Nature Issues Expression of Concern Over Autism Study Data Duplications

A 2017 study published in Nature that linked maternal gut bacteria to autism-like traits in mice has received an editorial expression of concern due to apparent data duplications, The Transmitter reported.

Key Clinical Takeaways:
• The 2017 paper, cited more than 680 times, reported that certain maternal gut bacteria increase the chances of overactive neurons and atypical behavior in mouse pups exposed to maternal immune responses.
• An expression of concern issued last week notes recurring identical values for two sociability parameters across multiple datasets presented in 10 different figures, with original raw data files lost during lab relocations.
• Investigators state the core conclusions remain unchanged and supported by subsequent research, while scientific integrity sleuths maintain that incorrect data must be compared against correct raw files to evaluate the true impact on findings.

Data Duplications Discovered After Lab Relocation Losses

The 2017 research investigated how specific maternal gut bacteria heighten the probability that pups exposed to their mother’s immune response in the womb develop patches of overactive neurons in their brains alongside atypical behavior. According to the editorial addendum, data for two parameters measuring pup sociability display recurring identical values across multiple datasets spanning 10 distinct figures. Because the primary raw data documents went missing during lab moves, the addendum explains that these problems cannot be resolved or checked.

The review of the paper followed scrutiny initiated in September 2025, when molecular biologist and integrity sleuth Sholto David flagged a duplicated image on PubPeer. Investigator Jun Huh, associate professor of immunology at Harvard Medical School, responded on PubPeer that same month to state his team had inadvertently repeated the image and would contact the journal. However, after reviewing the research for that original image, Huh decided to examine all data in the paper to check for other mistakes. In doing so, Huh discovered a series of duplicate numbers in the data, could not verify the source, and notified Nature, according to Dan Littman, professor of molecular immunology at New York University and an investigator on the paper. Littman noted that this action speaks to Huh being thorough and honest.

“As is the case when any concerns are raised, we investigated the matter carefully, following our established processes,” Francesca Cesari, chief biological, clinical and social sciences editor for Nature, said 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 authors agreed with the expression of concern except for first author Sangdoo Kim and Soyoung Ha, neither of whom responded to communications or could be reached for comment.

Independent Assessments of Core Concepts and Integrity Concerns

“The study’s conclusions remain unchanged and have since been corroborated by subsequent studies,” Jun Huh said. Huh referenced a 2018 study by University of Virginia researchers along with two later investigations originating from his own laboratory to corroborate the initial results. John Lukens, professor of neuroscience at the University of Virginia and an investigator on the 2018 study, stated that the core concepts have been reproduced. “I don’t think [the data duplications] should take away from what was reported in the original paper,” Lukens said, crediting the work as highly influential for the field.

Conversely, Mu Yang, scientific-integrity sleuth and director of the Mouse NeuroBehavior Core at Columbia University, argued that the duplicated data “are critical for the conclusions,” adding that the extent to which the duplications change the findings “can only be evaluated by comparing incorrect data with correct raw data.” Sholto David noted that he reached out to Nature directly in September 2026 to notify editors that the image flagged a year prior had still not been addressed.

When it was released in 2017, the study represented a major milestone in comprehending potential environmental triggers for autism, remarked Sarkis Mazmanian, a microbiology professor at the California Institute of Technology who did not participate in the study. Back then, scientists did not fully grasp the pathway that causes a pregnant woman’s immune system to generate interleukin-17a—an inflammatory protein specifically linked to a higher risk of autism. The work pinpointed a possible gut microbiome mechanism and laid the groundwork for subsequent studies in the field, Littman observed. Six experts declined requests for comment on the paper’s influence and the impact of the data duplications, and Mazmanian did not respond.

Broader questions persist regarding gut-brain research in autism. Kevin Mitchell, professor of genetics and neuroscience at Trinity College Dublin, who co-authored a perspective on the subject in January, stated that the literature is inconsistent and that immune-activation protocols remain unreliable regarding offspring behavior. Mitchell noted that phenotypes reported in the cortex occur quite frequently in wild-type mice, pointing to earlier data duplications in notable gut-brain studies covered by The Transmitter earlier in the year.

Disclaimer: The information provided in this article is for educational and scientific communication purposes only and does not constitute medical advice. Always consult with a qualified healthcare provider regarding any medical condition, diagnosis, or treatment plan.