Topic: Wrestling with the rosids II: too big to nail--challenges in conducting comprehensive analyses in the angiosperms
Presenter: Douglas E Soltis
With the exception of grasses, perhaps no other angiosperm clade plays a more important terrestrial ecological role than rosids, with 87,500 species and a worldwide distribution. Rosids are hyper-diverse in morphology, habit, and chemistry. Most dominant forest trees are rosids, and many arctic/alpine systems are rosid-dominated. The rosid radiation 115 – 93 Mya represents the rapid rise of angiosperm-dominated forests, dramatically altering the fate of other terrestrial lineages (e.g., insects, amphibians, ferns) that co-diversified in their shadow. Using the rosids as an exemplar, we illustrate how problematic it is to conduct comprehensive distribution and phylogenetic analyses for major clades of angiosperms. This observation contrasts with well-studied (and much smaller) animal groups (birds, mammals) for which near-complete sampling is available, enabling researchers to examine in detail the “layers of life”. Despite the ecological, paleobotanical, economic, and genomic importance of rosids, our knowledge of this clade is remarkably limited along any metric. Efforts to understand rosid biology have been hampered by the lack of a comprehensive phylogeny as well as disparate and scattered resources. We have found only ~29,400 of ~87,500 species with phylogenetically usable DNA data in GenBank (34%). This is in stark contrast to better-sampled clades such as vertebrates (56%) and butterflies (42%). Current DNA sampling is not random but highly biased toward certain subclades, with major portions of the tree remaining dark. The most poorly sampled larger subclades are Myrtales, Malvales, and Celastrales (25-32% coverage). Using comprehensive biodiversity databases with centuries of accumulation (e.g., the Catalogue of Life, OTL, GBIF, iDigBio, and GenBank), we wanted to assess if it is possible to conduct comprehensive phylogenetic and biogeographical distribution analyses for the rosids. We found that only 12,393 species have at least 30 spatial records, and these are biased toward the temperate zone with tropical species poorly characterized, meaning that attempts at niche modeling are severely hampered. Despite the biological importance of the rosid clade and centuries of significant study, available data for synthetic studies of rosid biodiversity are too limited for meaningful analysis, emphasizing how little we know about even ‘well-known’ clades. This also hinders grand synthesis given that other lineages have diversified in response to the rosids.
Keywords: phylogeny, Genomics, Biogeography, ecological niche modeling
Wrestling with the rosids II: too big to nail--challenges in conducting comprehensive analyses in the angiosperms