Tetrad analysis in plants and fungi finds large differences in gene conversion rates but no GC bias

GC-favouring gene conversion enables fixation of deleterious alleles, disturbs tests of natural selection and potentially explains both the evolution of recombination as well as the commonly reported intragenomic correlation between G+C content and recombination rate. In addition, gene conversion disturbs linkage disequilibrium, potentially affecting the ability to detect causative variants. However, the importance and generality of these effects is unresolved, not simply because direct analyses are technically challenging but also because previous within- and between-species discrepant results can be hard to appraise owing to methodological differences. Here we report results of methodologically uniform whole-genome sequencing of all tetrad products in Saccharomyces , Neurospora , Chlamydomonas and Arabidopsis . The proportion of polymorphic markers converted varies over three orders of magnitude between species (from 2% of markers converted in yeast to only ~0.005% in the two plants) with at least 87.5% of the variance in per tetrad conversion rates being between species. This is largely due to differences in recombination rate and median tract length. Despite three of the species showing a positive GC-recombination correlation, there is no significant net AT→GC conversion bias in any of the species, despite relatively high resolution in the two taxa ( Saccharomyces and Neurospora ) with relatively common gene conversion. The absence of a GC bias means that: (1) there should be no presumption that gene conversion is GC biased, or (2) that a GC-recombination correlation necessarily implies biased gene conversion, (3) K a / K s tests should be unaffected in these species and (4) it is unlikely that gene conversion explains the evolution of recombination. GC-biased gene conversion has many implications for evolutionary genetics, but its prevalence is unknown. Here, direct sequencing of tetrads in four species shows large differences in gene-conversion rate but no evidence of overall GC bias.