LAUSR

Clematis is one of the largest genera of Ranunculaceae with many phylogenetic problems left to be resolved. Clematis species have considerable genome size of more than 7 Gbp, and there was no whole-genome reference sequence published in this genus. This raises difficulties in acquiring nuclear genome data for its phylogenetic analysis. Previous studies based on Sanger sequencing data, plastid genome data, and nrDNA sequences did not well resolve the phylogeny of Clematis. In this study, we used genome skimming and transcriptome data to assemble the plastid genome sequences, nuclear single nucleotide polymorphisms (SNPs) datasets, and single-copy nuclear orthologous genes (SCOGs) to reconstruct the phylogenetic backbone of Clematis, and test effectiveness of these genome partitioning methods. We also further analyzed the discordance among nuclear gene trees and between plastid and nuclear phylogenies. The results showed that the SCOGs datasets, assembled from transcriptome method, well resolved the phylogenetic backbone of Clematis. The nuclear SNPs datasets from genome skimming method can also produce similar results with the SCOGs data. In contrast to the plastid phylogeny, the phylogeny resolved by nuclear genome data is more robust and better corresponds to morphological characters. Our results suggested that rapid species radiation may have generated high level of incomplete lineage sorting, which was the major cause of nuclear gene discordance. Our simulation also showed that there may have been frequent interspecific hybridization events, which led to some of the cyto-nuclear discordances in Clematis. This study not only provides the first robust phylogenetic backbone of Clematis based on nuclear genome data, but also provides suggestions of genome partitioning strategies for the phylogenomic study of other plant taxa.