I'm happy to announce the publication of our paper "Harvard Personal Genome Project: lessons from participatory research" in Genome Medicine, a general retrospective of the current Harvard Personal Genome Project. There is growing interest in participatory research and data sharing (with either participants or researchers), and the Harvard PGP has concrete experience relevant to these conversations. In this paper we share our experiences with the hope of encouraging and assisting others interested in similar research models.

In the rest of the post I’ll give some personal summary, written colloquially. Although many observations seem obvious or trivial, remember that hindsight is 20/20! This wasn’t always what people expected, and that’s why it’s important to share our experiences.

What happens when your project gives participants data access?

One thing that doesn’t happen (at least, not yet): the sky doesn’t fall. Our report explores this question in the form of quantifying the communications we receive from participants. Only one in ten participants that received access to whole genome data through the PGP followed up with us asking any questions about that data – and the majority of those questions were regarding file formats and additional files, not interpretation.1 For a project whose public image is "Personal Genomes!" (and whose participants are extremely engaged and interested in genomes), I think this is a very modest amount of "customer support". This should be reassuring to any project considering providing genetic data access to its participants.

One thing that does happen: when you tell participants you plan to give them data, they wonder where the data is. Over a quarter of our communications are from participants eager to be sampled and receive data. Research timelines are slow. Sometimes analyses fail. It’s really hard to tell someone "Your sample from 14 months ago is still sitting in the freezer".2 Participants are used to medical tests and commercial products which return results in days or weeks. But those tests are fast because commercial/clinical providers have developed standard, polished processes. Researchers have a much fuzzier idea of where they're going – that’s why it’s research! – and things almost never go as smoothly as the researcher imagines they will.3

Unexpected positive consequences of public participatory research

What happens when you have a public project and don’t promise anonymity? Participants talk to each other! They create forums, they write books, they report on their experiences. They can even complain about you – a powerful empowerment of participants.

But researchers, don't shy away from this scary new world where participants become people. There are some significant benefits for your work when participants become part of your community. An engaged cohort wants to be studied! They go out of their way to give you samples. They give you data. They find errors. They perform follow-up investigations and tell you what happened. You get participants who don't merely expect you to study them, they study themselves.

1: Regarding that interpretation – our interpretation process is semi-automatic and creates fairly limited, highly technical summaries (these genome reports are publicly shared on GET-Evidence).

2: Part of why it's hard is because the answers are often embarrassing, e.g. "Alexandra started that project, but she moved to Tokyo and nobody can read her notes". (This is hypothetical! I think the Harvard PGP has done much better than this! But it's not an unusual story.)

3: Researchers are also generally optimistic about timelines. If the researcher says it should take two weeks, expect it to take six. Maybe ten. If you think about it, it's just human nature – to want to do the research in the first place, the researcher must be excited and optimistic about it!