Data from: Testing genome skimming for species discrimination in the large and taxonomically difficult genus Rhododendron

Standard plant DNA barcodes based on 2-3 plastid regions, and nrDNA ITS
show variable levels of resolution, and fail to discriminate among species
in many plant groups. Genome skimming to recover complete plastid genome
sequences and nrDNA arrays has been proposed as a solution to address
these resolution limitations. However, few studies have empirically tested
what gains are achieved in practice. Of particular interest is whether
adding substantially more plastid and nrDNA characters will lead to an
increase in discriminatory power, or whether the resolution limitations of
standard plants barcodes are fundamentally due to plastid genomes and
nrDNA not tracking species boundaries. To address this, we used genome
skimming to recover near-complete plastid genomes and nuclear ribosomal
DNA from Rhododendron species and compared discrimination success with
standard plant barcodes. We sampled 218 individuals representing 145
species of this species-rich and taxonomically difficult genus, focusing
on the global biodiversity hotspots of the Himalaya-Hengduan Mountains.
Only 33% of species were distinguished using ITS+matK+rbcL+trnH-psbA. In
contrast, 55% of species were distinguished using plastid genome and nrDNA
sequences. The vast majority of this increase is due to the additional
plastid characters. Thus, despite previous studies showing an asymptote in
discrimination success beyond 3-4 plastid regions, these results show that
a demonstrable increase in discriminatory power is possible with extensive
plastid genome data. However, despite these gains, many species remain
unresolved, and these results also reinforce the need to access multiple
unlinked nuclear loci to obtain transformative gains in species
discrimination in plants.