Herbarium phylogenomics: Resolving the generic status of the enigmatic Pseudobartsia (Orobanchaceae)

The millions of herbarium specimens in collections around the world provide historical resources for phylogenomics and evolutionary studies. Many rare and endangered species exist only as historical specimens. Here, we report a case study of the monotypic Pseudobartsia yunnanensis D. Y. Hong (=Pseudobartsia glandulosa[Bentham] W. B. Yu & D. Z. Li: Orobanchaceae) known from a single Chinese collection taken in 1940. We obtained genomic data of Pseudobartsia glandulosa using high‐throughput short‐read sequencing, and then assembled a complete chloroplast genome and nuclear ribosome DNA region in this study. We found that the newly assembled three plastid DNA regions (atpB‐rbcL, rpl16, and trnS‐G) and nuclear ribosomal internal transcribed spacer (nrITS) of Pseudobartsia glandulosa were more than 99.98% similar to published sequences obtained by target sequencing. Phylogenies of Orobanchaceae using 30 plastomes (including 10 new plastomes), using both supermatrix and multispecies coalescent approaches following a novel plastid phylogenomic workflow, recovered seven recognized tribes and two unranked groups, both of which were proposed as new tribes, that is, Brandisieae and Pterygielleae. Within Pterygielleae, all analyses strongly supported Xizangia D. Y. Hong as the first diverging genus, with Pseudobartsia D. Y. Hong as sister to Pterygiella Oliver + Phtheirospermum Bunge (excluding Phtheirospermum japonicum [Thunberg] Kanitz); this supports reinstatement of Pseudobartsia and Xizangia. Although elements of Buchnereae‐Cymbarieae‐Orobancheae and Brandisieae‐Pterygielleae‐Rhinantheae showed incongruence among gene trees, the topology of the supermatrix tree was congruent with the majority of gene trees and functional‐group trees. Therefore, most plastid genes are evolving as a linkage group, allowing the supermatrix tree approach to yield internally consistent phylogenies for Orobanchaceae.