Editor’s note 3/17: Harvard Professor Henry Louis Gates wrote about Kasia’s research and “hidden African ancestry” in the The Root. Check it out.

Scientists have long used DNA to inform our understanding of big epochs of human change and migration. But what about the smaller changes, can DNA tell us something about recent human history?

diversity photo 23andMe researcher Katarzyna “Kasia” Bryc, who is also a postdoctoral research fellow in David Reich’s lab at Harvard Medical School, gathered anonymous aggregated data from our customers to look at the mix of African, European and Native American ancestry in the United States. What she discovered was an illuminating genetic portrait of the U.S. that both confirms some of what we know about America’s social history but also other things that are surprising and new.

“Perhaps one of the greatest discoveries stemming from recent advancements in genetic genealogy is that previously held assumptions about race and identity are being brought into question,” Harvard Professor Henry Louis “Skip” Gates Jr. wrote recently. “It was always known throughout American history that there was at least some intermixing of races, but only now, through DNA testing, can we see the extent to which this actually occurred.”

Gates was writing about something we’ve reported on in the past, the “hidden” African ancestry among white Americans.  Because of the ugly history of slavery, most African Americans have at least some European ancestry. But what is less known is how many individuals who identify as being of solely European ancestry – or white – also have African ancestry.  

Note that we use the term “hidden,” because these individuals self-identify as white and because of their very small percentage of African ancestry, may be unaware. But some of these individuals may indeed know about that African ancestry. Kasia’s research confirmed in more detail 23andMe’s earlier findings, and she looked at how that mix of ancestry for self-identified whites, African Americans and Latinos differed from state-to-state. Taken together Kasia’s research helps fill in more detail of how migration, as well as social and racial divisions shaped America.

Hidden African Ancestry Redux

A few years ago, Joanna Mountain, 23andMe Senior Director of Research, presented preliminary findings at an American Society of Human Genetics conference looking at the numbers of individuals who identified as white, but who had what was likely “hidden African ancestry.”African Ancestry for Self-Identified Whites In an update to that work, our researcher Kasia Bryc found that about about 4 percent of whites have at least 1 percent or more African ancestry.

Although it is a relatively small percentage, the percentage indicates that an individual with at least 1 percent African ancestry had an African ancestor within the last six generations, or in the last 200 years. This data also suggests that individuals with mixed parentage at some point were absorbed into the white population. Looking a little more deeply into the data, Kasia also found that the percentage of whites with hidden African ancestry differed significantly from state-to-state. Southern states with the highest African American populations, tended to have the highest percentages of hidden African ancestry.

In South Carolina  at least 13 percent of self-identified whites have 1 percent or more African ancestry, while in Louisiana the number is a little more than 12 percent. In Georgia and Alabama the number is about 9 percent. The differences perhaps point to different social and cultural histories within the south.

The DNA of African Americans

Most African Americans know they have some European ancestry. It’s a vestige of slavery. But what can be surprising is what that mix of ancestry can say about social history in America.African Ancestry in African Americans Previous published studies estimate that on average African Americans had about 82 percent African ancestry and about 18 percent European ancestry. But in self-identified African Americans in 23andMe’s database, Kasia found the average amount of African ancestry was closer to 73 percent.

Kasia found significant differences in state-to-state comparisons. African Americans in the northern and western states have more mixed ancestry than those in the southern states. African Americans living in South Carolina have the highest proportion of African ancestry, about 84 percent, compared to those living in any other state. A significant percentage of African Americans, more than 5 percent, had at least 2 percent Native American ancestry. This is much higher than previous estimates or data from the US Census.

Latinos in the USA

Looking at the mix of European, African and Native American ancestry, self-identified Latinos in the 23andMe database had the most diversity in their ancestry. But again, just as the mix of ancestry among African Americans and whites differed from state-to state, so too did Latinos, which again sheds more light on their social history.Native American Ancestry in Latinos On average Latinos had about 70 percent European ancestry, 14 percent Native American ancestry and 6 percent African ancestry.

The remaining ancestry is difficult to assign because the DNA is either shared by a number of different populations around the world, or because it’s from understudied populations, such as Native Americans. Obviously that large “unassigned” percentage means that those “averages” could be higher. As with African Americans, looking at the regional and state-to-state numbers for self-identified Latinos, the differences are striking. Recent studies have shown that countries across South America, the Caribbean, and Mexico have different profiles of genetic ancestry molded by each populations’ unique history.

