 37%
      -IVC/GH: 18%
      -Nigeria: 0%

      Chiana Kaleveo
      -Benin: 42%
      -Mali: 18%
      -IVC/GH: 36%
      -Nigeria: 4%

      I feel like GUR from Ghana and Burkina would tend to have higher Benin/TG, and although so far we only know one Senoufo results from IVC, ( that has a much higher Mali), I think the Senoufo would likely be different in general to other with less Benin/TG and Nigeria and higher Mali. The western you go the more the Mali % would likely increase.

      Concerning Ghanaians, I noticed the Ashanti among all their other Akan groups have higher frequencies to have some Mali %.

      Liked by 1 person

      • Yes i have also seen this tendency towards “Benin/Togo” for northern Ghanaians. Especially after this current update. It makes me wonder about the 179 (!) new samples Ancestry has been adding for the “Benin/Togo” region. As usual they are not saying anything about the actual ethnic origins of their samples… But perhaps this time they have also included samples from northern Benin/Togo who would not from the usual Gbe speaking groups in the south?

        Btw what kind of Ivorian samples do you think FTDNA is now using? I have been trying to find out more about their refernce populations. But I can’t find anything more than this:

        https://learn.familytreedna.com/family-finder-autosomal-testing/myorigins-family-finder-autosomal-testing/myorigins-population-clusters/

        Like

        • Yeaaa I also think that they possibly included some samples from Northern Benin and Togo in the Benin/TG for this update, but I remember also the very few GUR results I ve seen prior this update, especially from northern GH had considerable Benin/TG, and a Senoufo results you had prior the 2018 update also had higher Benin/TG than IVC/GH, which made me think if GUR ethnic groups are distantly related to GBE speaking groups.

          Concerning FTDNA, I think they are probably using Kru samples for the IVC/LIB region and Mande samples for Guinea/Sierra Leone. For the GH/Benin/Togo region though, I think they maybe added Ghanaian Akan to the GBE speaking groups. What do you think? Hopefully we will get to see more African results on FTDNA to see if their results align with this supposition.

          Liked by 1 person

          • Yes we need to see more African results. I have actually already seen a few, but from other parts of the continent. Yesterday I also got my own results back. And frankly they were a mess, haha! They gave me 17% North African LOL!!! Also my European part is designated as mostly Central European for some weird reason…

            I suspect that FTDNA’s algorithm doesn’t know how to deal properly with people of more complex or recently mixed background. A bit similar to Ancestry’s homogenizing algorithm but clearly worse in performance in many cases haha. My Cape Verdean DNA cousins on FTDNA do better though. Their African breakdown is consisting of practially only Senegal, Gambia & Guinea-Bissau. I have also seen some impressive results of Angolans and Brazilians. FTDNA has created 3 different clusters to describe strictly Central African DNA, which is the first time ever I think for a commercial DNA testing company!

            Like

            • I really wish Liberians were as accurately defined as Sierra Leonans. It looks lke “Mali” is picking up the Mande-speaking groups, and that maybe ‘Ivory Coast & Ghana’ are picking up the Kru-speakers? I’m now interested in how the language-isolate Gola-speakers – the largest Atlantic-speaking group, I think – would show up on Ancestry? Do you have any Gola samples? In any case, Atlantic-speakers seem to be a very small minority of Liberians. The Gola interest me, though, because they are some of the earliest to settle this area.

              BTW, is there any percentages about the number of people who speak Kru and Mande languages in Liberia? I’m curious to see how makes up more of the population.

              Like

              • Yes the Gola people are very interesting! I once posted the results of a person with partial Gola background. You can find it on this page linked below. If you scroll down to the Liberian section you will also find a listing of websites which mention a breakdown of languages/ethnic groups etc. Especially the Canadian (Cefan) website is very detailed. It is in French though:

                https://tracingafricanroots.wordpress.com/ancestrydna/west-african-results-part-1/

                Btw Liberians usually get very accurate results on 23andme of over 90% for “Ghanaian, Liberian and Sierra Leonean”, incl also recent ancestral locations mentioned within Liberia. Even more so after the latest update, in which 23andme seems to have chosen for a smoothing or homogenizing algorithm, similar to the one used by Ancestry. It will basically cause people to end up being described as nearly 100% of which ever ancestral component is well represented in the reference dataset of the company.

                Like

  4. Wondering how well this fits with average results for AAs now on Ancestry I know this is from 23andme

    Like

    • In an upcoming blogpost I will review that study by 23andme (“Genetic Consequences of the Transatlantic Slave Trade in the Americas”).

      Like

  5. Interesting read as usual. Of the current update, my European ancestry (largely English & German) was nearly identical as the previous update, and in keeping with my genealogical research. I’m one of the rare ones for which ‘Scotland’ didn’t overtake the rest of my British genes. They don’t seem to know what to do with Scotland. In all honesty, so much of Scotland over the past hundreds of years is, genetically, populated by Galicized Brits; I’d simply add it (and Wales) with England called it “Great Britain. The true Gaels on the west coast and Highlands might end up getting predominately “Ireland” as a region, but it’s better than much larger ‘England’ being smoothed over into ‘Scotland.’ I’d be very interested in where they got the reference populations for the British Isles from. While genetic research has found ‘tribes’ in Great Britain, they don’t really follow current political boundaries. For instance, the North and South Welsh are more similar to their eastern English neighboring regions than they are to each other. There isn’t any real “Welsh” genetic footprint. Anyway, my ‘Norway’ has stayed pretty consistent through 3 updates, which I find fascinating. I’ve always assumed it was old Viking blood via my English ancestry, but it’s stayed too consistent and beyond trace levels to be that far back I’d think. In that case, I have no idea where it comes from in my genealogy.

