Wastewater DNA reveals what communities eat

During the Covid-19 pandemic, health officials in many places tracked the activity of the virus by analysing sewage, the rationale being that infected individuals shed the genetic material of the agent in what goes down the toilet. We’re doing the same thing right now actually to keep tabs on other infections like polio. But could we go a step further and use the same samples and also ask another important health question: what are people eating? Because, as it turns out that, enough of the DNA of the meat, fruits and vegetables that we’re consuming ends up in wastewater to provide a snapshot of the dietary habits of a community, which could be used to guide and test both food and nutrition policies. Michelle Kirtley is with the Edible Atlas Lab at Duke University and she hopes the approach will help to close a gap in the way we currently collect this sort of nutritional information…

Michelle - The problem is that the way nutrition work is done, the way that we figure out what people are eating, has been really done the same way for decades. And it has a lot of inaccuracies. Mostly we find it out by asking people what they eat. And if I asked you, like, what did you eat yesterday? You would probably remember some things, but not others. And, the people I'm asking tend to be in nutrition research, the people who have the time to sign up for clinical research, not the people who are most at the margins, who are burdened with their time. There are lots of populations that are missed with traditional nutrition research. So we were trying to design a system that could measure what people eat objectively.

Chris - How? How have you gone about that?

Michelle - So everything you eat is alive, or at least most of the things. It once was alive, mostly made of plants and animals. And it turns out that a lot of that DNA survives digestion and shows up in your stool, in your waste. So what we're able to do is get the DNA out of your stool and read the sequences and separate out the different kinds of things that you've eaten in your stool.

Chris - Is this physically people sending you stool samples, or can you go downstream, for want of a better phrase? And could you measure a household by looking at the sewer pipe?

Michelle - Well, and that's exactly what we did. So for years, we had done this work with individual stool samples. And the problem with that is it's not any more comprehensive than what I told you earlier. And so we thought, well, what if we could do this comprehensively and capture a whole neighbourhood or a whole city by looking at the waste that ends up in wastewater, in the sewage?

Chris - And you can get an accurate impression just from the DNA reads there, can you?

Michelle - It is a little bit like being at the 100,000-foot level because it's an aggregate of what everybody has eaten in that community. So it's a little bit like looking at a satellite photo of a particular area. So we can be accurate, but not always telling what an individual meal was eaten at a particular house because we're taking a whole community snapshot at one time.

Chris - Is it quantitative in the sense that if I ate one grape and then you had a sample, you'd say, well, Chris eats grapes. So Chris has had one of his five a day. In fact, I skimped on the five a day. I had one grape, but you'd say I'd eaten grapes.

Michelle - If you gave us your individual stool sample, we might be able to see it if you ate one grape because the technique is quite sensitive. At the level of a whole community's wastewater, that's something that we're still trying to work out exactly how specific it is. We know that it can be quite sensitive. How specific it is, is something that we're still working on.

Chris - Would it confuse my grapes with a bottle of red wine? Because arguably one is gonna be better for your health than the other. So I would put grape DNA down the loo, whether I'd consume grapes or red wine, wouldn't I?

Michelle - You would. And in fact, you might even put it down the sink and not have eaten it at all because maybe it all goes into the same wastewater. So that is in fact correct. So corn is one of the best examples of this. We have corn on the cob, we have corn in tortillas, we have corn starch. All of them have DNA leftover and we can't, at the level of DNA, tell the difference between them. What we can do though, is look at what kinds of plants and animals tend to ride along together. And that's something that we're trying to do to work out whether something, let's say corn, is it most likely to have been corn starched in a processed food, or is it more likely to have been corn that was eaten in, let's say, a more healthy way. And we're trying to look at the combinations of foods that people eat to kind of get at that question.

Chris - I can see why this would work if you have a very homogeneous society where people tend to all eat the same sorts of diets. But if you've got a very mixed up community of certain people are gonna eat certain things, other people are gonna eat other things, when you put the whole lot together, you might get what looks to your assay like a really good balanced diet because it's all the food groups represented. But were we to zoom in from your 100,000 foot view in on individual households, you might find some of them are existing on burgers and chips only. And actually all the vegetable signals are coming from the house next door where they're all vegan.

Michelle - Well, that's exactly right. But the interesting thing about people is they tend to live in places where there are other people who are a lot like them. So, some of what we saw in our study is that areas that are more urban, even though it's exactly as you say, there are people eating healthy, presumably right next door to people who maybe have to consume food in a more convenient or fast way. We can actually see the difference between a whole rural community and an urban one, or one that was in the mountains and one that was in the coast. So even though it's a big soup and we can't disentangle what you or your neighbour ate, we can still see community level differences because communities themselves have their own character and flavour.

Chris - How did you validate this? So you know that when you get these signals, they actually do map onto what people really are eating.

Michelle - That's a great question. And we have done this, in part by in Durham, which is where our university is. We were able to compare the signals that we saw in the sewer water with the signals from individual stool samples that we had on hand from other clinical studies that we had done. And we were able to see that the same basic profile of plants and animals were in both sets of samples.

Chris - Does it look practical though? Can you see if you go to government and policymakers with the data you now have, do you think they would be receptive to this? And do you think this would be a valuable tool in dealing with nutritional inequalities across countries?

Michelle - We have been exploring that at a variety of scales, more recently going at the neighbourhood level to try to see if we can get at the question of food access, whether you have access to a grocery store or maybe too much access to fast food restaurants or convenience stores. Change a whole community's diet profile and whether we can use that data to help shape food access policy. There's a community organiser in a neighbourhood in Durham that was promised a grocery store as part of a public housing project that was done, and the grocery store never materialised. So they are trying to raise money for a grocery store to come to their neighbourhood so that the neighbourhood can have healthy food. And they approached us to see if we could measure the food in the wastewater before and after the grocery store was put in so that they could potentially show policymakers, hey, look, the grocery store made a difference.