Early detection of aquatic invaders using metabarcoding reveals a high number of non-indigenous species in Canadian ports

Aim Invasive species represent one of the greatest threats to biodiversity. The ability to detect non-indigenous species (NIS), particularly those present at low abundance, is limited by difficulties in performing exhaustive sampling and in identifying species. Here we sample zooplankton from 16 major Canadian ports and apply a metabarcoding approach to detect NIS. Location Marine and freshwater ports along Canadian coastlines (Pacific, Arctic, Atlantic) and the Great Lakes. Methods We amplified the V4 region of the small subunit ribosomal DNA (18S) and used two distinct analytic protocols to identify species present at low abundance. Taxonomic assignment was conducted using BLAST searches against a local 18S sequence database of either (i) individual reads (totalling 7,733,541 reads) or (ii) operational taxonomic units (OTUs) generated by sequence clustering. Phylogenetic analyses were performed to confirm the identity of reads with ambiguous taxonomic assignment. Results Taxonomic assignment of individual reads identified 379 zooplankton species at a minimum sequence identity of 97%. Of these, 24 species were identified as NIS, 11 of which were detected in previously unreported locations. When reads were clustered into OTUs prior to taxonomic assignment, six NIS were no longer detected and an additional NIS was falsely identified. Phylogenetic analyses revealed that sequences belonging to closely related species clustered together into shared OTUs as a result of low interspecific variation. NIS can thus be misidentified when their sequences join the OTUs of more abundant native species. Main conclusions Our results reveal the power of the metabarcoding approach, whilst also highlighting the need to account for potentially low levels of genetic diversity when processing data, to use barcode markers that allow differentiation of closely related species and to continue building comprehensive sequence databases that allow reliable and fine-scale taxonomic designation.