Self-incompatibility (SI) is a genetic mechanism that allows plants to enforce outcrossing by rejecting self-pollen and pollen from close relatives. In the Brassicaceae, SI is sporophytic and controlled by the self-incompatibility locus (S-locus). The S-locus harbors two tightly linked genes SRK and SCR, which encode the female and male SI specificity determinants, respectively. S-locus heterozygotes often only express the S-specificity of the more dominant allele, and at the pollen level such dominance relationships are mediated by small RNAs (sRNAs). The S-locus is thus an example of a locus under strong balancing selection, where dominance modifiers have evolved.

In this thesis, I investigate the consequences of S-locus dominance for plant mating system evolution and allopolyploid speciation. I further investigate evolutionary conservation and sequence-level effects of dominance relationships among S-alleles. For this purpose, I used the crucifer genus Capsella as a model system.

First, I demonstrated that targeted long-read sequencing results in structurally accurate assemblies of full-length S-haplotype sequences, and that indel errors in such assemblies can be corrected using short reads. Second, I investigated the genetic basis of loss of SI, the first step in the evolution of self-fertilisation, in the self-compatible (SC) Capsella orientalis. I found that loss of SI was dominant and mapped to the S-locus, where C. orientalis harbored a fixed coding frameshift deletion in SCR that is likely to lead to loss of male specificity. I further identified a sRNA-based dominance modifier that is associated with dominant suppression of recessive SCR alleles. Taken together, these results suggest that loss of SI in C. orientalis involved a dominant S-haplotype, suggesting that dominant haplotypes may be favored under conditions that select for loss of SI. Third, I show that a dominant S-haplotype may also have contributed to the shift to SC in the widespread allotetraploid Capsella bursa-pastoris. Fourth, I showed that dominance relationships at the S-locus are largely conserved between the SI outcrossing species C. grandiflora and Arabidopsis halleri which diverged ~8 Mya. I also found that dominant S-haplotypes accumulate more transposable elements than recessive S-haplotypes, in line with expected sequence-level consequences of S-locus dominance. In sum, this thesis provides new insights into the broad conservation of dominance hierarchies at the Brassicaceae S-locus, and the role of dominant S-alleles in allopolyploid speciation and plant mating system shifts.