Twittering Toilets and Phenomic Death Wishes
Unlike the first human genome sequence, which (rather mysteriously) we are able to “complete” every few years it seems, we may never be able to achieve a complete human phenome.
The totality of human physical traits, from the molecular to the behavioral levels, contains just too much information. There are many reasons why setting out to collect a complete human phenome is ill-advised, but even shrinking the goal to a much more manageable-sounding level like 0.01% of a human phenome will still face significant challenges, some economic, some technological, some social-ethical-legal-personal-and-everything-else.
The economic reasons are the most straightforward. Simply put, there are more data points in one human phenome than there are dollars to chase them. The technological reasons are slightly more interesting, boiling down to a lack of cheap tools for gathering good data. No doubt many of the major technological barriers to collecting a good human phenome will acquiesce to our geek prowess. As our technologies get “better, faster, cheaper” our phenomes will become more accessible and transparent. Our ability to collect information will rise, but so will our data quality standards.
Ten years ago, if we wanted to know how many poisons or toxins Bill Moyers had been exposed to over his life, the most technologically innovative and economically feasible solution might have been to ask him to respond to a questionnaire, something along these lines:
- Do you use sunscreen when you go to the beach?
- Have you ever suffered from carbon monoxide poisoning?
- Did you eat paint chips as a kid?
- Do you use pesticides on your lawn?
- How many times have you chemically permed your hair?
- How many times per month do you eat tuna fish?
Today if we want to learn about the environmental exposures Bill Moyer has experienced over his life — many of which he may have no way of knowing even if you asked him — we might supplement (or replace altogether) that questionnaire with a laboratory test on his blood, urine, hair, or other biological samples. In 2003, the “Human Toxome Project” did just that and apparently found quite a number of nasty substances inside Mr. Moyers:
In many ways a “Personal Phenome Project” may only become remotely approachable as technologies become ubiquitous and passive. The keyword here I think is passive. Let the technology do the work, not me. You want to know how many hours of sleep I get per night on average? I would be more open to sharing a datastream from sensors in my bed than keeping a written journal. The latter takes work and I already work too much. Just check the datastream from my computer versus the one from my bed and you’ll understand what I mean.
The ability to turn off datastreams or redact their outputs will likely become the standard prerequisites for sharing. Just like with a questionnaire, which inherently provides me with control over my information by allowing me to skip questions or give dishonest answers, I would want to be able to control the datastream emanating from said bed.
One day your toilet will probably be able to twitter if you let it. Your iPhone may include an environmental monitoring system that continuously collects data on air quality or radiation levels. Your kitchen sink or a filling in your tooth might monitor and report on water quality. Function may give fashion a run for its money when earrings can monitor blood sugar, cholesterol, and stress.
The ethical-legal-social-personal reasons for the impossibility of a complete human phenome are the most interesting. Technology will make information infinitely cheap to collect compared to yesterday. But there will still be secrets and the need to keep them private. And that can be a very good thing. One of the first ten participants in the Personal Genome Project summed this up perfectly the other day when she said that she would much prefer her genome be publicly available than the contents of her personal email.
“Radical transparency” or “baring it all” in the name of scientific or medical discovery has its limits. I suspect that people will adopt “medical personas” that will be quite distinguishable from their other personas, both online and off, e.g. family, friends, professional, Second Life, Facebook, housemates, neighbors, strangers, Match.com, IRS.gov, etc. Each of these personas are in some way “identity silos” that would demonstrate a wide array of preferences for sharing sensitive information if they were compared. These silos serve their own purposes. Setting out to link them all together to gain a single omniscient view of “the data” sounds like a bad idea.
One topic I hope to explore more is the limits of transparency, especially as they relate to phenomes and medical personas. For example, I wonder whether some people might be open to donating sensitive parts of their phenome to science only after death. When I say “sensitive data” I mean traits or behaviors that may create shame or stigmatize people while alive, such as addiction, tobacco use, sexuality, STDs, illicit drug use, cosmetic neurology, psychological traits, etc.
A phenomic death wish does seem a little creepy at first (especially when its called a “phenomic death wish” rather than a final testament or a donation to science). In a similar vein, Thomas Goetz wondered the other day whether genomes are included in the deal when people donate organs to science. An interesting topic to explore more I think. Where there’s a will, there’s a way!