Ancestry’s 2022 update: some bright spots, but too little too late!

In this blog post I will provide a review of Ancestry’s latest update.1 As always I will keep my focus on the African breakdown featured in Ancestry’s so-called Ethnicity Estimate. In fact Ancestry has released several updates earlier this year. Eventhough some positive developments are taking place overall speaking I am not impressed. Because I just know that Ancestry is able to do so much more! It almost seems that Ancestry is going out of its way in order to avoid making any truly breakthrough update for their customers of West/Central African descent…

***ivc-ghana-updatedAncestry is introducing two new African regions with this update. These are however either minimal or absent for practically all Atlantic Afro-descendants. The good news is that the predictive accuracy of “Ivory Coast & Ghana” is finally being restored to its former level from the 2013-2018 version! At last providing a reasonable indication of DNA associated with that region and neighbouring countries (especially Liberia!). Something which Ancestry managed to screw up during their previous updates..

***

My main qualm with Ancestry is that genetic communities for pinpointing West & Central African lineage are still missing! Even when earlier this year 23andme demonstrated that such a feature can be implemented already. Introducing no less than 25 ethnolinguistic genetic groups from Africa (see this blogseries). To be sure this tool based on matching strength is still a work in progress. But at least 23andme has set things in motion and stepped up their game! Something which Ancestry for some strange reason is still refusing to do…

Contents of this blog post:

  1. African breakdown for Africans after the 2022 update
    • Ancestry’s African Reference Panel
  2. African breakdown for Afro-descendants before and after the 2022 update
    • Best African breakdown since 2013-2018!
  3. Ethnicity Inheritance and Chromosome Painter
    • How can Ancestry finally take things to the next level?
  4. Screenshots of African updated results

***

1) African breakdown for African AncestryDNA testers

Table 1.1 (click to enlarge)

stats-african-2022Based on the updated results for 145 African AncestryDNA testers from 36 countries, across the continent (*=scaled to 100%). Keep in mind that the regions might also refer to DNA which is found in neighbouring countries. However in most cases the results are strongly in line with actual origins. Follow this link for my spreadsheet containing all the individual results. Compare also with this overview from 2020.

***

Although these “ethnicity” updates are becoming standard practice it is understandable that many people still tend to get confused. Do make an attempt to inform yourself properly without being dismissive right away. Despite my critical stance I am still convinced that you can get valuable ancestral clues from Ancestry’s African breakdown. The first place to get more clarification is Ancestry’s Ethnicity Update FAQ. Also don’t forget to consult the usually helpful maps which are included with your results. Although these maps are still only meant to be indicative!

Another way of making sense of it all is simply by looking into the results of African DNA testers to get an idea what to expect. Furthermore it will also be useful to contrast the results of any given group from the Afro-Diaspora with historical plausibility. This is something I have done for almost 10 years already (see this section). My motto remains the same: each DNA test as well as each update should be judged on its own merits.

It is also crucial to realize that macro-regionally speaking the changes will usually be far less drastic than on first sight. Differences in results are often merely representing shifts between neighbouring regions. Which is logical and in some ways inevitable given genetic overlap. For more insight compare with my previous findings from 2020 (based on the results of usually the same persons featured in my current survey):

This new update by Ancestry in fact has not resulted in that much change. From the two new regions being introduced only “Nilotic Peoples” seems like a useful addition. Although interpretation depends on the exact background of the 233 samples being used by Ancestry to determine these “Nilotic” scores. I am guessing that aside from South Sudanese samples also Kenyan Maasai samples are now included in Ancestry’s Reference Panel. Which might explain the somewhat exaggerated “Nilotic” scores for my Tutsi survey participants.2 Still this represents a major advance which only leaves the central Sahel region (Niger, Chad, Central African Republic) as one of the few remaining gaps within Ancestry’s African breakdown.3

The new “Nigeria East-Central” region is a big disappointment which I will discuss in greater detail in a separate blog post:

“Ivory Coast & Ghana” back on track again!

Map 1.1 (click to enlarge)

Ivc Ghana Map

***

For the already existing regions a gradual improvement can be seen on most fronts. Especially for “Senegal” and “Southern Bantu”. Interestingly “Mali” seems to have decreased quite a lot for my Hausa-Fulani survey participants (29% on average in 2020 vs. 13% in 2022). While “Benin/Togo” is slightly increasing for Yoruba people (19% in 2020 vs. 23% in 2022). It also appears that Ancestry’s homogenizing or “smoothing” algorithm has loosed up a bit because “Khoisan, Aka & Mbuti” scores are starting to appear again for Central and Southern Africans. Albeit with minimal amounts and still clearly underestimating such lineage, except for South African Coloureds.

We can also observe a spectacular return of predictive accuracy for the “Ivory Coast & Ghana region”. In particular for Akan people (55% in 2020 vs. 88% in 2022). This is certainly good news because this also has a positive impact on the overall West African breakdown for other people. Still it should be pointed out this merely implies a return of the results people were already able to receive when Ancestry’s pioneering West African breakdown was first available in its 2013-2018 version. At that time my Akan survey participants also had the exact same average of 88% “Ivory Coast/Ghana” (see this overview and also this blogpost).

