Molecular genetics and population genomics of the maintenance and breakdown of the floral polymorphism tristyly

Tristyly has evolved independently in several flowering plant families and functions to promote outcrossing by disassortative mating. The floral polymorphism is susceptible to evolutionary breakdown resulting in frequent transitions from outcrossing to predominant self-fertilization. By integrating experimental studies with analyses of genomic data and population genetic computer simulations, my Ph.D. thesis investigates the genetic and population genomic basis for the maintenance of this genetic polymorphism, and its breakdown to selfing in the annual aquatic Eichhornia paniculata (Pontederiaceae).;The transition from outcrossing to selfing in E. paniculata is associated with multiple, independent, long-distance colonization events from N.E. Brazil to the Caribbean and Central America, and a reduction in the effective population size of selfing lineages. The joint influence of these genetic and demographic factors was associated with several genomic consequences, including an increased frequency of effectively neutral mutations, more efficient removal of very strongly deleterious mutations, and the down-regulation of ∼1000 genes. These findings suggest that substantial differences in patterns of muttion and gene expression may accompany transitions from outcrossing to selfing in plants.;Genetic crosses and progeny tests conducted under glasshouse conditions demonstrated strong linkage between the S and M loci governing tristyly, with the loci separated by a map distance of 2.7 cM. Controlled crosses using semi-homostylous mid- and long-styled selfing variants from the Caribbean (ssMM) and Central America ( ssmm) combined with genetic mapping of a backcross progeny indicated that style length and anther height localized to the center of linkage group 5. This finding strongly suggests that this region contains the M locus. Population genomic data was consistent with the hypothesis that the M locus is maintained by frequency-dependent selection. Finally, QTL mapping studies also demonstrated that the two semi-homstylous variants of E. paniculata from the Caribbean and Central America were governed by different sets of modifier genes unlinked to the M locus supporting their independent origins. My thesis research provides novel insights into the biology of tristyly, genetic architecture of reproductive adaptation, and the genomic consequences of mating system transitions. It provides a contemporary genomic perspective complementing the earlier seminal studies of Darwin and Fisher on tristyly.