Abstract
With the increasing availability of whole-genome assemblies, interest in the genomic consequences of diverse reproductive modes has grown. Reduced efficiency of selection in asexually reproducing species is often cited as a major consequence and is frequently investigated using the ratio of nonsynonymous to synonymous substitutions (dN/dS). However, many studies do not give sufficient attention to the fundamental differences between nonsexual reproductive modes, which likely have substantial effects on the efficiency of selection. We reviewed studies that compared dN/dS ratios of sexual and asexual relatives across 20 taxa. Asexual reproductive modes were classified into two categories: Meiosis without interchromosomal mixis (automixis and premeiotic doubling), i.e. modified meiosis that restores ploidy without outcrossing, and apomixis, i.e. production of unreduced eggs that develop without meiosis and without interchromosomal or intrachromosomal mixis. We further included two modes of reproduction which are not sexual and not strictly asexual, but may be seen as such at first glance: Fissiparity/clonal growth, i.e. the formation of new modules/ramets by splitting of somatic tissue without meiosis or mixis, and selfing, i.e. meiosis followed by fertilization of the gametes by the same individual. Consistent with theory, fissiparous lineages generally exhibited elevated dN/dS ratios, whereas automicts and apomicts showed no consistent pattern, even among taxa with identical reproductive modes. We discuss limitations of commonly used phylogenetic analyses using maximum likelihood (PAML) branch and branch-site models, which can obscure subtle differences, and propose different avenues to assess detection of selection. We propose that the heterogeneous patterns in asexual lineages reflect lineage-specific consequences of asexuality, model limitations, and our incomplete understanding of cytological mechanisms underlying different nonsexual reproductive modes.