Harvard Personal Genome Project: lessons from participatory public research

Abstract

Background: Since its initiation in 2005, the Harvard Personal Genome Project has enrolled thousands of volunteers interested in publicly sharing their genome, health and trait data. Because these data are highly identifiable, we use an 'open consent' framework that purposefully excludes promises about privacy and requires participants to demonstrate comprehension prior to enrollment.

Discussion: Our model of non-anonymous, public genomes has led us to a highly participatory model of researcher-participant communication and interaction. The participants, who are highly committed volunteers, self-pursue and donate research-relevant datasets, and are actively engaged in conversations with both our staff and other Personal Genome Project participants. We have quantitatively assessed these communications and donations, and report our experiences with returning research-grade whole genome data to participants. We also observe some of the community growth and discussion that has occurred related to our project.

Summary: We find that public non-anonymous data is valuable and leads to a participatory research model, which we encourage others to consider. The implementation of this model is greatly facilitated by web-based tools and methods and participant education. Project results are long-term proactive participant involvement and the growth of a community that benefits both researchers and participants.

Figure 1

Status of 2,294 accounts created June 2012 to December 2013. About half of all accounts created on our site eventually complete the enrollment process to become participants (1,143 users, 50% of all accounts). Of the 1,151 accounts that did not complete the enrollment process, the majority (674 users, or 59% of incomplete enrollments) stopped at the enrollment examination stage.

Figure 2

Participant-initiated communications. Our website offers participants a ‘Contact Us’ button. From June 2012 to December 2013, we received 579 emails from participants. Few support requests derived from the return of genome data (3.6% of all emails, representing 11% of participants receiving genome data). Most of these were inquiries regarding data formats and additional files for the participants’ own analyses of their data (2.9%), rather than requests for additional interpretation on our part (0.7%).