The simultaneous conservation of species richness and evenness is important to effectively reduce biodiversity loss and keep ecosystem health. Environmental DNA (eDNA) metabarcoding has been used as a powerful tool for identifying community composition, but it does not necessarily provide quantitative information due to several methodological limitations. Thus, the quantification of eDNA through metabarcoding is an important frontier of eDNA-based biomonitoring. Particularly, the qMiSeq approach has recently been developed as a quantitative metabarcoding method and has attracted much attention due to its usefulness. The aim here was to evaluate the performance of the qMiSeq approach as a quantitative monitoring tool for fish communities by comparing the quantified eDNA concentrations with the results of fish capture surveys. The eDNA water sampling and the capture surveys using the electrical shocker were conducted at a total of 21 sites in four rivers in Japan. As a result, we found significant positive relationships between the eDNA concentrations of each species quantified by qMiSeq and both the abundance and biomass of each captured taxon at each site. Furthermore, for seven out of eleven taxa, a significant positive relationship was observed between quantified DNA concentrations by sample and the abundance and/or biomass. In total, our results demonstrated that eDNA metabarcoding with the qMiSeq approach is a suitable and useful tool for quantitative monitoring of fish communities. Due to the simplicity of the eDNA analysis, the eDNA metabarcoding with qMiSeq approach would promote further growth of quantitative monitoring of biodiversity.
Conspicuous loss in biological diversity is a current and fast-emerging global problem resulting in changes of ecosystem functions1,2,3. To effectively reduce biodiversity loss, a number of previous studies have argued for the importance of assessing and conserving ecosystem health using species abundance, evenness and richness as indicators4,5,6. Accordingly, researchers and resource managers have made efforts to quantitatively estimate biodiversity based on conventional survey methods, such as direct capture and visual census7,8,9. While these survey approaches provide us with valuable data, they also require a large amount of effort, time and expertise, which limits the feasibility and continuity of the research itself and the reliability of the data10,11,12,13. Additionally, especially for endangered species, sampling activities in direct capture surveys may damage target-species populations and/or their habitats14,15,16. To overcome these difficulties in the traditional survey methods, new approaches for accurate and effective biodiversity monitoring are being explored17,18.
Environmental DNA (eDNA) analysis has been rapidly developed over the past decade as an alternative and/or complementary biomonitoring tool and has become widely used for organisms from various taxonomic groups19,20,21,22,23. Environmental DNA analysis enables us to estimate the presence of organisms via only the collection and detection of cellular materials shed from them in environments, including soil, water, and air24,25,26. The success and high expectations of biomonitoring based on eDNA analysis can also be seen in the rapid growth of the number of publications18,27. Moreover, a meta-analysis has shown that 90% of the 63 studies reported the positive relationships between eDNA concentration in a water sample and the abundance and/or biomass of aquatic organisms15. These results suggest that the eDNA analysis can be further developed and used in the future as a quantitative estimation tool for biodiversity.
Environmental DNA analysis for detecting macroorganisms can be technically categorized into two methods, i.e., species-specific detection and metabarcoding; however, some challenges remain in the quantitative detection of eDNA in both methods28. The species-specific detection method using species-specific primers/probe and a real-time PCR system is currently a major eDNA quantitative method. Nevertheless, the development of a species-specific detection system or multiplexing assays of them is time-consuming and costly and requires prior knowledge and assumptions about the species present within a study site16,28,29. Thus, it is unsuitable for the quantitative detection targeting multiple species. Against this background, researchers have recently begun to explore the possibilities of quantitative analysis on eDNA metabarcoding using universal primers targeting taxonomic groups30,31,32,33. The eDNA metabarcoding enables us to identify community composition including multiple target taxa. However, the number of reads for each taxon output from the high-throughput sequencer is difficult to treat as quantitative information because it can easily vary between samples due to several problems such as PCR inhibitions, primer bias, and library preparation bias15,20,34,35.
