the data analytics associated with all those technologies along the way. In fact we think the project can be the impetus for the development of a lot of new technologies.
Lee has talked about this before: We want a chip that can make measurements of 50 proteins from each of, say, 50 major organ systems in the human body. So you’d have 2500 measurements of protein levels with just a little microfluidic device, just out of a droplet of blood. There are already companies in that space of small-volume blood measurements. Theranos is a good example, they have an agreement with Walgreen’s.
Also, wearable devices. I wear [one] that monitors calories, steps, heart rate, sleep—deep sleep and REM sleep—and perspiration. There’s a lot in wearable technology that we’re adopting. It would be huge if we can get something to measure calorie intake and, even grossly, the percentage of sugars versus fats versus proteins, for example.
There are technologies we will adopt on the ‘omics side, and also on the wearable device/lifestyle side of things. Things to measure environmental exposure, that would be another one.
X: How do you retrospectively square the adoption of those new technologies with measurements you’ve already taken?
NP: That’s right at the heart of the challenge. With every new technology there will be an evaluation of how well it matches up with old technology. We don’t want to be stuck in today’s technology the whole time, for 20 or 30 years. We think 20 years from now we’ll look at the amount of data we’re collecting today and just think that it’s puny and not very good. There has to be an evolution process.
Consider the microbiome: We’re doing 16S sequencing, which tells you the identity of all microbes that are there and the density, but you don’t get all the genes out of all the microbes. We’ll know with reasonable detail what species were there. When we do the full “metagenomics” in a few years—we’re going to switch as soon as we think it’s a reasonable cost— we’ll get all the genes [and] be able to map those genes back to those particular species.
It’s additive in that sense. Sometimes the new data will make the old data better, sometimes it will just supplant it.
X: Are there lessons to learn from famous longitudinal health studies like the Framingham Heart Study, which has been following residents of Framingham, MA, since 1948?
NP: The issue of transitioning technology, as you brought up, is a really serious one. Another is keeping track of individuals. Framingham was centered around a pretty small town with a reasonable amount of continuity. We like to say this is a digital-age Framingham study, but there is no Framingham location. People are moving around a lot. This study isn’t the center of their lives, we hope it adds value and they get excited about it, but the issues of retention and keeping track of people is a big one.
X: You mentioned opening the data set to others. How open, exactly?
NP: That’s not fully determined yet. Part of it won’t be until we figure out how the study is funded. Let me give a few principles. One: there’s personal health information involved here, so the rawest data has to be carefully protected. There are all kinds of legal regulations you have to follow. For the 100 pioneers we’ve put a lot of safeguards into place, and we hired a company to try and break in. They were unable to do it.
X: Back to understanding the early transition from wellness to disease. You’ve written about redefining what wellness means. What kind of new definitions might we have in 30 years?
NP: We want to understand better the concepts of wellness and resiliency. There are multiple dimensions of this. On one axis there’s fitness. You could have a really good fitness score. But then perhaps you get diagnosed with cancer, so your wellness would be low.
Or you can think about how close your system is to failure or breakdown in a particular way. Diabetes is an interesting case. Type 2 diabetes is a huge epidemic, so you can see aspects of wellness that go down earlier: obesity happens quite a lot, but you could also see early harbingers like insulin resistance. If your system isn’t past that breaking point but you can see the trend toward it, that would have a negative impact on wellness.
Another aspect is mental health: happiness, and self-reported well-being. You might be able to find molecular relationships with those things when people report being in good health, good moods, good energy, or when they don’t.
X: Identifying early transition to disease seems to be the way forward in Alzheimer’s, but should we assume that’s the right plan of attack on other diseases?
NP: We don’t really know. But take this example of someone who was developing osteoporosis at a young stage, took calcium, and reversed it. An executive at Microsoft was developing osteoporosis in his 30s, getting pain, and had problems with his joints. The outlook is bleak, there’s a good chance of being in a wheelchair and incapacitated over time. But he finds he has a known variant in a gene that means his uptake of calcium is much worse than the normal person. So he starts taking 20 times the amount of calcium to force it
