This week for BioNews, I report on a genetic astrology ancestry test that claims to “tell you where your DNA was forged, and is accurate to home village with a time resolution of the past 1,000 years.” Researchers named the test Geographic Population Structure, or GPS, presumably to convey a sense of satnav-esque accuracy.
The method was published to a fanfare of overblown press releases and uncritical media coverage.
“What we have discovered,” spun one of the press releases, “is a way to find not where you were born – as you have that information on your passport – but where your DNA was formed up to 1,000 years ago by modeling these admixture processes.”
At the same time, the researchers founded a company called Prosapia Genetics which tells you your supposed ancestral homes in exchange for money.
But it doesn’t work so well. Some customers found their ancestral genomes stuck slap-bang in an ocean:
This is because the tool averages between the locations of genomes for which it has geographical data. Apparently it’s meant to work this way. Dr Eran Elhaik – one of the study’s co-authors – wrote on Prosapia’s forum: “when you have British and Chinese parents you will be predicted to Iraq, as it is in the middle of the two gene pools.”
Many in the know have dismantled the science far better than I ever could. Debbie Kennett writes a wonderfully comprehensive account. Pseudonymous blogger Dienekes Pontikos explains some of the wrongness in the tool amid hints of plagiarism. And Joe Pickrell – who peer-reviewed the study – posted a summary of his problems with the paper.
According to the Prosapia website, the scientists are developing an extended tool – GPS2 – that will apparently show two parent groups that formed your DNA, rather than just one. (If you trace back 10 generations – around 300 years, for sake of argument – you have up to 1,024 genealogical ancestors but you contain DNA from only maybe 120 of them. See Luke Jostins’ blog for an explanation.)
Anyhow, I’m writing a longer comment piece for BioNews that will be out next Monday and will go into more detail there.
In the meantime, have a read: http://www.bionews.org.uk/page_416034.asp