23andMe Says It's Not (Now) in the Designer Baby Business

Posted by Pete Shanks October 4, 2013
Biopolitical Times
Part of the flowchart<br>in the patent application

On September 24, the direct-to-consumer genetic testing company 23andMe was awarded a patent for "gamete donor selection based on genetic calculations." The idea is that people using sperm or egg providers would choose from the gametes on offer based on the probability of getting the kind of child they want. The examples given in the application of the traits of hypothetical children that might be of interest include:

Height, eye color, gender, personality characteristics and risk of developing certain types of cancer

There is a helpful flowchart (part shown here) to explain how the technology would be used to evaluate the possible matches between the client (recipient) and the available donors, and so to identify preferred donors, given that it would be dealing with statistical information rather than certainties.

That, of course, points up the biggest flaw in the process. Even a direct-to-consumer gene-testing company is compelled to acknowledge that there can be no guarantees. But they don't make clear that for the vast majority of "traits of interest," the associations between particular alleles and particular phenotypes are actually rather weak. (A few conditions, including single-gene ailments and some inherited forms of cancer, can be identified and avoided through pre-implantation genetic diagnosis or carrier screening, but they are rare.) The associations were weak four years ago, when the patent was filed, and they still are.

But the company is, after all, in the genetic prediction business, and the patent in question referred to a "hypothetical child." So it's hardly surprising that when news of this patent broke, most journalists made the same call. Nature News and MIT Technology Review had "designer babies" in the headline. Wired called it a "designer baby system." The Los Angeles Times labeled it a "designer baby patent." New Scientist referred to "planning a baby."

Perhaps the most thoughtful immediate response was published in Genetics in Medicine just before the media got hold of the story. Four European bioethicists (Sigrid Sterckx, Julian Cockbain, Heidi C. Howard and Pascal Borry) considered the patent under the title:

"I prefer a child with …": designer babies, another controversial patent in the arena of direct-to-consumer genomics

It's a short paper in which the authors largely restrict themselves to raising questions and noting controversy. They point out, for instance, that the U.S. Patent Office does not seem to have questioned whether such techniques "were appropriate subject matter for a patent" although it did invoke "morality" when considering the human/animal chimera patent application made by Stuart Newman and Jeremy Rifkin. (That was eventually rejected as too human to be patentable.) And the authors ask whether 23andMe's customers would be disconcerted by this use of their collected data; there was an outcry last year when the company was granted a patent for a test for Parkinson's disease.

As soon as the news stories began to appear and 23andMe realized that people were drawing the obvious conclusion, it began backtracking hard. It now claims that the patent relates to "one of the tools we offer individuals as part of their genetic exploration [which] offers an engaging way for you and your partner to see what kind of traits your child might inherit from you." Nothing there about choosing donors. The company's blog explains:

At the time 23andMe filed the patent, there was consideration that the technology could have potential applications for fertility clinics so language specific to the fertility treatment process was included in the patent.

To say "included" is disingenuous; the process description in the patent is entirely about the selection of gametes from strangers. The post continues:

But much has evolved in that time, including 23andMe's strategic focus. The company never pursued the concepts discussed in the patent beyond our Family Traits Inheritance Calculator, nor do we have any plans to do so.

So that's all right then. But wait, there is a kicker in that same statement:

23andMe believes that patents should not be used to prevent individuals from accessing their genetic data or its interpretation.

So would the company license that patent? It doesn't say. The Center for Genetics and Society has called on 23andMe to use its patent to prevent others from exploiting this technology. That's not unheard-of: the Newman/Rifkin application was explicitly intended to prevent development of chimeras. Thus far, 23andMe has not explicitly ruled out its future use.

Many of those who commented on the 23andMe patent, including CGS, law professor Lori Andrews, and geneticist Daniel MacArthur, expressed doubt about the practical feasibility of most trait selection. But even if scientifically dubious, the project of "breeding better babies" is, obviously, morally and socially fraught.

A few philosophers, such as Julian Savulescu, actively campaign for it as a moral imperative. In opposition stand CGS and many others. Harvard's Michael Sandel put it well when he said that trait selection "edges close to eugenics." This debate has up till now been theoretical. It's getting all too real.

Previously on Biopolitical Times: