Under the selfish centromere model, costs associated with female meiotic drive by centromeres select on interacting kinetochore proteins to restore Mendelian inheritance. We directly test this model in yellow monkeyflowers (Mimulus guttatus), which are polymorphic for a costly driving centromere (D). We show that the D haplotype is structurally and genetically distinct and swept to a high stable frequency within the past 1500 years. Quantitative genetic analyses reveal that variation in the strength of drive primarily depends on the identity of the non- D centromere, but also identified an unlinked modifier coincident with kinetochore protein Centromere-specific Histone 3 A (CenH3A). CenH3A has also experienced a recent (<1000 years) selective sweep in our focal population, consistent with ongoing interactions with D shaping its evolution. Together, our results demonstrate an active co-evolutionary arms race between the DNA and protein components of the meiotic machinery, with important consequences for individual fitness and molecular divergence. 

These are F₂ genotype-phenotype data used for mapping of loci contributing to differences in the strength of heterospecific vs. conspecific centromere-associated drive by the driving D haplotype at MDL11 in Mimulus nasutus x M. guttatus hybrids. The %D_fem phenotype is the inferred female transmission of D allele (vs. nondriving d from M. nasutus or D^- from M. guttatus) to selfed F₃ progeny, assuming Mendelian transmission through male function. Only F₂s with at least 12 F₃ progeny genotyped were included in our analyses, but all genotyped individuals were used for estimation of D transmission to F₂s and are included for completeness. The CenH3A genotype (NN, GG, NG₇₆₇ and NG₁₆₀ for the two classes of interspecific heterozygotes) are from an exon-primed intron-spanning PCR marker that could distinguish all 3 parental alleles. The genome-wide mapped genotypes (coded as N, G, H) were derived from multiplex-shotgun-genotyping, window-calling, and mapping protocols described in Flagel et al. 2018 (PLOS Computational Biology). Sheet 1 has the genotype-phenotype matrix, sheet 2 the marker positions on the genetic linkage map and the v1 contigs of M. guttatus, and v1 windows can be matched to v2 and newer genome assemblies (with gene annotations) via the information in sheet 3.

The genotype-phenotype values provided are those analyzed, including missing data (as empty cells) as well as data excluded from the phenotypic analyses (individuals with >12  F3 progeny). Further useful details of the genotyping and phenotyping protocols are given in the paper.