So many genes, so close to home

On occasion, I like to feign refinement by butchering great literature. I stole the title of this comment piece from a Raymond Carver story that deals with themes of identity, isolation and disconnection. I stole that description from the dust-jacket.

I console myself in the knowledge that 'great artists steal' (that one's from Picasso although, rather brilliantly, may be misattributed), but what about scientists? The worth of a scientist is measured in ideas. It's important to be first and to be original.

So I wonder what will come from an accusation of plagiarism flagged up over a recent study. The study in question describes a method of tracing your DNA back through time, up to a millennium ago, to find a single home from which your ancestors came (see BioNews 752). The researchers called it Geographic Population Structure, or GPS for short.

Genealogist and blogger Dienekes Pontikos (a pseudonym) wrote that, back in 2011, he invented some of the methods on which GPS is based. This accusation calls into question the novelty - and hence, scientific merit - of GPS. Dr Eran Elhaik, co-creator of GPS, defended himself in the blog post's comments, claiming his team independently invented the method.

Some commenters also linked Dr Elhaik and his co-author Dr Tatiana Tatarinova to Prosapia Genetics, a private company selling the GPS test. Their suspicions were prompted by the fact that Prosapia was incorporated one day before the GPS study was published. When I called him for comment Dr Elhaik said that neither he nor Dr Tatarinova are affiliated with Prosapia and have received no royalties for their work. He told me: 'We made no money out of GPS'.

Fair enough. I don't want to spend this article raking over internet rumours, so let's look at the science.

Dr Elhaik told me his work 'challenges some of the basic notions of our scientific thinking'. The current view, he says, is that modern humans evolved from three or more 'races' (his word). I think he means sub-species, probably including Neanderthals and Denisovans. According to him, his work counters the view of this 'racial model' (his words) that historical populations remained isolated from one another. Just a cursory flitting through the literature shows that many models of genetic ancestry account for events like gene flow and migration, so I'm not sure about this argument.

The study prompted wild headlines amid deep sighs from those who know better. Professor Mark Thomas, an evolutionary geneticist at University College London, told BioNews that GPS is 'interesting but crude'. He explained: 'The method is probably better described as "human provenancing" since it is aimed at finding the location where a person's genotype is most likely to occur'.

GPS matches your DNA to one or more clusters of reference genomes held in its database, averaging between the known locations of each cluster. For example, if one of your parents comes from the UK and the other from China, GPS might suggest your ancestral home lies halfway between, near Iraq. Some people who paid for tests from Prosapia Genetics were disappointed that GPS placed them in the middle of oceans, with a few seeking refunds.

Dr Elhaik said: 'People who are surprised to find themselves in the middle of the ocean or large water reservoirs need to discuss these results with their parents'. I'm not sure how that helps.

There is also a deeper issue with GPS. The further you look back through time, the larger your dynasty. Going back one millennium, assuming a generation lasts 30 years, you have 34.4 billion forebears. (The real number will be much lower because of accidental - or non-accidental - inbreeding.)

To make matters more convoluted, your number of genetic ancestors - people from whom you inherited DNA - is much smaller than the number of genealogical ancestors. Dr Graham Coop, a geneticist at the University of California, Davis, estimated that each of us inherits only around 0.00001 percent of our genomes from any single predecessor living 1,000 years ago.

All these ancestors probably didn't dwell in the same place for all that time, either. This means very few people will have come from a single ancestral village. Regardless, whether GPS's predictions has any bearing on geography is up for debate - perhaps genetic clusters are just not geographically meaningful.

This saga is a battle of ideologies and conceptual understandings of history. It's about the use (and misuse) of statistics and data, clouded by complicated methods, to give easy answers to potentially unanswerable questions.

Then again, given how over-hyped studies can seduce an uncritical media, is it really so surprising that we are willing to fall for simple stories about our past?