    As for my African ancestry, I’m not sure quite what to think. I think them figuring out the true signifance of Nigeria in African Americans’ backgrounds was a good change in the previous update. But beyond that, things get more confusing as is usual with AncestryDNA. Benin & Togo is back as a significant source of ancestry for me, but I’m still against this even being a region as it’s not genetically distinct enough. I am almost certain what 19% ‘Benin & Togo’ is picking up for me is actually the Ewe in the far southeast of Ghana, which would explain why my top African region way back was 22% Ivory Coast/Ghana. My ‘Cameroon, Congo & Western Bantu’ is pretty similar to the previous updates (10% vs 14%). My Mali is sigificantly less than the previous update (5% vs. 9%), but all-in-all across all updates it’s been between 5% and 9% which isn’t a huge range or swing. I obviously have significant – but minimal – ancestry from this region, though where in this region I really don’t know.

    Anyway, just on some general news I’ve found yet another Igbo match, so I feel pretty comfortable ancestrally relating myself to them. lol This has been consistent across all updates. It seems as if at least a whole half of my African matches are Igbo.

    As for changes in regions, I’d dismantle ‘Benin & Togo’ and give Benin to ‘Nigeria’ and Togo to ‘Ivory Coast & Ghana.’ They can rename those two regions if they’d like.The problem with the region is that the Yoruba cross/intrude singificantly into Benin, and the border divides the Ewe in Togo. I’d argue that there really isn’t a ‘Benin & Toto’ genetic community. As for ‘Mali,’ I’m not sure how I feel grouping it with the coastal countries. Yeah, there are significant populations of Mande-speaking people in both Sierra Leone and Liberia, but they leave ‘between’ or amongst the unrelated Atlantic (Sierra Leone) and Kru (Liberia) speaking peoples that are seen as indigenous to the region. I’m not sure how to do this region, but grouping Sierra Leone and Liberia with large presumably non-genetic Mende populations doesn’t strike me as the best solution. And then this pushes up against what ‘Ivory Coast & Ghana’ should be measuring. Strikes me that it should be centered on the cross-border Akan’ people, but maybe include the cross-birder Kru peoples in the west of Ivory Cross and in Liberia and cross-border Ewe in southeastern Ghana and in Togo.

    Liked by 1 person

    • While genetic research has found ‘tribes’ in Great Britain, they don’t really follow current political boundaries. For instance, the North and South Welsh are more similar to their eastern English neighboring regions than they are to each other. There isn’t any real “Welsh” genetic footprint”

      Indeed, although I am curious to know how far Ancestry can take this project really there are some inherent obstacles. And this should also be very relevant for all those of us who wish to see their African DNA results specified by ethnic group. Most times this will just be impossible! In particular when wanting to make the distinction between neighbouring ethnic groups who are bound to have a great deal of shared DNA. Even if not always acknowledged or remembered in collective memory. Because much of these genetic similarities will go back to ancient times. Although also reinforced by recent inter-ethnic unions.

      Really I think a combination of more generic but still reasonably defined regions based on admixture combined with detailed genetic communities based on DNA matches works better. Intriguingly I am regularly seeing West Indian results featuring specific Irish, Scotish and even Welsh genetic communities. But not so for African Americans.

      My Mali is sigificantly less than the previous update (5% vs. 9%), but all-in-all across all updates it’s been between 5% and 9% which isn’t a huge range or swing. I obviously have significant – but minimal – ancestry from this region, though where in this region I really don’t know.”

      I have a hunch this might be mostly correlating with your Liberian ancestry. Which is already validated also by two robust Liberian DNA matches (9 & 12 cM). Before this update Liberian DNA was predominantly described by “Mali” (75% in my survey based on 12 samples). After this update however Liberians are shifting again towards “Ivory Coast/Ghana” as was also the case in Ancestry’s original 2013-2018 version. In this survey “Mali” was at 56% on average. And “IVory Coast/Ghana” increased to a level of 38%. But it can get even higher, this goes especially for Kru Liberians, see also the screenshots in the last section. For Sierra Leoneans however (esp. Mende) it is still “Mali” which gets reported with amounts of >90%.

      Anyway, just on some general news I’ve found yet another Igbo match, so I feel pretty comfortable ancestrally relating myself to them. lol This has been consistent across all updates. It seems as if at least a whole half of my African matches are Igbo.”

      Yes this goes for many African Americans and also West Indians like Jamaicans. As I have said before this is partially also to do with the circumstance that Nigerians and especially Igbo ones are most likely overrepresented within Ancestry’s customerdatabase. Because it seems that there are simply more Nigerians testing than other West/Central African nationalities. Nigerians are afterall also by far the biggest West/Central African migrant group in the USA (see this chart).

      However such an outcome certainly also has significant ancestral implications and is greatly in line with historical plausibility! Have you started looking into which side of your family these matches are to be associated with already?

      Like

      • As much as I can. Since I have my father and maternal grandmother, I’ve been able to narrow some of these downs with the huge cavaet that with matches as distant as these some of them may very well be my paternal side without matching either of them. The general pattern I’ve been able to ascertain is that my Igbo (and related ethnic Nigerian matches) must mostly be on my maternal side, and very few match my father or grandmother. And as far as I can tell, my Bantu matches so far are almost exclusively my paternal side, both his paternal and maternal sides.

