Effective ecosystem conservation and resource managements require quantitative monitoring of biodiversity, including descriptions of species composition and temporal variations in its abundance. Therefore, long-term, frequent and multispecies time series are invaluable in conservation ecology, but at the same time, they are usually time, effort and cost-consuming. Recently, environmental DNA (eDNA) metabarcoding approach has been rapidly developing as an effective tool for qualitative biodiversity monitoring, but the quantitative capacity of eDNA metabarcoding approach is still limited. In the present study, we added internal standard DNAs (i.e., amplified and quantified short DNA fragments from five fish species that have never been observed in a sampling region) to eDNA samples, which were weekly collected from a marine ecosystem in Maizuru-Bay, Japan (from April 2015 to March 2016, N=52), and performed eDNA metabarcoding analysis to identify fish species and quantify fish eDNA copy number simultaneously.
Using the number of sequence reads and added amounts (copy numbers) of the standard DNA, a standard curve was drawn for each sample (R2 of the standard curves were > 0.9 in general), which was used to convert the sequence reads of eDNA from fish species inhabiting in the study area to the number of eDNA copy. The converted eDNA copy numbers showed positive and significant correlation (close to 1:1 relationship) with eDNA copy quantified by quantitative PCR (qPCR), suggesting that eDNA metabarcoding with internal standard DNAs enabled the quantification of eDNA as accurately as qPCR. In some eDNA samples that contained a significant level of PCR inhibitors (i.e., the internal standard DNAs could not be sufficiently amplified), the number of eDNA copy quantified by qPCR was much smaller than that quantified by eDNA metabarcoding, suggesting that qPCR might underestimate the copy number when PCR inhibitors are contained. As the eDNA metabarcoding detected ca. 70 fish species, 70 quantitative fish eDNA time series were obtained by a single run of Illumina MiSeq. Our method will improve the efficiency of time-series data generation, and would potentially contribute to more effective ecosystem management.