Genome scans for the whole family: 23andMe research director talks about genotyping her kids

One of the ethical quandaries raised by direct-to-consumer genetic testing is the possibility that customers may send in DNA samples for analysis from other people who haven't provided informed consent - prospective spouses, for instance, a la Gattaca - and then use that genetic information for nefarious purposes. In the past, personal genomics company 23andMe has responded to this possibility by arguing that it's not really the company's responsibility to prevent its customers from performing illegal acts; they have also (quite convincingly) pointed out the difficulties of extracting 2 mL of saliva from an unwilling victim without arousing suspicion.

However, there is one class of potential customers who are legally unable to provide informed consent, but who could usually be convinced quite easily to give up a tube full of spit: customers' children. And in this case 23andMe's view seems to be extremely relaxed - they even have a non-judgmental warning about the difficulty of collecting the required saliva from infants under 3.

In fact, judging from a recent entry on 23andMe's corporate blog The Spittoon, the company is actively encouraging customers to bring their own kids into the happy 23andMe family. This is a bold move, and one that is likely to provoke considerable controversy (and outrage). But rather than shying away from this debate 23andMe seems determined to confront the issues head-on.

The post is by Stanford Assistant Professor Joanna Mountain, who heads a genetic anthropology research group and also serves as senior director of research at 23andMe. In the post Mountain describes the process of registering her two sons (aged 7 and 12) for analysis of several hundred thousand genetic markers at 23andMe - and while she acknowledges the ethical issues associated with analysing children's genetic data, such concerns don't seem to have weighed heavily on her mind:

The question of whether and under what circumstances children should be genotyped has been debated for decades, and this ethical discussion is a fascinating and important one. As a scientist however, I chose to take a pragmatic approach.

Mountain frankly discusses the major issues associated with genotyping her children: the possibility of revealing non-paternity (thankfully not an issue for either of her sons, although she admits to some momentary discomfort as the results came in), and of course the chance that the scan might reveal some seriously nasty disease predisposition.

The latter situation is unlikely for the current version of the 23andMe test, which reports only on common variants with a generally small effect on disease risk, and deliberately excludes the APOE variants associated with Alzheimers disease. However, uncovering serious disease risk will become more of an issue as personal genomics companies move into large-scale gene sequencing (as Navigenics will be later this year, and 23andMe keeps hinting at) - such detailed testing could reveal genetic defects associated with a near-certainty of horrific diseases such as Huntington's.

As you would expect from the excerpt above, Mountain takes a matter-of-fact attitude regarding such a possibility:

And what if we are surprised one day, with news that one or both of the boys is at high genetic risk for a serious disease? I predict that we'll use that information somehow. We'll learn more about the disease, we'll study all the prevention and treatment options. We'll find out about the latest research, and take steps to prevent anyone from feeling that they "caused" this by handing down a mutation. We'll be proactive. It'll be a family project.

There is a genuine ethical quandary here: adults can choose whether or not they wish to know about their genetic risk of a serious disease (particularly one for which there is no effective treatment), but children can't make this decision in an informed sense - and once they have that knowledge, it can never be undone. In a sense, Mountain could be seen as violating her sons' autonomy to remain ignorant of their genetic disease risks.

However, Mountain notes that she only allows her sons to access their data under her supervision - and indeed that it is 23andMe policy that children are not allowed to "own" their own accounts. This approach provides the option for parents to filter out any potentially seriously harmful results until their children are old enough to decide for themselves whether or not they wish to know.

Overall, Mountain argues that the process of investigating her sons' genetic makeup has been a rewarding and educational experience for both them and their parents, and that the benefits outweigh the potential risks and ethical objections. And for what it's worth - and of course I speak as someone without children, and as someone with a pretty laissez-faire view of genetic information - I think her overall argument is convincing. But there's also an interesting story between the lines here that's worth expanding on.

