Our health is a combination of genetics and environment. Maybe someone’s genes make them vulnerable to high blood pressure, for example, but by watching what they eat — in effect, controlling their body’s environment — they can keep their numbers within normal levels.
Right now, we know a lot about the genetics side of this combination, as an explosion of research has yielded incredible detail about people’s genetic profiles. We also have insight into how our internal bacterial environments — the microbiome — impact our health. But the environmental piece of the puzzle is still fuzzy. We don’t measure all the chemicals we encounter each day, from the microscopic fungi on a walk to the car exhaust on a highway.
That is, most people don’t.
Michael Snyder, a Stanford biologist and pioneer in genomics, does. For the past several years, Snyder has been wearing a device he invented that measures the environment around him. It’s part of his quest to learn how the environment impacts our health by studying what he calls people’s “exposomes,” or the various air particles, pollutants, viruses, and more that we come into contact with each day.
“I think the way we do medicine now is very primitive, compared to what it could be.”
In a recent paper in the journal Cell, Snyder and his colleagues describe what they’ve learned from affixing 15 people with these air-monitoring devices for up to 890 days. Each device is about the size of a big matchbox, and contains filters that trap particulates, chemicals, and microbes from the air around it. Medium talked to Snyder about the study, the exposome, and his own self-monitoring discoveries.
This is not your first foray into detailed self-monitoring. A few years ago you were monitoring your own blood over the course of 14 months, and you detected the onset of your own diabetes, right?
Michael Snyder: Yes. The monitoring started when I was doing genome sequencing and other profiles — like gene expression — on myself, 8 years ago. I was using myself as a test subject to get the technology going — I didn’t know I’d be interesting! Then I got diabetes. My genome predicted it, and I got the disease over the course of this profiling. These types of measurements helped me catch the onset of the disease and gave a much more detailed picture of people’s health.
With these techniques — like regular testing of blood for changes in gene expression and other readouts — we’re following a group of 109 folks now, many of them for four-and-a-half years or longer. We added wearables about five years ago to continuously monitor physiology, things like heart rate and blood oxygen levels. It helped me figure out when I got Lyme disease, actually. I was with my brother in rural Massachusetts putting up fences, and two weeks later I flew to Norway. When you fly, your blood oxygen levels will drop, but they usually recover after you land. Mine didn’t. And my heart rate was abnormally high. I got a low-grade fever and went to a doctor, who told me I had a bacterial infection. I told him I thought it was Lyme. He recommended penicillin, but I said I think I need doxycycline, which is what you take for Lyme. I measured myself when I got home, and sure enough, I was Lyme positive. It was a perfectly controlled experiment, because I’d given blood before I left and I was negative then.
We want to bring big data into health; that’s the motivation. I think the way we do medicine now is very primitive, compared to what it could be. Everybody is focused on disease, but we want to focus on health and transitions to disease. We’ve written algorithms based on the data we’ve collected, and we now think we tell can when you get sick before you realize it, because your heart rate goes up. We’ve shown that on myself and three other people. We’re now trying to set up a 1,000-person study to learn more.
How does your new research — which focuses on what you call the “exposome,” meaning the sum total of everything you are exposed to — fit into this picture?
That area has been a big hole. We know you’re affected by your genes and your environment, but nobody is capturing the environment individually. Nobody carries around something on their sleeve to monitor their exposure.
We took a standard high-end air monitor and re-engineered it. The monitor has a pump that sucks up about one-fifteenth of the air you breathe. We put a submicron filter on the monitor to collect all the particulates in the air. Under that we have a cartridge with a chemical absorbent. We take that filter and elute off the particulates, and sequence, incredibly deeply, the DNA and RNA that’s there. Then we match it up against a custom database with 40,000 species of microorganisms, viruses, plants, and animals. We can see exactly what you’re getting exposed to from the biological side. Then from the chemical absorbent, we elute that off and run it through a mass spectrometer. We see all the chemical structures.
The study has just over two years of data on the biologicals; the chemicals we did for only a few months, but nonetheless we learned a ton.
What did you find?
The first thing we learned is the exposome is vast. There were more than 2,000 species, from bacteria to my pet guinea pig, registered during my own two years of profiling. Even the guy or gal who wore it for three months for the study was exposed to over 1,000 species. There were close to 3,000 chemical features detected in the whole study.
Second is that the exposome is dynamic. It varies a lot. How much of the variation is regional or seasonal? For the part we could figure out, location is the number one factor, especially for the chemicals. The time of year is another important factor. We sampled four people living in the Bay Area — me, and people in Sunnyvale, Redwood City, and San Francisco. We profiled them over the same month, and everybody’s different. The person in San Francisco had sewer sludge bacteria in their samples; there are definitely parts of San Francisco that don’t smell so good. Every time I go to Monterey, I get a fungal exposure. Location really matters.
DEET is everywhere, which surprised me; it was in all the samples. There’s a few carcinogens, like the solvent diethylene glycol. A limitation of our study is we don’t know the absolute amounts of exposures; we know relative amounts. That’s something we are working on to pin down. This was really just a survey to see generally what we are exposed to.
You noticed that whenever pyridine — a chemical used in paint — appeared in samples, fungi numbers were low. You hypothesize that unbeknownst to us, pyridine is an antifungal?
Right, that was interesting. Pyridine is a nasty chemical. But you can argue this certain ways. If you are very allergic to fungi, which some people are, maybe for them it’s good to have pyridine, although that is a different potentially detrimental exposure.
Were exposures connected with people’s health?
This is not in the paper, but there’s some correlation with health. We’re still trying to sort this all out.
Your eosinophils — a type of white blood cell — are actually a measure of allergic response. We can correlate my eosinophils with what exposures are out there. I thought I was probably most allergic to pine, but the correlation was actually better with eucalyptus. One in five Americans has allergies or asthma. It’s useful to know what triggers this.
In California, I’m in eucalyptus heaven! I’m not going to cut down a eucalyptus tree, although if a tree ever had to go in my yard that would be the first one to go. Chemical exposures you could try to track and get rid of.
What’s next for this research?
We’re going to get these devices on more people — we will try to get inexpensive devices — just to get this out there so lots of people can do this. Ultimately, we have to take samples and analyze them offline. That will be true for a while. But once we figure out what might be most impactful on people’s health, then we’ll try to set up real-time personal monitors for those things.
I would argue this is the first map of the human exposome, like the first genome map. We see what’s there, and then we try to understand how it affects your health. As we get more devices out, we will be able to make more associations between allergies and exposures. It would take long-term monitoring to understand the effects of toxins. But I do think we need that data.