Evolutionary patterns of host type and chasmothecial appendage morphology in obligate plant parasites belonging to Cystotheceae (powdery mildew, Erysiphaceae)

The chasmothecial appendages of Erysiphaceae are considered to function in the overwintering strategy and evolve morphologically in line with transitions of different host type. However, the evolutionary patterns and relationships of these traits have not yet been verified using statistical models based on phylogenetic information. We aimed to clarify the evolutionary process of host type and appendage morphology in Cystotheceae using phylogenetic comparative methods (PCMs) and to evaluate the evolutionary relationship of these traits. The ancestral state estimation of host types showed that the deciduous type is the most ancestral in Cystotheceae, and the herb or evergreen types evolved secondarily four times and twice, respectively. Branched- or circinate-type appendages were estimated to be the most ancestral, and the mycelioid and rudimentary types evolved secondarily thrice and once, respectively. The results of the random forest analysis showed that the host type was predictable from the phylogeny and appendage morphology. The ancestral state estimation suggested that simultaneous transitions of the host type and appendage morphology occurred at several ancestral nodes. These results suggest some functional relationships between host type and appendage morphology, but there was no statistical support for an overall trend in evolutionary dependence between these traits. Our results demonstrate the utility of PCMs in the study of trait evolution in Cystotheceae, which can be applied to a broader phylogeny of powdery mildews to elucidate the evolutionary relationship and functional causality of phenotypic traits.

Automated summary

The study clarified the evolutionary process of host type and appendage morphology in Cystotheceae using phylogenetic comparative methods, showing functional relationships between host type and appendage morphology, but lacking statistical support for an overall trend in evolutionary dependence between these traits.