Research reviewed by Gram found that DNA from stool can detect some foods infants eat, with nuts showing about 1.9 times higher odds of matching parent reports. The 2026 study of 73 Dutch infants suggests stool testing may one day complement questionnaires, though it is not yet ready for everyday use.

Parents often struggle to remember exactly what their baby ate, which makes diet research difficult. A new study in The American Journal of Clinical Nutrition tested whether DNA left behind in stool can reveal what infants eat. Researchers analyzed stool samples from 73 Dutch infants at 9 and 12 months old, plus samples from 25 mothers, and compared the results with parent questionnaires. According to research reviewed by Gram, the stool tests matched parent reports for several foods, including nuts and potatoes, and for formula feeding. Stool analysis is not yet ready for everyday use, but it may someday give scientists a more objective way to track early childhood nutrition.

Key Statistics

A 2026 study of 73 Dutch infants published in The American Journal of Clinical Nutrition found that food DNA in stool was about 1.9 times as likely to match parent-reported nut consumption.

A 2026 cohort analysis of infant stool samples found that formula feeding detected in stool was linked to about 4 times higher odds of being reported by parents.

In the same 2026 study, matched stool and questionnaire food pairs showed statistically stronger associations than unmatched pairs, with a p-value of 1.6 in a trillion.

The Quick Take

  • What they studied: Can DNA from food found in a baby’s stool show what the baby has eaten, and does it match what parents report?
  • Who participated: 73 infants from the Dutch Lifelines NEXT cohort, sampled at 9 and 12 months old, for 139 stool samples in total. Samples from 25 mothers were included for comparison.
  • Key finding: Nine food types showed a link between DNA found in stool and what parents reported. Nuts had the clearest link: infants whose stool showed nut DNA were about 1.9 times as likely to have been reported as eating nuts. Formula feeding showed the strongest link, with odds about 4 times higher.
  • What it means for you: This is an early research step, not a home test. It suggests stool DNA may become a useful, non-invasive way to check infant diet in future studies, which could make diet tracking more accurate than relying on memory alone.

The Research Details

The researchers worked with families taking part in the Lifelines NEXT cohort, a large Dutch health study that follows children from early life. They collected stool samples from infants at 9 and 12 months of age and from mothers. Each sample was tested with FoodSeq, a newer tool that searches stool for small pieces of DNA from foods. Each food DNA signal was then compared with what the parents had reported on a food frequency questionnaire, a form that asks how often a child eats certain foods.

The team used statistical models designed for repeated measurements from the same children over time. These models account for the fact that the same infant contributed more than one sample. They looked at whether a food found in the stool was more likely to appear on the questionnaire, and whether the link was stronger when the stool and questionnaire matched for the same food.

The researchers also examined the gut microbiome, the community of bacteria living in the intestines, to see whether questionnaire answers lined up with bacterial patterns that would be expected from those foods. This extra check helped them judge whether the stool-based food signals were consistent with the broader biology of digestion.

Diet studies in babies usually depend on parents remembering and reporting what their child ate. Memory is imperfect, and parents may not know exactly what was in a meal or snack. A method that reads food signals directly from stool could provide an independent check on those reports. Comparing the two sources side by side shows whether they agree, and where they do not.

The study was published in a respected nutrition journal and used a well-established cohort with repeated samples over time. The statistical methods were chosen to handle repeated measurements from the same infants. However, the sample was relatively small, with 73 infants, and some results only reached the usual threshold for statistical significance after adjustment. The findings should be treated as promising early evidence that needs replication in larger groups.

What the Results Show

The researchers found nine meaningful links between food DNA in stool and foods that parents reported. Five of these are highlighted in the study: nuts, potatoes, dairy, formula feeding, and pork. The strongest and most reliable link was for nuts. Infants whose stool contained nut DNA were about 1.9 times as likely to have been reported as eating nuts, and this result met the study’s threshold for statistical confidence.

The links for potatoes, dairy, formula, and pork were also strong, with odds ratios from about 2 to 4, but they fell just short of the strictest confidence cutoff. This means they point in a promising direction but should be confirmed in future work. Formula feeding showed the largest effect, with odds about 4 times higher when formula DNA appeared in stool. This makes sense because formula is a very consistent, easily detected source of dietary material.

Another notable result was that matched pairs, meaning cases where the stool test and questionnaire agreed on a specific food, showed much stronger associations than mismatched pairs. The statistical difference was extremely large, which suggests the agreement was not due to chance. The researchers also found that infant stool contained fewer distinct food groups than mothers’ stool, which fits with babies eating a narrower range of foods.