        As I’ve described before, from both genealogy and the genetic communities I have on Ancestry; there is very strong evidence that most of my maternal African ancestry comes from colonial Maryland and Virginia. My maternal grandmother has deep, deep South Carolina roots who didn’t even leave the state until after the abolition of slavery. If they had been sold it, was intrastate. My paternal grandmother has very specific genetic communities (Greater Georgetown and South Carolina North Central). It’s weird how specific these are because we’ve been able to trace back to both Georgetown and North Central South Carolina (the Pee Dee region).

        A wrench has been thrown in my paternal grandfather’s ancestry, though. Given some very close matches for me and my father, it seems that my grandfather’s parents were very likely not his biological parents. In addition to my father having my paternal grandmother’s ‘South Carolina North Central’ community, he (and I) both have ‘East Texas, Arkansas & Louisiana’ and even more specifically I have ‘Arkansas-Lousiana Border.’ I don’t know the slaving history of this region much, so I’m not sure if my African ancestors came in through places like New Orleans or Texas or something.

        Liked by 1 person

        • The general pattern I’ve been able to ascertain is that my Igbo (and related ethnic Nigerian matches) must mostly be on my maternal side, and very few match my father or grandmother. And as far as I can tell, my Bantu matches so far are almost exclusively my paternal side, both his paternal and maternal sides.”

          That is a great basis already! What I personally find to be a very promising avenue of follow-up research is to ask all your African matches to upload on Ged-Match. As this will enable you to indentify the exact location of the shared DNA segments you have in common with them. You can then move on to see if any of your other DNA matches are also sharing the same segment with you (triangulation). If these other matches happen to be African Americans with decent family trees it may allow you to zoom into a shared ancestor which goes beyond your grandparents. And keeping in mind all the usual caveats the odds will be high then that his particular family branch may be plausibly associated with the background of your African match.

          You will not always succeed with such requests though. Because regrettably Gedmatch has been getting a bad rep lately. But still only 1 African match on Gedmatch might already give you valuable clues for further research! Also you might find additional African matches on Gedmatch actually. Although from my experience not that many are to be expected (<5).

          I don’t know the slaving history of this region much, so I’m not sure if my African ancestors came in through places like New Orleans or Texas or something.”

          That’s quite complex indeed. I do not know that much either about Texas or Arkansas specifically. But from what I’ve read it was mostly Domestic Slave trade by way of both South Carolina & Virginia into New Orleans and then across state borders taking place in the 1800’s. As well as actually illegal slave trade directly from Africa. Galveston being notorious. It is very tricky though to estimate the relative share of people coming in this way. Prior to that Louisiana itself already had a very diverse slave trade, also by way of the West Indies. With direct Trans-Atalantic Slave Trade actually playing a relatively minor role. See also:

          https://tracingafricanroots.wordpress.com/2015/09/25/louisiana-most-african-diversity-within-the-united-states/

          Like

          • I found a few articles stating that the slave trade into Galveston, texas especially brazoria alot of the slaves were illegally smuggled in through cuba. As far as Arkansas goes I couldn’t find much of anything.

            Liked by 1 person

          • I’m actually a bit confused about that last sentence. In the link, it clearly shows most of Louisiana’s slaves during the height of the trade (54%) were African-born with 45% being American-born. That means, of course, that most of the state’s slaves were not the result of domestic trade. I assume that means that they either came straight from Africa or that they came from the Carribean after being “seasoned.” While I guess that’s minor compared to Virginia or South Carolina, I wouldn’t call it minor. Am I not understanding you correctly?

            Liked by 1 person

            • Those %’s are relating to the Louisiana Slave Database which covers the period 1719-1820. So it’s not really a snapshot for a given year. But merely saying something about those 29,769 captives which happened to have been documented within the database during one century.

              It is an incredibly valuable source. But most of the Domestic Slave Trade into Louisiana would have taken place afterwards (1820-1865). Afterall this relatively late timeperiod saw the sharpest increase in Louisiana’s slavepopulation. See for example this chart:

              The African-born slaves could indeed very well have been brought over by way of the West Indies in many cases. The odds may have been almost 50/50. See also this chart which covers the period up till 1810. Again this is mainly before the Louisiana Purchase!

              You can see that the total number of slave imports for Louisiana by way of either Africa or West Indies (~25,000) is relatively subdued when compared with the estimated 1 million people having said to been impacted by Domestic Slave Trade in total. I do not know the exact numbers for Louisiana itself but you can do the math yourself when knowing that Louisiana’s slave population was said to be 19,926 in 1795 while in 1860 it was a staggering 331,726 !

              https://www.whitneyplantation.org/history/slavery-in-louisiana/

              Like

              • Ah, okay. I’d mistakenly assumed Louisiana’s peak slave importation period happened much earlier than it actually did.

                BTW, the finding of this unfamiliar – and very specific – subregion in my ancestry has gotten me to focus on the frequency of biological families being broken up during the trade. How an enslaved child may have been sold off from his or her biological parents and sent hundreds of miles away and considered an entire other set of parents his or her mother and father, a kind of informal adoption.

                It doesn’t exactly explain the story of my great-grandfather, who was born well after slavery and whose parents were probably born after it’s end, too. But it does explain how African Americans have close DNA connections to places they’ve never heard of in their family lore. Of my predominantly African lines, I’ve only been able to trace one of my paternal grandmother’s lines back to the period of enslavement using both DNA and genealogy. There are tons of other lines that are tracing back to family names and locations I’ve never heard of, giving strong evidence to the idea that there may have been informal adoptions or non-parental events.

