In 2013 AncestryDNA updated their Ethnicity Estimates to include a very detailed breakdown of West African ancestry. A pioneering initiative when compared with the rather basic regional within-Africa resolution being provided by other DNA testing companies. Obviously 100% accuracy cannot be guaranteed and should also not be expected at this stage. Given the inherent limitations of DNA testing in general and more specifically the restricted availability of African reference populations. See AncestryDNA Regions for a more detailed assessment. Still I find the results very fascinating and insightful for anyone wanting to learn more about their African roots. Here’s a link with more info:
Survey of the African breakdown on AncestryDNA
Since the update I have been collecting AncestryDNA results in a spreadsheet in order to conduct a survey of the African origins being reported by AncestryDNA for African Americans as well as other Afro-descended nationalities. Scroll down to the bottom of this page for an overview of the main statistical findings of my survey. At this moment of writing it contains at least 1,377 results for 34 nationalities, among which 515 are for African Americans. For easier comparison I have recalculated everyone’s original African percentages as mentioned in the Ethnicity Estimates so that they add up to 100%. In other words the scores mentioned in my speadsheet represent fractions of a person’s total African ancestry and not fractions of their entire ancestry. Below is a link to the spreadsheet which contains all the results. Besides a tab for the main statistics there are several other tabs on the bottom of the sheet where the individual results can be found, sorted according to nationality.
These are the links to the blog posts I have written sofar in an attempt to analyze and illuminate the African breakdown for each group:
- Afro-Diaspora AncestryDNA results: A Comparison (2016)
- African AncestryDNA results (2015)
- African American AncestryDNA results (2015)
- Anglo-Caribbean AncestryDNA results (2015)
- Cape Verdean AncestryDNA results (2015)
- Dominican AncestryDNA results (2015)
- Haitian AncestryDNA results (2015)
- Latin American AncestryDNA results (2015)
- Puerto Rican AncestryDNA results (2015)
As the name already implies this blog is dedicated to Tracing African Roots. However many if not most Afro-descendants actually also have additional non-African ancestry. And for some people this part of their DNA might also be interesting to explore further. I have therefore started a new survey featuring the AncestryDNA results of persons from all over the world. In order to improve correct interpretation of AncestryDNA’s regions by comparing results with persons from verified backgrounds. These are the links to the blog posts i have written sofar:
- Asian, Pacific & Native American AncestryDNA results (2017)
- European AncestryDNA results (2017)
AncestryDNA’s ethnicity estimates have been updated several times now. On this page I am dealing exclusively with AncestryDNA version 2 which was current between September 2013 and September 2018. All matters being discussed on this page are therefore not pertaining to recently updated results (version 3). In my opinion the new version regrettably has been a downgrade rather than providing any meaningful improvement. Which is why I have discontinued my AncestryDNA survey. For more details see:
- Did Ancestry kill their African breakdown? (part 1)
- Did Ancestry kill their African breakdown? (part 2)
- Did Ancestry kill their African breakdown? (part 3)
AncestryDNA’s ethnicity estimates have been updated again starting in October 2019. This 2019 update has restored most of the major flaws from the previous 2018 update. However I do not think it offers any substantial improvements for the most part. At least not when compared with the original African breakdown from the 2013-2018 version. See this blog post for more detailed discussion:
My AncestryDNA survey remains discontinued therefore. At least when based on regional admixture. I am currently shifting my attention towards DNA matches and also how these may correlate with admixture. See also:
- DNA matches reported for 50 Cape Verdeans on AncestryDNA (2018)
- The Mozambique connection on Ancestry & MyHeritage (2019)
- African DNA matches reported for 30 Jamaicans on Ancestry (2020)
For a short recap of my final survey findings read these blog posts:
- Fula, Wolof or Temne? Upper Guinean AncestryDNA results 2013-2018
- Akan or Ewe? West African AncestryDNA results 2013-2018
- Igbo, Yoruba or Hausa-Fulani? Nigerian AncestryDNA results 2013-2018
- Final summary: North & East African AncestryDNA results 2013-2018
- Final summary: Central & Southern African AncestryDNA results 2013-2018
In the spreadsheet referenced above i have compiled and recalculated the AncestryDNA results which have kindly been shared with me or otherwise collected by me on the internet from public websites. The greater part of my African American data was gathered on these three threads posted on the 23andme forum and the Ancestry Support Community:
- Share & Discuss Ancestry DNA’s Ethnicity Estimate’s African Refined Results (posted Oct. 21, 2013)
- AncestryDNA breakdown of SSA for African Americans & other Afrodescendants (posted Apr. 17, 2014)
- Post your African breakdown (incl. Trace Regions!) (posted Mar. 21, 2014)
In addition i was also kindly invited by many people on Ancestry.com to view their Ethnicity Estimates. I like to extend my gratitude to anyone who participated in these threads as well as anyone who granted me access to their results on Ancestry.com. Your willingness to contribute to my research has been really valuable to me! I have mainly used acronyms or nicknames in my spreadsheet in order to safeguard everyone’s privacy. If anyone should have second thoughts about being included in the spreadsheet or would like to be renamed, just send me a PM and i will act accordingly.
