Inference of higher-order conifer relationships from a multi-locus plastid data set1

We reconstructed the broad backbone of conifer phylogeny from a survey of 15–17 plastid loci and associated noncoding regions from exemplar conifer species. Parsimony and likelihood analyses recover the same higher-order relationships, and we find strong support for most of the deep splits in conifer phylogeny, including those within our two most heavily sampled families, Araucariaceae and Cupressaceae. Our findings are broadly congruent with other recent studies, and are inferred with comparable or improved bootstrap support. The deepest phylogenetic split in conifers is inferred to be between Pinaceae and all other conifers (Cupressophyta). Our current gene and taxon sampling does not support a relationship between Pinaceae and Gnetales, observed in some published studies. Within the Cupressophyta clade, we infer well-supported relationships among Cephalotaxaceae, Cupressaceae, Sciadopityaceae, and Taxaceae. Our data support recent moves to recognize Cephalotaxus under Taxaceae, and we find strong support for a sister-group relationship between the two predominantly southern hemisphere conifer families, Araucariaceae and Podocarpaceae. A local hotspot of indel evolution shared by the latter two conifer families is identified in the coding portion of one of the plastid ribosomal protein genes. The removal of the most rapidly evolving plastid characters, as defined using a likelihood-based classification of substitution rates for the taxa considered here, is shown to have little to no effect on our inferences of higherorder conifer relationships.