Improved efficiency in eDNA metabarcoding of benthic metazoans by sieving sediments prior to DNA extraction

Abstract Environmental DNA (eDNA) metabarcoding can be used to rapidly characterize the taxon assemblage of benthic communities in sediments and thus has high potential to complement routine regulatory monitoring of benthic impacts. However, when using DNA extracted directly from subtidal sediments, only a small proportion of metazoan reads are obtained regardless of which available universal primers are used. Developing new metazoan‐specific primers for broad taxonomic amplification is very challenging and may not solve this problem; here, we investigate whether sieving sediments prior to DNA extraction provides a solution. The effect of sieving was tested on 84 sediment samples collected from two salmon farms. Average percentage of metazoan reads was 19.53% and 17.10% in nonsieved samples, and 81.03% and 89.92% in sieved samples at the two sites. Sieving effectively removed Ochrophyta taxa (e.g., seaweed and phytoplankton), but did not remove pelagic metazoans. Average percentage of benthic metazoan reads in sieved samples was 4.1 times of that in nonsieved samples (47.29% versus 11.46%) at one site and 5.7 times (20.03% versus 3.52%) at another. Sieving increased the number of benthic metazoan amplicon sequence variants by 27.67% and 51.30% at the two sites. Relative abundances of only a small fraction of benthic metazoan phyla showed significant differences between sieved and nonsieved samples. These differences can be taken into account when designing a benthic monitoring approach using eDNA metabarcoding. Since sediment sieving can be conducted either in the field or laboratory, and it allows many more samples to be multiplexed on one MiSeq run without compromising sequencing depth of benthic metazoan reads, we suggest this method can greatly increase the efficiency of eDNA metabarcoding for assessing benthic environments