What do people eat? Plant and animal DNA in sewage provide clues

DNA signatures of plant and animal species remain behind even after they’ve been eaten, digested, excreted, and flushed down the toilet.

• 0:00Speaker 00

  WNYC is supported by Bayer. Science is about asking questions, testing, trying again. It's not easy, but it's relentless. It evolves, and that's why science delivers. Learn more at sciencedelivers.com.

• 0:17Speaker 01

  Hey, it's Flora, and you're listening to Science Friday. Imagine you wanted to get an idea of what people are eating in a given community. You could look at grocery store sales, but that doesn't capture everyone and everything they ate. You could do surveys, but not everyone will respond. And if they do, maybe they don't remember exactly what they ate, or they'll selectively block out that sleeve of Entenmann's chocolate frosted doughnuts. So my next guest is working on another way of tracking community nutrition from a backend data source: sewer water. Dr. Lawrence David uses genetic data to study nutrition, and his research group came

• 0:58Speaker 01

  up with this innovative, if maybe a little bit gross, technique to analyze the DNA in sewage to identify what plants and animals people are eating. He's calling his research project The Edible Atlas. Lawrence, welcome to Science Friday.

• 1:14Speaker 02

  Hi, nice to be here.

• 1:16Speaker 01

  Thanks for being here. Tell us, right off the bat, what foods show up? I mean, what are you finding?

• 1:23Speaker 02

  Oh, you see everything, things like rice, bananas, wheat, soybeans, a lot of the really common foods. And then we see a lot of things that you wouldn't necessarily think we eat a lot of, like spices and herbs, so like rosemary and sage, or garlic and ginger. So we see almost everything, it seems.

• 1:44Speaker 01

  How granular can you get? Do you get fish, or can you get catfish?

• 1:49Speaker 02

  We can see all kinds of different fish, and it depends on how genetically related they are. We see maybe 50 different species of fish. It's many more than I thought we ever could. But on the other hand, some plants that are genetically very similar, we actually have a hard time telling apart. So things like broccoli, cauliflower, things that are closely related actually kind of look the same to us.

• 2:16Speaker 01

  Do all foods show up equally as well in the data?

• 2:20Speaker 02

  Hmm, that's a good question. Some things, depending on how we prepare them, don't always show up. So the thing that comes to mind is coffee. Coffee doesn't seem to turn up as much as, at least, I know I'm doing my part to put it into the wastewater, and we're not seeing it. But tea shows up pretty well. And I think that might have to do with how, when you steep tea, you're actually drinking a little bit of those leaves.

• 2:47Speaker 01

  I have to say, I'm surprised that DNA can survive my stomach and intestines and the sewer. Is it a hardy molecule? I mean, how much of that DNA is surviving?

• 2:59Speaker 02

  Yeah, DNA is hardier than you might think. So our digestion, though, is pretty efficient. We estimate that probably about 99% of the DNA that we eat gets digested and broken down. The thing is that we eat a lot of it. So we estimate that people are eating maybe up to a gram of DNA a day, so billions of copies of the genes that we sequence. And so even if 99% are degraded, there are still millions of copies per person that can be ultimately amplified out and sequenced. And also, when we go back, anthropologists have done this, and archaeologists, you can

• 3:42Speaker 02

  get DNA from poop samples that go back hundreds of years, sometimes further than that. So DNA can really stick around.

• 3:51Speaker 01

  That's interesting. Okay, walk me through the logistics for how you do these experiments. What should I picture?

• 3:58Speaker 02

  Depending on where you're sampling, someone might go, say, to a wastewater treatment plant, right as the sewage is entering, someone will scoop out some of that wastewater, and then it gets brought--

• 4:13Speaker 01

  When you say "some of that," gallons, a teaspoon?

• 4:17Speaker 02

  Liters, so a decent amount. You can do it two different ways. You can scoop out liters or suck out liters of fluid. You can do it all at once, or you can do it over the course of, say, 24 hours. And the other way you do it is that you kind of lower down something very absorbent, and that sort of sucks up a material and DNA.

• 4:37Speaker 01

  Is the wastewater treatment plant onboard, or can you just sort of like pop open a manhole cover and do this?

• 4:44Speaker 02

  You can do both. You should probably ask for permission if you're going to do either of those things, which we generally do. But yeah, I think that we've been really delighted with how open and helpful a lot of people who work with wastewater have been in letting us take a peek.

• 5:04Speaker 01

  Is it like Christmas morning, when you run it through the sequencer?

• 5:08Speaker 02

  It is. When things come out, it's always fascinating to us if we see something we haven't seen before. And honestly, just if the experiment works. I really didn't think it would work in the beginning, or I didn't know what to expect. And so when we saw it, it was amazing.

• 5:24Speaker 01

  Why didn't you think it would work?

• 5:26Speaker 02

  I think for the reasons that you brought up, right? You kind of assume that a lot of the DNA should be digested. And then, as it passes through things like the sewer and wastewater, you think that it might be broken down. And then finally, one of the things that we were doing in our pilot experiments was piggybacking on the samples that got collected in the pandemic for tracking things like viruses. And I had thought that maybe the same kinds of sampling procedures of looking for viruses won't work for food. But it turned out they did.