According to 23andMe’s data, the United States has its own set of unique regional histories. For example, some Latinos have no discernible  Native American ancestry, while others have as much as 50 percent of the ancestry being Native American. Latinos in states in the Southwest, bordering Mexico – New Mexico, Texas, California and Arizona – have the greatest percentage of Native American ancestry. Latinos in states with the largest proportion of African Americans in their population – South Carolina, Louisiana and Alabama – have the highest percentage of African Ancestry.

Every Picture Tells A Story

America is a crossroads of cultures and our DNA reflects this. Kasia’s research sketches a picture of a few of those cultures but is far from offering the full genetic portrait of the people of the United States. Still her work illustrates the power of DNA – and the 23andMe database – to tell us about ourselves and our history.

*Apologies to Bryan Sykes for our appropriation of the title of his 2012 book, “DNA USA: A Genetic Portrait of America.”  A great read, the book is both a travelogue and a series of short profiles of people with different ancestral backgrounds, who have had their DNA tested.

  • Draillad

    Where can I get the numeric data underlying those maps? Has this research been published somewhere or will it be?

    • Scott23H

      Yes Kasia plans to publish on this and with much more detail. We’ll let our readers know when that happens.

      • Draillad


      • Saul Bravo

        any word about it yet?

        • Scott23H

          Saul, The paper has been submitted. Usually the process takes a few weeks or months depending on the publication. Once we know the timing of publication we will post that on the blog.

  • George McKaig

    It would be beneficial to African American Genetic Genealogy researchers if 23andme & Kasia could develop a chart such as found here: http://www.plosgenetics.org/article/fetchObject.action?uri=info:doi/10.1371/journal.pgen.1000519.g005&representation=PNG_M

    The above chart was developed by Patterson, Reich and others at Harvard. This chart’s methodology could easily be replicated by 23andme researchers to more precisely map for a 23andme African American customer their African American Female Slave Ancestor admixture event with a European American Male Slaveholder Ancestor between 1770 and 1870.

    This above mentioned chart shows a correlation between ancestry proportion and estimated time since admixture in the sample African Americans. Each grey point shows an
    estimate of the time λ since admixture corresponding to one of 935 analysed
    African American individuals.The red line shows sliding averages of 20
    individuals, binned according to increasing African ancestry proportions. The
    peak slavery period from 1800 to 1865 (roughly ~Gen 6 to ~Gen 8) is discernible. This
    investigations of ancestry inference reveals a correlation between time since admixture and ancestry proportion across individuals. The mean estimated time across individuals was 6.62 generations. However, different individuals showed admixture time estimates ranging from 1.25 generations to 13 generations.

    In addition to developing and publishing a chart such as this, it would be beneficial if 23andme could develop for each of their African American customers an admixture time estimate in generations and plot that point on a chart as mentioned above. 23andme spends alot of resources in developing Neanderthal percentage estimates for their majority white customer base… so why not do some things that really would benefit their black African American customer base?

    • Scott23H

      Kasia is indeed able to estimate the time of admixture.
      This blog offers really just some highlights of her research. She hopes to publish on that and offer more details then. Her estimates for admixture regarding African admixture as well as Native American admixture seems to track well with what we know about the history of slavery as well as of migration. Thanks for your comments.

  • John Doe

    I’m African American myself, and have not taken the test yet but would love to. My father’s family is from Texas and my mother’s family is from Florida. I find it quite astonishing that the average African ancestry in the southern states is around 75-80 percent African, which would mean that the average African American in the south would be about 20-25 percent European. I have seen African American’s post their admixture estimates from other companies, and it seems that most of them end up in the 80-90% range and only some of them in the 70-80% range and a few were above 90%. Why is there a huge difference in estimates? My sister did a DNA test from AncestryDNA and she ended up being 88 percent African and 10 percent European which I presume would be quite a low level of admixture relative to the average. By looking at the map, it seems round about the same level of admixture of an average African American from South Carolina, where most of the descendants of the Gullah Islanders migrated to. Among this group, very little racial mixing between Europeans and Africans took place. Whats interesting about my family though, especially on my moms side, is that many of her family members have blatantly non-African features. For example my grandfather had “high yellow” skin, red hair and freckles and looking at pictures of his mother you will see naturally straight black hair and a light skin complexion. There is a family myth that she was half Seminole, but I think its far more likely that she was a mulatto. No one knows. All seven of his brothers and sisters have light skin.The same is true for my grandmother who is dark, but all of her other eight siblings have cocoa brown to light skin. Light skin definitely dominates on my mothers side. On my fathers side its more of a mixed bag.

