Abstract
Optimizing environmental DNA (eDNA) metabarcoding to maximize species detection while minimizing costs and effort remains a challenge, particularly in high-throughput workflows. In the marine intertidal zone off Point Conception, California, we compared eDNA richness accumulation between filtration approaches that differed in filter type and downstream processing, field and laboratory replicates, and independent and pooled polymerase chain reaction (PCR) replicates. Ten 10-cm2 capsule filters and 10 5-cm2 syringe disc filters were processed using a standardized laboratory workflow for 12S ribosomal ribonucleic acid (rRNA) (marine vertebrates) and 18S rRNA (metazoans and phytoplankton) metabarcoding, with 10 independent PCR replicates per filter. Because capsules generated larger lysate volumes than discs, the fixed extraction input volume under the standardized workflow resulted in less lysate being processed. A follow-up study pooled 10 PCR replicates for 12S rRNA. When comparing individual PCR replicates, individual filters, or all filters aggregated, discs detected greater marine vertebrate richness than capsules. Probabilistic models indicated that varying effective volume alone could produce similar richness accumulation patterns to those observed, though other filter characteristics were unaccounted for. Minimal differences in richness accumulation were detected between filters for 18S rRNA, and modeling indicated that this was not explained by higher initial copies of 18S rRNA. Meanwhile, vertebrate richness accumulation was similar when adding PCR replicates from the same sample or independent field samples across the short, relatively homogenous intertidal transect. Pooling PCR replicates had no detectable difference on richness compared to sequencing PCR replicates independently: a single pooled PCR replicate detected the same vertebrate richness as the sum of 10 independent PCR replicates. These results show that, under the standardized workflow and conditions tested here, eDNA programs may reduce costs and effort without detectable loss of richness by using filtration approaches aligned with extraction input and pooling technical replicates.