Scientific investigators have identified 230 reports of unexpected protein expression across 221 research papers. In these cases, genetically modified cell lines appeared to produce proteins despite lacking the genes needed to make them. Journals are beginning to retract or revise affected papers, but researchers say the problem is probably much larger.
“This problem is not going away,” says Thierry Souci, a cancer geneticist at Uppsala University in Sweden. Souci notes that several other examples appeared in a now-retracted paper but were not included in the latest analysis. “This is just the tip of a huge iceberg.”
Unexpected proteins in modified cell lines
A cell line is a population of cells derived from a single source and grown in the laboratory for experiments. Some cancer cell lines have been genetically modified with mutations or gene deletions that remove controls on cell division. The resulting cultures can continue dividing, allowing researchers to study tumour progression and test potential treatments.
MCF-7 is one such cell line. It lacks the gene for caspase-3, a protein involved in cell death. A 2009 warning1 noted that several studies using MCF-7 nevertheless reported caspase-3 expression in Western blots. This imaging technique identifies proteins according to their molecular weight.
The latest analysis was published as a preprint on October 21 in the Zenodo data repository.2 Sholto David, a scientific investigator based in Arbroath, UK, found that the issue had persisted and appeared in other cell lines.
David compiled a list of 20 commonly used cell lines with genetic mutations or deletions that prevent them from expressing particular proteins. He then used Google text and image searches to look for Western blot images that appeared to show those proteins. He also searched a collection of 140,000 papers, mostly biomedical studies, that he had previously downloaded and screened for image-integrity and other problems.
His analysis found 230 reports of so-called phantom proteins in 221 papers published since 2004. The most common error was the reported presence of caspase-3 in MCF-7 cells, identified in 49 papers. The cell-death protein BAX was reported in Jurkat cells in 47 papers, while the signalling protein STAT3 was reported in PC-3 cells in 35 papers.
Possible explanations
David proposes several possible explanations for the apparently impossible protein-expression patterns. Some could result from honest mistakes: researchers might have misidentified a cell line or contaminated it with another cell line. Alternatively, an antibody used in a Western blot might have been nonspecific and bound to a different protein with the same molecular weight.
David says he encountered the unusual protein patterns while working on a project that alerted hundreds of studies to the use of incorrect antibodies in experiments.
Another possibility is that some papers were produced by paper mills, companies that manufacture and sell scientific papers. David says that some papers in his collection contain features associated with such operations, including duplicated images and text templates resembling those in other papers. He says it is not clear what proportion of the papers might fall into this category.
Nature contacted several authors of papers containing the anomaly for comment, but none responded.
Source: www.nature.com












