Improved <scp>COI</scp> taxonomic assignment workflow enhances detection of invertebrates in environmental <scp>DNA</scp>

Abstract Environmental DNA (eDNA) metabarcoding is an increasingly popular and powerful tool in ecology, offering a non‐invasive method to assess biodiversity and monitor ecosystem health across diverse environments. While the mitochondrial cytochrome c oxidase subunit I (COI) gene is one of the most widely used markers for studying eukaryotic community dynamics, its taxonomic assignment process often suffers from low confidence and inconsistent results. In this study, we employed COI metabarcoding to investigate changes in invertebrate communities in Cultus Lake (British Columbia, Canada), over a 36‐month time series. To enhance taxonomic accuracy, we developed an enhanced COI workflow incorporating: (1) bacterial‐inclusive reference databases, (2) protein‐level alignment and cleanup procedures, and (3) taxon‐specific nucleotide refinements, thereby accounting for bacterial sequences, codon degeneracy, pseudogenes, and other erroneous sequences. Compared to 18S rRNA metabarcoding of the same samples, our improved COI workflow detected greater invertebrate diversity in Arthropoda and Rotifera, provided higher taxonomic resolution, and showed stronger correlations with microscopic zooplankton counts. Based on these findings, we recommend COI metabarcoding with our enhanced workflow for eDNA‐based studies of aquatic invertebrate communities.