Diet diversity: DNA metabarcoding unlocks potential for personalized plant nutrition

07 Jul 2023 --- Researchers developed a scalable tool to assess dietary plant intake and diversity by sequencing DNA in stool, which they say opens the door for using breakthroughs in DNA sequencing to monitor and improve nutrition.

With “molecular barcodes” – regions of genomic biomarkers – the researchers can identify a food species through its DNA sequence from the residual pool of food-derived DNA in human excrements. A similar method is used to detect plant DNA in environmental research. 

Co-author Lawrence David, associate professor at Duke University, US, tells Nutrition Insight he hopes the study will help build an appreciation “that new advances in genomics have the potential to help us address long-standing challenges in nutrition research.”

Currently, scientific research on diet relies on dietary assessment through self-reported questionnaires, which have random and systematic errors. 

David also sees the research can help “deepen our understanding of the interactions between health, lifestyle and diet that shape the diversity of food species we consume.”

“[The research recognizes] that we don’t just consume ingredients and nutrients; rather, we ultimately construct our diets from different plant and animal species.” 

He expects the method could scale to allow one lab to track tens of thousands of study participants in a year. 

David expects the method could scale to one lab tracking tens of thousands of participants per year.Reference database  
The study, published in Proceedings of the National Academy of Sciences, concludes that the tool may allow researchers to develop a more robust definition of healthy dietary diversity. 

Implicitly, it can more specifically reflect the amount and quality of food consumed by identifying foods eaten in large quantities or those that are not industrially processed – such plants don’t have degraded DNA. 

“We can envision it being used to provide individuals with data on their food intake without having to take the time to fill out dietary surveys or food tracking apps,” says David. “For this kind of data return, I imagine commercial entities involved in personal nutrition might find the technique helpful.” 

The researchers extracted DNA from the samples through a Polymerase Chain Reaction (PCR) protocol, which they compared to a reference database of 791 sequences from 468 plant species. 

They state this database provides “nearly complete” coverage of foods consumed in Western diets but will need to be updated for global cohorts. 

With the tool, they detected 187 unique sequence variants that represent 146 taxa from 73% of prominent food crop families in the reference database. 

“We can detect at least 400 different plant and animal species,” explains David. “While I think it’s possible for traditional dietary assessments to detect this variety of species, doing so with typical surveys could require long questionnaires.” 

He adds that identifying food ingredients in finished products may be challenging for study participants. “For example, I can’t always taste and distinguish all the different spices I eat at a restaurant, and some food additives like guar don’t have easily detectable tastes.”  

Diet diversity 
The researchers used dietary richness – measured as the number of plant taxa per sample – to identify the diversity of foods consumed, since public health nutrition guidance often includes advice to “eat a variety of foods.” 

Diet diversity was measured by the number of plant taxa per sample.“At least in some cohorts we have worked with,” explains David, “we have observed that the number of different plant species that we detect in individuals’ stools is positively correlated with conventional metrics of dietary quality, like the Healthy Eating Index and the Food Variety Score. That may be because these scores go up when people consume a variety of fruits, vegetables, greens and grains.” 

Food quality and diversity have been linked to several health benefits. For example, earlier this year researchers estimated that poor diet resulted in 14.1 million new diagnoses of Type 2 diabetes in 2018. 

“Complementing” self-reported diet constituents
Asking individuals to self-report their food intake is currently the “best method feasible for tracking diet in most studies of free-living people,” David highlights. 

“Still, some drawbacks that have been associated with self-reports include errors and biases associated with memory. Self-reports also might require minimum levels of cognition and literacy, may not be tailored for minority or ethnic foods and place time burdens on individuals with busy schedules.” 

“Our genomic technique doesn’t face all of these drawbacks, so that we can envision it as a useful complement to dietary self-reports.” 

The researchers explain that the tool can provide alternative means of assessment for populations where self-reporting is limited because of participant burden or resource constraints, such as in adolescents. 

In terms of costs, David says the test costs around US$60 per stool sample (independent of shipping). “That’s more expensive than some apps or surveys that participants fill out themselves, but cheaper than some services where interviewers are hired to collect diet data.”  

Depending on people’s gastrointestinal transit time differences, the tool could summarize food intake from 24h to multiple days. 

The researchers note the tool can complement self-reported dietary assessments.Validating data 
The study’s authors tested the accuracy of the developed tool on 1,029 fecal samples of 324 participants, which they note is a scale comparable to validation studies in nutritional epidemiology. 

The researchers tested the tool on five groups, comparing DNA results with restricted food choices or reporting them in a survey. 

A weight loss group chose foods from a digital menu, while a controlled feeding group consumed food from four recurring daily menus. Two larger groups of adults ate their typical diet and reported intake in a food frequency questionnaire, while a group of adolescents did not submit dietary survey data. 

In the adolescents group, the metabarcoding tool detected 111 plant taxa. Over 70% of the group ate four specific taxa – from the wheat, chocolate, corn and potato families.

Tool implementation and next steps 
David recognizes several limitations of the tool. “Our current study only focused on dietary plants and did not directly measure the form that food comes in; that is, how was the food species prepared and cooked?”

“It is possible to track dietary animals, and we are developing those approaches now.” 

“We are also investigating different approaches for inferring how much food species were processed, and those experiments are underway,” he concludes. 

By Jolanda van Hal