Challenge to accurate quantification of eDNA through metabarcoding is an important frontier of eDNA-based biomonitoring, and some approaches have recently been developed36,37,38. Particularly, the quantitative MiSeq sequencing approach (hereafter, qMiSeq approach) developed by Ushio et al.38 allows us to convert the sequence read numbers of detected taxa to DNA copy numbers based on a linear regression between known DNA copy numbers and observed sequence reads of internal standard DNAs. One of the advantages of the qMiSeq approach is that the copy number can be calculated considering the sample-specific effects of PCR inhibition and library preparation bias because a sample-specific standard line is obtained by adding internal standard DNAs to each sample. Accordingly, Ushio et al. (2018)38 reported the successful quantitative monitoring of eDNA derived from multiple fish species in coastal marine ecosystems by combining the qMiSeq approach and the fish universal primer, MiFish-U39. A further verification experiment also showed a positive correlation between the estimated eDNA concentration and the sound intensity of fish40. These results suggested that the qMiSeq approach is a promising technique for quantitative eDNA-based monitoring of fish communities. However, none of the previous studies have been able to compare the observation data with the estimated eDNA concentrations by species and study site. The accumulation of comparative studies on the use of the qMiSeq approach will contribute to the further development and application of quantitative eDNA metabarcoding and maximize its potential as a quantitative monitoring tool for biodiversity41.
The objective of this study was to evaluate the performance of the qMiSeq approach as a quantitative monitoring tool for fish communities. Here, we compared the qMiSeq approach to the capture-based survey with an electrical shocker at four river systems. Specifically, the following two aspects were examined to determine whether there is a significant positive relationship between the fish eDNA concentrations obtained by the qMiSeq approach and the capture data: (1) within sites (including multi-species data) and (2) within each taxon (picked up from multi-site data). These two examinations will provide an indication as to whether comparisons of eDNA concentrations obtained by the qMiSeq approach between multiple samples for the community level and in each species respectively lead to reasonable results. Finally, we discussed the usefulness of the qMiSeq approach for quantitative monitoring of fish community.
Overview of qMiSeq approach results and the comparison of eDNA concentrations quantified by the qMiSeq approach with those estimated by qPCR
The iSeq paired-end sequencing (2 × 150 bp) for the 27 libraries (21 field samples, four cooler blanks, two PCR negative controls) yielded a total of 4.71 million reads (Q30 = 95.2; PF = 67.3%). Negligible sequence reads were detected from PCR negative controls (maximum 75 reads; Table S5), which were ignored in the subsequence analyses. The sequence reads of internal standard DNAs in field samples and cooler blanks had a significant positive relationship with the copy numbers (lm, P < 0.01, R2 > 0.93; Table S4). The coefficients of the linear regressions varied from 174.1 (IN st. 4) to 900.8 (IN cooler blank) (Table S4), suggesting that sequence reads are proportional to the DNA copy numbers in a single sample. Thus, we converted sequence reads of detected taxa in each sample to the DNA copies using the coefficient of the sample-specific regression line. In the comparison of DNA concentrations of the three taxa calculated by qMiSeq approach and species-specific qPCR, a significant positive relationship was found for each taxa (linear regression analysis; P < 0.001, R2 = 0.81, estimate 0.32 for C. temminckii; P < 0.001, R2 = 0.99, estimate 0.58 for C. pollux ME; Figs. 1, 2).
Comparison of species richness observed by qMiSeq and electrical shocker
The qMiSeq approach consistently detected more species than the capture-based survey using an electrical shocker (Fig. 3). On the other hand, one or two taxa that were detected by the electrical shocker were not detected by the qMiSeq approach (i.e., false negative) in several study sites: HS st.3, Liobagrus reinii; FK st. 4, Opsariichthys platypus and Oryzias latipes; FK st. 5, Cobitis matsubarae; FK st.6, Squalidus gracilis gracilis and Tachysurus nudiceps; IN st. 5; S. gracilis gracilis. The false negative result in qMiSeq results for Cobitis matsubarae at FK st. 5 was excluded from all subsequent analyses as it was likely due to a lack of reference DNA database (see Discussion for detail). On the other hand, no false negatives were observed at 16 out of a total of 21 sites. At almost all sites, Rhinogobius spp. was detected by both or either of methods, but it was excluded from subsequent analyses because they contain many unknown haplotypes and many closely related species that cannot be discriminated and/or detected based on the target region of MiFish primer. Also, at IN st. 5, Cyprinus carpio were detected by both methods, but were excluded from subsequent analyses because five individuals (approximately 50 cm, total length) were also visually identified immediately upstream of the study section. Their eDNA almost certainly flowed into the study section and was considered to be noise in this study. Most of the fish taxa detected only by the qMiSeq method had the following characteristics that would cause false-negative results in a capture study; a rare native species (e.g. Anguilla japonica in YK st. 1, 4, 5, 6), non-dominant invasive species (e.g. Channa sp. in FK st. 5 and 6), larger maximum body length (e.g. Cyprinus carpio in FK st. 5, 4, 3) (Table S5).