                Liked by 1 person

  6. Hello, Felipe. I’ve been binge reading your articles and I thank you for your work. I am adopted so I went on a road to find biological relatives and I found more than I ever thought I would. I reconnected with my biological father through matching my aunt(his sister) then uploaded my raw DNA results to FamilyTreeDna and matched with my first cousin who through emails, I was able to reconnect with my dad.

    So because of this, I was able to confirm a lot of things about my family history on both sides. I read how Pre Civil War Black Population Descendants have roots in either Virginia or South Carolina. My Genetic Community from AncestryDNA is Early Virginia African Americans and both sides of my lineage(including my aunt), most of my matches who have Sub Saharan African Descent has that Community even if they have other Communities and even other Communities like Southwest Ohio African Americans(I don’t have that Community even though I was born in Cincinnati and my mom’s mom has roots in Maysville, Kentucky- a migration reference point-AncestryDNA’s migrations has that Community’s origins from Virginia.

    So with AncestryDNA’s updated results in mind, I have this question for you but I’m going to post my new results that I got in June first:

    Nigeria 37%(Down from 45%)
    Cameroon, Congo and Southern Bantu People 18%(Up from 17%)
    Benin and Togo 11%(Up from 5%)
    Ivory Coast and Ghana 6(up from Ghana 1%)
    Mali 4%(Down from 7%)
    Senegal 1%(Down from 2%).

    I’m not posting my European Admixture which is 22%(20% British Isles and 2% Norwegian) or the new region I got(Yucatan Peninsula, 1%) because it’s not relevant to my question. Here’s my question: Do you believe that the Yorubas are part of Creolization that you wrote about with regards to Antebellum Virginia?

    I ask because Benin is usually associated with Yoruba population and from the stats, I don’t see high amount of slaves from Bight of Benin that is seen from Bight of Biafra. As you know, most of those who are Pre Civil War Black Population Descendants aka African Americans had their Benin and Togo estimates go up including me. When I read the section for Benin and Togo, one of the things they stressed is that 28% of those from there have similar DNA profiles to Nigerians(as well as 43% with population from Ivory Coast and Ghana). So wouldn’t that imply that the Yorubas are part of the Creolization of Virginia and that Benin and Togo is possibly underestimated as far as the Virginia population goes? Or could it come from intermixing with other Pre Civil War Black Population Descendants who might not have been from Virginia via Domestic Slave Trade?

    I know I said one question but a couple others came up as soon as I was typing. Again, thank you for all your work on this blog.

    Liked by 1 person

    • Hi Curtis, thanks a lot for your comment and your interest in my blog! I am really glad to hear that your DNA testing journey has been so gratifying already. I wish you continued succes in your ancestral quest!

      About your question:

      Do you believe that the Yorubas are part of Creolization that you wrote about with regards to Antebellum Virginia?

      Like you said historically speaking the likelihood of substantial Yoruba lineage among AA’s is rather small. And even less so for Virginia. This can be roughly inferred from the share of direct slave trade coming in from the Bight of Benin into the USA (around 3% for Virginia, see this page). However one also has to take into account inter-colonial slave trade by way of the West Indies. My latest blog post happens to deal with this very topic!

      For Virginia it is estimated that this so-called Intra-American Slave trade was around 10% for the whole Slave Trade period (not counting Domestic Slave trade though!). So overall speaking Trans-Atlantic Slave trade was still far more important (90%). But much of this Intra-American Slave trade was happening in the early 1700’s or even earlier. And so you could have had founding effects as a result. African captives were routed mostly by way of Barbados into Virginia. And actually slave trade with Bight of Benin at that time was much more frequent. For Barbados & Jamaica combined an estimated share of around 26%!

      On the other hand we also know that actual Yoruba captives at that early time may not have been that common. Instead most captives coming in from the Bight of Benin would have been Gbe speaking people (Aja, Fon etc.). Most Yoruba captives entered the Americas because of the break-up of the Oyo empire in the 1800’s. This possibility is therefore irrelevant for Virginia which did no longer have any direct Trans-Atlantic Slave Trade already after the American revolution in 1776!

      So really that increase of “Benin/Togo” for yourself could mean several things. In fact it could also merely be associated with Ghanaian DNA which Ancestry currently still finds difficult to categorize. Frankly I would not directly associate it with any Yoruba lineage per se.

      However in case you still want to know if you have any Yoruba lineage my advise would be to look into your African DNA matches. Most African Americans will have a far greater share of Igbo DNA matches, but some do also have Yoruba ones. The next thing to do then is to see if these matches are robust and legitimate, judging by shared DNA amount and if such matches happen to be shared also with any close relatives of your biological family. If this is not the case you could be dealing with a so-called population match. Indicative of southern Nigerian lineage but not per se ethnically specific due to great overlap in the overall genepool of that area.

      I intend to do a survey among African Americans and their Nigerian DNA matches on Ancestry to see how things balance out between Igbo , Yoruba and other Nigerian matches. So keep your eye out on that!

      Like

  7. I’ve found mine and my family’s 23andMe results to be a lot more accurate and reliable than Ancestry’s. Here’s a screenshot of my most recent Ancestry DNA results. For reference my family is half African American (from North Carolina) and Half Afro-Caribbean (mostly Trinidadian and Antiguan with distant redleg ancestry from Barbados and distant ancestry from St. Maarten).

