A New Source of Potential Error in Scientific Research

In today’s Times, Nicholas Wade reports on a potentially serious, besides costly, problem for biomedical researchers: Human DNA Contamination Seen in Genome Databases. He writes:

Nearly 20 percent of the nonhuman genomes held in computer databases are contaminated with human DNA, presumably from the researchers who prepared the samples, say scientists who chanced upon the finding while looking for a human virus…

The full report was published yesterday in PLoS One. The investigators, based at the University of Connecticut, screened for a common human sequence in 2,749 non-primate public databases – NCBI, Ensembl, JGI, and UCSC – and found 492 were contaminated with human DNA. Affected sequences included include bacterial, fish, plant and other genomes.

The implications are broad because if the findings in this report are true, scientists throughout the world have drawn inferences and conclusions and published papers based on incorrect DNA sequence information. As the PLoS authors write in their introduction:

The danger in the propagation of errors in scientific discourse has been demonstrated in cases of both scientific fraud as well as incorrectly described or referenced experiments in reviews [1], [2].

abstract double helix (Wikimedia Commons)

Sound familiar? Think of Lies, Damn Lies and Medical Science, as I considered in a December post on the Decline Effect and other problems that cast doubt on research findings we take for granted.

What happened is likely that, over the years and at many separate institutions, researchers handling cells from which DNA would be extracted, or perhaps just handling the DNA and doing sequencing and other experiments with that, contaminated the specimens with their own genetic material. This is a real headache for researchers, or should be.

Yesterday an on-line colleague and patient advocate, Trisha Torrey, (via a Twitter conversation) to the “HeLa bomb” as recounted in Rebecca Skloot’s The Immortal Life of Henrietta Lacks. In that, Skloot describes 1960s researchers who realized that the cells they’d been using for cancer research experiments were contaminated in vitro. HeLa cells tended to grow so rapidly, they’d sometimes overwhelm other cultures growing in nearby petri dishes or flasks. Once scientists realized that the cells they were using weren’t what they thought they were, and that HeLa cells weren’t all the same due to acquired mutations, their results became questionable.

My take is that researchers need to take care, and not make assumptions, such as “the cells I have received for analysis from a colleague’s lab are the cells they are said to be on the label.” Vials get mislabeled, sometimes. Cultures get contaminated by bacteria and fast-growing cell lines. And now it’s evident that at least some published genomes are incorrect.

But also – and more generally, we should constantly be questioning and checking and reviewing our methods and reagents, and whatever forms evidence, especially when results are surprising (think Madoff) and/or have implications for patient care and therapy, because errors do really happen in all realms of medical and scientific research.

Related Posts:

Leave a Reply


This site uses Akismet to reduce spam. Learn how your comment data is processed.