At the same time we can also see that many Liberians are scoring substantial “Ivory Coast/Ghana” amounts as well. Although their group average is behind what it originally was during the 2013-2018 version (73% in 2018 vs. 52% in 2022). It appears that especially for northern Liberians “Mali” remains prevailing while for Liberians with a Kru background more convincing “Ivory Coast/Ghana” scores can be seen (even 100% in one case!). For (Mende) Sierra Leoneans it is still so-called “Mali” all the way. Which most likely indicates that Ancestry is still using them as additional samples to define their “Mali” region. Otherwise this blogpost from 2018 still applies:

 

Ancestry’s new Reference Panel

Table 1.2 (click to enlarge)

Reference Panel 2022a

***

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 2022 update. See also:

Some observations:

  •  The proportion of African samples in Ancestry’s Reference Panel remains stuck at 6% (4431/68714). In 2020 this share was also 6% (2856/44703). Despite increase in actual numbers this relative proportion has remained marginal during each update since 2018. Clearly demonstrating the minor importance of African samples within Ancestry’s evolving Reference Panel.
  • The improvement of the predictive accuracy of “Ivory Coast & Ghana” appears to have been caused by a better alignment of samples from “Benin & Togo” and “Ivory Coast & Ghana”. Just speculating but especially the large decrease of samples for “Benin & Togo” (-151) appears to be the result of better quality control? I suppose better late than never 😉
  • The new “Nilotic Peoples” and “Nigeria East-Central” regions are based on a robust number of samples (resp. 233 and 471). This certainly provides a good basis. However as I have said in previous reviews of Ancestry’s Reference Panel: more is not always better! It is also about careful selection, overall composition and balancing of the various samples. As well as the algorithm being applied of course.  See also:
  • In typical fashion Ancestry is still not revealing the exact background of their African samples (beyond nationality). Even when such details are crucial for proper interpretation! Ancestry is not alone in being hesitant when providing such details. This also goes for many other DNA testing companies. Especially the ethnic but also provincial backgrounds of their reference samples can be very relevant but are often not clarified. However Ancestry should take a cue from 23andme which is becoming more transparent in this regard (see this link for 23andme’s African Reference Panel and this link for 23andme’s African genetic groups).

***

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

Table 2.1 (click to enlarge)Stats Diaspora 2022aBased on the updated results for 50 AncestryDNA testers from across the Afro-Diaspora. The West African breakdown is arguably in the best shape it has ever been. The improved coverage of “Ivory Coast & Ghana” being the main cause. But also “Senegal” and “Nigeria” levels seem more realistic now. Still some imperfections remain. In particular the”Cameroon, Congo & West Bantu” scores for Brazilians and Puerto Ricans continue to appear to be inflated. The individual results can be seen by way of my spreadsheet.

***

In order to judge how Ancestry’s 2022 update has impacted the results of people across the Afro-Diaspora I have yet again performed a small but still comprehensive survey for various Afro-descended nationalities (n=50). And I must say the results are more in line with historical plausibility than ever before. Mostly due to Ancestry finally fixing the “Ivory Coast & Ghana” defect (after FOUR years!) Some observations:

  • “Senegal” no longer seems to be structurally underestimated for African Americans and West Indians. This had been the case since 2018 already. Actually for Cape Verdeans and most Hispanic Americans the predictive accuracy of “Senegal” has been on point again since 2019. Steadily improving ever since. Mostly at the expense of their “Mali” scores. Take note however that for African Americans and West Indians “Mali” scores have remained more or less the same! Probably because for them these scores are more so related with Liberian and Sierra Leonean DNA.
  • After a long delay (since 2018!) a reasonable estimate of Ghanaian and related DNA is yet again available on Ancestry. In line with historical expectations “Ivory Coast & Ghana” scores have been jumping up for especially Barbadians, Jamaicans and African Americans. The average level they are now showing is remarkably close to my original surveyfindings from 2013-2018 (see this overview). The only key difference is that in 2022 Sierra Leonean DNA is unlikely to be included in any “Ivory Coast & Ghana” score. Instead it will be covered by “Mali”. Liberian DNA right now being in limbo between the two. Although Kru lineage is very likely to be mostly ending up in “Ivory Coast & Ghana”.
  • The change in “Benin & Togo” scores is not too drastic and also not consistent for all groups. But generally speaking the direction of change makes sense when being aware of slave trade levels from Bight of Benin. Especially the increase of “Benin/Togo” for Brazilians and Haitians, while for most others it is decreasing. Still a very substantial level remains on display for Barbados!
  • Also the “Nigeria” levels seem pretty realistic at this moment. Especially for African Americans and West Indians I suspect that there is no longer much reason to assume these scores are heavily overstated, as they used to be in previous versions.
  • Central African DNA continues to be detected well by Ancestry. However due to Ancestry’s oversmoothing algorithm in some cases it might be inflated for some people. Especially Brazilians and Puerto Ricans. Also interesting to see a somewhat higher level of “Southern Bantu” appearing for Brazilians and Hispanic Americans. In line with historical expectations and also reminiscent of previous “Southeastern Bantu” scores being highest for them with the 2013-2018 version.
  • The scores for the new regions are mostly non-inexistent or at a trivial level of <1%. Intriguingly I did observe one survey participant from Puerto Rico who scored 1% “Nilotic Peoples”. Probably within the margin of error. But still not entirely impossible that this could be a genuine indication. But as always this needs to be backed up by additional corroborating evidence (esp. associated DNA matches!). However it find it more telling that all previous scores from Northeast Africa have now disappeared. A reminder that such exotic trace regions can be quite short-lived, just like fool’s gold 😉
  • Scores for “Nigeria East-Central” were a bit more more frequent but still 80% of my survey group (40/50) did not receive any “Nigeria East-Central” score at all. All group averages are around 1% or less. As can be seen in Table 2.1. The highest individual score in my survey was only 3% for a Haitian. Granted: such minimal scores can still provide valid indications. But for me personally these paltry statistics are a clear sign of how little added value this new region brings. Much more meaningful improvements for the West African breakdown could have been initiated by Ancestry! For more discussion see:

 

Best African breakdown since 2013-2018!

Table 2.2 (click to enlarge)

AA Macro-RegionalThis chart shows my survey findings for ten African Americans whose results I have been following since 2018 and even before. Notice that the 3-way breakdown has actually remained consistent. In fact most of the major changes appear to represent internal reshuffling within any of the three main macro-regions.