    • Scott23H

      Hi John,
      Other published studies have given relatively wide ranges in estimates. Some of that has to do with sample bias. We feel fairly confident with our estimates – this is a very large sample size. At the same time, Kasia looked at our own study to see if there was any bias in our own sample that would be skewing the number. She attempted to account for that in these estimates.
      You are right that the Gullah is a special population and there the African ancestry percentages are in the upper 90 percent.

  • Patricia Weiss

    I am not sure why the admixture of White and African American would necessarily go back to slavery time. It is not uncommon for these two races to have children together now and has been common for many generations. I really enjoy my 23andMe membership.

    • Scott23H

      The time element can be determined by looking at the length of different admixed segments in the DNA.
      You are correct that now it is not unusual for people of mixed ancestry but this is showing admixture that occurred several generations back.

  • Scott23H

    I can’t answer that question. AncestryDNA has a different model of estimating ancestry then 23andMe.
    All I can say is that we have a very good and robust system for making the estimates that we make.

    • John Doe

      Oh ok. Last question. Do you have plenty of African samples and how confident are you of those samples, because don’t you have this African program where you’re trying to get Africans from specific countries to take it for free.

      • Scott23H

        We do have a lot of African Americans who are part of 23andMe. And we are confident in the samples. Although it’s not discussed in detail in the blog, Kasia did account for bias in our sample. (We have a lot of customers from California for instance.)
        We also have a good number of African American customers, although we’d like to attract more. This helps in many ways, but probably most importantly with research around conditions that are specific to African Americans. There is a big need in diversity in genetic research of disease.
        The African program is something different. That was an effort to attract people who are from Africa to 23andMe. That will improve the reference populations and hopefully give finer details about where in Africa people trace their African ancestry.

        • John Doe

          So Kasia found that the average African admixture in African Americans is only 73 percent. That is an unusually low number compared to others I’ve seen. This would equate to the average european admixture hovering around 25 percent which seems quite high. I As you said in the article, post studies have shown numbers around 82 percent with euro admixture around 18 percent. Just by looking at the map as well, you can see that the average Average ancestry in the south (except south Carolina), which has the highest concentration of blacks is below 80 percent. Since many customers come from California do you think it skews the average a bit?

          • Scott23H

            Kasia looked specifically at whether the geographic location of our customer base was skewing the numbers. She attempted to account for that possibility in making the calculations and feels quite confident that it is a good representation of the percentages. We have a very large sample size that is still very geographically diverse, unlike many of the studies in the literature. At the same time there’s more work that she is doing before publishing on these findings. Thanks for the comments.

          • Sharon37

            Oh my goodness, what is your obsession with white adnpmixture in African Americans. Just fork over the money for the test and find out. Also, no two family member has the same European percentage. Be proud of who you are. These DNA tests are changing how we see ourselves but it shouldn’t change how we value our heritage in the first place. We are still humans just more closely related than previously thought.

  • KULL612

    Do you know if Kasia is going to give more attention to the African American and Native American mixture when she publishes her report?
    In order to estimate the timeline for instance African American and Native American mixture, does she used the longest segments from a parent’s contribution or does she adds all of them together and come up with an estimate? Or does she use the smallest segment for her determinations?
    I had both parents tested also, and I get at least 5 chromosomes that are contributed to me from my mother. Thank You

    • Scott23H

      Kasia will be looking at learning about the timeframe of admixture using publicly available methods that use information from *many* individuals to learn about the population course of mixture. She is not looking at individual level data.

  • George McKaig

    In regards to the 23andme interpertation about “hidden” African ancestry admixture events in self identified white European Americans taking place “solely in the USA” and “within the past 6 generations (180 years)”, I think this section of the 23andme blog needs further clarification and additional research: “Although it is a relatively small percentage, the percentage indicates that an individual with at least 1 percent African ancestry had an African ancestor within the last six generations, or in the last 200 years.”

