A 2017 Nature paper linking maternal gut bacteria to autism-like traits in mice received an editorial expression of concern last week because of apparent data duplications.
The study, which has been cited more than 680 times, found that certain maternal gut bacteria increase the chances that pups exposed to their mother’s immune response in the womb will have patches of overactive neurons in their brains and atypical behavior.
But the data for two parameters used to measure sociability in the pups have recurring identical values across multiple datasets presented in 10 different figures, according to the expression of concern. These issues cannot be addressed or verified because the original raw data files were lost during lab relocations, the addendum continues.
“The study’s conclusions remain unchanged and have since been corroborated by subsequent studies,” says investigator Jun Huh, associate professor of immunology at Harvard Medical School.
But Mu Yang, a scientific-integrity sleuth and director of the Mouse NeuroBehavior Core at Columbia University, says these data “are critical for the conclusions,” and the extent to which the duplications change the findings “can only be evaluated by comparing incorrect data with correct raw data.”
In an email to The Transmitter, Huh listed two subsequent studies from his lab and a separate 2018 study from researchers at the University of Virginia that support the 2017 study’s findings.
“The core concepts in the paper have been reproduced,” says John Lukens, professor of neuroscience at the University of Virginia and an investigator on the 2018 study. “I don’t think [the data duplications] should take away from what was reported in the original paper,” which he credits as highly influential for the field.
The expression of concern notes a correction for a duplicated image that was flagged on PubPeer in September 2025 by Sholto David, a molecular biologist and well-known sleuth who was awarded millions in a separate whistleblower lawsuit for identifying issues in published research. Huh responded on PubPeer that same month, saying his team had inadvertently repeated the image and that he would contact the journal to correct it. (David says he reached out to Nature directly in September 2026 to notify editors that the image had still not been addressed.)
But in reviewing the research for the original image, Huh decided to go through all the data in the paper in case there were other mistakes that had slipped through the cracks, which speaks to Huh’s “being thorough and honest,” says Dan Littman, professor of molecular immunology at New York University and an investigator on the paper. In doing so, Huh discovered “a whole bunch of duplicate numbers” in some of the data but could not verify the source and notified Nature, Littman says.
“As is the case when any concerns are raised, we investigated the matter carefully, following our established processes,” said Francesca Cesari, chief biological, clinical and social sciences editor for Nature, in a statement provided through a spokesperson. “Because the authors have been unable to locate all the original raw data files we determined, in consultation with the Springer Nature Research Integrity Group, that issuing an EEoC was the most appropriate course of action.” All the authors agreed with the expression of concern, except for first author Sangdoo Kim and Soyoung Ha, neither of whom responded to Nature’s communications or could be reached for comment for this story.
T
he research marked “a significant advance in understanding potential environmental contributors for autism,” Sarkis Mazmanian, professor of microbiology at the California Institute of Technology, who was not involved in the work, told The Transmitter in 2017.The mechanism that led the mother’s immune system to produce interleukin-17a, a specific inflammatory protein associated with an increased likelihood of autism, was not well understood when the study was published in 2017, according to the paper. The research identified a potential factor in the gut microbiome and became the basis for subsequent work on the topic, Littman says.
Six experts declined The Transmitter’s request for comment on the paper’s influence on the field and the potential impact of the data duplications on the paper’s conclusions; Mazmanian, along with more than 20 others, did not respond to the request.
More broadly, research linking the gut microbiome to autism is plagued by conceptual and methodological flaws, says Kevin Mitchell, professor of genetics and neuroscience at Trinity College Dublin, who co-authored a Neuron perspective on this issue in January. He says immune-activation protocols are unreliable when it comes to the offspring’s behavior, and “the ‘phenotypes’ reported in the cortex occur quite frequently in wildtype mice.” The Transmitter covered other data duplications in notable gut-brain studies earlier this year.
“That literature is all over the place,” Mitchell says.