***

Many people seem to think that each update on Ancestry only leads to more random changes. But as I have always maintained the macro-regional breakdown has stayed pretty much the same. To make more sense of my Afro-Diasporan survey findings across the years I have been using an additional more basic macro-regional framework divided into:

The chart above clearly demonstrates this. It features the group averages for 10 African American AncestryDNA testers I have been keeping score of since at least 2018 up till now. A small but representatively chosen survey group (see this spreadsheet for their individual results). By first looking into their group averages in the original version (2013-2018) and then after the 2022 update you can get a good idea of the main impact the update had. It can be confirmed that the 3-way breakdown is practically the same, give or take a few %’s. This was already true in 2019 (see this chart and also detailed discussion in section 3 of this post).

I am using the 2013-2018 results as a base line because in my opinion the 2013-2018 version used to provide the best African breakdown Ancestry had on offer untill they blew it with their disastrous update in 2018.  Based on how the data on a group level matches up with historical plausibility as well as the results of actual African AncestryDNA testers. Subsequent updates up till now have not been able to fully amend things.

But with this 2022 update for the first time I am inclined to say that also going by the regional averages it seems that Ancestry’s African breakdown has come full circle again for Trans-Atlantic Afro-descendants! After four whole years… To be fair for many Africans (incl. Cape Verdeans) actually Ancestry did already provide clear improvements over their original 2013-2018 version, starting with the 2019 update. However due to their more complex and intricately mixed African origins for most Afro-descendants the updates between 2018 and 2021 have often not been beneficial…

Again see footnote 1 for a full overview of all the reviews I have done on Ancestry’s updates across the years. Just for a very quick round up. The 2018 update was the biggest slap in the face because of vanishing “Nigeria” and “Senegal” scores and super inflated “Benin/Togo”, “Cameroon/Congo” and also “Mali” scores. In 2019 Ancestry tried to remedy this but imperfectly so. Causing much confusion because of all the major flipflops. Also this time “Ivory Coast & Ghana” got derailed completely because Ancestry made the ill-advised decision to split this region.

The 2020 and 2021 updates took small steps in improving things but were still leaving intact some of the major imperfections.  With “Nigeria” and “Benin/Togo” being slightly overstated and “Ivory Coast/Ghana” being seriously underestimated. Only with this 2022 update it seems Ancestry is finally able again to offer a reasonably close fit for historically plausible regional roots within Africa. Despite remaining shortcomings such an outcome is not something to carelessly brush aside when wanting to Trace African Roots!

I would have to collect more data for better judgment but right now I can already establish that based on group averages the estimates for “Ivory Coast & Ghana” as well as “Nigeria” will be more reasonable than before. And this will be of particular benefit for African Americans and West Indians. When comparing with the 2013-2018 version it appears to me that the average “Benin/Togo” level is now also more in tune with historical expectations. Although for African Americans this is still higher than recorded Trans-Atlantic slave trade with Bight of Benin (around 3%). But this can be explained by other plausible historical scenarios (see this blogpost).

Another advantage over the 2013-2018 version is the greater accuracy in differentiating between Southeast African and Eastern African DNA. The former “Southeastern Bantu” region was often actually detecting Central African or Southwestern Bantu DNA (see second part of this blogpost). But right now this region has been replaced competely by the much more appropriate regions of “Southern Bantu” and “Eastern Bantu”. These regions are still being reported at times for African Americans and other Afro-Diasporans. However usually with minimal amounts of 1%. From what I have seen scores above 5% will be very exceptional indeed. Which is in agreement with expectations. Either way in most cases I do suspect these scores are picking up on something meaningful and worthy of follow-up research.

With each update since 2018 Ancestry has steadily been expanding its African regions to the east.  Northeast African DNA currently is being specified with no less than 4 regions: “Nilotic Peoples”, “Somalia”, “Ethiopia & Eritrea” and also “Egypt”. The fact that these four regions almost never get reported for Atlantic Afro-descendants is of course not surprising. After all they are almost exclusively of West & Central African descent with only a minor additional Southeast African component. Something which can be corroborated not only by history and genetics but also by way of cultural retention.

Nearly ten years ago in 2013 AncestryDNA revolutionized the world of DNA testing with a pioneering West African breakdown. Obviously this original 2013-2018 version had several shortcomings. But still it represented a major advance at that time and offered unrivaled West African regional specificity when compared with other commercial DNA testing companies during that same time period as well as any third-party analysis such as available on Ged-Match, DNA Land etc.. See also these links:

With this 2022 update Ancestry is finally (after 4 years!!!) providing some (modest) improvement over this 2013-2018 version. However I don’t think this is cause for much celebration given that Ancestry could have done so much better in previous years already. And of course in the meanwhile the competition has not stood still. Arguably 23andme in particular has taken the lead, not only in providing reasonable regional estimates in line with historical plausibility. But also by spearheading their feature of Genetic Groups which can link you to 25 African ethnolinguistic groups.4

***

3) Ethnicity Inheritance and Chromosome Painter 

Figure 3.1. (click to enlarge)

Ethnicity InheritanceThese are the new Ethnicity Inheritance results for an African American. This new tool can be insightful in many ways. In this particular case notice how it appears that one of the parents has passed on a much greater chunk of “Nigerian” DNA. Although this info should be interpreted carefully, it could still be very valuable for sorting out associated Nigerian DNA matches. Confirming or indicating which family line they fall on.

***

Earlier this year Ancestry already introduced two new features which are also based on regional admixture a.k.a. “ethnicity”. These new features are interrelated because Ancestry uses the same new technique for it called: Sideview. Basically Ancestry is using your DNA matches and the regional admixture they share with you to predict which DNA segments you have inherited from each parent. Amazingly they are able to do this without either one of your parents having tested with Ancestry! This allows Ancestry to come up with hypothetical DNA inheritance results from your parents. As well as providing an overview of regional admixture for each one of your chromosomes and how this might relate to either one of your parents.

