April 8, 2024 -- A lightbulb moment hit as Lawrence David was chatting one day with an ecologist who studies the microbiomes and diets of large herbivores in the African savanna. David was envious. He'd been studying the human microbiome, and this ecologist had tons of animal statistics that were way more specific than what David had obtained from people.
"How on earth do you get all these dietary data?" David recalled asking. "Obviously, he didn't ask the animals what they ate."
All those specific statistics came from DNA sequencing of animal scat scooped up from the savanna.
Indeed.
Depending on when you read this, you may have the DNA of more than a dozen plant species, plus another three or four animal species, gurgling through your gut. That's the straight poop taken straight from, well, poop.
David and colleagues are analyzing the DNA in human feces to better understand digestion and the links between diet and health, potentially paving the way to treatments for diet-linked diseases.
Diet, DNA, and Feces
Everything we eat (except vitamins, minerals, and salt) came from something that was living, and all living things have genomes.
"A decent fraction of that DNA" goes undigested and is then excreted, said David, a PhD and associate professor of molecular genetics and microbiology at Duke University in Durham, NC.
"We are using DNA sequencing to reconstruct what people eat," David said. "We try to see if there are patterns in what people eat and how we can measure them by DNA, or kind of genetic forensics." Then they connect that data to health outcomes like obesity.
A typical person's excrement probably contains the DNA of 10-20 plant species and three or four types of animal DNA. "And that's the average person. Some people may have more like 40 types at any given time," David said.
Studying DNA in human feces has potential applications in research and in clinical settings. For instance, it could help design personalized nutrition strategies for patients, something that's already being tested. He hopes that DNA information will help "connect patterns in what people eat to their microbiomes."
One big advantage: Feces don't lie. In reconstructing someone's diet, people either forget what they ate, fudge the truth, or can't be bothered to keep track.
"Patients report the fruit they ate yesterday but not the M&Ms," said Neil Stollman, MD, chief of the division of gastroenterology at Alta Bates Summit Medical Center in Oakland, CA.
Some people can't write it all down because they're too old or too young — the very people at highest risk of nutrition-associated disease, said David.
Fetching and Figuring Out Feces
It's a lot of work to collect and analyze fecal matter, for ethical, legal, and logistical reasons. "And then there's sort of an ick factor to this kind of work," David said.
To get samples, people place a plastic collection cup under the toilet seat to catch the stool. The person then swabs or scoops some of that into a tube, seals the top, and either brings it in or mails it to the lab.
In the lab, David said, "if the DNA is still inside the plant cells, we crack the cells open using a variety of methods. We use what's called 'a stomacher,' which is like two big paddles, and we load the poop [which is in a plastic bag] into it and then squash it — mash it up. We also sometimes load small particles of what is basically glass into it and then shake really hard — it is another way you can physically break open the plant cells. This can also be done with chemicals. It's like a chemistry lab," he said, noting that this process takes about half a day to do.
There is much more bacterial DNA in stool than there is food DNA, and even a little human DNA and sometimes fungi, said David. "The concentration of bacteria in stool is amongst the highest concentrations of bacteria on the planet," he said, but his lab focuses on the plant DNA they find.
They use a molecular process called polymerase chain reaction (PCR) that amplifies and selectively copies DNA from plants. (The scientists who invented this "ingenious" process won a Nobel Prize, David noted.) Like a COVID-19 PCR test, the process only matches up for certain kinds of DNA and can be designed to be more specific or less specific. In David's lab, they shoot for a middle ground of specificity, where the PCR process is targeting chloroplasts in plants.
Once they've detected all the different sequences of food species, they need to find the DNA code, a time-consuming step. His colleague Briana Petrone compiled a reference database of specific sequences of DNA that correspond to different species of plants. This work took more than a year, said David, noting that only a handful of other labs around the country are sequencing DNA in feces, most of them looking at it in animals, not humans.
There are 200,000 to 300,000 species of edible plants estimated to be on the planet, he said. "I think historically, humans have eaten about 7000 of them. We're kind of like a walking repository of all this genetic material."
What Scientists Learn from Fecal DNA
Tracking DNA in digested food can provide valuable data to researchers — information that could have a major impact on nutritional guidance for people with obesity and digestive diseases and other gastrointestinal and nutrition-related issues.
David and Petrone's 2023 study analyzing DNA in stool samples, published in the Proceedings of the National Academy of Sciences (PNAS), showed what — and roughly how much — people ate.
They noticed that kids with obesity had a higher diversity of plants in them than kids without obesity. Sounds backward — wouldn't a child who eats more plants be a healthier weight? "The more I dug into it, it turns out that foods that are more processed often tend to have more ingredients. So, a Big Mac and fries and a coffee have 19 different plant species," said David.
Going forward, he said, researchers may have to be "more specific about how we think about dietary diversity. Maybe not all plant species count toward health in the same way."
David's work provides an innovative way to conduct nutrition research, said Jotham Suez, PhD, an assistant professor in the department of molecular microbiology and immunology at Johns Hopkins Bloomberg School of Public Health.
"We need to have some means of tracking what people actually ate during a study, whether it's an intervention where we provide them with the food or an observational study where we let people eat their habitual diet and track it themselves," said Suez, who studies the gut microbiome.
"Recall bias" makes food questionnaires and apps unreliable. And research suggests that some participants may underreport food intake, possibly because they don't want to be judged or they misestimate how much they actually consumed.
Stollman, an expert in fecal transplant and diverticulitis and a trustee of the American College of Gastroenterology, sees many patients with diverticular disease who could benefit from this type of research.
"One of the core questions in the diverticular world is, what causes diverticular disease, so we can ideally prevent it? For decades, the theory has been that a low fiber diet contributes to it," said Stollman, but testing DNA in patients' stools could help researchers explore the question in a new and potentially more nuanced and accurate way. Findings might allow scientists to learn, "Do people who eat X get polyps? Is this diet a risk factor for X, Y, or Z disease?" said Stollman.
Future Clinical Applications
Brenda Davy, PhD, is a registered dietitian and professor in the Department of Human Nutrition, Foods, and Exercise at Virginia Tech. She conducts research investigating the role of diet in the prevention and treatment of obesity and related conditions such as type 2 diabetes. She also develops dietary assessment methods. More than a decade ago, she developed one of the first rapid assessment tools for quantifying beverage intake — the Beverage Intake Questionnaire — an assessment that is still used today.
"Dietary assessment is necessary in both research and clinical settings," Davy said. "If a physician diagnoses a patient with a certain condition, information about the patient's usual dietary habits can help him or her prescribe dietary changes that may help treat that condition."
Biospecimens, like fecal and urine samples, can be a safe, accurate way to collect that data, she said. Samples can be obtained easily and noninvasively "in a wide variety of populations such as children or older adults" and in clinical settings.
Davy and her team use David's technology in their work — in particular, a tool called FoodSeq that applies DNA metabarcoding to human stool to collect information about food taxa consumed. Their two labs are now collaborating on a project investigating how ultraprocessed foods might impact type 2 diabetes risk and cardiovascular health.
There are many directions David's lab would like to take their research, possibly partnering with epidemiologists on global studies that would help them expand their DNA database and better understand how, for example, climate change may be affecting diet diversity and to learn more about diet across different populations.
