High proportions of bacteria and archaea across most biomes remain uncultured

A recent paper by Martiny argues that "high proportions" of bacteria in diverse Earth environments have been cultured. Here we reanalyze a portion of the data in that paper, and argue that the conclusion is based on several technical errors, most notably a calculation of sequence similarity that does not account for sequence gaps, and the reliance on 16S rRNA gene amplicons that are known to be biased towards cultured organisms. We further argue that the paper is also based on a conceptual error: namely, that sequence similarity cannot be used to infer "culturability" because one cannot infer physiology from 16S rRNA gene sequences. Combined with other recent, more reliable studies, the evidence supports the conclusion that most bacterial and archaeal taxa remain uncultured.