Nuclear phylogenomics provide evidence to clarify key morphological evolution and whole-genome duplication across rosids

Rosids, comprising 90,000–120,000 species, form a large clade of angiosperms, including extensively studied families with many economically and scientifically important plants. They are also ecologically important, dominating many temperate and tropical ecosystems. Great progress in understanding rosid phylogenetic relationships has facilitated evolutionary studies, but phylogenetic uncertainties remain. To construct a more comprehensive nuclear phylogeny with expanded taxon coverage at the familial levels, we generated 203 new transcriptomes and two shotgun genomes. Along with other available data sets, our sample includes 419 eudicots, including 316 rosids, representing 83 families and all 16 rosid orders. Compared to the 1KP study, our highly resolved rosid phylogeny provides strongly supported internal relationships for one additional order and 16 families. We uncovered cytoplasmic-nuclear discordance for several deep rosid relationships with possible evidence of hybridization/gene flow and incomplete lineage sorting. By tracing ancestral states of morphological characters, we revealed putative floral evolutionary trends in some major clades. We detected strong evidence for 27 putative whole-genome duplication (WGD) events distributed across 20 rosid families, including five novel WGDs. Additionally, our expanded taxon sampling allowed for revised phylogenetic positions of several previously reported WGD events. Most of the supported WGDs correspond to origins of families or large subclades and occurred near times of geological and global climate upheavals, including those at the Cretaceous–Paleogene boundary. Our findings support the idea that large-scale genomic changes and key morphological innovations might have contributed to adaptive evolution and increased biodiversity in rosids.