Fish community structure
The dissimilarity of the fish community structure among study sites obtained by qMiSeq approach was shown in the nonmetric multidimensional scaling (NMDS) biplot (Fig. 4a). A total of 20 research sites (FK st. 1 was excluded due to absence of detected taxa) were categorised into five clusters by k-means clustering. The clusters tended to be divided by the down and up stream. Cluster 3 and 4 consisted only of the study sites at Yokomichi River.
Comparison of the qMiSeq eDNA concentrations with the capture data by electrical shocker
When all sites were used for analysis, both the number and biomass of captured fishes had significant positive correlations with the qMiSeq eDNA concentrations (Kendall rank correlation test, Ps < 0.0001, τ = 0.467 for abundance and τ = 0.469 for biomass; Fig. 4b, Table S6). We separately conducted the Kendall-rank correlation test for the five clusters based on classifications by k-means clustering because the relationships between observed DNA concentrations and capture data are expected to vary among them. As a result, significant positive correlations were also found for all site clusters (Kendall rank correlation test, both the number and biomass, P < 0.01 for all site clusters; Fig. 4b, Table S6). The Kendall’s τ value varied widely among clusters (0.315 to 0.598 for abundance; 0.359 to 0.601 for biomass). Moreover, for seven out of eleven taxa which were detected from three and more sites by each of qMiSeq approach and electrical shocker, we found significant positive relationships between the qMiSeq eDNA copy numbers at each site and the number of captured individuals and/or the total biomass (GLM; Fig. 5, Table 1). When eleven taxa were grouped into four benthic fish taxa and seven non-benthic fish taxa, there were significant positive relationships between the qMiSeq eDNA concentration and both the number of captured individuals and the total biomass (GLM; Table 1).
eDNA metabarcoding detected higher species richness than electrofishing survey
We found that eDNA metabarcoding with the qMiSeq approach could detect not only almost all the taxa captured using the electric shocker but also other taxa that could be false negatives in the capture surveys. This result is consistent with several previous studies that have reported advantages of eDNA metabarcoding compared with traditional capture-based methods in terms of detection sensitivity on biodiversity assessment13,39,42,43,44,45,46,47,48,49,50,51. On the other hand, we should always be aware of the possibility of false negatives. In this study, although almost all taxa were detected, a total of six taxa were missed by eDNA metabarcoding at the sampling site level (Squalidus gracilis gracilis, Pseudobagrus nudiceps, Liobagrus reinii, Cobitis matsubarae, Opsariichthys platypus and Oryzias latipes). Three main causes of false negative results have been suggested in eDNA metabarcoding: (1) low concentration of target DNA, (2) PCR drop-out because of primer mismatch and (3) incompleteness and inaccuracy of the reference sequence database10,11,52. It is difficult to identify the causes of the false negatives, as these factors can combine to influence the results. However, S. gracilis gracilis, L. reinii and O. latipes were only captured by one individual at each site, so low DNA concentrations are likely to be one factor in the false negative results. As for other species and possibilities, future expansion of the reference sequence database will allow a more detailed examination.
The significant positive correlation between eDNA concentrations quantified by qMiSeq and qPCR
The significant positive relationship between eDNA concentrations quantified by the qMiSeq approach and species-specific qPCR indicated that the qMiSeq approach successfully estimated the number of sequence reads of each taxon to the eDNA copy number. For both taxa considered, the eDNA copy numbers calculated by the qMiSeq approach tended to be smaller than those estimated by qPCR. This observation was consistent with those reported in the previous study38. Additionally, eDNA metabarcoding has been recognised to be less sensitive than species-specific detection due to the effects of technical issues such as primer bias, PCR bias and the potential reduced amplification efficiency in tailed-PCR28,53,54. However, this characteristic would not be a serious problem when we compare the eDNA concentrations quantified by the qMiSeq approach on a species-by-species basis.
