Research into the genetic causes of disease have been overwhelmingly done only on majority-European populations. For example, a 2009 review found that just 4% of the participants in published genome-wide association studies had non-European ancestry. 23AndMe (which I was unaware about until yesterday) is trying to change the game. They currently only have 1,000 African-American samples in their database(or 1.2% of their total database), even though black Americans are 13% of the U.S. population. As a comparison, they have 56,000 samples of mostly Northern European ancestry, 3,500 Hispanic samples, and 3,400 South Asian samples). Part of this gap is cost ($400 for a test). However, many (most?) African Americans who get such tests instead use African Ancestry’s service (which has 25,000+ samples). People also probably don’t know about 23AndMe’s services either.
As Booker Rising mentioned yesterday, 23AndMe is teaming up with Professor Henry Louis Gates, Dr. Rick Kittles (from PBS’ “African American Lives” series) and Harvard University in order to test 10,000 African Americans and thus dramatically boost the diversity of their genomic database. Of course, there is already controversy. Some folks argue how to address fraud, as lily-white folks may try to secure a free test. As Razib Khan points out, African American genetic admixture is pretty clear, with Africans/pure blacks and European/lily whites sticking out like sore thumbs. Of course, there are also white folks complaining about black folks getting a free test. My response: (1) African-American genomic data is of higher value because far less is known about it; and (2) there’s far more black resistance to scientific research due to history (e.g., the infamous Tuskegee Institute study). Thus, incentive offer (free test to learn about your ancestry and health genome) in order to secure data.
The health data sounds iffy to me, perhaps because it’s only been done on white folks. While there will be some overlap with African-Americans, I’d take the health info.
What I find interesting is the ancestry painting feature and Relative Finder service. With ancestry painting, they look at your 22 bi-parentlly inherited chromosome pairs (the 23rd one determines your gender) one segment at a time — there are 1 million+ markers — and determines for each stretch whether it was most likely inherited from Africa, Europe (Eurasia), or Asia.
Erica Baker, an African-American tech professional and blogger in California, has written about her experience (disclosure: Erica’s employer is an investor in 23AndMe, but she paid for her test on her own). Above is her ancestry painting. Green is for African, blue is Eurasian (called only “European” here, even though it encompasses folks ranging from South Asians to Iraqis to Swedes), and orange is Asian/Native American (called only “Asian” here).
There has been some controversy about the ancestry painting because the reference population for the African test are all Nigerian Yorubas and the reference population for the “European” test are Utah whites. East Africans who have no significant Bantu admixture (e.g., Somalis, Eritreans, Ethiopians) are showing up as much more “European” than they actually are because the ancestry painting algorithm doesn’t track their East African ancestry. In addition, the “European” contribution that they do have comes from West Asia (more specifically, Arabia) or North Africa, and not the Northern European sector found in Utah whites. 23AndMe is gonna need to add some East African and southern African populations as reference points.
They also have something called a Relative Finder. At left is Ms. Baker’s Relative Finder results. Some 26 people share DNA segments with her, five of which are featured at left who are predicted to be her fourth cousins. This looks like a very valuable tool for African Americans in getting names to those anonymous non-black ancestors that we’ve all got in our family tree by connecting family genealogy records and locations. Erica agrees, calling it a “complete game changer” for African American genealogy.
Now, some of my readers are already saying that genomic research is a conspiracy to wipe out the black race. If educated black folks believe such, then 23AndMe has its job cut out for it in mapping out the genomes of racially mixed populations such as African Americans which span two or even three continents (versus usually one for the vast majority of the white folks they’ve been testing). Not to mention the fact that 93% of all human genetic diversity lies in Africa alone. It writes: “Note that this project poses challenges beyond our previous studies that were conducted in subsets of the 23andMe membership with European ancestry:
- Because of the time depth of human history in Africa, genetic diversity is higher in African populations and in groups that trace much of their ancestry to Africa. Because most genotyping chips were developed based on information about non-African genetic variation, these chips do not capture as much of the genetic variation in African Americans as they could. We will begin by using the current 23andMe v3 chip. We may decide at some point to conduct additional genetic analyses; data from the current chip may help us determine who to follow up with for any additional analysis.
- Most African Americans trace ancestry to multiple continents, which poses particular complications for genome wide association studies. We will use approaches such as admixture mapping, which takes advantage of this pattern, but we are also excited to leverage our strengths in statistical genetics to find novel ways to tackle this challenge.