In order to make the results intercomparable, regardless of non-African admixture, i have scaled the African part of the AncestryDNA breakdown to 100%. The scaling formula i used is very simple and can be verified from within the spreadsheet by clicking on any cell featuring a regional score and then viewing the calculation in the function bar (fx) in the upper left corner. All other Excel formulas i used throughout the sheet and especially in the tab “Stats” can also be verified in this same way.
Basically i applied the following formula: percentage for a given African region divided by percentage of total African amount, as mentioned in the Ethnicity Estimates. For example, someone with a total African amount of 80% and a “Senegal” score of 20% would be entered into my sheet as 20/80 = 25% “Senegal” out of his total African ancestry. The same regional ratio of 25% could be obtained for a person of only minor African descent who according to AncestryDNA scores 4% “Senegal” out of a total African amount of 16% (4/16). In this manner the regional proportions of the African breakdown become instantly visible for everyone and can also be counterchecked with 100% Africans for example.
In order to rule out any entry errors i created an extra column which automatically adds all regional shares up. This column named “Sum” acts as a built-in check as normally it should show 100%. In some cases it still doesn’t precisely add up to 100% however, due to either incomplete or incorrect data. Sometimes it’s also because the original Ethnicity Estimates on AncestryDNA themselves don’t add up to exactly 100% (usually because of the <1% Trace Regions).
Some other features of the spreadsheet:
- Regions which were reported with the highest amount by AncestryDNA are bolded in green cells. The frequency of these regions being ranked #1 is also mentioned in the tab “Stats”.
- The column named “∑ Top 2” features the sum of the two biggest regional scores (already scaled). Combining the shares of the two main regions provides a rough measure of how homogenous or rather heterogeneous a person’s African breakdown might be.
- I have used abbreviations for some of the regions and in some cases i renamed them:
- “Ghana/Ivc.” = “Ivory Coast/Ghana”
- “Camr./Congo” = “Cameroon/Congo”
- “SE Bantu” = “Africa Southeastern Bantu”
- “Pygmy/San” = “Africa South-Central Hunter-Gatherers”
- “North Africa”= “Africa North”
- In the tab Stats+ i make use of an additional (macro)-regional breakdown into: Upper Guinea, Lower Guinea and Central Africa. This was done in order to comply with what’s common practice in slave trade literature. And also to analyze the regional patterns from a related yet slightly different perspective. It is based on very rough proxies and imprecise delineations. “Central Africa” should be read as also including Bantu speaking ancestry from Southeast Africa. “North Africa” was excluded to enable a more evenhanded comparison with Hispanic Americans and Cape Verdeans.
- I have also entered additional ancestral details for many individual results. Whenever such information was available and deemed to be relevant. Generally this info is to be found in appropriately named columns either to the far left or far right.
My survey consisted out of entering the African breakdown of individual AncestryDNA results according to nationality and also USA state origins. Subsequently group statistics have been calculated and blogposts have been written in an attempt to provide more analysis and context to the data. I basically started this research for two key reasons:
1) Obtaining a better understanding of everyone’s personal results. Generally speaking I find that it’s only by comparing with each other under the same settings that we can derive deeper meaning from our DNA test results. Without any benchmarks so to speak it’s more difficult to get a good grasp of how your personal results might possibly be interpreted and also ultimately how reliable they would be. It is my hope that at the very least the way the African regional amounts have been presented in the spreadsheet has been helpful in adding some more perspective for those who participated.
