DNA barcoding insufficiently identifies European wild bees (Hymenoptera, Anthophila) due to taxonomic problems, genus-specific barcoding gaps, and database errors

Recent declines in insect abundances, especially populations of wild pollinators, pose a threat to many natural and agricultural ecosystems. Traditional species monitoring relies on morphological character identification and is inadequate for efficient and standardized surveys. DNA barcoding has become a standard approach for molecular identification of organisms, aiming to overcome the shortcomings of traditional species monitoring. However, its efficacy depends on the completeness of reference databases. Large DNA barcoding efforts are (almost entirely) lacking in many European countries, and such patchy data limit Europe-wide analyses of precisely how to apply DNA barcoding in wild bee identification. Here, we advance towards an effective molecular identification of European wild bees. We conducted a high-effort survey of wild bees in Slovenia, a country where central Europe meets the Balkan peninsula, and DNA barcoded all collected morpho-species. For global analyses, we complemented our DNA barcode dataset with all relevant European species and conducted a global analyses of species delimitation, general and genus-specific barcoding gaps, and examined the error-rate in DNA data repositories. We found that i) a sixth of all specimens from Slovenia could not be reliably identified, ii) species delimitation methods show numerous systematic problems, iii) there is no general barcoding gap across all bees, iv) the barcoding gap is genus-specific, but only after curating for errors in DNA data repositories. Intense sampling and barcoding efforts in underrepresented regions and strict curation of DNA barcode repositories are needed to enhance the use of DNA barcoding for identification of wild bees.