Current laboratory protocols for detecting fish species with environmental DNA optimize sensitivity and reproducibility, especially for more abundant populations

Abstract Analysing environmental DNA (eDNA) in seawater can aid in monitoring marine fish populations. However, the extent to which current methods optimize fish eDNA detection from water samples is unknown. Here, we test modifications to laboratory components of an eDNA metabarcoding protocol targeting marine finfish. As compared to baseline methods, amplifying a smaller proportion of extracted DNA yielded fewer species, and, conversely, amplifying a larger proportion identified more taxa. Higher-read species were amplified more reproducibly and with less variation in read number than were lower-read species. Among pooled samples, 20-fold deeper sequencing recovered one additional fish species out of a total of 63 species. No benefit was observed with additional PCR cycles, alternative primer concentrations, or fish-selective primers. Experiments using an exogenous DNA standard to assess absolute eDNA concentration suggested that, for a given proportion of a DNA sample, current laboratory methods for metabarcoding marine fish eDNA are near to maximally sensitive. Our results support the unofficial standard collection volume of one liter for eDNA assessment of commonly encountered marine fish species. We conclude that eDNA rarity poses the main challenge to current methods.