Structure-informed evolutionary analysis of the meiotic recombination machinery.

Despite being essential for fertility, many proteins involved in meiotic homologous recombination have diverged rapidly. The evolutionary forces driving this divergence remain mostly unknown, in part because of challenges in accounting for the interplay of sequence changes with constraints imposed by proteins' structures and physiological roles. Here, we explore strategies to more sensitively detect signatures of positive or relaxed selection by integrating evolutionary analyses with structural and functional information, using meiotic recombination proteins in four taxa-primates, rodents, birds and budding yeasts. By mapping selection rate estimates onto predicted protein structures, we characterized protein regions likely to have experienced positive selection. We further identified subtle sequence variation within protein domains that are well conserved generally because of structural constraints. To detect sequence variation masked by these constraints, we analyzed selection at structurally matched residues, comparing homologs across different lineages as well as between meiosis-specific and generalist paralogs. These approaches identified lineage- and paralog-restricted enrichment of non-synonymous substitutions that may indicate loss of functional constraints and/or adaptive innovation. Finally, we used cross-species complementation experiments in