The future of participant-driven genomic research

Over at the 23andMe blog The Spittoon, company co-founder Linda Avey expands on her vision for a novel model of genomic research, in which personal genomics customers contribute their genetic and health data to fuel research into the inherited and environmental triggers for disease.

This is a model that 23andMe has been building towards for a long time. In May last year the company launched 23andWe, a cutely-named effort to obtain detailed health and trait data from their existing customers through online surveys which could then be combined with genetic data to find novel gene-trait associations.

In today's post Avey formally announces a drive to recruit 10,000 Parkinson's disease patients for a targeted analysis of the genetic and environmental causes of this disease, which is backed by the Michael J. Fox Foundation and The Parkinson's Institute, and will apparently be largely funded by Google co-founder Sergey Brin.

The model is simple: Parkinson's patients get access to the full features of a 23andMe genome scan for $25 (compared to the standard retail price of $400), in return for participation in online surveys. The company will then combine their recruits' genetic and survey data to look for unexpected associations with disease status, using volunteers from among their existing customers as controls.

So, will 23andMe uncover new genes for Parkinson's disease?

It's certainly possible; previous genome-wide association studies
(GWAS) of Parkinson's have been seriously under-powered (just a few hundred
cases and controls), so there's every chance that there are still low-hanging
common risk variants that could be captured with a sample of 10,000 cases. Building an active community of participants may also provide substantial advantages in terms of making it easier to explore environmental risk factors and to chase down promising leads with further rounds of targeted surveys.

However, there are also caveats: the self-reported disease status and health data from
participants will never be as rigorous as information collected in a clinical
setting; and it may well turn out that the bulk of the genetic risk for
Parkinson's is due to rare genetic variants that will be
essentially invisible to 23andMe's common variant-focused chip.

All in all, though, my guess is that this study will contribute at least something of value to our understanding of the etiology of Parkinson's. But this will perhaps prove to be less important than the precedent that this study sets for future genomic research: perhaps I'm just a sucker for 23andMe's hype, but I think the model of participant-driven research is a novel and powerful one, that will increasingly come to dominate the genomic research landscape.

23andMe has unique advantages in a study of this type: it provides an incentive for patients to enroll in the study (access to information about their ancestry and genetic disease risk); motivation for enrollees to continue to participate (as new features and updated association data are added); and a slick interface and experienced team for returning data back to participants. Few academic consortiums can match this set-up; yet it seems inevitable that participants in genomic research will increasingly demand some return (in terms of genetic information) for their participation in large-scale studies.

This really hit home to me during a presentation by the NHGRI's Len Biesecker at the AGBT meeting last month. Biesecker presented on the ongoing ClinSeq project, during which he noted the ethical and logistical challenges of data return for academic genomics consortiums. For genomic researchers to generate genome-wide genetic data - with health implications well outside the current study - and then not return it to participants seems frankly unethical, but few if any genomics consortiums are in any way equipped to return whole-genome data to participants in a useful format.

I immediately wondered: why don't genomics consortiums simply take advantage of the databases and graphical interfaces already generated by personal genomics companies? Basically, once a study is complete the researchers can send anonymised data to a personal genomics company and provide each participant with a unique, anonymous login code; then the research subjects can access their complete genomic data through the company's interface.

This seems like a winning solution all round: academic scientists can fulfil their ethical obligations in terms of data return without attempting to build their own complex interfaces; research participants get their data back in a useful format; and personal genomics companies get an additional, stable stream of revenue for little more than the cost of some data storage and bandwidth.

So, will we see future funding proposals for disease genomics projects include an extra allocation to pay a personal genomics company to return genomic data to participants? I suspect it will take a while before academic researchers overcome their (quite understandable) suspicion of the consumer genomics industry; but the combination of the ethical imperative of data release and the logistical challenges of building their own data release infrastructure may well force their hand.

Subscribe to Genetic Future.