PLoS Biology: mutation rates of E. coli and yeast genes do not depend on gene importance
A multi-parameter analysis of mutation accumulation data for wild-type E. coli and baker's yeast found no lower mutation rates in genes for essential or abundant proteins.
A team led by Laurence Hurst, with Sofia Radrizzani as lead author, analysed accumulated spontaneous-mutation data for wild-type Escherichia coli and baker's yeast Saccharomyces cerevisiae. The work, published in PLoS Biology, tested whether the mutation density of a gene (mutations per base pair) depends on how important the gene is plos.org
PLoS Biologyplos.orgOpen source ↗. Radrizzani describes the dataset as the largest and cleanest set of mutations collected to date.
The idea under test
The proposal that important genes mutate less often dates to 1972 and has resurfaced periodically, most recently in 2023. In its selectionist form, genes of high importance are protected from mutation, in prokaryotes partly by lying on the leading strand of DNA replication, while genes needed for rapid adaptation mutate more often, partly by lying on the lagging strand. Earlier replication attempts gave contradictory results: none fully allowed for covariates, and many used strains lacking mismatch repair (MMR).
The authors used wild-type strains and a multiple-parameter analysis. Gene importance was captured by protein level and essentiality; in E. coli the strand relative to DNA replication was added. Observed correlations and partial correlations were compared with values expected from trinucleotide-controlled simulations.
What the data showed
In neither species did genes for more abundant or essential proteins have lower mutation rates plos.org
PLoS Biologyplos.orgOpen source ↗. In E. coli, more highly transcribed genes may have slightly lower rates, which could be consistent with a selectionist model, but this was not seen in yeast. In E. coli there was no difference in mutation rate between genes on different strands, and the same lack of difference was found in other prokaryotes. The degree of purifying selection did not predict mutation density either.
The authors also report a strong artefact in the standard off-the-shelf statistical approach and note that results from MMR-deficient strains are discordant. They conclude that the trends are nearly uniformly unsupportive, inconsistent between species and of minuscule effect size, so they present no challenge to the convention that mutation is random.
According to Hurst, earlier observations of fewer mutations in important genes may reflect selection removing harmful mutations before they can be analysed; he calls these "ghost mutations". The only reliable factor the authors identified was replication timing: genes copied later in DNA replication tend to mutate more often.
