Four years ago, a team of researchers led by a heavyweight in the field of microbiology made a stunning claim: Cancers have unique microbial signatures that could one day allow tumors to be diagnosed with a blood test.
The discovery captured the attention of the scientific community, as well as investors.
A prestigious journal published the research. More than 600 papers cited the study. At least a dozen groups based new work on its data. And the microbiologists behind the claim launched a startup to capitalize on their findings.
Since then, the work has suffered multiple setbacks.
The paper was retracted in June following criticisms by other scientists who questioned the methodology and said the findings are likely invalid. Support for the startup has dried up. And published research that relied on the study’s data might have to be corrected or retracted.
The events illustrate the far-reaching ripple effect of flawed science.
“It has polluted the literature,” said Steven Salzberg, a computational biologist at Johns Hopkins University, whose critique, written with other colleagues in the field, led to the study’s retraction.
When the study landed in Nature in 2020, the sterling reputation of the paper’s senior researcher contributed to the spotlight trained on the work.
Rob Knight, the University of California San Diego professor who led the study, is widely regarded as a pioneer of big-data microbial analysis. His résumé lists multiple awards and prestigious fellowships at scientific societies, two books and a TED Talk. More than a decade ago, when the federal government embarked on a massive effort to document the role of microbes in human health, Knight built a reputation as a key participant and contributor.
In a statement through a university spokesperson, Knight said the retraction was warranted but that the paper’s major conclusions are true.
In the Nature study, he and his co-authors reported that 32 different cancers, from prostate tumors to skin melanomas, harbored unique combinations of microbes, chiefly bacteria and viruses, that acted as fingerprints for each type of tumor. The idea had clinical significance: A blood test could allow physicians to use evidence of microbes detected in the blood as a proxy to diagnose the cancers.
With data from more than 17,000 samples from over 10,000 people with cancer, the paper was hailed as a tour de force for pulling off complicated analyses on a vast data trove.
“It was a highly ambitious study and a ton of data,” said Susan Bullman, a cancer microbiologist at University of Texas MD Anderson Cancer Center in Houston who studies bacteria infiltrating colorectal tumors and wasn’t involved in the study. “If there was a group that could attempt to do that, it would be Rob Knight’s group.”
Over the past decade, scientists including Bullman had gathered evidence that tumors of the gut and intestines, where microbes are abundant, regularly harbored bacteria. The Knight study was the first to survey cancers across the human body and show that microbes existed in combinations that were unique to each type of tumor.
“It was certainly the first with this striking of a finding,” said Curtis Huttenhower, a computational biologist at the Harvard T.H. Chan School of Public Health who studies bacteria that live in people. Huttenhower has collaborated with Knight on other work but wasn’t involved with this study. “It was both surprising that microbes were always detectable, and it was surprising how well they could differentiate cancers.”
The year before the paper was published, Knight and two other authors on the study founded a company, Micronoma, to build blood tests to detect microbial genetic material and to catch cancers early.
Micronoma raised at least $17.5 million from SymBiosis, an investment firm based in Bentonville, Ark., and Seerave Foundation, a philanthropy started by an oil-industry veteran who began investing in cancer and microbiome research after his wife was diagnosed with metastatic melanoma.
The company announced a collaboration with a University of New South Wales team in 2021 to hunt for microbial markers for liver cancer, part of a government grant for 4 million Australian dollars (equivalent to about US$3 million), and a partnership with a group at NYU Langone Health to look at lung cancer, part of a National Institutes of Health grant for more than $703,000 in 2023.
Meanwhile, the FDA gave “breakthrough” designation to Micronoma’s device to test blood for lung cancer, which put the device on an accelerated track for development and approval.
But independent researchers had begun raising alarms.
One red flag was that some microbes the researchers flagged as components of cancer signatures weren’t known to exist in humans, prompting further scrutiny.
Salzberg, the computational biologist, and a team analyzed a handful of the cancer types and didn’t find most of the bacteria reported in the Nature study. Their analysis, published in October 2023 in the journal mBio, stated that the “near-perfect association between microbes and cancer types reported in the study is, simply put, a fiction.”
Among the errors, according to the critique, the UC San Diego team had incorrectly deployed a genomic tool built by Salzberg’s lab to match tumor data to microbial sequences.
“It wasn’t a close call,” Salzberg told The Wall Street Journal. “This data is completely wrong.”
In an emailed response, Knight said Salzberg’s critique alleges but doesn’t prove mistakes and pointed to a paper he and his co-authors published in the journal Oncogene in February that acknowledged the critique and presented a new analysis they say validates their central claim.
When Nature retracted the original study in June, it cited Salzberg’s critique and noted that all of the paper’s authors agreed to its retraction.
As criticisms of the paper mounted, Micronoma’s partnerships dissolved.
NYU Langone ended its association with Micronoma last year, according to a spokesperson at NYU. The University of New South Wales’s collaboration is no longer active, according to one of the lead scientists there. Seerave Chief Executive Officer Manuel Fankhauser declined to provide details about its involvement with Micronoma. SymBiosis didn’t respond to requests for comment. But according to Darryl Garrison, Micronoma’s former chief operations officer, funding dried up.
In April, the company emailed the chair of its scientific advisory board to say it was planning to shut down.
“The retraction certainly didn’t help,” said Dr. Martin Blaser, a Rutgers University professor and chair of the board.
Knight and UC San Diego referred questions about Micronoma to Eddie Adams, the company’s former CEO and president. Adams didn’t respond to multiple requests for comment.
Meanwhile, research groups that, according to Salzberg, relied on the study’s data are dealing with the fallout, as are the journals that published the work.
The publishers or journal editors of eight studies have begun reviewing the papers, according to spokespeople at Springer Nature, Frontiers, Wiley and the American Association for Cancer Research. The corresponding authors of these studies didn’t respond to requests for comment.
Lingeng Lu, a chronic-disease epidemiologist at Yale School of Public Health, said the conclusions of his research, a narrow look at gut bacteria, weren’t affected by the error, though the study credits the Knight team for data. An Elsevier spokesperson said the journal is in touch with Lu to confirm the paper holds up.
Eytan Ruppin, a cancer researcher at the National Cancer Institute, is reanalyzing his own results for a 2022 Nature Communications paper that used the data; an editor’s note has been appended to the study.
A reanalysis is also under way for a second paper that Ruppin co-authored published that same year in the journal Cancer Cell that used data published in the Knight paper.
“We are hoping that our key conclusions may still hold,” Ruppin said. “But that of course remains to be seen.”
Write to Nidhi Subbaraman at nidhi.subbaraman@wsj.com