Calculated eDNA concentration by qMiSeq reflected both fish abundance and biomass in each study site
It was worthy of note that both abundance and biomass of inhabiting fishes had positive correlations with the qMiSeq eDNA concentrations. This study is the first to demonstrate the relationships between the quantitative capture data of fishes and their eDNA concentrations quantified using eDNA metabarcoding. Meanwhile, Kendall's rank correlation coefficients were low in both abundance and biomass (τ < 0.5), possibly because differences in eDNA release and/or degradation rates among taxa inhabiting the study sites, capture efficiency with electric shockers, and environmental conditions could affect the results.
The rank correlation coefficients (Kendall’s τ) tended to be improved by dividing the samples into site clusters based on the dissimilarity of fish community compositions, compared to when all study sites were considered. The results suggested that the study site clustering based on fish community compositions has mitigated the effects of the differences of the eDNA ecology and/or dynamics such as release rate and/or degradation rates for each taxon and dispersion processes on the relationship between qMiSeq eDNA concentration and fish capture data. In contrast, at some clusters, Kendall’s τ values were decreased. Particularly, in cluster 5 sites, Kendall’s τ values decreased for both abundance and biomass and were smallest in all site clusters. The cluster 5 consisted only of downstream sites (IN st. 5, 6 and FK st. 5, 6), and vegetation covered large areas of the channel, mainly along the river bank (Figure S1). Relatedly, the taxa with highly biased relationships between qMiSeq eDNA concentrations and capture data had ecological characteristics such as a preference for vegetation (Pungtungia herzi) and rock crevices (O. obscura), hiding in the sandy bottom (Pseudogobio esocinus esocinus) and high swimming ability (C. temminckii)55,56. The environmental condition and these ecological characteristics would have strongly influenced capture efficiency in the electrical shocker survey. Taxa hiding in vegetation, rock crevices and sandy bottom were difficult to capture because electric shocks cannot reach them, or they faint in unseen crevices. Fish taxa with high swimming ability were able to quickly escape from the investigator’s position or hide under vegetation. In this study, it was difficult to conduct an exhaustive fish capture by closing the study site section due to the limitation of labour, manpower and time for the survey. These facts may indicate that differences in the capture efficiency of the electrical shocker survey have influenced the results of this study, although the magnitude of the impact could depend on the environmental conditions and the ecological characteristics of taxa that inhabit at each study site.
qMiSeq eDNA concentration reflected both fish abundance and biomass in each detected taxon
For seven taxa out of eleven taxa with higher detection frequency in both methods, we demonstrated that the qMiSeq eDNA concentration was significantly related to the abundance and/or biomass. Additionally, qMiSeq eDNA concentration and both the number of captured individuals and total biomass were significantly related when the eleven taxa were grouped into benthic or non-benthic fish taxa. It was consistent with the results of many previous studies showing positive relationships between eDNA concentration and the abundance and/or biomass of target species (90% of relevant 63 papers published by 202015. This observation supports the potential and usefulness of the qMiSeq approach as a quantitative monitoring tool of biodiversity beyond simply identifying species presence. Our result in this study was significant for two reasons.
First, to the best of our knowledge, there have been no studies comparing eDNA concentrations of various macroorganism taxa obtained by quantitative metabarcoding methods, such as the qMiSeq approach, with quantitative capture data for each taxon. To the future implementation of quantitative eDNA metabarcoding methods for quantitative biodiversity monitoring, we should continue the detailed and repeated research efforts to determine whether the results obtained from conventional survey methods and eDNA-based methods yield similar measurements for the taxa of interest20. We believe that this study was the first step in this process. Second, it is worth mentioning that the comparisons of eDNA concentrations and capture data in this study have been performed using pooled eDNA concentration data for each species derived from multiple different samples. In general, the direct comparison of sequence reads count among different samples to quantitatively estimate biodiversity lead to erroneous conclusions because the number of sequence reads will vary among samples depending on the effects of PCR inhibition, primer bias, library preparation bias and sequencing depth, etc.36,38. In most cases, researchers used the proportional abundance of sequence read counts in a sample for each detected taxon instead of a direct comparison of their sequence read counts e.g.31,33,57. However, as differences in the number of taxa detected per sample inherently bias the proportion of sequence read counts (i.e. an increase in the total number of species would reduce the proportional abundance), it is difficult to apply to quantitative comparisons of target taxa among sites, seasons and years with different community compositions28. In contrast, although the qMiSeq approach also cannot mitigate the effects of primer bias (see below), it allows avoiding the effects of PCR inhibition, library preparation bias and sequencing depth38. We reasonably compared the eDNA concentrations of each taxon from multiple samples and demonstrated the usefulness of the analytical advantages of the qMiSeq method for quantitative eDNA metabarcoding for macroorganisms.
