Abstract
The use of environmental DNA (eDNA) methods for community analysis has recently been developed. High-throughput parallel DNA sequencing (HTS), called eDNA metabarcoding, has been increasingly used in eDNA studies to examine multiple species. However, eDNA metabarcoding methodology requires validation based on traditional methods in all natural ecosystems before a reliable method can be established. To date, relatively few studies have performed eDNA metabarcoding of fishes in aquatic environments where fish communities were intensively surveyed using multiple traditional methods. Here, we have compared fish communities' data from eDNA metabarcoding with seven conventional multiple capture methods in 31 backwater lakes in Hokkaido, Japan. We found that capture and field surveys of fishes were often interrupted by macrophytes and muddy sediments in the 31 lakes. We sampled 1 L of the surface water and analyzed eDNA using HTS. We also surveyed the fish communities using seven different capture methods, including various types of nets and electrofishing. At some sites, we could not detect any eDNA, presumably because of the polymerase chain reaction (PCR) inhibition. We also detected the marine fish species as sewage-derived eDNA. Comparisons of eDNA metabarcoding and capture methods showed that the detected fish communities were similar between the two methods, with an overlap of 70%. Thus, our study suggests that to detect fish communities in backwater lakes, the performance of eDNA metabarcoding with the use of 1 L surface water sampling is similar to that of capturing methods. Therefore, eDNA metabarcoding can be used for fish community analysis but environmental factors that can cause PCR inhibition, should be considered in eDNA applications.