Gedmatch’s HarappaWorld admixture test answers my West African results question

I have been covering the challenges of African descended DNA analysis in this series of posts. The children of African descent, particularly those in the Americas and the Caribbean do not have a straightforward genetic inheritance. Through the centuries we have inherited a smorgasbord of genetics from wildly different populations. Putting our non-African inheritance to one side, slaves came from many parts of the African continent. And this is at the heart of my conundrum with the generously donated and free admixture tests available on Gedmatch.

I have been scratching my head over why my Gedmatch admixture test results were so completely skewed to West Africa. The results bore no resemblance to my West African heritage shown on a more comprehensive DNA test taken a year ago via Genebase.

I had my suspicions, namely that the people who have given their time to developing these free analytical tools could only create their analytical tools with publicly available admixture data sets. In other words, they had to work with what they could get. The fault, if there was one, lay in the data sets with which they were working. But I needed proof.

The smoking gun came in the form of the Harappa World test on Gedmatch. I am grateful that the developer of Harappa has been so transparent on the blog he created about this test.

I ran the Harappa test and you’ll see the results below:


S-Indian 0.54%
Baloch 3.43%
Caucasian 6.91%
NE-Euro 15.13%
Siberian 0.12%
Papuan 0.59%
American 0.94%
Beringian 0.07%
Mediterranean 12.04%
SW-Asian 1.82%
San 0.68%
E-African 2.62%
Pygmy 2.86%
W-African 52.25%

Again, West Africa is wildly out of proportion to what I know to be true. So I went to the Harappa blog ( to investigate. And I found background information on the test that would prove to be a goldmine.

The test uses populations from around the world. Which is great for those of us with complex non-African genetic inheritances. I could get an overall sense once again of just how mixed my admixtures really are. Again, looking at the table above, the non-African results are pretty much in line with what I know about my admixture makeup already. There are some omissions to be sure, but those populations weren’t included in the data sets for this test – it’s always a good idea to really read the background information for these tests to understand what world populations have and haven’t been included and what each test has been designed to measure.

Group results that form this test are shown in the bar chart below (click on the images  below to see a larger image).

harappaworld-admixture-1harappaworld-admixture-2harappaworld-admixture-3harappaworld-admixture-4harappaworld-admixture-5harappaworld-admixture-6harappaworld-admixture-7If I’m understanding this bar chart correctly, populations from the Caribbean and the US are influencing the results for West Africa – and not by a little bit either. The inclusion of either of these populations would drastically skew results for West Africa. Together, the weight they place on this result is dramatic. The Caribbean and the US should be their own categories, and not included within West African results. African descended peoples in either region are the children of Africa, not the progenitors.

Not all slaves who arrived in the Americas and the Caribbean came from West Africa, which is another point to consider. Yes, a sizable percentage of African slaves did come from this region. However, one shouldn’t assume that just because you’re African descended and live in these two regions that the larger part of your African DNA inheritance comes from West Africa. I’m a living example of this with the majority of my African DNA arising from Northwest Africa (Tuareg & Berber) with regions such as North, Central and East Africa contributing far more than my inheritance from either South and West Africa.

Again, kudos to Harappa’s creator for his candor and transparency. In his own words:

“…the admixture components do not necessarily represent real ancestral populations. Also, the names I have chosen for the components should be thought of as mnemonics to ease discussion. I chose them based on which populations in my data these components peaked in. They do not tell anything directly about ancestral populations. The best way to look at these admixture results is by comparing individuals and populations.”

This, in the end, answered my question. I would advocate that such an important qualifier should accompany the test itself.

Casting an eye down the bar chart provided above, making note of the various tribes who comprise results for West Africa, there are many who influence this result. Some I have to question. I will take the Kongo, as an example. The larger part of the Kongo admixture (86%) is attributed to West Africa. In actuality, historically, this is a Northeast African tribe. Which is odd as Northeast Africa is only shown as contributing to 5% f this tribe’s admixture. In my Gedmatch DNA test, the Kongo tribe is attributed to Northeast Africa and accounts for roughly 22% of my admixture from Northeast Africa (Egypt is the largest contributor from this region of Africa at 53% via Genebase).

Again, this isn’t a criticism of the person behind the Harappa DNA analysis test. It’s more to do with the originator of the data set that was produced.

I have a strong feeling that this is the reason behind the skewed West African results for the tests I’ve done to-date on Gedmatch.

I have one more Gedmatch test to run and post about. However, what I will say at this point is that a more refined understanding of African admixtures is sorely needed for these kinds of admixture analysis tests. Personally, I would love for someone who understands the intricacies of Africa, its peoples and African admixtures to develop a Pan-African admixture test that is every bit as comprehensive and detailed as the European and Eurasian focused tests found on Gedmatch. That would be an amazing thing indeed.

