“Earlier this year, molecular biologists announced that 20 per cent of nonhuman genome databases are contaminated with human DNA, probably from the researchers who sequenced the samples,” Technology Review‘s The Physics ArXiv blog said on Thursday.

“Now, the human genome itself has become contaminated. Bill Langdon at University College London and Matthew Arno at Kings College London say they’ve found sequences from mycoplasma bacteria in the human genome database.”

Holy Zuul, Batman! These are stealth pathogens unaffected by many common antibiotics and can cause serious diseases — autodialing CDC!

“These mycoplasma genes are clearly successful in reproducing themselves in silico,” the post further warned. “One possibility is that we’re seeing the beginnings of an entirely new kind of landscape of infection. Here, genes that can masquerade as human (or indeed as other organisms) can successfully transmit themselves from one database to another.”

Huh? I decided to reality-check this with Arno and Langdon, co-authors of More Mouldy Data: Virtual Infection of the Human Genome.

“Some people have totally misunderstood the blog, and are thinking that we meant human genomes (i.e. DNA in living human cells, in humans) are becoming infected with mycoplasma and this will take over the world,” said Arno reassuringly in an email. “This is not what we meant! I think this was caused by the less than scientific language used in that blog, and a few of the more ‘creative’ assertions there — nothing wrong with a few hypotheses, though.”

“The gene sequences chosen when they were designed are built into them and cannot be changed,” Langdon added. “The results of thousands of such gene chips are widely disseminated via the Internet. Once a tissue sample is contaminated, routines are used to destroy it immediately in the hope of preventing the mold spreading to other samples. Some laboratories routinely sterilise all their experimental glassware every 6 weeks. They do not ‘disinfect’ their computer databases.”

There’s actually a potential benefit, though, Langdon said: “If a sample is contaminated with mycoplasma, there is probably no point to using the data collected from it. [But] if the gene chip design measures the presence of the mold gene, it gives a cheap automatic way of detecting mycoplasma contamination.”

Hmm… maybe there’s a germ (pun intended) of an idea here for a new kind of multi-pathogen detector?