Weill Cornell Medical College's Christopher Mason and his colleagues recently reported in Cell Systems that the New York City subway system is teeming with bacterial life. But some took issue with their inclusion in their report of finding DNA that appeared to be from Bacillus anthracis and Yersinia pestis.
Ed Yong at Not Exactly Rocket Science notes that Mason "downplayed the significance of these results" and said there was nothing to worry about. Still, Columbia University's Ian Lipkin tells Yong that the researchers didn't present "persuasive evidence" for the presence of these pathogens in the subway.
"The genetic footprints they report are not specific for the agents that cause anthrax or plague; they are also found in other common bacteria that are not associated with disease," Lipkin adds.
At MicroBEnet this week, Mason notes that he and his colleagues only provided evidence that DNA from these species might be present, highlights that the paper said it was doubtful the bacteria were alive, and adds that they had further hedged their bets by including phrases like "if truly present."
And as Yong and Mason both point out, those species might not even really have been there. Some snippets of DNA Mason's team found best matches in regions of the B. anthracis and Y. pestis genomes, but those sequences could also be present in other bacteria.
In his new blog post, Mason says they included these findings because all their computational tools seem to indicate their presence at low levels and because anyone downloading their data and analyzing it with, for instance Galaxy, MetaPhlAn, and Blast, would uncover those results easily. He notes, though, that the findings are "only as good as our references."
Indeed, he writes that comparing their data to the DNA Database of Japan also gave Y. pestis, for example, as the top hit, though the European Nucleotide Archive Sequence Search database came up with a better match in Enterobacter hormaechei, which Mason notes is also a potentially pathogenic bacterium. Only using Blast in their paper was a limitation, he says.
"It is a clear problem that one can use different methods and to get dramatically different, unique answers from the same data, or complete data non-reproducibility," Mason writes. He adds that he "grossly underestimated how the press would sensationalize these results."