LA JOLLA (September 9, 2025)—Malnutrition is responsible for more than half of all deaths in children under the age of five worldwide. Those who survive can still experience lifelong consequences like cognitive and developmental delays, impaired academic performance, economic instability, and negative maternal health outcomes. This enormous public health issue demands solutions. The latest studies point to gut microbiome—the diverse bacteria, viruses, and other microbes living in our intestines—as a great place to start.
Salk Institute researchers searched for links between undernutrition (a form of malnutrition), microbiome health, and childhood growth in a group of toddlers from Malawi, an African nation with an especially high incidence of child stunting (35 percent). They collected fecal samples from eight children over the course of nearly a year to identify microbial patterns associated with child growth. The study found that children whose gut microbial genomes changed more over time tended to have poorer growth, suggesting that microbiome stability may be an important sign of good gut health.
The researchers also used this new dataset to establish the first-ever pediatric undernutrition microbial genome catalog. The resource contains the full genetic profiles of 986 microbes—collectively called a "pangenome"—found in the fecal samples. This will be a critical public health resource for predicting, preventing, and treating malnutrition. The team also established a novel workflow to create this catalog, which saves researchers time and money while preserving data accuracy. Their method could be used to build genome resources for other health conditions, monitor environmental and agricultural microbiomes, track biodiversity, and enable metagenomic research in remote locations.
The findings, made in collaboration with colleagues at Washington University School of Medicine in St. Louis and University of California San Diego, were published in Cell on September 9, 2025.
"Despite a decade of research linking the microbiome with malnutrition, the genetic and biological factors have remained a mystery due to a lack of resolution on the microbes in the gut," says senior co-corresponding author Todd Michael , research professor at Salk. "By using cutting-edge genome sequencing and pangenomic approaches in a longitudinal design, we were able for the first time to pinpoint specific microbial changes linked to poor growth, opening the door to new diagnostics or therapeutics that could help address a crisis impacting more than 150 million children worldwide."
What we know about the link between malnutrition and the microbiome
One of the first studies to draw a causal link between the microbiome, diet, and severe malnutrition was published in 2013. Researchers transplanted microbiota from severely malnourished children into mice, fed them Malawian-like diets, and watched those mice lose weight like their human counterparts. This result traced a direct line from microbiome health to malnutrition. One author of that paper, Mark Manary, a professor of pediatrics at WashU Medicine, is a co-corresponding author on the new Salk study.
In this latest study, the Salk researchers zoomed in on undernutrition, a type of malnutrition that results from poor nutrient uptake due to either an inability to process nutrients effectively or a nutrient-lacking diet. One broadly used metric of undernutrition is length-for-age scores (LAZ), which track children's heights compared to population-derived expectations for their age and sex.
A low LAZ indicates insufficient growth for the child's age, and a consistently low or worsening LAZ over time is often associated with chronic gut inflammation or environmental intestinal dysfunction. Chronic gut inflammation, scientists have found, can result from dysfunctional microbes impairing the body's ability to process and absorb nutrients.
Between the evidence that microbiome health and malnutrition directly impact each other, and research pointing to dysfunctional microbes as one cause for worsening undernutrition, the Salk team had two new goals: 1) create a comprehensive library accounting for the vast variety of gut microbiota present in children with worsening and improving LAZ, and 2) evaluate if the genetic content of the bacteria are predictive or associated with undernutrition.
Establishing a novel microbiome library
Geneticists piece together genomes using two types of technologies, called "short-read" and "long-read" sequencing. Short-read sequencing breaks DNA into many small fragments that are 50 to 300 base pairs long, while long-read sequencing breaks DNA into fewer, larger fragments that are 5,000 to 4,000,000 base pairs long. Once broken up, the genome can be reassembled like putting together a puzzle. A long-read puzzle, as you might imagine, is much easier to put back together—like completing a 10-piece puzzle rather than a 1,000-piece one.
