DNA barcoding of Rhododendron (Ericaceae), the largest Chinese plant genus in biodiversity hotspots of the Himalaya–Hengduan Mountains

The Himalaya–Hengduan Mountains encompass two global biodiversity hotspots with high levels of biodiversity and endemism. This area is one of the diversification centres of the genus Rhododendron, which is recognized as one of the most taxonomically challenging plant taxa due to recent adaptive radiations and rampant hybridization. In this study, four DNA barcodes were evaluated on 531 samples representing 173 species of seven sections of four subgen-era in Rhododendron, with a high sampling density from the Himalaya–Hengduan Mountains employing three ana-lytical methods. The varied approaches (NJ, PWG and BLAST) had different species identification powers with BLAST performing best. With the PWG analysis, the discrimination rates for single barcodes varied from 12.21 % to 25.19% with ITS < rbcL < matK < psbA-trnH. Combinations of ITS + psbA-trnH + matK and the four barcodes showed the highest discrimination ability (both 41.98%) among all possible combinations. As a single barcode, psbA-trnH per-formed best with a relatively high performance (25.19%). Overall, the three-marker combination of ITS + psbA-trnH + matK was found to be the best DNA barcode for identifying Rhododendron species. The relatively low dis-criminative efficiency of DNA barcoding in this genus (~42%) may possibly be attributable to too low sequence divergences as a result of a long generation time of Rhododendron and complex speciation patterns involving recent radiations and hybridizations. Taking the morphology, distribution range and habitat of the species into account