Nature to retract a highly influential Cancer Microbiome Study
Claiming that a blood sample’s microbial mix can indicate the presence of a tumor, the study inspired further research and led to the founding of a company to commercialize the findings.
A highly influential paper that claimed to have identified specific microbial signatures associated with various human cancers has been retracted after journal editors concluded that the findings were unsupported1.
The research employed machine-learning algorithms to detect correlations between resident microorganisms in the human body and the presence and type of cancer, with some claims of near-perfect accuracy. Published in Nature in 2020, the study inspired numerous follow-up studies and led to the establishment of a company to commercialize its results. However, scrutiny from critics revealed methodological errors, prompting the journal to issue an editorial note in 2024 and ultimately retract the paper2.
“Expert post-publication peer review of the issues raised and the authors’ responses has confirmed that some of the findings of the article are affected and the corresponding conclusions are no longer supported,” states Nature’s retraction notice. and all authors agreed with the decision and saw it as the only proper response.
Rob Knight, a microbiologist at the University of California San Diego and a leader in microbiome studies, expressed his team's belief in the major conclusions of the original manuscript, asserting that future research with improved methods and comprehensive data would support their claims.
The retraction is already impacting related research. Eytan Ruppin from the National Cancer Institute, who used the 2020 paper’s data set for a 2022 study in Nature Communications, plans to discuss with the journal whether to redo the analysis with corrected data or retract his work as well. This shows the need for researchers to be aware before using others’ data even if it was published in a prestigious journal like Nature.
Knight and colleagues had compared DNA in blood and tissue samples from cancer patients and healthy individuals with a database of known microbial DNA. Their machine-learning models purportedly predicted the presence and type of cancer from blood samples. In 2019, Knight and Greg Sepich-Poore co-founded Micronoma to commercialize the research, securing substantial funding and developing an experimental lung cancer blood test that received a “breakthrough device” designation from the U.S. FDA.
However, UK researchers noted anomalies in the Nature paper's findings, such as seaweed bacteria being indicative of bladder cancer. Steven Salzberg, a computational biologist at Johns Hopkins University, and colleagues detailed major data analysis errors, including improper filtering of human DNA and computational mistakes that created false patterns. Initially, Knight’s team dismissed these criticisms, but Nature eventually added a note to the article acknowledging concerns about the data and conclusions.
Salzberg welcomed the retraction, stating it was the only proper response. Other researchers, like Lesley Hoyles of Nottingham Trent University, also supported the retraction, appreciating that concerns were acknowledged and the authors agreed.
The future of technology based on the retracted findings is uncertain. Former Micronoma CEO Sandrine Miller-Montgomery claimed their products were unaffected. Still, the company has been silent since last year, and her successor Eddie Adams announced a sale of scientific equipment, suggesting potential struggles.
Researchers stressed that the retraction does not invalidate the field of cancer microbiomes. Ivan Vujkovic-Cvijin from Cedars-Sinai Medical Center urged balanced perspectives on tumor microbiomes. Hoyles emphasized the need for the microbiome community to remain open to reassessing research as understanding of bioinformatics and machine-learning limitations evolves.
Journal retracts influential cancer microbiome paper. (n.d.). Retrieved June 27, 2024, from Science.org website: https://www.science.org/content/article/journal-retracts-influential-cancer-microbiome-paper
Poore, Gregory D., et al. "RETRACTED ARTICLE: Microbiome Analyses of Blood and Tissues Suggest Cancer Diagnostic Approach." Nature, vol. 579, no. 7800, 2020, pp. 567-574, https://doi.org/10.1038/s41586-020-2095-1.