Recombination in Maize is Stable, Predictable, and Associated with Genetic Load: A Study of the US and Chinese Maize NAM Populations

Among the fundamental evolutionary forces, recombination arguably has the largest impact on the practical work of plant breeders. Varying over 1,000-fold across the maize genome, the local meiotic recombination rate limits the resolving power of quantitative trait mapping and the precision of favorable allele introgression. The consequences of low recombination also theoretically extend to the species-wide scale by decreasing the power of selection relative to genetic drift and thereby hindering the purging of deleterious mutations. In this study we used genotyping-by-sequencing (GBS) to identify 136,000 recombination breakpoints at high resolution within United States and Chinese maize nested association mapping populations. We find that the pattern of crossovers is highly predictable on the broad scale, following the distribution of gene density and CpG methylation. Several large inversions also suppress recombination in distinct regions of several families. We also identify recombination hotspots ranging in size from 1kb to 30kb. We find these hotspots to be historically stable and, when compared to similar regions with low recombination, to have strongly differentiated patterns of DNA methylation and GC content. We also provide evidence for the historical action of GC-biased gene conversion in recombination hotspots. Finally, using Genomic Evolutionary Rate Profiling (GERP) to identify putative deleterious polymorphisms, we find evidence for reduced genetic load in hotspot regions, a phenomenon that may have considerable practical importance for breeding programs worldwide.