Social engineering and economics in the genome age
Mountain's depiction of the testing and analysis as a positive, interactive process serves as a striking contrast to the visceral revulsion I often encounter regarding the notion of genetic testing of children. Of course, that's presumably precisely the message that 23andMe wants people to hear as they read Mountain's story, because there's a strong commercial incentive for 23andMe and other personal genomics companies to convince customers to bring their family into the fold.

In addition to simply selling more tests, exploring genetic information from multiple family members creates a much richer and more rewarding experience for customers who might otherwise lose interest after a few months of poking around in their own genomes. There's the fun of tracing specific traits through the family tree, and there's also a much larger potential interest pool whenever 23andMe's database of genetic associations is updated: hearing about a new risk variant for schizophrenia might not hold much interest for the average customer - but it might if they can immediately check to see if their crazy uncle Harry (or their moody adolescent son) is a carrier.

This extra long-term value will keep customers coming back for additional tests - large-scale sequencing, then whole genomes, and then epigenomes, and who knows what else. But most importantly, it creates an incentive for customers to recruit new family members, who then recruit other branches of the family tree, ultimately perhaps creating the sort of passive Facebook-style self-marketing product that most companies can only dream about. That keeps the revenue streaming in, and also helps to build up the customer database that 23andMe needs to start carrying out its own genetic association studies.

However, to create such a market 23andMe needs to do some non-trivial social engineering: the company needs to tear down the engrained, visceral opposition of large sections of society towards the notion of both genetic testing of children and of sharing genetic information with others, and replace it with a positive view.

Mountain's post is one small step towards this goal. While the scientist's rather cold rationalisation of her choices won't appeal to everyone, it may well strike a chord with the well-educated, scientifically-minded humanists (the Wired readers, if you will) who have always represented a big chunk of 23andMe's core demographic. The sheer coolness of Mountain tracing the descent of her sons blue eye alleles to her own father is a powerful attraction to such potential customers - if they can be convinced to overcome any taboos that stand in their way.

Of course, this is an easier audience to convince than the rest of society, but it's a start, and there have been hints that the general social environment is becoming more permissive to such attitudes. If the Personal Genome Project succeeds in its goal of demystifying genomics and encouraging the open sharing of genetic information the balance will tip even further in 23andMe's favour.

The dangers
I should emphasise that genotyping your kids isn't always a great idea. The major danger, of course, is that rather than treating the process as a learning exercise some parents might use seriously incomplete genetic information as a basis for pushing their kids down specific life paths, perhaps to their considerable detriment.

This isn't merely an abstract possibility: a Singaporean company, DNA Destiny, is already marketing a "DNA Kids Innate Talent Test". Details are scarce on the company's website, but a recent article in Asia One states that the test "can identify up to 33 genetic traits, such as intelligence, emotional quotient (EQ) and entrepreneurship".

The prospect of parents making long-term decisions about their childrens' careers based on such tests is a worrying one: right now, the genes underlying these traits are essentially completely unknown (with most reported findings in the scientific literature being either unreplicated or subsequently falsified), so the company will be providing a costly, scientific-sounding but ultimately meaningless prediction to parents. To be fair, this is probably not greatly worse than the hundreds of arbitrary ways that well-meaning parents make decisions that influence their child's future - but it's certainly not the direction that I'd like to see personal genomics head.

DNA Destiny is a small and amateurish fish, and there's certainly no evidence that 23andMe or the other major players in personal genomics would ever stoop to this kind of ugliness - they shouldn't be judged on the basis of the behaviour of the bottom-feeders in their industry. However, it will be important for everyone involved in the industry, including regulators, commentators and the companies themselves, to make sure such activities are publicly shamed and marginalised. If 23andMe plans to actively encourage parents to genotype their kids it will be important for them to be leaders in educating their customers - and the broader public - about the dangers of such behaviour.

Edit 30/10/08: Section on the future of genomics removed to avoid distractions from the main theme of the post - to be reposted elsewhere.

Subscribe to Genetic Future.