Generic image

Dear Professor Gates:

According to 23andMe, my brother and I each have 0.1 percent sub-Saharan African ancestry. But instead of being at the identical spot on the chromosome, his segment and my segment are next to each other. How would we estimate the date that the African-American ancestor entered our family tree? Would we combine those two segments to get 0.2 percent and use that to estimate recency? We think we know which line of our family this relationship comes through, but we do not know how far back it would be. —Dorothy Estelle


To answer your questions, we turned to 23andMe population geneticist Katarzyna “Kasia” Bryc, Ph.D., who not only read your questions but also reviewed the test results for you and your brother that you sent her. Her response follows:

So there are two questions here. The first is whether the African ancestry is real, and the second is how you would estimate how many generations separate you from your African ancestor based on the data.

Do These Results Truly Point to Recent African Ancestry?

Unfortunately, there is no general rule to guide the interpretation of your results. One needs to take into account the ancestries that are present in the individual, how differentiated the ancestry is from the background, whether the signal might be stemming from shared population history and whether the estimate still holds if different confidence thresholds are tried.


In the case of you and your brother, both of you have almost entirely Northern European ancestry with just that tiny bit of West African, and your brother’s African DNA is still estimated as West African even under the “conservative” confidence threshold of 90 percent. Typically, European and West African DNA are genetically quite distinct. Our recent 23andMe paper in the American Journal of Human Genetics looked at self-reported European Americans who carry low levels of African ancestry (1-2 percent), and showed several lines of evidence validating that our estimates were indeed picking up on recent African admixture.

However, in your case you only have 0.1 percent African ancestry and only one short segment, which is not much! That your brother likewise carries some African ancestry argues that our estimates are picking up on some true recent ancestry, but it’s not the same segment of DNA. If both segments came from one parent carrying one long African segment, then for both you and your brother to inherit those two African segments, together with your particular IBD (identical by descent) matching on that chromosome, might require as many as four recombinations—a recombination being the exchange of genetic material between chromosomes that happens in each generation. If each of your parents carried only one shorter segment, it would require only two recombinations, which is more likely the case on such a short chromosome. But two very short segments are more likely to be picking up on some genetic pattern by chance, rather than necessarily reflecting a recent ancestor from Africa. Hmmm.

So in summary, I’m not sure if those segments are indicating a recent African ancestor. I would suggest that you look for African ancestry in other close relatives to really get a sense of whether the reports that you and your brother received are indeed picking up on an ancestor from Africa.

How Does One Count the Generations?

So if we assume that the ancestry assignments are correct, then next comes the question of when you and your brother might have last had an African ancestor. On average, people inherit about half their DNA from each parent, a quarter from each grandparent and so forth. Based on that calculation, you might have had an ancestor about 10 generations ago. However, the challenge is that there is a lot of randomness in what you inherit from a particular ancestor, so 0.1 percent could come from an ancestor anywhere from seven generations or many more generations ago!

For an explanation of the randomness of what you inherit from an ancestor, I like to point people to a nice blog post from the Coop Lab for Population and Evolutionary Genetics at the University of California, Davis, looking at how many genetic ancestors a person has.

The reason that we at 23andMe report on those 0.1 percent results is that our algorithm does a very good job estimating ancestry for each small window of the genome, and we feel that having the full results, without censoring, allows people to get the most information—though, of course, interpretation of the results can be hard.

Henry Louis Gates Jr. is the Alphonse Fletcher University Professor and founding director of the Hutchins Center for African and African American Research at Harvard University. He is also chairman of The Root. Follow him on Twitter and Facebook. Kasia Bryc is a population geneticist from 23andMe.


Send your questions about tracing your own roots to