Potentially both tools should enable you to make valuable advances in your ancestral research, incl. Tracing African Roots. Once more illustrating that regional admixture DOES matter! Naturally given correct interpretation and with sufficient context being provided.5 From my experience the Ethnicity Inheritance tool is most useful at this moment. The Chromosome Painter is still in Beta phase. And frankly it doesn’t measure up to comparable tools available with other DNA testing companies, such as 23andme.

For people whose parents are of clearly different background Ethnicity Inheritance should be most accurate. I have seen this confirmed by many people already. And it is certainly true for me as well. Because my personal Ethnicity Inheritance results were spot on and very much in agreement with my father being Dutch and my mother Cape Verdean. For people whose parents belong to the same or a closely related background it is less clearcut. Because when your parents show overlapping ethnic estimates it becomes more tricky to disentangle. Mainly due to random inheritance. But also because from what I have seen Ancestry is not always consistent in its ethnic estimates across generations.

Despite such imperfections I do think this tool has great potential especially with new updates. It can be very useful already for tracking down regional admixture which is unique or more prevailing for one of your parents. This will often be the smaller scores you might have, incl. also non-African scores. But as shown in the screenshot above this might also apply for your main or even primary regions.

In fact Ancestry also provides a useful detailed comparison with the actual percentages. And there is also an option to set the parents on maternal/paternal, in case you have solid indications for this. Even better when you have tested both of your parents or even also your grandparents because that way you can really start assigning certain distinctive regional scores to certain family lines. Going up your familytree and therefore getting closer to the first ancestor who would be the original source of that regional score. I can also see how this would work really great when you combine with any of your African DNA matches and are able to apply triangulation. See also:

Figure 3.2. (click to enlarge)

Chromosome PainterAncestry’s Chromosome Painting still has many shortcomings. However especially for distinctive regional admixture this tool could be useful already. Especially in this case quite likely the genetic footprints of a Malagasy lineage are highlighted for an African American. Which could prove to be very valuable when combined with  DNA Painting and triangulation of your associated DNA matches. 

***

As I have already pointed out Ancestry’s Chromosome Painting is still in beta phase. Although certainly promising it should be mentioned that currently Ancestry’s Chromosome Painter is clearly unequipped to provide the same powerful and detailed analysis as provided by 23andme. In fact 23andme has been providing a similar tool for more than ten years already. So to say that Ancestry is kind of late to the party is a huge understatement… Also FTDNA provides a similar tool. While MyHeritage has a Chromosome Browser (without specifying ethnicity).

From comparing my own results on both 23andme and Ancestry it is apparent that 23andme provides much greater detail, incl. crucial segment information. And also covering the X-chromosome. Furthermore 23andme’s algorithm seems less homogenizing than on Ancestry. Entire chromosomes often being assigned to the same regional admixture category on Ancestry. While on 23andme a more scattered overview is given which often will be more in line with reality. The same observations have been made by many other people. Actually the reluctance of Ancestry to implement a fullfledged Chromosome Browser (which does include segment info) still seems to be unchanged.

Seemingly for privacy reasons Ancestry continues to refuse reporting highly significant details about the DNA segments you share with your DNA matches. As I have been arguing for several years already Ancestry should atleast start showing the ethnicity/admixture of shared DNA segments with your matches. This can be very useful for many purposes. For example when reviewing your Afro-Diasporan matches it will tell you if you happen to share an African or European ancestor. And zooming into shared African lineage it will also make it more easier to single out at least a particular shared African region of provenance. Especially when combined with the location of your DNA segments this could truly be  beneficial for Ancestry’s customers. For more reviews and discussion see these links:

 

How can Ancestry finally take things to the next level?

Africa next levelIn order to really step up their game Ancestry should refocus on West & Central Africa. As this will be beneficial for the greatest number of their Afro-descended customers. Principally African Americans but also other Atlantic Afro-descendants (Latin Americans, West Indians and Cape Verdeans). Source image.

***

____________________

“I really think Ancestry should finally begin expanding their Genetic Communities tool into West & Central Africa. […] 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? ” (Fonte Felipe, 2020)

“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.” (Fonte Felipe, 2020)

____________________

Throughout the years I have made numerous suggestions for improvement. In 2018 I already made a plea to include West/Central African genetic communities on Ancestry. Also during my review of Ancestry’s 2020 update I discussed this in great detail. Because I believe such an introduction (when carried out correctly) will be a game changing event for people who are in most pressing need of finding out their ancestral origins. Regrettably Ancestry does not have a good track record in dealing with their customers’ feedback. They might ask for your opinion after each update. But this appears to be mere lip service… Action still speaks louder than words!

In 2013 Ancestry was very much at the forefront of providing a West & Central African breakdown which was truly beneficial for Ancestry’s largest segment of Afro-descended customers. African Americans and other Atlantic Afro-descendants (West Indians, Latin Americans and Cape Verdeans). However with subsequent updates during 2018-2022 Ancestry has not really moved beyond what it had already achieved in its 2013-2018 version. The predictive accuracy of the West & Central African breakdown was being played around with. Something which only now with the 2022 update seems to be fully corrected.

To be fair new regions have been created for Southern and Eastern Africa. However these regions are hardly ever being reported for African Americans and other Atlantic Afro-descendants. And therefore they are of little to no use. Despite the addition of likewise marginally appearing “Nigeria East-Central” it is still true that there are only 6 African regions which really matter for Atlantic Afro-Diasporans. At least when going by double digit scores (“Senegal”, “Mali”, “Ivory Coast & Ghana”, “Benin & Togo”, “Nigeria” and “Cameroon, Congo & Western Bantu” ). See also Table 2.1.