    Like

    • What kind of genetic communities did you get on Ancestry? Currently 23andme doesn’t yet have recent ancestral locations in place for African Americans. So in that respect atleast Ancestry still has the upperhand. But otherwise I agree that 23andme might often be more reliable than Ancestry, especially when it comes to trace amounts of admixture.

      Like

  8. I’m curious about purported European ancestry in AAs in light of us having also some Fulani and Native American ancestry. Even if less than recent European ancestry
    Do you think some of the European is misread Fulani?

    Like

    • No, I don’t see any particular reason for that being the case. European DNA can be distinguished quite accurately from Native America DNA. There is some genetic overlap between Southwest European (Iberian) DNA and North African DNA that’s true. But judging from actual Fula results this issue is no longer showing up. See also the Fula screenshots posted in the last section of this page. You can verify for yourself none of them show any European regions. “North Africa” showing up for AA’s could in some cases be inherited by way of Fula ancestors. However such scores among AA’s are uncommon & minimal from my observation.

      Like

  9. I’ve noticed a consistent pattern of people being aware of the “Nigerian” and Bantu origins of AAs even back in the 19th century.

    Like

    • Yes, you forwarded me those links already earlier. Like I said then:

      there will almost always be a degree of euro-centric bias you will need to take into account. But still I think most of these testimonies do provide valuable insights. On the other hand sometimes generalizations were just uncritically copied from one writer to the other and also from one generation to the other. So you do always have to be careful. Thankfully nowadays we have other means to corroborate historically documented references to the specific ethnic origins of the Afro-Diaspora.”

      Have you read this book btw? I plan on reading it shortly:

      Exchanging Our Country Marks: The Transformation of African Identity in the Colonial and Antebellum South by Michael Gomez.

      Like

  10. Yes I’ve read it. It would along with other evidence I’ve seen. Seem to be in alignment broadly with AncestryDNA results. Oh and in regards to the Fulani. I just read an article about Fulani from Burkina Faso saying they had inherited their ability to digest milk from Europeans. Fulani the study claimed are a mix between Sub-Saharan + NorthAfrican who themselves mixed with Europeans.
    I’ve also read that Native Americans have West Eurasian admixture + Siberian.

    Like

    • Yes I am aware of those research findings. I actually discussed the Burkina Faso Fulani study (Vicente et al. (2019)) on this page (scroll down to section 3 as well as section 5 and see also note 6):

      https://tracingafricanroots.wordpress.com/west-african-results-ii-2013-2018/

      These are mostly (very) ancient DNA markers being picked up when performing a certain type of specialized admixture analysis geared toward finding exactly such finergrained genetic distinctions (“deep ancestry”). It is important to keep in mind that commercial DNA testing companies generally do not perform the same type of analysis!

      Instead on Ancestry as well as 23andme a customized algorithm is being employed which focuses on longer DNA segments. Ignoring/homogenizing any minor degree of genetic diversity to be found within those DNA segments. This is sometimes also called oversmoothing. The intention being to describe one’s ancestry within a relatively recent timeframe of around 500-1000 years.

      Again when looking into actual Fula results on Ancestry (after the 2020 update) you will not see this ancient genetic overlap between North Africa and Southwest Europe. In Ancestry’s 2013-2018 version you could see it though. But only to a minimal extent of usually around 2% in the trace regions. Also the precision/recall for Native American DNA as well as actually African, Asian and European DNA is known to have been very high for several years already. This is why admixture analysis on a continental level is said to be the most accurate.

      This circumstance of any potential ambiguity has been used by many people wishing to explain their subdued or even absent Native American scores. But really just simply looking into the current results of actual Native American descended people (incl. from North America) will show you that its predictive accuracy is on point.

      Even if it were true that some parts of European DNA have been misread in the past for African Americans for this genetic overlap reason I cannot imagine it having been substantial or even surpassing noise level (<1%). Afterall their genuine Fula lineage and even more so Native American admixture is usually going be greatly diluted! AA’s hardly ever show any North African admixture as a result or “byproduct” of Fula ancestry. And even if it does show up for a few people it will almost always be limited to 1% from my observation. So why should you expect any misread Southwest European % among AA’s to be any higher when in fact North African admixture among Fula people is much more clearly detectable?

      Like

      • I actually would not expect it. However, I’m just trying to be ready for any fall-out from family who might not accept the results. I finally bit the bullet and took 3 different DNA tests 23andme, AncestryDNA and AfrorootsDNA. I took the last so I can show my results to you since so far I don’t think anyone else. Has. I have no idea how to judge their accuracy. I do have roots in Georgia and South Carolina. So maybe a bit higher than average Bantu or Upper Guinean/ Senegambian? Truth be told I’d like to see some African results from their company.

