Mutation rate, whether per locus or genome-wide, features as a variable in many calculations that underlie evolutionary genetic hypotheses. But a true experimental estimate of mutation rate is difficult to obtain. To this end, several groups have used 'mutation-accumulation' lines in both Drosophila melangaster and Caenorhabditis elegans: in an attempt to minimize selection pressure, genetically identical flies or worms are maintained for many generations in a benign environment. Such assays are long and tedious: for one collection, scientists even separated each worm in each generation from everyone else to avoid any selection pressure induced by the stressful dating scene, and instead, self-fertilize.

Previous studies have used fitness-based assays to assess the number of mutations that developed in each line, yielding an indirect estimate of mutation rate. For a more direct estimate, Denver and colleagues have now revisited a collection of these special worm lines and sequenced more than 4 Mb of loci scattered around the genome. In a recent paper in Nature, they report the total haploid genomic mutation rate to be approximately 2.1 mutations per genome per generation — an estimate that is an order of magnitude higher than previous best guesses, and 2 orders higher than the indirect estimate from the same collection (although the previous estimates referred to deleterious mutation rate, whereas these authors estimate total mutation rate). Not only that, but the more frequently observed mutations were insertions, in contrast to reports based on pseudogenes that indicated that most naturally occurring mutations in worms are deletions.

So, with one study, Denver and colleagues prompt the entire field to rethink the process of mutation over time and our measurement of it. It does not take long for this to generate controversy: in a thought-provoking News and Views piece, Rosenberg and Hastings speculate on the mechanisms at work in the Denver study. Either previous estimates were wrong because they only detected mutations that produce phenotypes, as Denver et al. would suggest, or the new study uses methods that push the worms to develop more mutations even in the absence of deleterious selection, maybe even through triggering stress-response pathways. Future studies might have to bring the worms out of their posh retirement to settle the question.