Gene Importance Does Not Predict Mutation Rates in Bacteria and Yeast

Between-gene variation in mutation rates is not predicted by the selective importance of genes, according to a multiple parameter-controlled analysis of wild-type microbial genomes. Researchers Sofia Radrizzani and Laurence D. Hurst evaluated whether natural selection actively protects vital genomic sequences from DNA changes, finding that putative markers of genic importance fail to forecast mutational density. The findings show that genes encoding essential proteins or more abundant cellular products do not experience lower rates of mutation, confirming that mutational processes conform to long-standing evolutionary conventions rather than adaptive protection models.

The established convention that genetic mutation occurs as a random process had previously faced challenges from selectionist hypotheses proposing that natural selection shapes mutation rate variation between different genes. Under these adaptive models, vital genes were claimed to be shielded from mutations in prokaryotes by being situated on the leading replication strand, whereas genes required for rapid adaptation were hypothesized to exhibit elevated mutation rates on the lagging strand. While these original selectionist claims sparked controversy following several replication failures, previous replication studies failed to account fully for underlying covariates and frequently relied on data collected from mismatch repair knockout strains.

To address these limitations, the study examined amassed mutation accumulation datasets from wild-type Escherichia coli, the primary focal organism for adaptive claims, and the yeast Saccharomyces cerevisiae. The authors implemented an analytical framework controlling for multiple parameters to predict between-gene variation in mutational density, measured as mutations per base pair. Alongside relevant covariates, the investigators evaluated selectively relevant parameters, specifically essentiality and protein levels to represent genic importance, as well as replication strand orientation in Escherichia coli. The statistical assessment evaluated deviations of correlation and partial correlation values against baseline expectations generated by trinucleotide-controlled simulations.

The analysis revealed that neither essentiality nor elevated protein abundance correlates with reduced mutation rates in either Escherichia coli or Saccharomyces cerevisiae. Although more highly transcribed genes in Escherichia coli displayed slightly lower mutation rates, a pattern potentially consistent with selectionist expectations, this relationship was absent in yeast. Furthermore, Escherichia coli exhibited no difference in mutation rates between genes located on the leading versus the lagging strand, mirroring results confirmed across other prokaryotes. The degree of purifying selection acting on genes likewise failed to predict mutational density across the analyzed genomes.

The authors also identified a substantial statistical artefact that arises when using off-the-shelf analytical methods, while noting that mutational patterns derived from mismatch repair-deficient strains are discordant with wild-type observations. Because the observed empirical trends are nearly uniformly unsupportive of adaptive mutation models, inconsistent between divergent species, and characterized by miniscule effect sizes, the researchers conclude that existing between-gene variation provides no challenge to the convention of random mutation.

Subject of Research: Between-gene variation in mutation rates and genic importance in microorganisms

Article Title: The importance of genes does not predict their mutation rate

Article References: Radrizzani, S., & Hurst, L. D. (2026). The importance of genes does not predict their mutation rate. PLOS Biology, 24(11), e3004089. https://doi.org/10.1371/journal.pbio.3004089

Image Credits: AI Generated

DOI: 10.1371/journal.pbio.3004089

Keywords: mutation rate, genic importance, Escherichia coli, Saccharomyces cerevisiae, mutation accumulation, mutational density, purifying selection, leading strand, lagging strand, mismatch repair, protein abundance, essential genes, DNA replication, selectionist models, trinucleotide simulations, covariate control, partial correlation