Using the qMiSeq approach, although a significant relationship was observed between qMiSeq eDNA concentration and capture data for 7 taxa and benthic or non-benthic fish taxa group, the obtained values varied widely between and within the taxon. It is obvious that the relationship between qMiSeq eDNA concentration and abundance and/or biomass of target taxon observed by field experiments can be obscured by biological and non-biological factors, which is the limitation of analytical method and survey design15,20. In this study, the environmental conditions such as water temperatures and flow velocity at the time of water sampling varied widely among study sites (Table S1), and there may have been associated differences in the characteristics of the ‘ecology of eDNA’ within each taxon (e.g., its origin, state, fate and transport)52,58,59,60. Additionally, as mentioned above, the limitations of capture survey using electrical shocker was likely to have caused significant fluctuations in the observed values. Furthermore, we also need to mention the effect of primer bias, which is an analytical limitation of eDNA metabarcoding. Primer bias is recognised as one of the main causes of the distortion of the relative abundance of amplified DNA in a sample in DNA metabarcoding data34,61,62,63. The MiFish-U primers used in this study has been proven to have high primer universality by many previous studies e.g.11,59,64. Thus, we believe that the distortion of results due to primer bias was minimised in this study.
We here compared the eDNA concentrations quantified by the qMiSeq approach and the results of the capture-based survey using an electrical shocker to evaluate the performance of the qMiSeq approach as a quantitative monitoring tool for fish community. We found positive relationships between the qMiSeq eDNA concentrations and both the abundance and biomass of each captured taxon. Together, our results suggested that eDNA metabarcoding with the qMiSeq approach is a suitable and useful tool for quantitative monitoring of fish community. The simplicity of eDNA analysis will reduce barriers forassessing changes in species abundance, evenness and richness and will facilitate the collection of valuable information for better understanding biodiversity changes.
Materials and methods
Overview of survey design and ethics statement
Field surveys including capture survey and water sampling were conducted in four rivers in western Japan (Fig. 1a, Fig. S1); Yokomichi River (YK; 11th September 2019), Hisakane River (HS; 12th September 2019), Fukuchi River (FK; 8th November 2019) and Ino River (IN; 15th November 2019). The three (HS) or six (YK, FK and IN) study sites were set for each river. Details of each survey site were shown in Table S1. The overview of the survey design is shown in Fig. 1b. Water sampling for eDNA analysis and capture survey with an electrical shocker was performed for each study site. In Japan, the usage of fish in research is permitted by the current laws and guidelines regarding animal experiments. All surveys were performed with attention to animal welfare. All captured fish were resuscitated after measurement and returned to the survey site.
Water sampling and filtration for eDNA analysis and Capture survey by electrical shocker
At each study site, we collected 1 L of surface water using a bleached bottle from the center line of stream at the downstream end of each site (a total of 21 field samples, Fig. 1b). We added benzalkonium chloride (1 mL, 10% w/v;65, Fujifilm Wako Pure Chemical Corporation, Osaka, Japan) to each water sample to preserve eDNA and transported the sample bottles to the laboratory under refrigeration. As a cooler blank for each sampling day, 1 L of deionized water with benzalkonium chloride (1 mL, 10% w/v) was placed into the same cooler box and thereafter treated in the same way as the collected water samples (a total of four cooler blank samples). The water samples were vacuum filtered using GF/F glass fibre filters (diameter: 47 mm, mesh size: 0.7 μm; GE Healthcare Japan, Tokyo, Japan) within 36 h after sampling. All filter samples were stored at –20˚C until DNA extraction.
After the water sampling, we conducted fish samplings (up to 45 min per site) using backpack electroshocker (Smith-Root, LR-20B, USA) and dip nets. Three survey lines were set at each survey site and electric shocks were applied along each line (Fig. 1b). The collected fishes were identified and released within the study reach except for the fishes we could not identify in the field.
