Evolution of self-fertilization in plants

Association between selfing-rate modifier and inbreeding depression

Because selfed offspring are generally less vigorous than outcrossed offspring (inbreeding depression), early naturalists thought that the evolution of self-fertilization must be maladaptive. However, Fisher (1941) proposed a two-fold transmission advantage of selfing; two sets of genes are transmitted to each offspring through selfing, whereas a single set is transmitted through outcrossing. This strong selection for selfing (automatic selection) leads to the fixation of completely selfing genotypes unless there is a counteracting selective force such as severe inbreeding depression. More recent coevolutionary models, where mating system and level of inbreeding depression are allowed to evolve simultaneously, have shown that the dynamics of mating-system evolution can be more complex than previously suspected. However, many aspects of these coevolutionary models remain untested.

One untested aspect of the coevolutionary theory was the existence of an association between a modifier locus controlling selfing rate and viability (inbreeding depression). Using an annual plant endemic to California, Gilia achilleifolia (Polemoniaceae), as a model system, I showed that the degree of herkogamy (spatial separation between anther and stigma) influences selfing rate of individual plants within a population with isozyme analysis (Takebayashi and Delph in prep). Considering herkogamy to be a selfing rate modifier, we provided the first evidence for an association between a floral trait controlling the selfing rate and level of inbreeding depression within a population (Takebayashi and Delph 2000). However, a number of questions remain, such as the genetic basis of this association, which I plan to investigate next.

Is selfing an evolutionary dead-end?

At a higher taxonomic level, I examined current phylogenetic data to determine if self-fertilization is an evolutionary dead-end. This study was motivated by the question: why do most plants outcross (only 20-25% of plant taxa are predominantly selfing) if the advantage of selfing (automatic selection) is so large? One potential answer is that selfing may have a short-term transmission advantage, but that it is an evolutionary dead-end in the long term. This is because the loss of genetic diversity may preclude adaptive evolution, including reversion from selfing to outcrossing. My review of the literature and re-analysis of existing data, published in American Journal of Botany as a Special Invited Paper (Takebayashi and Morrell 2001), suggests that this old hypothesis is well supported in light of current evolutionary genetic theory. Furthermore, it appears that most of the current phylogenetic data are in concordance with the hypothesis; that is, selfing taxa seem to be in terminal clades. However, the large-scale phylogenetic data that could be used to quantitatively assess the hypothesis are not yet available. In collaboration with Dr. Peter Morrell, I have started a preliminary phylogenetic analysis of Gilia species to overcome the pitfalls which we identified in previous studies.

Polygenic inheritance of selfing-rate modifier

Using a simulation model, I investigated the effects of polygenic inheritance of the selfing rate on the consequence of mating-system evolution. This study was motivated by the discrepancy between current models and empirical observations in plants. Most models describing the evolution of selfing rate assume that selfing rate is controlled by a single locus, whereas, in reality, many floral traits influence selfing rate and each of these traits is polygenic. I found that under polygenic inheritance, the optimal selfing rate may not be always realized if the population size is finite (Takebayashi and Repasky in prep). Furthermore, deviations from the optimal selfing rates depended on the underlying genetic architecture; that is, how multi-locus genotype get mapped to the phenotype (selfing-rate).