Ancestry.com has finally updated their corporate page to say that their DNA database is “over 10 million people”. Until this Labor Day weekend (Sep 3), the page said “almost 10 million”. Based on past growth—they added 2 million testers between February and April 2018—they should have crossed the 10 million threshold back in May or June. Why they haven’t promoted such a major milestone is anyone’s guess.
Regardless, I’m happy for any excuse to update the graph of autosomal DNA database sizes. “Over 10 million” is a geek’s nightmare of imprecision; it could mean anything from 10,000,001 to the entire human population. For the graph, I used 10 million for AncestryDNA, even though I’m sure that’s an underestimate. I also got an updated number from GEDmatch.
Based on past trajectories, there been an apparent slow-down at AncestryDNA. If I extrapolate the growth rate from the previous two AncestryDNA reports (from February and April) to today, they should be at almost 12.8 million. Perhaps their database really is nearly 13 million and their description of “over 10 million” is too modest. Alternately, they may have experienced a real decrease in growth rate. Without inside information, there’s no way to tell which is the case at the moment.
GEDmatch does seem to have experienced a real slow-down, however. If the growth from February to May had continued, they would have about 1.2 million people today. The official value is 1,080,000, or 10% less than expected. Two possible explanations are that (1) the slow-down at AncestryDNA is real and is having downstream effects on GEDmatch, or (2) people are shying away from GEDmatch now that they know law enforcement is using the database for criminal investigations. Of course, both possibilities may be at play or even other factors I haven’t considered.
I did similar calculations for each of the major databases and summarized the results in this table:
|Company||Estimated size||Official size|
|AncestryDNA||12,777,778||> 10 million (Sep 2018)|
|23andMe||8,468,239||> 5 million (Feb 2018)|
|MyHeritage||1,739,130||1,400,000 (May 2018)|
|GEDmatch||1,203,292||1,080,000 (Sep 2018)|
It will be interesting to see whether these projections are accurate.