DNA extraction from the filter sample
The DNA on each filter sample was extracted following the method described by Environmental DNA Sampling and Experiment Manual ver 2.166 with minor modifications. First, each filter sample was placed in the upper part of the Salivette tube (SARSTEDT AG & Co. KG, Nümbrecht, Germany). The 220 µL of extraction solution containing 200 μL Buffer AL and 20 µL of proteinase K in DNeasy Blood & Tissue Kit (Qiagen, Hilden, Germany) were added onto each filter and incubated at 56˚C for 30 min. After incubation, the Salivette tube was centrifuged at 5,000 × g for 1 min. 220 μL Tris–EDTA (TE) buffer (pH 8.0; Nippon Gene Co., Ltd., Tokyo, Japan) was added onto each filter and re-centrifuged at 5,000 × g for 3 min. After that, 200 μL of ethanol was added to the solution in the bottom part of the Salivette tube and mixed well by gently pipetting. The whole of the mixed solution was transferred to a DNeasy Mini spin column, and the DNA was purified according to the manufacturer's protocol. The DNA was finally eluted in 100 μL Buffer AE and stored at − 20˚C.
eDNA quantification: Quantitative metabarcoding with qMiSeq approach
The qMiSeq approach38 and MiFish-U primers39 was performed to identify fish taxa and quantify their eDNA concentrations simultaneously. The qMiSeq allows us to convert the number of sequence reads into DNA copies without being affected by differences in PCR efficiency by obtaining a sample-specific standard line using internal standard DNAs. The paired-end library preparation with a two-step PCR was performed in 12 µL of reaction mixture according to the described method in Tsu ji et al.67 (See appendix for details). Briefly, in first-round PCR was performed with MiFish-U-F/R primers39 (Table S2), which can amplify fish DNA and three internal standard DNAs (5, 25, 50 copies per reaction, respectively: Table S3). Internal standard DNAs were added to only field samples and cooler blanks. PCR negative control with ultrapure water instead of both eDNA sample and standard DNA mix was added in all first PCR runs. The second-round PCR was performed to add index-sequence and adapter-sequence for the Illumina sequencing platform (Table S2). The indexed products of the second-round PCR were pooled, and the target bands (ca. 370 bp) were excised using 2% E‐Gel SizeSelect Agarose Gels (Thermo Fisher Scientific). The prepared DNA libraries were sequenced by 2 × 150 bp paired-end sequencing on the iSeq platform using the iSeq 100 i1 Reagent v1 cartridge (Illumina, CA, USA) with 30% PhiX spike-in.
The bioinformatics analysis was performed using the PMiFish pipeline11 (https://github.com/rogotoh/PMiFish). Briefly, first, low-quality tails were trimmed from each sequence read, and the paired-end reads were merged. Then, primer sequences were removed, and identical sequences were merged using UCLUST (USEARCH v10.0.240, Edgar, 2010). The merged sequences with 10 or more reads were assigned to the taxonomy using local BLASTN search with the reference database including all inhabiting freshwater fish taxa around the study sites (MiFish DB ver. 37) and the sequence of used three internal standard DNAs. The top BLAST hits with a sequence identity ≥ 98.5% were applied and used in further analyses. For Odontobutis obscura only, sequences detected with sequence identity ≥ 90.0% were also used in the analysis, as they are known to exist in highly genetically differentiated regional populations. As all study sites were in freshwater areas, any saltwater fishes detected were excluded from subsequent analyses. For internal standard DNAs, to obtain a sample-specific standard line, linear regression analysis (lm function in R version 3.6.0 software) was performed using the obtained sequence reads and their known copy numbers (the intercept was set as zero). For each sample, the number of eDNA copies per litre for each taxon was calculated as qMiSeq eDNA concentration using each sample-specific standard line: qMiSeq eDNA concentration = the number of iSeq sequence reads/regression slope of the sample-specific standard line (Table S4). If a taxon was detected from a cooler blank, the DNA copy number observed in the field sample minus the DNA copy number of the relevant taxon observed in the cooler blank was used in the analysis.
eDNA quantification: real-time quantitative PCR (qPCR) for two taxa
The real-time qPCR for Candidia temminckii and Cottus pollux ME were performed using a StepOne-Plus Real-Time PCR system (Applied Biosystems, FosterCity, CA, USA). The eDNA samples from the Yokomichi River were omitted from the comparison examination because they were used in another study and there was no remaining volume for analysis. Species-specific primers for each target species were developed in this study (Table S2), and their specificity was tested by in silico and in vitro tests. All qPCR reactions were performed in a total 15 µL volume and triplicated (See Appendix for details of real-time qPCR conditions and species-specific primer development).
