Response to No gene-specific optimization of mutation rate in Escherichia coli

TL;DR

This paper defends the evolution of local mutation rate variation in E. coli against critiques, clarifies the influence of mutational processes, and reanalyzes data to support the original findings.

Contribution

It provides a theoretical and empirical rebuttal to claims against mutation rate optimization, reaffirming the evolution of mutation rate variation within genomes.

Findings

Long-standing mutation-modifier theory supports local mutation rate evolution.

Caution is needed when comparing data from different mutational contexts.

Reanalysis of data undermines previous conclusions about mutation rate associations.

Abstract

In a letter published in Molecular Biology Evolution [10], Chen and Zhang argue that the variation of the mutation rate along the Escherichia coli genome that we recently reported [3] cannot be evolutionarily optimised. To support this claim they first attempt to calculate the selective advantage of a local reduction in the mutation rate and conclude that it is not strong enough to be favoured by selection. Second, they analyse the distribution of 166 mutations from a wild-type E. coli K12 MG1655 strain and 1,346 mutations from a repair-deficient strain, and claim to find a positive association between transcription and mutation rate rather than the negative association that we reported. Here we respond to this communication. Briefly, we explain how the long-standing theory of mutation-modifier alleles supports the evolution of local mutation rates within a genome by mechanisms acting on sufficiently large regions of a genome, which is consistent with our original observations [3,4]. We then explain why caution must be exercised when comparing mutations from repair deficient strains to data from wild-type strains, as different mutational processes dominate these conditions. Finally, a reanalysis of the data used by Zhang and Chen with an alternative expression dataset reveals that their conclussions are unreliable.