A gut microbiome may have hundreds of species or strains, like a single puzzle box with many smaller puzzles inside. Using long-read sequencing means opening that box to find 200 10-piece puzzles, rather than 200 1,000-piece ones. The Salk team pieced together long-read puzzles from the fecal samples of eight toddlers across a spectrum of improving and worsening LAZ scores.
"Longitudinal sampling and measurement, five times over 11 months, allowed for a unique assessment of both within- and between-child change over time in the microbiome and growth," said co-senior author Kevin Stephenson, assistant professor at WashU Medicine. "These data can offer insights otherwise obscured in simple cross-sectional analyses."
With long-read sequencing, the team collected 50 times more complete microbiota genomes than would have been possible with short-read sequencing.
"This would not have been possible with short-read technology," says first author Jeremiah Minich, a postdoctoral researcher in Michael's lab. "We found the most efficient, accurate, and cost-effective long-read workflow, applied that workflow to analyze 10- to 20-fold more human samples than anyone has analyzed before, and came out on the other end with a critically important genome resource for undernutrition."
The final genome resource contains 986 complete microbial genomes, dozens of which are entirely novel. With this comprehensive library established, the next step was finding microbiota patterns specific to undernutrition.
What the team found
The researchers used novel pangenome comparison tools, partially developed in Michael's lab, to quickly scour their new library of 986 microbiota. Remarkably, within a given genus (one classification step above species), they found genetic differences in bacterial genomes between children with improving versus worsening growth in four genera (Bifidobacterium, Megasphaera, Faecalibacterium, and Prevotella).
But more interesting than these specific bacteria was an observation about bacterial genome diversity over time.
"Our analysis showed that children with improving growth had stable microbial pangenomes within species, while those with growth faltering had unstable microbial pangenomes," says Manary. "It may then be possible to assess gut health and collect that crucial public health data by measuring gut microbiome genetic diversity."
What's next for malnutrition microbiome research
The study accomplishes four incredible feats across laboratory technology and public health. First, the team collected 10- to 20-fold more human clinical samples than any previous study in the field. Second, they assembled the first longitudinal, pediatric undernutrition microbial library, which contains 986 complete microbiota genomes. Third, they identified specific bacteria and genes amongst species linked to undernutrition and found microbial genome instability over time was associated with poor child growth. And finally, they optimized a long-read sequencing workflow that can now be applied across scientific disciplines.
"When applied in remote, field-based molecular laboratories, the genome sequencing and pangenomic approaches we developed can deliver real-time insights not only into pandemic surveillance, antibiotic resistance, and infectious disease, but also into agricultural productivity, environmental monitoring, and biodiversity conservation," says Michael. "It's a powerful technological advance that expands the reach of genomics and sets a new standard for scientific research in the field."
Other authors include Nicholas Allsing, Nolan Hartwick, Allen Mamerto, and Tiffany Duong of Salk; M. Omar Din, Caitriona Brennan, Lauren Hansen, and Rob Knight of UC San Diego; Michael Tisza and Daniel McDonald of Baylor College of Medicine; Kenneth Maleta of Kamuzu University of Health Sciences in Malawi; Justin Shaffer of California State University; and Emilly Murray of Salk and Scripps Institution of Oceanography.
The work was supported by the NOMIS Foundation, Tang Genomics Fund, National Science Foundation, and U.S. Agency for International Development.
About the Salk Institute for Biological Studies:
Unlocking the secrets of life itself is the driving force behind the Salk Institute. Our team of world-class, award-winning scientists pushes the boundaries of knowledge in areas such as neuroscience, cancer research, aging, immunobiology, plant biology, computational biology, and more. Founded by Jonas Salk, developer of the first safe and effective polio vaccine, the Institute is an independent, nonprofit research organization and architectural landmark: small by choice, intimate by nature, and fearless in the face of any challenge. Learn more at www.salk.edu .