Map 3.1 (click to enlarge)

GC - East AfricaThis is a screenshot of where Ancestry’s latest African genetic community is to be found. Not located in West or Central Africa as may have been expected given that Ancestry’s Afro-descended customers will overwhelmingly trace back their African origins to these parts. When will Ancestry get its priorities right? When will we have Genetic Communities for West & Central Africa?

***

In the meanwhile Ancestry has actually already implemented a handful of  genetic communities which are directly linked to Africa. Although not all of them are picking up on strictly African DNA: Morocco & Algeria“; Asians in South Africa“; South Africa, European Settlers“; Portuguese Islanders in the Eastern U.S.”. Despite the labeling this last community is actually referring to Cape Verdeans! Ancestry continues to ignore that Cape Verde has been an independent island state located in West Africa since 1975.6 A gross oversight which is in clear violation of how Ancestry supposedly is concerned about “cultural sensitivity”

In 2021 Ancestry introduced a new genetic community for East Africans. 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 this year’s novelity “Nilotic Peoples” this new community will practically always be absent for Atlantic Afro-descendants. I find this really disappointing because for me it underlines once again that further specification of West and Central African DNA seems to be last on Ancestry’s priority list. This new East African community shows a wide grouping of countries, going from South Africa to Ethiopia. But this range does not make any genealogical sense from what I have seen.7 So even for the very few East African descended customers on Ancestry who do get assigned to this community it has little use…

I am fully aware that Ancestry’s management might not be very concerned about this issue.8 Even when Ancestry’s Afro-descended paying customers will overwhelmingly trace back their African origins to West & Central Africa. And therefore a refocus on West & Central Africa is not only in great need but also fully justified. Going just by the number of people who are poised to benefit from it. Again aside from African Americans this will also include Latin Americans, West Indians and Cape Verdeans. Minorities to be sure but all combined representing a substantial share of Ancestry’s customers who do not deserve to be continuously neglected!

To keep it short and simple I will once more suggest two basic improvements which should already be feasible from my understanding. Given Ancestry’s ample resources, incl. probably the biggest number of African tested customers. And additionally also their widely acclaimed Sorenson (SMGF) database which includes thousands of unique African DNA samples.9 Of course where there is a will there is always a way 😉 Hopefully Ancestry can find inspiration in 23andme’s recommendable initiative to provide a new feature which can link you to 25 African ethnolinguistic groups.

  • 1) Refocus on providing greater specificity for West & Central Africa, by providing historically relevant regions & genetic communities for the Atlantic Afro-descendants. By far the greatest segment of Ancestry’s African descended customers! These should at least include separate regions/communities for:
    • Atlantic and Mande speaking groups in Upper Guinea; also taking into account the widespread Fula migrations across the Sahel!
    • Sierra Leone and Liberia. Preferably decoupling these countries from the so-called “Mali” region.
    • “Mali” region should be redesigned to exclusively cover DNA from the western Sahel, incl. also Burkina Faso.
    • Ewe/Gbe speaking people and Akan people.
    • southwestern Nigeria, southeastern Nigeria and northern Nigeria. To be related to genetic affiliations with resp. the Yoruba, Igbo and Hausa.
    • Cameroon, preferably focusing on distinctive groups not found in neighbouring countries
    • Bakongo people from DRC Congo/Angola to be contrasted with groups from interior DRC/Angola/Zambia.
    • Madagascar/Mozambique. These countries are not West or Central African but still historically very much relevant for Atlantic Afro-descendants! Clearly more so than the “Eastern African” genetic community…
  • 2) Provide adequate and helpful context which will help people to interpret their results. Avoid meaningless infotainment texts and misleading labeling which might confuse people about their true lineage!

See also:

To end on a positive note I want make it clear that I do appreciate the effort that has gone into developing the new Ethnicity Inheritance tool! I have also already acknowledged that Ancestry’s African breakdown after this 2022 update is arguably in its best shape since the 2013-2018 version. Another redeeming aspect is that Ancestry has finally seen the light when it comes to providing a helpful map for interpreting their “Khoisan, Aka & Mbuti Peoples” region. This map used be an embarrassing illlustration of how Ancestry became increasingly sloppy since its 2018 update…Hopefully things have now turned for the better.10

4) Screenshots of African updated results

Just limiting myself to a few results which feature the updated “Ivory Coast & Ghana” region as well as the new regions “Nilotic Peoples” and “Nigeria East-Central”. See my spreadsheet for a complete overview of individual results.

***

GHANA (Ashanti)

***

GH - Ashanti (slkd)

***

GHANA (Akan & Ga-Adangbe)

***

GH - Akan + GA

***

GHANA (Ewe)

***

GH - Ewe (dt)

***

GHANA (North: Chamba)

***

GH (north)

***

LIBERIA (Kru?)

***

LIberia - Kru

***

LIBERIA (Kru + Kpelle)

***

LIBERIA (yar)

***

LIBERIA (northern)

***

LIBERIA (north)

 

***

NIGERIA (Northeast/Taraba state: Hausa-Fulani)

***NG - Haua-Fulani (ssad)

***

NIGERIA (Igbo)

***

NG - Igno (go)

***

NIGERIA (Yoruba)

***

NG - Yoruba (GO)

***

ANGOLA 

***

ANG )katia)

***

ZAMBIA

***

ZAMB - (mw)

***

1/2 SOUTH SUDAN (Nuer) & 1/2 Euro-Canadian)

***

Z.SUD - Nuer (Eur-Canadian)

***

UGANDA (1/2 Nilotic: Kakwa & 1/2 Central Sudanic: Aringa)

***

UGA - Ben

***

UGANDA (?)

