Two decades of DNA barcoding in the genus Usnea ( Parmeliaceae ): how useful and reliable is the ITS?

We present an exhaustive analysis of the ITS barcoding marker in the genus Usnea s.lat., separated into Dolichousnea , Eumitria , and Usnea including the subgenus Neuropogon , analyzing 1,751 accessions. We found only a few low-quality accessions, whereas information on voucher specimens and accuracy and precision of identifications was of subpar quality for many accessions. We provide an updated voucher table, alignment and phylogenetic tree to facilitate DNA barcoding of Usnea , either locally or through curated databases such as UNITE. Taxonomic and geographic coverage was moderate: while Dolichousnea and subgenus Neuropogon were well-represented among ITS data, sampling for Eumitria and Usnea s.str. was sparse and biased towards certain lineages and geographic regions, such as Antarctica, Europe, and South America. North America, Africa, Asia and Oceania were undersampled. A peculiar situation arose with New Zealand, represented by a large amount of ITS accessions from across both major islands, but most of them left unidentified. The species pair Usnea antarctica vs. U. aurantiacoatra was the most sampled clade, including numerous ITS accessions from taxonomic and ecological studies. However, published analyses of highly resolved microsatellite and RADseq markers showed that ITS was not able to properly resolve the two species present in this complex. While lack of resolution appears to be an issue with ITS in recently evolving species complexes, we did not find evidence for gene duplication (paralogs) or hybridization for this marker. Comparison with other markers demonstrated that particularly IGS and RPB1 are useful to complement ITS-based phylogenies. Both IGS and RPB1 provided better backbone resolution and support than ITS; while IGS also showed better resolution and support at species level, RPB1 was less resolved and delineated for larger species complexes. The nuLSU was of limited use, providing neither resolution nor backbone support. The other three commonly employed protein-coding markers, TUB2 , RPB2 , and MCM7 , showed variable evidence of possible gene duplication and paralog formation, particularly in the MCM7 , and these markers should be used with care, especially in multimarker coalescence approaches. A substantial challenge was provided by difficult morphospecies that did not form coherent clades with ITS or other markers, suggesting various levels of cryptic speciation, the most notorious example being the U. cornuta complex. In these cases, the available data suggest that multimarker approaches using ITS, IGS and RPB1 help to assess distinct lineages. Overall, ITS was found to be a good first approximation to assess species delimitation and recognition in Usnea s.lat., as long as the data are carefully analyzed, and reference sequences are critically assessed and not taken at face value. In difficult groups, we recommend IGS as a secondary barcode marker, with the option to employ more resource-intensive approaches, such as RADseq, in species complexes involving so-called species pairs or other cases of disparate morphology not reflected in the ITS or IGS. Attempts should be made to close taxonomic and geographic gaps especially for the latter two markers, in particular in Eumitria and Usnea s.str. and in the highly diverse areas of North America and Central America, Africa, Asia, and Oceania.