Strengths and weaknesses of DNA-based monitoring: Assessing macroinvertebrates in 18 Finnish streams with metabarcoding and morphology
1) DNA metabarcoding holds great promise for assessment of stream ecosystems with macroinvertebrates. However, few large-scale studies have compared the performance of DNA metabarcoding with that of routine morphological identification. 2) We tested metabarcoding using 18 macroinvertebrate samples from Finland using four primer sets. The samples were collected in 2013 and identified based on morphology as part of a Finnish stream monitoring program. Morphological identification was performed to the taxonomic level at which identification was reliable following standardized protocols. 3) We identified over twice the number of taxa, with greater species-level resolution, using DNA metabarcoding than morphology-based identification. For each sample, we detected more taxa by metabarcoding than by previous morphological methods, and all four primer sets showed similarly good performance. There was a significant linear correlation between sequence abundance and the number of taxa in each sample, but the scatter was up to two orders of magnitude. Ecological status assessment indices calculated from morphological and DNA metabarcoding datasets were mostly similar, with a few exceptions. With the recent drop in sequencing costs per sample, both methods identification are currently equally expensive. 4) We used actual samples for monitoring to demonstrate that DNA metabarcoding can achieve similar results and better taxonomic resolution than current morphological identification methods. Metabarcoding has thus already become a viable and reliable invertebrate identification method for stream assessment. However, to unlock the full potential of DNA metabarcoding for ecosystem assessment key problems in current laboratory protocols and reference databases, specified in this work, will require further attention.