Catching small fish in a big pond: optimizing fish environmental DNA detections

Biomonitoring of teleost fish using environmental DNA metabarcoding is frequently performed using the 12S ribosomal RNA marker gene, especially with the MiFish primer set. These primers have associated laboratory and bioinformatic processing challenges that can generate spurious, incomplete, or inaccurate results. Here, we address each challenge to minimize off-target amplification and maximize the accuracy of observed species richness and community composition of the target taxa (teleost fish). We first tested the effects of amplification and sequencing library preparation protocols. We next tested infrequently used parameters in the common bioinformatic software program DADA2 using both environmental samples and mock communities with known composition. Our results support removing known off-target amplicons prior to sequencing and demonstrate that certain parameter settings can maximize the number of fish amplicon sequence variants generated bioinformatically. Furthermore, while changes in OMEGA_A and pooling parameters may yield additional amplicon sequence variants and previously undetected taxa, these amplicon sequence variants generally do not improve biodiversity estimates and are more likely to have associated uncertainty in taxonomic annotations. Based on our results, we recommend the following approaches using the MiFish 12S marker for marine biomonitoring: 1) Incorporate amplicon size selection prior to sequencing, 2) Consider increasing OMEGA_A and investing in manual curation of results.