2) Exploring the regional African origins of people from across the Afro-Diaspora. Attempting to establish how much the AncestryDNA results on a grouplevel can already (despite limitations of sample size) be correlated with whatever is known about the specific regional African roots for each Afro-descended nationality mentioned in my spreadsheet.
During my research I have come across many fascinating findings which possibly might provide profound insights. They are described in more detail in the various blogposts referenced above. With the most comprehensive summary to be found in: Afro-Diaspora AncestryDNA results: A Comparison. One overriding theme has been the sheer regional diversity on display in almost all the African compositions reported by AncestryDNA. Only a few persons and even fewer nationalities showing a clear and consistent pull towards just one single predominant African region. As a consequence it seems fair to state that: when tracing your African roots expect to find ancestors from multiple African regions and also from multiple African ethnic groups. This essential piece of knowledge has often been obscured by history for many Afro-descendants. Some finding it easier to imagine just one single place in Africa to identify as one’s ancestral homeland. Or just one single African “tribe” they wish to be matched with. In almost all cases this will however prove to be a big misconception. An intricate mix and unique blending of various ancestral lineages from throughout Western, Central and Southeastern Africa being much more likely. See also:
Going beyond any possible data-entry errors on my part and the inherent limitations of DNA testing (see also AncestryDNA regions) I would like to point out the following:
- Future updates by AncestryDNA will undoubtedly produce new and hopefully improved data. The analysis performed by AncestryDNA, even when pioneering and already very valuable in my estimation, is only a first rudimentary step in deciphering one’s origins within Africa. The current nine African regions might very well have different ethnic implications for each separate Afro-descended nationality but also in individual cases.
- I am naturally inclined to seeing the glass as half full rather than half empty. When reviewing the current AncestryDNA results I have preferred to look for whatever informational value may be obtained already from them. Rather than let any imperfections lead to an overhasty dismissal. To me that would be like throwing out the baby with the bathing water. I have nonetheless maintained a critical stance and attempted to point out any inconsistencies or limitations of the data whenever I could.
- The sample size for most groups seemed robust enough to warrant some further generalizing analysis. Keep in mind i am just offering my personal interpretations in the blogposts i have written. My research findings are not set in stone but rather based on the (limited) information available to me. Also other people might come up with different conclusions based on the same information.
- For the purpose of my research i have focused on group averages to get a better grip on the underlying patterns. However it should be stressed that individual variation is a given and it’s not my intention to deny it in any way. Every one in my spreadsheet will have unique family trees including lineages from various parts of Africa. Because of random inheritance some of this regional ancestry might not even get expressed in DNA markers detected by AncestryDNA. Especially for people of minor African descent. Whichever regional ancestry is being reported will often be showing up as either more or less prominent than the group averages. This is only natural.
- Determining the largest regional components, on average, for each Afro-descended nationality has been a primary research effort. As they can be considered to have the highest reliability at this stage and might also be confirmed independently by historical sources. Some regions, especially when turning up as Trace Regions, will have been termed by me as “minor” or even “insignificant”. But of course this is only strictly relatively speaking. The actual ancestors behind single-digit regional percentages are not of minor importance themselves. They might still evoke some personal interest for anyone who receives the results. I naturally respect this. Still I also think it’s prudent to at least acknowledge the predominant ancestral components you’re made up of. As the people associated with these greater parts of your ancestry will have contributed the most to who you are, at least genetically speaking.
- During my research up till now I have in no way been affiliated with Ancestry.com. I have been intrigued from the start by the African regional specification provided by AncestryDNA. I personally find it more insightful than anything presently on offer by both commercial DNA testing companies as well as any thirdparty analysis such as available on Ged-Match.
For more detailed analysis of this data see the blogposts referenced above or also the dropdown menu below AncestryDNA. For an up-to-date overview of my ongoing survey see:
- AncestryDNA results for Africans and the Afro-Diaspora (only African breakdown)
- AncestryDNA results for Africans (group averages, incl. for selected ethnic groups)
- AncestryDNA results for Afro-Diaspora (group averages)
- AncestryDNA results for Asians, Native Americans and Pacific Islanders (group averages)
- AncestryDNA results for Europeans (group averages)
***(click to enlarge)