The researchers found that questionnaire-reported foods that had stronger links to particular gut bacteria also tended to agree more often with the stool DNA results. This suggests that the food signals in stool are linked to real digestive activity, not just random noise. It also gives scientists a second way to judge whether a parent’s report is likely accurate.

The study also showed that infant stool contains a narrower range of detectable food groups than maternal stool. This is consistent with the simpler diets of babies around 9 to 12 months old. Understanding these differences could help researchers design diet tests that are tailored to children rather than simply adapted from adult methods.

Diet assessment in young children has long relied on parent questionnaires and food diaries, which are known to contain recall errors and social bias. Researchers have also explored biomarkers in blood or urine for specific nutrients, but these typically measure a few compounds rather than a broad range of foods. Stool DNA analysis is a newer approach that builds on advances in gut microbiome research and DNA sequencing. This study adds evidence that food DNA can be detected in infant stool and that it partly matches what parents report, which supports the idea that the method could complement existing tools.

The study included only 73 infants, which is small for drawing firm conclusions about diet in general. Several food associations did not reach the strictest statistical threshold, so they may be due to chance until confirmed. The stool test detects only foods that leave enough DNA behind and can be told apart by genetic markers, so many foods will not be captured. Cooked, processed, or heavily mixed foods may be harder to detect. The participants were from one cohort in the Netherlands, so results may not apply to infants with different diets or cultural eating patterns. Finally, the tool is still in development and needs further validation before it can be used widely.

The Bottom Line

Research shows that stool DNA can detect some infant foods, including nuts and formula, and that these signals partly match parent reports. Confidence is moderate at this stage because the study is small and several findings need confirmation. Parents should not change feeding decisions based on this research, and no home stool test is available. Researchers and clinicians may consider stool-based methods as a supplement to questionnaires in future studies.

Researchers studying early childhood nutrition, pediatric dietitians, and public health teams designing infant diet studies should pay close attention. Parents may find the research interesting, but it does not offer a practical tool yet. Parents of infants with allergies or unusual feeding needs should continue to work with their pediatrician, since this study does not provide clinical guidance.

Stool DNA diet tools are still in the research phase. Larger studies and validation work will likely take several years before the method could be used in routine research. Clinical or consumer use would require additional testing and regulatory review, so it is unlikely to be available in the near term.

Frequently Asked Questions

Can a baby’s poop show what food they ate?

Partly. A 2026 study found that DNA from some foods left in infant stool matched parent reports, with nuts and formula showing the clearest links. The method is still experimental and not available as a home test.

How accurate are parent food logs for babies?

Parent questionnaires depend on memory and can contain errors. Stool DNA analysis offers an independent check, and the 2026 study found that matched stool and questionnaire results were more strongly linked than mismatched ones.

What is FoodSeq and how does it work?

FoodSeq is a newer tool that reads small pieces of food DNA from stool samples. It identifies food groups by genetic markers, and in a 2026 study of 73 infants it detected signals for foods such as nuts, potatoes, dairy, and pork.

Should I change my baby’s diet based on this research?

No, this research does not provide feeding advice. It is an early study of a diet measurement method, so parents should continue following guidance from their pediatrician or a registered dietitian for infant nutrition decisions.

Want to Apply This Research?

  • Log each food your baby eats, including the food name, amount, and time of day, for a set period such as 7 days. Note whether the food was homemade, store-bought, or formula. Compare these logs with any stool or diaper observations you choose to record, but treat them as personal notes, not diagnostic data.
  • Use the app to set a reminder to log meals right after feeding, rather than trying to remember at the end of the day. Accurate, timely entries are the same principle that makes questionnaire data more reliable, which is the problem this research is trying to solve.
  • Keep a running food log through the first year and review it monthly to spot patterns, such as which foods your baby is eating regularly. Share summary reports with your pediatrician at well-child visits. As stool-based diet tools mature, the app can be updated to accept lab results alongside your logs.

This article summarizes an early research study on a method for measuring infant diet and is for general information only. It is not medical advice and does not replace guidance from your pediatrician, pediatric dietitian, or other qualified health professional. The stool-based diet tool described is still in development and is not available for clinical or home use. Always consult a healthcare provider before making changes to your baby’s diet.

This research translation is published by Gram Research, the science division of Gram, an AI-powered nutrition tracking app.

Source: Assessing dietary intake in infants through fecal sample analysis in the LLNEXT Cohort. , The American journal of clinical nutrition (2026). PubMed 43039144 | DOI
Topics
infant nutrition stool DNA diet assessment baby food tracking FoodSeq gut microbiome infant feeding food frequency questionnaire