I’m fortunate. I found a DNA testing service provider that answered this question for me already. My wish is for others to have a refined and accurately reflective picture of their own African genetic inheritance.

I have one last Gedmatch test to report on. It’s the EthioHelix K10 Africa Only Admixture test. In many ways, I saved the best Gedmatch admixture test for African descended peoples to last.

Books by Brian Sheffey:

Purchase link:
Purchase link:

10 thoughts on “Gedmatch’s HarappaWorld admixture test answers my West African results question

  1. It seems like you’re saying you’re an African American from the first paragraph which means you’re likely not of Tuareg or North African descent but rather west African. Some African Americans want to be East African and North African for some reason, but that’s not really the case.

    1. Rick. My DNA results are my DNA results. They are what they are. I have shared those results, shared screen grabs of those results, and written about them numerous times.

      It’s not ‘wanting to be’ anything. I didn’t say that my African DNA was entirely East or North African. Nor did I say it was entirely Tuareg. My African DNA comes from ALL four corners of the African Continent. My quibble is with the accuracy of free analytical tools.

      It’s kind if funny, though. I keep getting a chorus of criticism about anything other than West African DNA results. It’s funny because the more people of color in the US who post their DNA results in the myriad of genealogy and genetic genealogy Facebook groups show….varying amounts of Eastern, Northern, and Central African DNA.

      What you offer is an opinion. While I get what that opinion is based on, DNA testing is proving that people of colour have an incredible mix of trans-African DNA. The history books need some serious updating.

      1. Wow. I enjoyed the dialogue and your explanation @genealogyadventures. I just received my results and when I uploaded the raw dna to gedmatch it showed I had lots more Eastern Bantu and N african, almost 1/2 and more than 1/2 in some cases. I also had the Middle East which was new to me. I was wondering which admixture is best for African Americans. The Ethiohelix K10 + French doesn’t show Central Asia as other admixtures do. It just adda more North African. Interesting….

  2. First, let me say that I greatly admire what you have been doing here. You are definitely giving people some things to think about when they receive their results. Next… I think my ancestry might be even more confusing, even dumbfounding, than what you have shared of yours. I got my Ancestry results a few days ago and saw pretty much what I expected for a presumed “white” girl. There are known Native American branches to my tree, they just don’t show up in me (other than almond-shaped eyes & high cheekbones that look rather silly with my seriously pink skin). My report from Ancestry estimates that I am 47% Great Britain, 38% Europe West, 14% Ireland/Scotland/Wales, and 1% Finland/Northwest Russia. Going on family stories, I could tell anyone what I assumed my genes were, Mom’s family were mostly Irish & French immigrants & biological father’s people were reportedly German. OK, all good. The Finland/Northwest Russia threw me for a sec until I looked it up. I have read in multiple places that a lot of people who have known Native American ancestry see this pop in their reports instead of showing the Native American. OK, I can accept that. Then I find gedmatch. Oh boy, did that ever open a can of worms, lol. On a lark, I ran my raw data through ALL of the filters & services. Whew… Yes, I got tons of hits on the ancient Eurasian Hunter-Gatherer lines, but I also got a lot of things from some unexpected places. Some of them are pretty tiny fragments of dna on a few of my chromosomes, but then there are a few that, in my mind, are more significant. After doing a little research, I can totally understand having more than a little bit of dna from Northern Africa. What I don’t understand, though, is where I got even fragments of dna from not one, but from both of the larger-known Pygmy populations. Yes, both Mbuti and Biaka dna show up in my dna, as well as some (I think) sizable numbers of Khoi-San, Hadza, and Omotic. Well, being from southwest Louisiana, with very, very curly hair, I sort of figured there was at least a drop or two of some African blood in there somewhere from someone not mentioned in the tree for whatever reasons. My husband is having a field day with finding the Pygmy genes. He already calls me “midget” (affectionately, I assure you, and I’m 5’3″), this discovery is just adding fuel to his fire, haha. EthioHelix K10 + French isn’t the only place I got hits on African dna. It is also present in results from EurasiaK9 ASI, EurogenesK13, MDLP K11 Modern, puntDNAL K10 Dodecad V3, and HarappaWorld. So, they can’t all be wrong. I’m going to just take it as, “I have some of the most ancient genes in the world still alive and kicking in me.” I used to joke that I was an American Mutt. I suppose I was telling more truth than I knew. I suppose some of my distant ancestors really got around with their adventures.

    1. Karyn,
      If you really want to go wild on drilling down all your DNA info to the tiniest %, try using nMonte with Gedmatch spreadsheet results.
      None of it costs a time, just a little bit of tearing out your hair until you get the hang of it all.

      On Windows, you will download R, a program to run the nMonte script.
      You can find nMonte via google.
      Use open office if you need spreadsheet creation tool.

      It can take a while to get the hang of it all, but it was so cool to run all my gedmatch results anew with fresh way of calculating more that just 1-4 admix results in the oracles.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: