Thursday 3 January 2019

What we learned about fighting bad science by taking on a genetic ancestry testing company

The following blog post was written by David Balding and Debbie Kennett. It is based on an article written in collaboration with Mark Thomas and Adrian Timpson entitled The rise and fall of BritainsDNA: a tale of misleading claims, media manipulation and threats to academic freedompublished in the peer-reviewed journal Genealogy. In just a few weeks the article has achieved the distinction of being the most viewed article in the journal's history. As of today's date it has been seen 3,658 times, and 2,190 people have downloaded a copy of the article. The blog post was originally intended for publication in The Conversation. However, the piece was subsequently rejected because the website's lawyer considered that it was "potentially defamatory in its current state". The Defamation Act of 2013 includes a provision for matters of public interest and provides special privileges for statements published in peer-reviewed journals. We believe that there is a strong public interest in highlighting this story. It is important that academic debate is not stifled by legal threats. There is nothing in the blog post which is not already referenced in our peer-reviewed article. We have therefore published it below in its entirety. 

The worlds of academia and industry are getting closer than ever before. Academic scientists are encouraged to engage directly with industry through consultancy roles, and to commercialise their research through the creation of new enterprises. At the same time, research institutions encourage promotion of resulting new findings to a broad public through the news media.

These trends can lead to conflicts of interest. Media savvy companies can and do attract free coverage for their science-related business under the guise of a public interest science story. It is possible that universities could collude with this deception in their eagerness to attract media attention by allowing a scientist to use the university brand in media presentations, without acknowledging the business motivation.

Our new case study of the former consumer genetic ancestry testing company BritainsDNA, published in the journal Genealogy, sheds light on how conflicts of interests can play out in reality.

Genetic ancestry tests are important tools for genealogists when used in combination with documentary and historical records. Y-chromosome DNA (Y-DNA) tests can be used to trace a man’s paternal ancestry, while mitochondrial DNA (mtDNA) provides information about ancestry on the direct maternal line. There are also autosomal DNA tests (the autosomes are the chromosomes other than the X, Y and mtDNA, and contain most of your DNA) which are useful in finding matches with genetic relatives in a database. Autosomal DNA tests are now the most popular tests. Ancestry testing is a multi-million-dollar industry, and around 18m people have now tested worldwide.

Such tests can be very reliable to reveal ancestry in recent generations. However, once you go beyond about 10 generations back, only a small fraction of the DNA of ancestors will have contributed to a living individual’s DNA. So while there’s a lot of research on human history through DNA, there is little that can be said that is specific to the customer. That means these tests cannot be used on their own to determine exactly where you came from.

The case of BritainsDNA
BritainsDNA was active before the growth of the autosomal DNA databases and focused on Y-DNA and mtDNA testing. They were able to achieve substantial favourable coverage in newspapers, radio and television, with stories drawing questionable links for example between contemporary British people and the Queen of Sheba.

In another promotion, the public service Welsh-language TV channel S4C ran a five-part series called “DNA Cymru” investigating the question of “Who are the Welsh?”. To participate, members of the public were invited to buy a Y-DNA or mtDNA test from the company’s Welsh website.

But the results of this “research” were not published in a scientific journal. Instead viewers were regaled with stories about the ancestry of celebrities, for example, that their Y-DNA or mtDNA results indicated they were ancient Welsh, pioneers or Rhinelanders. Yet these descriptions are so generic that they apply to ancestors of almost anyone: they are essentially meaningless. Y-DNA and mtDNA comprise just 2% of our DNA, and convey very limited information about the history of a nation.

So how could this happen? Two principal actors in the company were a geneticist from The University of Edinburgh and an historian and former television executive who at the time held an unpaid position as Rector of St Andrews University.

The university roles of the company’s directors were used to lend credibility to the promotions. The media outlets did not seek the views of other scientists, who would have contested many of the claims. Few journalists have scientific training, which can allow sensationalised or unbalanced reporting. And while most scientists can be relied upon to be objective, journalists need to be aware that research-related commercial interests can affect scientists’ motivations.

Challenging the claims
We formed part of a small group of concerned scientists who tried to challenge this avalanche of marketing disguised as science. This was prompted by an interview on the prestigious BBC Radio 4 Today programme, which described a “massively subsidised” project to study the DNA of Britons as “bringing the Bible to life”. In fact, the interviewee was there to sell DNA tests and the BBC interviewer turned out to be an old chum. Our challenges were met with legal threats from the company, and resistance to acknowledging editorial failure from the BBC and other media.

Many scientists don’t speak out because of a fear of legal action. We were fortunate to have strong support from the then UCL Provost, and from many colleagues. So we decided to continue to challenge the misleading claims. We were also encouraged by the science writer Simon Singh, who had himself been sued by the British Chiropractic Association for critical comments made in The Guardian. Although the case against Singh was eventually dropped, he suffered years of personal stress and substantial unrecovered legal costs. We also received support from the charity Sense About Science, and we worked with them to prepare the pamphlet Sense about Genetic Ancestry Testing and an article on Sense about Genealogical DNA Testing.

The satirical magazine Private Eye was the only media outlet to see through the company’s misleading media campaign from the start. But eventually we had complaints upheld by the BBC, which also aired a radio documentary that partly corrected previous claims.

Over a period of years we got the upper hand. The company did not pursue its legal threat and eventually went out of business. With a move to genome-wide genetic data, containing more information than is available from Y-DNA and mtDNA, there is reduced scope for fanciful storytelling today. However, there remain problems with ancestry companies failing to reveal limitations of their analyses or to indicate uncertainty in inferences. The population labels that are used are not well defined and can conform to outdated notions of race and identity.

Our story has wider implications about the relationships between business and academia and the reporting of science stories in the media. We hope that our case study will be used to inform media training and education programmes, and that universities monitor the abuse of academic position to advance business interests. Most importantly, we hope that other scientists will be encouraged by our experience and will not be afraid to speak out against bad science.

Further reading
Academics pan Melrose-based DNA business  an overview of our paper from Ewan Lamb on the Not Just Sheep and Rugby blog.
Talking Headlines with Debbie Kennett - My interview with Talking Headlines about our BritainsDNA paper, the lessons learnt and how to detect fake science news.