When I heard that 23andMe had filed for bankruptcy, I wasn’t alarmed, at first. I’d taken a test from the company, years after I did one for AncestryDNA at a genetics meeting – I hadn’t even remembered taking that first test.
I knew 23andMe would someday be in trouble, despite the yearly blitz of ads at holiday time, because of the ephemeral value of their DNA tests – there wasn’t a viable way to keep customers on the hook.
Privacy hadn’t bothered me. If someone wanted to know my SNP (genetic marker) profile, or the fact that I have more Neanderthal DNA than most people, so be it. The company wasn’t sequencing genomes, just identifying a few hundred thousand places in the genome where people vary in the DNA base – A, C, T, or G – a little like a map of state capitols rather than of all streets.
But then in late 2018, the first of what would turn out to be a dozen or more surprise half-siblings contacted me. I told the story of the long-ago sperm donor in this New York Times Modern Love podcast in 2022. Knowledge of my sudden new family had come from that forgotten AncestryDNA cheek swab I’d nonchalantly given at the genetics conference.
When my second half-sister Sara Jane appeared in January, 2019, we three confirmed our initial Ancestry findings with 23andMe. But when I saw Sara Jane’s photo on Facebook, it was like looking at my twin. Photos of us much younger are even more startling – at age 16 we were nearly identical.
So I will always be thankful for consumer DNA testing for this quite unexpected, shocking even, information about my origins. But keeping my DNA marker info in 23andMe’s database is now a liability.
Of Ancestry Pie Charts, Blob Maps, and Yellow Stars
Sara Jane was the first to text me the news that 23andMe had filed for bankruptcy, March 24. She’d already deleted her data when the probable fate of the company was announced about six months ago. She urged me to do it now, stat, not because of anything health-related, but because of the huge pie chart depicting our ancestry.
Unlike the multi-colored pie charts of rich African or European ancestry, ours is solid Ashkenazi Jewish, not depicting religion, but our place of origin. Ancestry.com uses colored blobs (which they call polygons, but they aren’t) superimposed on maps instead of pie charts. But however we looked at our origins, they were right on top of what is now Ukraine: 100 percent Ashkenazi ancestry.
I’d been so dismissive of my genetic markers and Neanderthalism that I’d forgotten all about the ancestry pie chart and blobs. So I deleted my 23andMe data.
I couldn’t help but wonder, might the DNA ancestry charts and maps become the equivalent of the yellow stars that Jewish people in Nazi Germany were forced to wear? In these days of political chaos, anything can happen.
My Curious Personal History with 23andMe
Deleting my DNA data reminded me of the dawn of 23andMe’s direct-to-consumer DNA testing. I was there. In 2012, this DNA Science post looked back at the beginning, and how my view evolved, from shock and concern, to acceptance and, finally, participation. I wrote:
“On a Thursday night in October 2007, I sat with hundreds of geneticists at the American Society of Human Genetics annual meeting in San Diego, so stunned that we ignored the free dessert. At a table in front of the crowd were several very nicely-dressed physicians and genetic counselors representing a trio of companies gearing up to offer, in the coming year, direct-to-consumer (DTC) genetic testing.
Yes, ordinary people would be able to send samples of themselves – spit, it would turn out – to companies that would charge fees to return results right to them, circumventing health care professionals. The companies had names much catchier than those of the biotech companies of the past two decades: 23andme, Navigenics, deCODE Genetics.
We were collectively shocked, and in retrospect, I’m not sure why. But the chatter in the hallways and elevators afterwards was decidedly negative.
What a difference half a decade makes!”
Five years on, I’d become familiar with the accurate, well-written information on 23andMe’s website, and taken many of their “research” questionnaires – some inane, but fun. When my genetic counseling patients showed me their reports from 23andMe, I’d discuss their risks and set up tests for specific conditions with medical testing labs, if warranted.
So now I’m sad to have deleted my data from a company that made so many people more aware of genetics, and brought together some unusual families. We eventually found our biodad thanks to sleuthing on Facebook. But I deleted my data in fear of antisemitism, should my boring ancestry pie chart and Ukrainian blobs fall into nefarious hands.
DNA Ancestry Testing: How the Sausage Gets Made
When covering 23andMe’s bankruptcy filing, the media tended to interview lawyers and other talking heads who don’t know much about biology. And I’ve seen little discussion of exactly how those pie charts and blobs hovering over Ukraine come to be.
The phrase “don’t let us get bogged down with how the sausage gets made” unfortunately applies to understanding how DNA ancestry testing works. So when adding an ancestry (genetic genealogy) chapter to my textbook Human Genetics: Concepts and Applications, I explained it. The approach is logical and brilliant.
