Caught in the genetic social network



Direct to consumer genetic testing is growing rapidly; 23andMe has hired Andy Page to help the company scale – especially since it aims at having one million members by the end at the year (currently, since its launch, 23andMe has tested over 180,000 people around the world). While most ethics discussion about personal genomics has focused on the impact on individuals (is the risk of misunderstanding or bad news so bad that people need to be forced to go via medical gatekeepers or genetics counsellors? is there a risk of ‘genomization’ of everyday health? and so on), the sheer number of tested people and their ability to compare results can result in interesting new ethical problems, as a friend found out.

My friend was first happy to discover that he was in the 99th percentile of Neanderthal genes. He rejoiced in finding lower risk for a bundle of diseases, and read up on the big elevated risk he had. Then he began to combine the data with his genealogical interests. As he remarked, this hobby has been completely transformed in recent years as more and more data come on-line, and people researching ancestors can link up on-line and compare notes. But with genomics data it also means that actual relatedness can be inferred. 23andMe allows “genetic cousins” to find each other. In one case he discovered one “genetic cousin” that helped him figure out the most likely parent of a known illegitimate ancestor: without 23andMe they would not have known the genetic link between their families, but knowing it it was an easy matter to figure out what happened back in the small town in 1855.

Then my friend was contacted by a genetic 3rd or 4th cousin: she knew he was a good “genetic sleuth”, and she wondered about her biological father. It took him less than an hour of searching through Facebook pages, genealogy sites and Spokeo, to figure out who the likely father was, and that he lives not too far away from the lady. At this point my friend had a dilemma: should he tell her? On one hand he didn’t want to reveal information against someone else’s will – maybe the father wanted to be left alone. Who knows what family confrontation might ensue? But wasn’t she morally entitled to that information? Maybe a reunion would be good? And the conclusion was not hard to get, given the right skills and publicly available information. My friend was a disinterested third party, without any personal stake in the situation.

In the case of adoption there exist Reunion Registries, also known as Mutual Consent Registries, where members of an adoption triad can register in the hope of being matched with somebody searching for them. In this case mutual consent is guaranteed. 23andMe are in a sense providing a similar service with allowing people to contact each other based on relatedness. But in this case the father was outside the consent. My friend was more like an intermediary (some states can appoint intermediaries to look into whether adoption-separated people would want to reconnect, preserving privacy if one of the parties refuses), with the power to break anonymity or not.

In the end, after having talked to friends about what he ought to do, he decided to give the cousin part of the package: pointers and information that would allow her (with some work) find the father. He felt he had done his own best approximation of the right thing in this morally ambiguous situation.

The genetic Facebook

In the end, the story shows that it is not just the genetic scan that matters, but the social and informational networks it is embedded in:

\”So who “outed” my cousin’s father? Did I do it? Or was it 23andMe? If both, would that be 23andMe and me? Or was it grandma, who went through the old family bible and recorded it all into a software family tree such as Ancestral Quest, then submitted the file to Ancestry dot com? Or was it Facebook, who gave me the place and date of my cousin’s birth, something she didn’t offer to me personally, but published it to the whole world? Or was it Spokeo, which is really little more than the modern analog of a phonebook except with a loooot more information in it?\”

As more and more people get their genes sequenced, put their data online, and interact with each other, the net will become ever finer. The inferential promiscuity of data is growing exponentially as it becomes more dense and people figure out new ways of using it.

Should 23andMe try to prevent it? By allowing genetically related people to contacting each other on their site 23andMe enables conundrums like the one above. They could abstain from it – but people are also sharing their genetic data outside (just visit OpenSNP). Regulation will not do much except slow things a fraction. One might argue that 23andMe has some responsibility to help counsel people about how they use the information, but wouldn’t this apply equally to Facebook, genealogy databases and libraries?

But if the responsibility is on each individual’s shoulders, then we better learn how to take it properly. In this case we might need to learn very quickly, if 23andMe meets their goals.


One interesting question raised by the old illegitimacy case is how long does privacy last after you’re dead? My friend had a common sense answer:

\”The guy from 1855 is fair game in my book. There is no one left alive who would have had any contact with him when he was alive. Reasoning: if he was about 18 then, and lived to be 90, that would take him to 1927, so anyone who would have had any kind of emotional bond would need to overlap by at least 7 or so yrs, so 1920 to now is 93, nah, all his secrets are fair game, all his base are belong to us.\”

I like the reasoning, although there might be some people leaving emotional bonds and moral obligations reverberating much further.

As I argued in some earlier talks, the future has an unfair advantage over us: it can see us nearly perfectly, while we cannot see it. Our future selves and descendants can invade our privacy in ways we cannot perceive, and quite likely cannot conceive of. Worse, they might have very different values than what we think are right and proper… and maybe they are even right about them. On the other hand, there can exist a kind of reciprocity over time, a “temporal golden rule” if you will: I will respect the values and preferences of past people as I expect to have the future respect mine.

It is not a strong or easily protected principle. But I think it is about as much as we can expect.