• 6:00Speaker 01

  What are some of the sort of top-level trends that you've found?

• 6:05Speaker 02

  I think that one of the things that just stands out to me, and that still just really fascinates me, is that when we count, it's got to be probably north of 100 different animal species across the cities that people are eating, and then maybe close to 200 plant species that show up just day to day in how we live our lives or how cities live. And I think I had never really counted that or thought in my head, like, you know, how many plants and animals does it take for people to eat in, say, a town across, like, a day? And I think there's something really cool and beautiful about that.

• 6:49Speaker 02

  There's this idea that diversity has been lost in our food systems. And actually, there's still quite a bit.

• 6:57Speaker 01

  What are some of the questions that you think this method could answer that the traditional ways of doing this research have not or could not?

• 7:07Speaker 02

  You know, I think one of the biggest has to do with food availability within cities or between cities. And I think one of the things, for example, that comes up is when recommendations go out that people should say, eat more fresh fruits and vegetables. Are people doing that? Are they doing it in season? And then who gets to do that? Does it matter, say, how wealthy a city or a neighborhood is? Does everyone seem to have equal access to those kinds of foods? Those are the kinds of questions we think, hopefully, now it'll be easier to get data on, as well as get data in almost, in a matter of days.

• 7:51Speaker 01

  Hmm. Do you feel like, are there privacy concerns here?

• 7:55Speaker 02

  Oh, yeah, definitely. I think this is measuring what people are eating, right? And that they might not necessarily, it's not like they signed a piece of paper saying that that's what they're doing. One of the couple of ways that we moderate that, one is that we make sure never to sample an area where there are fewer than 500 people, so that people can remain anonymous and we're not identifying anyone. And then, second, what we've done a lot of is tried to talk to people, kind of like we're talking now. We try to be very open about what we're up to, so that people, A, know, and B, we try

• 8:38Speaker 02

  to get as much feedback as we can from the communities near us for how to make this useful to them. And we're hoping that that ultimately kind of mitigates some of those concerns. But I think they're very well-founded, because honestly, if people are watching what I'm eating, I kind of want to know about that.

• 8:58Speaker 01

  Yeah. I mean, I wonder if one person was really into a particular food or a particular spice, would that show up in your data? Or do you really need people eating in bulk to get a signal? I guess that would require some kind of control experiment.

• 9:18Speaker 02

  Oh, and we do have those in mind. We have these protocols that we're prepping that are, they don't involve individuals eating things. Instead, we can go to a very scientific study bathroom and do things like blend up foods and then just flush them down the toilet. So that's how we're going to start analyzing those questions.

• 9:37Speaker 01

  Instead of having one person have a guava, you can just flush some guava down a toilet.

• 9:43Speaker 02

  Yeah, it turns out you can do that, too.

• 9:46Speaker 01

  Are you squeamish about poop?

• 9:49Speaker 02

  Very squeamish. You should have seen when our first kids were born, and I'm getting in there with the diapers, you would think I'd be better at it by now.

• 9:57Speaker 01

  Really? Well, then how did you get started down this path?

• 10:02Speaker 02

  I had a wonderful graduate advisor who set me up for a project when I was a grad student. And I, in a way, got my PhD in part by collecting, I think, what might be one of the world's largest personal collections of feces. I ended up collecting my poop every day for a year as part of a science project where, and this kind of helped teach us how to do what we're doing today. I was collecting my poop to look at my microbiome, so the collection of bacteria that live inside of our guts. And as part of that, A, I had to collect a lot of poop, and B, I had to measure what

• 10:45Speaker 02

  I was eating and write it down, because we thought that that would be impacting our microbiomes. And it took me 45 minutes a day to write down everything I was eating. This kind of taught me about why it's challenging to measure what even just one person is eating. And then the other thing that that study taught me was, in our stool or feces, is a lot of DNA. And so at the time, we were targeting the DNA from bacteria. But it turned out to actually be the same kind of technique to now target the DNA from plants and animals that we eat.

• 11:22Speaker 01

  That's fascinating. I'm still hung up on largest personal stool collection.

• 11:26Speaker 02

  There was one other scientist out, I think, in Colorado who sort of got even more samples. So hats off to him.

• 11:34Speaker 01

  Did you have a freezer at work that was marked, like, "Do not eat Lawrence poop only"?

• 11:41Speaker 02

  We had a code name for our project. It was called Project Omega Gold, because we didn't want to freak out the other scientists in the building who were studying, I guess, less gross things.

• 11:55Speaker 01

  Omega Gold.

• 11:57Speaker 02

  Yeah, so we could talk about it and not worry the other scientists.

• 12:02Speaker 01

  Dr. Lawrence David is an associate professor of molecular genetics and microbiology in the Center for Genomics and Computational Biology. He runs the Edible Atlas Project at Duke University. Lawrence, thanks for being here.

• 12:15Speaker 02

  No, thank you so much for your interest in allowing us to kind of share what we do with more people. Really, really grateful.

• 12:22Speaker 01

  Appreciate it. This episode was produced by Charles Burquist. If you're flush with questions or just ready to spew some feedback at us, call us, 877-4 Sci-Frid is our number. Thank you for listening. I'm Flora Lichtman.