***

UGA - leo

***

RUANDA (Tutsi)

***

RUA - Tutsi

***

KENYA (Kikuyu)

***

KEN - Kikiyu3

***

KENYA (Swahili)

***

KEN - Swahili

***

KENYA (Taita)

***

KEN - Taita

***

TANZANIA (Jita)

***

TAN - Jita (mm)

***

TANZANIA (Hangaze)

***

TAN - Hangaze

***

___________________________________________________________________________

Notes

1See these links for my previous reviews of updates on Ancestry:

2) Interesting to compare the “Nilotic” scores of Tutsi with the “East African” results which were given to them by 23andme in the period 2013-2018. And that time 23andme was mostly using Maasai samples. See also comment section.

3) Actually in 2021 Ancestry also introduced a new “Egypt” region. I have not included this in my survey because I do not have access to any Egyptian AncestryDNA results. I have also not observed any meaningful “Egypt” scores among my African & Afro-descended survey participants. Except for a few cases when a minimal trace amount of <1% was reported. I have included such scores with “North Africa”.

Of course the introduction of this region and also “Nilotic Peoples” is welcome in itself for people of such descent. But this addition is not helpful at all for understanding the West & Central African lineage of Atlantic Afro-Diasporans! As shown by my survey findings these new regions will usually be absent for Atlantic Afro-descendants or only appear at minimal noise level. Their African DNA almost exclusively to be traced back to West & Central African ancestors, except for some minor additional Southeast African lineage.

4) Although I tend to be critical in my assessment of 23andme there is no doubt in my mind that this company is currently best equipped to deliver trustworthy details on African origins in a robust manner. Going by regional admixture and haplogroup assignment 23andme’s trackrecord is already very strong. For many pundits genotyping by 23andme is simply the best around. Especially after the 2018 and subsequent updates it seems 23andme is taking a decisive lead on its competitors. Meanwhile Ancestry has only been putting out lacklustre updates in the last few years. The absence of any new genetic communities for West & Central Africa is really telling… See also:

5) 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. Especially after this latest 2022 update.

To be combined with any other ancestral clues you may have, especially DNA matches. This is now cleverly being applied by Ancestry in their two new tools: Ethnicity Inheritance & Chromosome Painting. However it it still a shame that in 2020 Ancestry deleted so many potentially very informative smaller African matches. Compensation in the form of new genetic communities for West & Central Africa is still not forthcoming…See also:

6) This is a quote from Ancestry’s website:

How do we name our regions?

“Representing the people of the world fairly is our priority. We go through several steps and work with a diverse network of outside scholars and experts to develop and review how our regions are named. We start with a set of maps that show us where we typically find a particular ethnicity region in the results of people who are native to an area. We also consider who will typically be getting a region in their results to make sure a name is broad or narrow enough to be a good fit. That gives us a starting point. From there we sometimes test various names with users local to an area or ask them for suggestions. Finally, we have a panel of outside subject-matter experts with local and scholarly expertise who review our names for both accuracy and cultural sensitivity.”

Of course this is all fine and dandy but words should matter beyond just making politically correct statements! I have reached out to Ancestry several times to inform them that their genetic community which is centered on people with shared Cape Verdean lineage is clearly mislabeled. However Ancestry is choosing to maintain a very akward and potentially offensive labeling. Using the name “Portuguese Islander”, even when Cape Verde obviously is an independent state in West Africa! At the same time Ancestry has been setting up more than a dozen specific genetic communities for actual Portuguese islanders in the Azores and Madeira. Which makes you wonder how much Ancestry is truly committed to listening to their (Afro-descended) customers. See also this comment of mine on Ancestry’s own blog, made in 2019.

7) I have kept track for whom this new “Eastern African” genetic community is being reported among my African survey participants. This includes persons from South Africa, Zimbabwe, Zambia, Ruanda, Madagascar, Tanzania, Kenya and Uganda. However unlike what is mentioned by Ancestry I have not yet seen this community being reported for people from Ethiopia or Somalia. Still this covers a huge area! Which is inhabited by hundreds of ethnic groups! Applying common sense they should not have close genealogical ties to people living almost on either side of the continent!

I was very therefore much surprised by this wide geographical range which frankly defeats the purpose of setting up a genetic community. Keep in mind that for Europeans Ancestry is able to zoom into very detailed micro level. For example currently for Ireland Ancestry is making a distinction for almost 100 (sub) communities! See this link. Despite not always being 100% accurate these results are still usually very much indicative and beneficial for one’s genealogical quest.

Ancestry does not provide any useful context in the community history for “Eastern Africa”. But I suspect that possibly this community might also be picking up on ancient genetic affiliations due to Bantu speaking migrations from Eastern Africa into Southern Africa. However possibly also the genetic effects of more recent movements of people in the region (incl. Swahili slave trade and the Ngoni migrations) are being reflected. Which is intriguing in itself. Also when comparing with the similar grouping of “Shona & Nguni” on 23andme (see this map). But naturally this should then be clarified for unexpecting customers… I really hope that when Ancestry eventually decides to also add genetic communities for West & Central Africa such careless handling will be avoided.

8) I take it for granted that DNA testing companies are bound to be profit-driven. However this does influence how much true committment to customer value for Atlantic Afro-descendants should realistically be expected. DNA testing for this group of people is no trivial matter as it may be for other people! And it is certainly not something to be played around with by companies who still seem to severely underestimate the wider impact of the product they are selling. In this regard the news reports about the current management of Ancestry (Blackstone) are not really encouraging. This investment group is even notorious in the Netherlands!

