The detection of butterflies with DNA metabarcoding in Malaise trap bulk samples

Abstract The application of DNA metabarcoding to insect bulk samples still requires some guidance on how to interpret the data provided, including taxon lists and read counts. To this end, it is particularly important to gain insights from samples taken under real field conditions. In our study, we compared butterfly data from Malaise trap samples identified by an expert and DNA metabarcoding. We calculated biomass, species richness, the mean detection congruence, and identified added and missed species records (detection of a species in a sample). We tested if the latter had a lower share of total biomass than the overlapping records, which could be one of the reasons for missing detections. Finally, the relative read count was plotted against the relative biomass to explore the relationship between the two variables. Across all samples, we found a high overall congruence of species, with 23 of 27 overlapping between methods. The mean congruence was 91%. Of the 126 species records, 111 overlapped, 20 were added and 15 were missed in the DNA metabarcoding. Missed records had a significantly lower share of total biomass compared to overlapping records. Most affected by added detections were Lycaenidae. A positive correlation was observed between the relative read count and relative butterfly biomass; however, the explanatory power of the read count was considered low. Implications for insect conservation Our study improves the understanding of the butterfly taxon lists and the read counts generated from DNA metabarcoding and provides guidance on how to use the data for conservation purposes.