        Liked by 1 person

  11. I found a 15cm match on ancestry with the surname Daboh. When Iooked it up it said Daboh is Sierra Leonean. My scores for Mali and Senegal are each 4% I also found a 9cm Yoruba surnames match. Aiyugbesi. I got 46% Nigeria and 3% Benin& Togo. So far no Ghana, Congolese or Igbo/Biafran matches. I do know that Yoruba were taken to the Gulf Coast and I did get Alabama,Georgia and South Carolina African Americans. I also got South Philippines, 1 % pretty sure that’s Malagasy related. I got Early Virginia AAs

    Liked by 1 person

  12. finally found 2 Igbo surnames @ 9cm shared on 1 segment so this is 3 Nigerian matches. Interestingly enough my Yoruba match @9cm has rather high western Bantu. I also noticed that my Sierra Leonean matches have elevated European. I suspect AA admix maybe Krio? they are @ 20 and 15cm. All of these are on Ancestry

    Liked by 1 person

  13. Great findings! Interesting about the match with the Yoruba surname. You should contact him to find out more details about his family background. Many Nigerians are actually of recently mixed ethnic background. And of course at times they may also not be aware of any ethnic intermingling dating back from several generations ago. As you suggested yourself the Sierra Leonean matches might also be through shared African American ancestors. Triangulation and focusing on the exact location of the shared DNA segment might offer more clues.

    Like

    • *Update. I have 3 Igbo matches. Surnames are Olisa,Okeke,and Ononye. 1 Yoruba match surname Aiyegbusi. A Sierra Leonean surnamed Davenport, who has no European ancestry. Elevated Mali as a confirmation. The elevated Western Bantu of my Yoruba match is interesting.

      Liked by 1 person

  14. Aiyegbusi Nigeria76%
    Cameroon, Congo & Western Bantu Peoples 15%.
    Part of my Yoruba cousins results. Still waiting to here back from him. I did not post his first name, which is also Yoruba. Interesting that he did not get much Benin.

    Liked by 1 person

  15. s .davenport Nigeria18%

    Cameroon, Congo & Western Bantu Peoples1%
    Senegal45%
    Mali 31%
    Benin & Togo 4%
    My Sierra Leone match. We share 8cms. She is 100% African. The Nigerian and Benin & Togo makes me suspect Krio. Alternative scenarios are that we share Upper Guinean ancestry ; my ancestors ended up in the Low Country, or if we connect via Nigeria it could be that I had maybe though not guaranteed an Igbo or related relative who left during the American revolution. Obviously just speculation on my part.

    Like

  16. Another Sierra Leone link
    This one has some European ancestry.
    I.D.
    Nigeria 12%
    Cameroon, Congo & Western Bantu Peoples 4%
    Scotland 2%
    Mali 54%
    Senegal 7%
    Benin & Togo 10%
    Sweden 1%
    Ivory Coast & Ghana 1%
    England & Northwestern Europe 9%
    I suspect again an African American link

    Like

  17. I just found 2 Liberians. Both have English surnames and more Nigerian than what your survey shows for typical Liberians but more Upper Guinean than I seem to have. My Ivory Coast & Ghana is only 2% on Ancestry. My Senegal and Mali total 8%. Now what is interesting and a head scratcher to me is my 12.7% Ghana,Liberia,& Sierra Leone on 23andme.

    Liked by 1 person

  18. Sometime back I said I’d give my AfrorootsDNA results. finally, here they are
    2.0%
    Yacouba 2.0%

    7.6%
    Gambian 7.6%

    7.9%
    Mende 7.9%

    10.1%
    Igbo/Yoruba 10.1%

    10.7%
    Bariba 10.7%

    3.5%
    Fon 3.5%

    17.8%
    Esan 17.8%
    2.1%
    Tsogo2.1%

    0.9%
    Orungu0.9%

    0.7%
    Eviya0.7%

    5.9%
    Ovimbundu 5.9%

    1.7%
    Benga 1.7%

    5.2%
    Kimbundu 5.2%

    2.0%
    Umbundo 2.0%

    1.9%
    Bapunu 1.9%

    0.8%
    Fang 0.8%

    Bateke 5.0%
    5.0%

    0.9%
    Luhya 0.9%
    1.5%
    Spanish 1.5%

    10.0%
    Western European 10.0%

    Liked by 1 person

  19. I got my AfrorootsDNA results the highest % is 17.8% Eshan, next is 10.7% Bariba and then 10.1% Igbo/Yoruba. They apparently cannot distinguish these two. They also gave me 3.5% Fon. The rest was mainly ethnic groups found in Gabon, The Congo’s or Angola,plus Liberia/Sierra Leone/Ivory Coast,and Kenya, the Luhya. 7.9% Mende 7.6% Gambian 2.0% Yacouba. I
    only got double digit scores for Nigerian ethnic groups. The Eshan link is interesting though suspect to me given that I got Early Virginia African Americans on Ancestry and 3/4 of my Nigerian matches all only 9cms are Igbo ( I assume) all my Nigerian matches are 100% African

    Liked by 1 person

  20. Thanks for sharing! Interesting that they are not making a distinction between Yoruba and Igbo scores. However they do make a separate mention of Esan. Even when all three are southern Nigerian populations with a great deal of genetic overlap and similarity…

    I have not seen that many AfroRoots results to make a balanced assessment yet. You might want to compare with these results on Youtube

    As a Black-owned DNA testing company it is certainly a praiseworthy initiative. From what I understand they had some starting up problems last year. So I have been inclined to give them some slack to see how they evolve. But at this stage I am tending towards the same opinion as what I have said about LivingDNA:

    I was very excited about LivingDNA’s African breakdown when they first announced it. LivingDNA is certainly to be commended for adding so many African Reference populations into their database. However it turns out that they made their new African breakdown overdetailed and at this moment they cannot live up to the hype they themselves created.

    Which is why I prefer a sketchy regional breakdown such as on 23andme above an overspecified breakdown which gives false hope about pinpointing specific lineage.”