A sausage is made of fatty animal flesh, binders, and spices. An ancestry pie chart or geography blob is derived and extrapolated from DNA data on groups of people that know their geographic origins, families that stayed in the same place.
Genetic genealogy traces origins using three data sources: DNA sequences from Y chromosomes, mitochondrial DNA, and markers scattered amongst the 23 chromosome pairs.
Y chromosome DNA sequences trace paternal lineage, and mitochondrial DNA sequences trace maternal lineage. That’s straightforward: only males have Y chromosomes and only females transmit mitochondria (the cell part that controls energy reactions and has its own DNA.)
The third level of genetic genealogy is to superimpose “ethnicity estimates” on geographic origins. The strategy uses shortcuts. Instead of entire genome sequences, ancestry testing harnesses large sets of sites among all the chromosomes where a DNA base – A, T, C, and G – varies among individuals. The variable site is called a SNP (single nucleotide polymorphism). A a profile of these single-base markers is a haplotype. The more closely related two individuals are, the more haplotypes they share.
But how do DNA patterns yield an “ethnicity estimate” like in the ads? “I’m 42 percent Italian, 24 percent Kenyan, 19 percent Danish, and 14 percent Scottish!” Where do the pie charts and blobs come from?
Ethnicity Estimates
SNP patterns in DNA from cheek swabs or spit samples are compared to databases of DNA from individuals who know their family histories back at least five generations to people who lived in the same place and had children only with others in their ethnic group. Traditional genealogy often provides the information on which the comparison populations are derived: documents, photographs, and family lore. It is critical to interpreting DNA findings.
In this way, years ago, AncestryDNA (launched in 2012) combined SNP profiles from a few thousand individuals with reported family origins to establish 26 “reference populations.” That was based mostly on Europeans; the number has grown and is more diverse today.
Now to the tricky part – the math. Consider the 26 populations.
To derive an ethnicity estimate, an algorithm systematically compares a customer’s SNP profile to 25 of the 26 reference populations at a time, omitting a different population at each iteration. If a “left out” population doesn’t alter the comparative sizes of the other 25 pieces of the pie, then the tested DNA profile is not from the omitted population.
Iterations reveal the populations in which the profile is found, and the associated geographical region. Often they jibe with historical evidence of ancient migrations. But how much does each constituent ancestral population contribute to a particular modern genome?
The frequencies of specific SNP patterns in the human population are used to derive the relative proportions of ancestry from around the world – such as 42 percent Italian, 24 percent Kenyan, 19 percent Danish, and 14 percent Scottish. That translates into the relative sizes of pie pieces or blob overlaps.
With my ethnicity estimate of 100% Ashkenazi, I’m something of a purebred. We are all so much alike – very distant cousins – because of our shared history of persecution.
I wrote in The Genomic Scars of Antisemitism: the Ashkenazim have survived an undulation of population bottlenecks, a choreography of hatred that has serially strangled the diversity of our gene pool since our origins in the Levant (Egypt, Cyprus, Iraq, Palestine, Syria, Jordan, Lebanon, and Turkey) during Biblical times. Wherever our ancestors wandered and then expanded, today we all likely descend from a few hundred individuals.
But we don’t need to know our DNA patterns to know why we’re so alike, genetically speaking. Our grandparents told us. My grandmother escaped a pogrom in what is now Ukraine to come here as a girl in 1905. She also escaped the Triangle Shirtwaist Factory Fire in 1911, because she had a cold that Saturday morning and stayed home.
By marrying within our population group, our genomes have remained very similar in sequence.
Why Delete 23andMe Data Now? Because It Can Happen Here
When 23andMe indicated it was headed for likely bankruptcy six months ago, some people advised Ashkenazim to delete their DNA data. But I wasn’t alarmed, until January 20, 2025, when the world turned upside down. Sinclair Lewis chillingly described what is happening now in his 1935 dystopian political novel It Can’t Happen Here. Wikipedia summarizes:
“Set in a fictionalized version of the 1930s United States, it follows an American politician, Berzelius “Buzz” Windrip, … who quickly rises to power to become the country’s first outright dictator (in allusion to Adolf Hitler’s rise to power in Nazi Germany), and Doremus Jessup, a newspaper editor who sees Windrip’s fascist policies for what they are ahead of time and who becomes Windrip’s most ardent critic.”
The US is a different country now, run by hatred and lies, and a frightening return of “othering.” Because I now fear that it CAN happen here, anything, I’ve removed my DNA data from 23andMe.
Shalom.