Utility of a large, multigene plastid data set in inferring higher-order relationships in ferns and relatives (monilophytes)

Premise of the Study: The monilophytes (ferns and relatives)--the third largest group of land plants--exhibit a diverse array of vegetative and reproductive morphologies. Investigations into their early ecological and life-history diversification require accurate, well-corroborated phylogenetic estimates. We examined the utility of a large plastid-based data set in inferring backbone relationships for monilophytes. METHODS: We recovered 17 plastid genes for exemplar taxa using published and new primers. We compared results from maximum-likelihood and parsimony analyses, assessed the effects of removing rapidly evolving characters, and examined the extent to which our data corroborate or contradict the results of other studies, or resolve current ambiguities. Key Results: Considering multifamily clades, we found bootstrap support comparable to or better than that in published studies that used fewer genes from fewer or more taxa. We firmly establish filmy ferns (Hymenophyllales) as the sister group of all leptosporangiates except Osmundaceae, resolving the second deepest split in leptosporangiate-fern phylogeny. A clade comprising Ophioglossaceae and Psilotaceae is currently accepted as the sister group of other monilophytes, but we recover Equisetum in this position. We also recover marattioid and leptosporangiate ferns as sister groups. Maximum-likelihood rate-class estimates are somewhat skewed when a long-branch lineage (Selaginella) is included, negatively affecting bootstrap support for early branches. CONCLUSIONS: Our findings support the utility of this gene set in corroborating relationships found in previous studies, improving support, and resolving uncertainties in monilophyte phylogeny. Despite these advances, our results also underline the need for continued work on resolving the very earliest splits in monilophyte phylogeny.