Discussion
We found that the abundance of eDNA in aquaria was highest just after fish were added and tended to decrease with time. This is consistent with a previous study which showed that the highest abundance of fish eDNA was at the start of experiments, due to the struggles during the acclimation of introduced fish to the environment (Andersen et al., 2012; Barnes et al., 2014; Takahara, Minamoto, Yamanaka Doi, and Kawabata, 2012; Jo, Murakami, Yamamoto, Masuda, and Minamoto, 2019; Sassoubre, Yamahara, Gardner, Block, and Boehm, 2016). However, we found that different mock fish communities had different peaks in eDNA abundance, with MC1 peaking on day 0, and MC2 peaking on day 1. The reason for this is not known, indicating that further studies of eDNA and fish behaviour are needed. The abundances of fish species obtained by HTS were very different from the absolute copy numbers of eDNA quantified with dPCR and qSeq. In general, PCR efficiency can affect the proportions of sequences in sequence libraries. PCR efficiency is affected by primer-target mismatches, composition (e.g., GC content) of sequences, and the bases adjacent to primers (Ben-Dov et al., 2012; Ruijter et al., 2009; Salipante et al., 2014; Sipos et al, 2007). Therefore, these factors may have influenced the proportions of sequences in our fish sequencing library, resulting in different proportions compared with in the original eDNA samples.
In this study, the values quantified by dPCR tended to be slightly higher than those by qSeq. It might be possible that these results are due to non-specific binding of the probes and subsequent digestion, resulting in the overestimation of dPCR. Another possible explanation might be that qSeq-MiFish-U-F did not hybridize to all target eDNA during single primer extension, causing underestimation. Indeed, the sequence of the qSeq-MiFish-U-F primer used for SPE in this study has two mismatches to the sequence of O. latipes .  In PCR, this mismatch resulted in low PCR efficiency; however, the mismatch disappeared once the primer synthesized template DNA because the synthesized DNA did not have a mismatch. In this study, we amplified DNA by a 50-cycle PCR, and thus the effect of the mismatch would not be significant and even DNA with two mismatch sequences was amplified and quantified. In contrast, the SPE reaction in the qSeq protocol occurs only once, and DNA is not amplified in the next PCR step if the first single-stranded DNA failed to be synthesized. Thus, since SPE would be more sensitive to mismatches in the template DNA than PCR, increasing the specificity of the SPE reaction would prevent qSeq from amplifying and quantifying environmental DNA containing mismatches.
This is the first study to employ the qSeq technique for quantifying eDNA in fish species. Using an aquarium experiment with five fish species, we demonstrated that qSeq could quantify eDNA from fish, verified by dPCR. Quantitative sequencing can simultaneously quantify multiple species by adding just one step to the standard iTag-sequencing procedure without the need to establish specific assays for individual target species like in dPCR or qPCR, which require design of primers and probes, and condition optimization. Comprehensive quantitative data sets of eDNA abundance can be obtained by applying qSeq to various environments and species, providing datasets that will deepen our knowledge of natural communities and the distribution of species.