    I do think if you ‘re not able to achieve such detailed ethnic resolution you should also not present the results as such. Instead group all closely related southern Nigerian groups together and label it “Southern Nigerian” for example. Would be even more useful if southwest Nigeria could be distinguished form southeast Nigeria. But if they are not able to do this either at this point, they should just admit it. Of course everyone loves all the new ethnic detail on display. But personally I think it is a more honest approach for DNA testing companies to not be more specific than their data supports.”

    See also these links to read more of my thoughts on LivingDNA.

    https://tracingafricanroots.wordpress.com/2019/12/11/the-mozambique-connection-on-ancestry-myheritage/#comment-7030

    https://tracingafricanroots.wordpress.com/2019/11/27/ancestrys-2019-update-back-on-track-again/comment-page-1/#comment-7443

    Like

  21. one of my matches is surnamed Olisa which apparently is a cognate with other South Nigerian names for God. seemingly found more to the west of Igboland In traditional Itsekiri religion, Oritse is the supreme deity and creator of the world. Among the other deities are Umale Okun, god of the sea, and Ogun, god of iron and war. Divination may be accomplished by men skilled in consulting the Ifa oracle, and ceremonies are performed to the ancestors on various occasions.Dec 12, 2016 http://www.britannica.com › … › Peoples of Africa My suspicion is that maybe I descend from a more westerly Igbo or Igboid group with perhaps more Edoid influence

    Itsekiri | people | Britannicawww.britannica.com › … › Peoples of Africa THE SUPREME BEING- ‘OSANUBUA’For the Edo people, the Supreme Being is traditionally called ‘Osanubuwa’ or ‘Osanubua’. This name can be broken into four parts for etymological clarification:
    ‘Osa’ is a contracted or shortened form of the name “Orise” or the Itsekiri “Oritse” meaning the source of all being.
    ‘N’o’ is “who” or “which” as we have in ‘Chineke’ among the Igbo.
    “B” means “carry” or “Sustain”.
    “Wa” or Uwa means the world or the universe”.
    ‘Osanubuwa’ base on the above etymological trace, mean “the Source of all beings who carries and sustains the world universe”4.

    This name portrays the Supreme Being as the creator and absolute sustainer of the universe. The controller of all other god’s or activities who serve as God’s will and help Him in the theoretic control and maintenance of the universe. They are the intermediaries between ‘Osanubuwa’ (God) and ‘Oria’ (Man); the world we can see and that we cannot see.https://www.africanholocaust.net/news_ah/edo_religion.html
    then we have
    Olisa | Igbo Nameswww.myigboname.com › entries › Olisa
    Find the meaning of Igbo names. Olisa. /Olisa/. Gender: m. (From the expression “Olisa-bulu-ụwa”). The Lord who bears the weight of the world.
    of course Yoruba Orisha

    Like

  22. all the little matches I got from Gabon curious. also Bariba and Fon. 3.5% Afroroots 3%. Ancestry &range 0-11% Benin/Togo
    Afroroots 10.7% Bariba. not sure if Bariba should count as more Nigerian or Beninois(e) . So far no Akan hits only 2% Ivorycoast/Ghana. Afroroots 2% Yacouba about 8% each Mende and Gambian. I do have some Liberian and Sierra Leonean matches, albeit with elevated Nigerian and Central African admix. Ancestry gave me AL,GA,SC AAs as a genetic community which makes sense. However 23andme assumes recent Jamaican ancestry so no Akan seems off. Overall I think Afroroots can give plausible proxies for now.

    Like

  23. I’ve been thinking about my roughly 11% Bariba match on AfrorootsDNA. This to me is intriguing due to neither of my genetic communities being tied to areas that received a lot of people from the Bight of Benin. Alabama, Georgia and South Carolina African Americans and Early Virginia African Americans. Ancestry gave me 3% Fon and this AfrorootsDNA gave me 3.5%. the Esan and Igbo/Yoruba scores I get. But the Bariba is curious. Any insight on what could be going on? Also being 7.6% Mende but only 4% Mali.

    Like

    • It could be various things really. The %’s will vary because of differences in algorithm as well as differences in refernce populations being used. As well as the relative composition of these refernce samples. Also the range of your ethnicity estimates are to be taken into account.

      It is indeed really interesting that AfroRoots seems to be using Bariba samples from Benin. But such ethnically specified scores are usually not to be taken too literally. One thing you might derive from this outcome is that this part of your DNA is not from either Central Africa or Upper Guinea. Most likely narrowing it down to the area in between Ghana and Nigeria. But any specification beyond that is bound to be based on shaky grounds.

      Like

      • I did get a link to Jamaica on 23andme maybe its Chamba or some other Gur ancestry. I know the 3.5% Fon they gave me could be just part of any southern Nigerian or generic Gbe speakers. 10.7% of this Bariba. 0-11% Benin&Togo but reported as 3%. Other than the Bariba loop. Pretty much Afroroots is in line with my 23andme and Ancestry results. South Nigerian, West Bantu,and (South )Mande.

        Like

  24. I have a match who I’m not sure if she is both Edo and African American or what. Her last name is Eduwensuyi. Took me a while to find out her last name is Edo. She is on Ancestry. On what we share it’s all African. If she does have any European it is not being shown. Her results are not complete when I add up the African it is not 100%.