    I disagree and affirm that those European Americans with .5% to 2.0% of African Ancestry “could quite possibly” had an ancestor(s) with African ancestry before they reached the shores of America. In other words, those admixture events “could quite possibly” have occurred in Southern Europe, North Africa, the Middle East, etc. rather than just solely within in the USA. For further information see: http://www.plosgenetics.org/article/info%3Adoi%2F10.1371%2Fjournal.pgen.1001373

    Facts are that 1 – 3% of African admixture events observed throughout Southern Europe has been dated to around 55 generations ago (~1600 years ago) consistent with the historically attested African slave trade practiced by the Roman Empire. Roman occupation of North Africa lasted until the 5th century AD and many of these admixed individuals migrated into Southern Europe. There is about 2.7% African admixture in Southern Italy versus only about 1.1% in Northern Italy.

    As noted here: http://www.plosgenetics.org/article/fetchObject.action?uri=info:doi/10.1371/journal.pgen.1001373.t002&representation=PNG_M, I believe that 23andme researchers have to consider “other African and European admixture events in Southern Europe and Northern Africa…. and not just the ones that took place in the USA”

    By doing so, the reader, and especially those European Americans with .5% to 2.0% of African Ancestry, could have a better perspective of what happened (and where) in their ancestry and the example ancestry of Natalie Baker, a white European American, who has about 1.5% African Ancestry. Her case was wrritten about by Skip Gates and quoted in this 23andme blog. http://www.theroot.com/articles/history/2014/02/traces_of_african_ancestry_in_a_white_person_the_likely_reason.html

    • Scott23H

      Thanks for the note. Without going into detail, I’m going to disagree with you. Kasia will get into some detail when she writes up her research regarding the admixture timetable. We are quite confident that we are not detecting African admixture from as far back as you suggest.

      • Chris Gray

        I agree because my 2% Native American is almost exactly 200 years old AND supported by a paper trail. Thing is that the comment said “I disagree and affirm that those European Americans with .5% to 2.0% of African Ancestry “could quite possibly” had an ancestor(s) with African ancestry before they reached the shores of America.” and most of the mixing actually happened in Colonial Va. even earlier. Much earlier

    • hernandayoleary

      There are two types of testing, autosomal and allosomal/ (y and mtdna), they seem to be relying on the one that only goes back a few hundred years–autosomal, not the same one that you are relying on which uses mtdna dn y dna and draws population projections. Dna says these are recent admixtures, at least since colonialism, they do not detect dna from 1600 years in autosomal testing. While its true some italians and southern europeans have mixture, it wasn’t happening in 1600

  • Chris Gray

    I notice the African DNA is higher for Aframs on ancestry.com. I did their test, but I have YET to get results after 5 weeks. I hear its a common complaint along with not being able to download or view your own data

  • Scott23H

    Hi Randall, Thanks for the note. Kasia will be submitting her paper for publication soon and it will have much more extensive data in it.

  • hlritter

    I have a primarily German ancestry going back 6 to 10 generations in the New World (depending on which family account, if either, is correct). All generations prior to mine were born in Bucks County, PA. Another presumably 1/8th comes from a Scottish GGF. There is no known Afro-American contribution. I just joined 23andMe and was intrigued to find that I have just over 4% sub-Saharan African DNA. For no particularly cogent reason, I feel pleased with this, and I’m curious if your results are mature enough to tell me what percentile of self-identified European Americans have a percentage this high.

  • Susan White

    I really want to do the genetic testing but haven’t had the funds. I self identify as Native American and have Caucasian ancestors as well. I find the research fascinating and would welcome any ethnicity in my family tree. I also have Ehlers Danlos syndrome, Mast cell activation issues and other issues. My doctor has told me I have “a lot of autoimmine issues”. The medical part of the genetic testing would be interesting for me also, especially since I suspect the MTHFR gene in my family. I have met several people with mixed racial ancestry who also have EDS, mast cell issues and or mitochondrial disease. It would be interesting for me to see if any of these health genes are related to particular groups but it doesn’t seem to be the case for EDS, MCAD etc. Thank you for the post.