9) The purchase of the Sorenson (SMGF) database enabled Ancestry to achieve its pioneering West & Central African breakdown in 2013. But I highly suspect that Ancestry’s SMGF database is still being underused. Especially given the sudden appearance of no less than 471 samples for the “Nigeria East-Central” region. Ancestry is not transparant about this (see this link). But according to the archived version of the former Sorenson website (see this link) they collected no less than 4,220 DNA samples from Nigeria! And also thousands others from other usually undersampled African countries (esp. Benin, Cameroon and Mali). Naturally this could turn out to be a true goldmine! Even if only half of these samples are available within the AncestyDNA set-up, incl. genetic communities! Serving as an alternative for relying on relatively rare African (migrant) customer samples. More discussion in my next blogpost:

10) “Khoisan, Aka & Mbuti Peoples” a.k.a. “Hunter-Gatherer” scores continue to be misleading or underestimated for many Africans. However at least Ancestry FINALLY got the map right 😀 This is something I have been pointing out since my first review of updates on Ancestry. Eventhough I personally do not even find this region very useful. Because it is of only marginal ancestral significance for most people, except South African Coloureds (see this blogpost). Still the sloppy maps created for this region were a disturbing indication that decent quality control by Ancestry has been severely lacking the last few years. Hopefully with this 2022 update a more professional and dedicated approach will now prevail among Ancestry’s staff members involved. In fact the map for “Southern Bantu” has now been corrected as well to also include Madagascar, as I pointed out already in my 2020 review

Map in 2018 (click to enlarge)

***

Map in 2019 (click to enlarge)

***

Map in 2020 (click to enlarge)

***

Map in 2022 (click to enlarge)

***

Advertisement

14 thoughts on “Ancestry’s 2022 update: some bright spots, but too little too late!

  1. I’m Glad they are finally starting to fix things. My Nigerian score went from the single digits originally, and shot up all the way to over 30%. After the first update. And my Ghana scores, which were the third highest, at like 14-16 percent,crashed down to zero…lol. it does seem like things have started to make more of an improvement now. My Ghana scores have now Rose back from the grave. Though have as high at 8 percent. My Nigerian scores took a haircut again. Though they aren’t as low as my original estimate.

    Here’s my update (African only)

    Benin togo : 22%
    Central Africa/western bantu: 18%
    Nigeria:18%
    Mali: 12%
    Ivory coast/Ghana: 8%
    Senegal: 6%

    Also, I have a extra little gift for your data base. :). A relative of mine several months ago recently took the ancestry test and got their results. This person is my half-brother, we are half through our mother side. 🙂 I will give his entire ancestral breakdown though instead.

    Nigeria: 32%
    Benin/Togo: 13%
    Cameroon/congo/western bantu: 10%
    Mali: 9%
    Ivory coast/Ghana: 6%
    Senegal: 9%
    England and northwestern Europe:16%
    Ireland: 1%
    Sweden/Denmark: 1%
    Norway:1%

    Liked by 1 person

    • Thanks, I actually still have your original breakdown from before the first update in 2018, lol! All in all I would say your updated results from 2022 have come full circle more or less when compared with the breakdown you had in 2018. With the minor differences to be considered as an improvement. And your results from in between 2018-2022 best to be forgotten haha!

      You had quite a prominent “Benin/Togo” score back then (31%). It is still in first place right now, but probably a more reasonable estimate this time. This should also go for your “Nigerian” score. In my screenshot you had 10% “Ivory Coast/Ghana” in the beginning. Most of it has now come back indeed! The difference is probably because of your increased so-called “Mali” score which actually also will include Liberian and Sierra Leonean DNA.

      Also interesting to see how your current 18% “Cameroon, Congo & Western Bantu” score is about the same as your previous “Cameroon/Congo” score of 15% added up with the “Southeastern Bantu” region (4%). This latter region has now been replaced by “Southern Bantu” and “Eastern Bantu” but these regions are quite uncommon for African Americans and rarely above 1%. Which should be according to historical expectations. So again a slight improvement.

      Have you checked out your Ethnicity Inheritance results already? As always not to be taken as gospel 😉 Still intriguing to see that according to Ancestry your “Mali” score and also your “Ivory Coast/Ghana” score have been entirely passed down to you by one parent. Who could be either your mother or father.

      This outcome should be interpreted carefully because just because you inherited all of this regional DNA from one parent does not rule out that your other parent might also still have additional lineage from these places. It’s just that because of genetic reshuffling you did not get that DNA from the other parent. Still could be very useful info when combining with any African DNA matches associated with these regions. Because it could give you clues then to which parental side such matches belong. Best thing though would be to test your parents.

      Like

  2. I’m glad someone was as angry as I was when I saw the Nigeria East Central region map. Ancestry purposely chose a region in Nigeria that includes all the the three major tribes Igbo, Yoruba and Hausa Fulani. It was a slap in the face and stupid to include.

    Liked by 1 person

    • Hi Lamont, very disappointing indeed because this was the first addition to the West African breakdown since 2013. Ancestry could and should have done so much better! They had more than enough time to prepare for it…

      Keep an eye out for my follow-up blogpost which will deal with this new region in greater detail. Including also how to interpret this region if it happens to show up for you. Although like I mentioned in this post already for most people “Nigeria East-Central” will be entirely absent and so this was truly a wasted opportunity!

      Feel free to share that upcoming post and also this one with Ancestry when they ask you for any feedback on their update.

      Like

  3. Hey Felipe!

    I commented a long time ago on one of your posts with my results from a Brazilian test company. My mother is Black, from one of the Blackest cities in Maranhão, while my father is white, from a proeminent family (a bit like gentry), and from Ceará. I finally tested with 23andme and I hope to test my mother with 23andme someday. Here it goes my results: https://imgur.com/a/CRMxJ4Z

    Liked by 1 person

    • Ola Daniel, muito obrigado por compartilhar! I remember indeed that you previously tested with MyHeritage. These results with 23andme are bound to be much more accurate! Especially you can be much more confident in your predominant “Angolan & Congolese” score, unlike the absurd “Kenyan” and “Somali” scores on MyHeritage. Also interesting to see the secondary score for “Nigeria”. Keep in mind that this might also include DNA from neighbouring Benin. Although due to so-called smoothing (putting related types of DNA in the same box) it might be that certain regions are being underestimated. While others are perhaps somewhat overstated. This is also still an issue on Ancestry

      For example I am a bit surprised about the minimal “Senegambian” given your mother’s Maranhão background. But perhaps Mandinga is just not in her lineage because she’s from a particular place. You mentioned Codó the other time right? Do you know if she’s aware of any Quilombola heritage? I would love to hear more about it!

      It will be very interesting to countercheck with your mother’s 23andme results. She might also get a more specified group mentioned for her African results (see this blogpost for several Brazilian examples). Please do let me know when you get them in! Or if you have the opportunity you might also want to test her with Ancestry instead. Because that way you will have a greater chance of getting linked to African DNA matches which will be very useful for zooming into specific ethnic lineage. Posso te ajudar com isso! And actually as described on this page Ancestry’s African breakdown is currently not that bad at all and possibly in the best shape it has been since 2018. So when it comes to just regional admixture the results should be quite close with 23andme.

      Congratulations anyways! Because your personal results are also already very fascinating of course. Incl. also the European breakdown and the Jewish, North African and Amerindian scores. I have visted both Maranhão and Ceará when I travelled across Brazil. Loved both places! Adoro Brasil e espero com todo meu coração que as coisas vão melhorar brevemente!

      Liked by 1 person

      • Please I can give you my results because there are a lot of misconceptions about us..

        By the way….what do you think is the real meaning or breakdown of “Eastern Bantu.”

        Thank you.

        Liked by 1 person

        • That would be great! In case you like to participate in my ongoing survey please follow these steps:

          • Sign in to Ancestry.com
          • Click the DNA tab and select Your DNA Results Summary
          • Click the Settings button on the right side of the page
          • Scroll down to Visibility and sharing
          • Click on DNA Test sharing and then Invite
          • Enter my Ancestry username which is FonteFelipe
          • Select role of viewer
          • Click the INVITE button

          About the “Eastern Bantu” region I actually think Ancestry has (partially) been using Luhya samples for it! Which would also explain why you have such a high score for it. However it is not intended as an exclusive marker of Luhya lineage of course! The highest score I have seen sofar is 98% “Eastern Bantu” for a person from Uganda, see also the screenshot section above. Unfortunately I do not know his ethnic background, but I do know his family is probably from the north. A Tanzanian with Kuria background scored 81% “Eastern Bantu”.

          Ancestry has gradually been adding more samples for this region. In 2018 it was introduced as “Eastern Africa’ with a sample size of 82. In 2019 it was renamed into “Eastern Bantu” and sample size became 91. In 2020 this increased to 138 and in 2022 we’re already at 177 samples! So that’s an increase of almost 100 since 2018. Regrettably Ancestry still refuses to give us more information about their exact background..

          For more details about Ancestry possibly using Luhya samples see this blogpost from 2018, scroll down to section 2: “New samples added from Kenya as well as Tanzania?”

          Did Ancestry kill their African breakdown? (part 2)

          Like

  4. Hey Felipe. Thanks for your great write up as usual. Did the 2022 update (or the update before that) make Igbos score for central africa region plummet? I remember in the past Igbos used to have a significant amount of central africa but from the 5 samples you posted above they’re almost 100% nigeria now.

    Like

    • Thanks a lot! Yes it was actually already with the 2019 update that the “Cameroon/Congo” scores for Igbo people mostly vanished. As well as the substantial “Benin/Togo” they used to get actually. Ever since that update they tend to get over 95% “Nigeria”. I rarely see Nigerians come up with scores higher than 10% for “Cameroon, Congo, Western Bantu”. Perhaps for some groups from near the border it could still happen though.

      Ancestry’s 2019 Update: Back on Track Again?

      Like

      • Ok tx Felipe. That means the cam/congo/western bantu that showed up in a relative’s profile is not just from SE Nigerian heritage then 🙂

        Like

        • Well I don’t know your relative’s background. Because that might also be a factor 😉 However generally speaking yes since the 2019 update Ancestry has provided a sharper delineation between “Nigeria” and “Cameroon, Congo, & Western Bantu”. Which should make it easier to distinguish also more roughly between West African DNA vs Central African DNA.

          However as also stated in this blogpost:

          Central African DNA continues to be detected well by Ancestry. However due to Ancestry’s oversmoothing algorithm in some cases it might be inflated for some people. Especially Brazilians and Puerto Ricans.”

          To a lesser degree this might also be true for many Haitians. From what I have seen for African Americans so far it looks to be in line though with historical expectations (going by group averages).

          Generally speaking what seems to be still at work is that Ancestry’s algorithm focuses on long stretches of DNA segments when assigning to a particular regional category. While shorter segments run the risk of being assigned to a neighbouring region which is more prevalent, on the same chromosome.

          So let’s say you have 1 Congolese ancestor from the 1800’s and 10 Nigerian ancestors from the 1600’s. Because of random recombination each generation the Nigerian segments will naturally become more diluted, fragmented and just smaller, on average. While the genetic inheritance of the relatively recent Congolese ancestor will be much more clearcut. With odds being high that also longer DNA segments will have remained. Now when Ancestry’s algorithm is having a look at your DNA it might very well decide that whenever a very small “Nigerian” DNA segment is adjacent to a longer “Congolese” DNA segment that it will just group them all together under the label “Cameroon, Congo, Western Bantu”. Possibly resulting in overstated estimates of Congolese lineage and underestimation of the more distant Nigerian ancestry.

          This should be more an issue for people whose regional origins are very different according to time period/century. For others whose African ancestors mostly arrived in the same time period I am guessing it should not really matter that much. This risk of so-called “oversmoothing” is not just on Ancestry btw but also on 23andme see this link:

          https://blog.23andme.com/articles/algorithm-gets-an-upgrade

          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 )

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