    Ethnicity Estimates
    You
    s. eduwensuyi
    46% Nigeria 34%
    29%Cameroon, Congo & Western Bantu Peoples
    28%
    4%Mali 2%
    4%Senegal 1%

    3% Benin & Togo 17%
    2% Ivory Coast & Ghana 5%
    5%Scotland 0% 2%Sweden 0%
    1% Wales 0%
    1% Southern Philippines 0%
    1%Northern Italy 0% 1% England & Northwestern Europe 0% 1% Indigenous Americas—North
    0%

    Like

    • Yes at times Ancestry will not show the complete breakdown of your DNA matches due to privacy settings. Only showing regions you have in common. Hopefully reaching out to this match will clarify things.

      Like

    • My Liberian cousin has Bassa and Ghanaian ancestry on her mom’s side with some recent German ancestry apparently, her dad’s side is Americo. One of the surnames via her Ghanaian is Okai, which might be Ga. What stands out in her results is How high her Western Bantu % is. I’m guessing this must be her dad’s side.

      Like

      • Interesting! Hopefully the two of you will eventually be able to find out how you are related. As it seems several possibilities exist.

        Ever since my Ancestry survey (2013-2018) but also currently on 23andme I have been struck by how Americo-Liberians as well as Krio Sierra Leoneans often will show distinctive regional scores. Reflecting their more complex African origins, at times also including Central African lineage! The fact that both Ancestry and 23andme are able to pick up on this is quite reassuring. Even when of course follow-research, incl. finding associated DNA Matches will always be needed.

        Liked by 1 person

  25. Another Autumn brings yet another update of Ancestry’s Ethnicty Estimates… Unfortunately much of what I discussed last year still stands. The African breakdown continues to be merely a sideshow on Ancestry. Hardly any meaningful positive changes for people of West & Central African descent in particular! Which is why I will not dignify this update with a full-blown review as I have done in previous years.

    To be fair overall speaking Ancestry does seem to have made some modest steps towards improvement also with their African breakdown. But like last year nothing game changing. And it is increasingly becoming frustrating when you are aware that Ancestry is able to do so much more! As proven by the detailed regional resolution Ancestry is offering for especially their customers of European descent.

    I will provide a brief summary of my main impressions further below. Although I have to stress that I haven’t performed any representative survey as I was accustomed to do. But first let me repeat these statements of last year.

    However the relative neglect of African & Afro-descended customer needs does go against Ancestry’s selfproclaimed goal to make their product experience inclusive for everyone. In my previous blog post I stated that Ancestry should seek to offer new tools geared to facilitate specialized research for Afro-descended customers. It should be clear that this update does NOT compensate for the loss of small African matches,

    As I have said several times now I really think Ancestry should finally begin expanding their Genetic Communities tool into West & Central Africa. This is needed now more than ever. Given that Ancestry’s deletion of smaller DNA matches (6-8 cM) has been harmful especially for Afro-descendants! Especially Nigeria and Ghana being promising first candidates, based on greater availability of customer samples. But when enabling DNA matching with samples from Ancestry’s Reference Panel possibly also additional genetic communities could be realized for Upper Guinea, Sierra Leone/Liberia and Cameroon/Congo? Such an expansion would be truly beneficial for Ancestry’s customers who aim to Trace African Roots. Ancestry would then truly be stepping up their game. In line with their self proclaimed goal to make Ancestry’s experience truly inclusive for everyone!

    1) new features: “Egypt” and new genetic community “Eastern Africa”. Which is of course welcome in itself for people of such descent. But this addition is not helpful at all for understanding the main lineage of Atlantic Afro-Diasporans! Just like last year’s novelity “Southern Bantu” these new regions will usually be absent for Trans Atlantic Afro-descendants or only appear at minimal noise level.

    Really disappointing as it underlines for me that further specification of West and Central African DNA still seems far from being a priority for Ancestry… The total number of African regions (mostly genetic communities for African Americans) has only marginally increased from 110 last year to 112. While Europe now has no less than 1180 regions. Compared to 838 last year, so quite an increase yet again on that front!

    2) New African samples were added to be fair. See overview below. But still the proportion of African samples in Ancestry’s current Reference Panel remains stuck at only about 6% (3553/56580). It should also be noted that more than half the increase of these African samples went into North Africa/Egypt.

    3) Predictive accuracy of most African regions also appears to be steadily improving for the most part. Probably mostly due to the addition of new samples into Ancestry’s Reference Panel. I don’t think anything has changed for Ancestry’s algorithm. Hence some small shifts in African breakdown for Diasporans, mostly in the right direction from what I have seen sofar. In particular “Nigerian” levels are less likely to be inflated. But still likely to underestimate Ghanaian lineage in particular.

    4) Just based on a few individual results: “Ivory Coast/Ghana” is on the increase for Akan Ghanaians and even more so for Kru Liberians. So it seems that Ancestry is gradually returning to the set-up of the 2013-2018 version. However still some way to go as “Ivory Coast/Ghana” levels are still far below what they used to be for Akan Ghanaians and Liberians. Which means that also for Afro-Diasporans this region’s tell-tale function of signalling such lineage is still not restored.

    5) Ancestry’s sloppiness when it comes to regional descriptions etc. is getting somewhat less but still leaves much to be desired. As discussed last year this could be seen especially from the maps included in your results. With this new update some corrections have been made for the maps of “Mali”, “Southern Bantu” and “Khoisan, Aka & Mbuti people”. However other maps such as “Nigeria” remain potentially disorientating.

    To sum it all up: Ancestry is still not stepping up their game and taking it to the next level! Basically still same story as after the update in 2018 …

    Suggestions for improving the African breakdown on AncestryDNA

    Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s