For the following analyses, the R ver. 3.6.0 software was used68. The significance level was set at 0.05 in all analyses. To examine whether the DNA concentrations quantified by the qMiSeq approach reflect those quantified by species-specific qPCR, the eDNA concentrations of C. temminckii and C. pollux ME quantified by the two methods were compared using the linear regression analysis (lm function in R). The differences in fish community compositions among sites were visualised using nonmetric multidimensional scaling (NMDS) with 500 permutations and cluster analysis with group average method. The Bray–Curtis dissimilarity index was used as the fish community dissimilarity in both NMDS and k-means clustering to use the DNA copy number as quantitative information. Based on k-means clustering, a total of 21 study sites were divided into five clusters (see Results). The number of cluster, five, was determined on the basis of Calinski criterion values (Fig. S2). The differences in eDNA release rate and capture efficiency are expected between study sites with largely different fish community compositions due to differences in geographical distribution, ecology, and behaviour of each species, which could distort the relationship between observed DNA concentrations and capture data. Thus, qMiSeq eDNA concentrations and capture data (abundance and biomass of each taxon) were compared using the Kendall-rank correlation test by classifying the data in two patterns as follows: (1) using data from all survey sites and (2) using each cluster site data classified by similarity of fish community. Here, the Kendall rank correlation test was used to examine whether the classification of sites based on the fish community would improve the correlation between qMiSeq eDNA concentrations and capture data by comparing τ values. Kendall’s τ value is a non-parametric measure of relationships between columns of ranked data, and it returns a value of − 1 to 1 (i.e. 0 is no relationship and 1/ − 1 are perfect positive/negative relationship). Moreover, for eleven taxa which were detected from three and over sites by each of qMiSeq approach and electrical shocker among all study sites, the pooled dataset of qMiSeq eDNA concentrations and capture data (abundance or biomass at each survey site) was created for each taxon. For each taxon, qMiSeq eDNA concentrations and capture data were compared using generalized linear models (GLM; the glm.nb function in the MASS package in R69 assuming that eDNA copy numbers follow a negative binomial distribution with log-link function (Table 1). Here we have chosen GLMs with log-link function as a number of previous studies have suggested that the tendency of the eDNA copy number increase to become gradual as abundance or biomass increases e.g.70,71. Additionally, the eleven taxa were divided into four benthic fish taxa group and seven non-benthic fish taxa groups, and the same analysis was carried out for each group as described above. Note, however, that data on O. obscura and O. hikimius, which are closely related and have almost identical ecology, were merged and treated as a single taxon.
Cardinale, B. J. et al. Biodiversity loss and its impact on humanity. Nature 486, 59–67 (2012).
Dornelas, M. et al. Assemblage time series reveal biodiversity change but not systematic loss. Science 344, 296–299 (2014).
Magurran, A. E. et al. Divergent biodiversity change within ecosystems. Proc. Natl. Acad. Sci. 115, 1843–1847 (2018).
Blowes, S. A. et al. Local biodiversity change reflects interactions among changing abundance, evenness, and richness. Ecology online, e3820 (2022).
Crowder, D. W., Northfield, T. D., Gomulkiewicz, R. & Snyder, W. E. Conserving and promoting evenness: Organic farming and fire-based wildland management as case studies. Ecology 93, 2001–2007 (2012).
Hillebrand, H., Bennett, D. M. & Cadotte, M. W. Consequences of dominance: A review of evenness effects on local and regional ecosystem processes. Ecology 89, 1510–1520 (2008).
Masuda, R. et al. Fish assemblages associated with three types of artificial reefs: density of assemblages and possible impacts on adjacent fish abundance. Fishery Bulletin, National Oceanic and Atmospheric Administration. 108, 162–173 (2010).
Miyazono, S., Patiño, R. & Taylor, C. M. Desertification, salinization, and biotic homogenization in a dryland river ecosystem. Sci. Total Environ. 511, 444–453 (2015).
Yonekura, R., Kita, M. & Yuma, M. Species diversity in native fish community in Japan: Comparison between non-invaded and invaded ponds by exotic fish. Ichthyol. Res. 51, 176–179 (2004).
Evans, N. T., Shirey, P. D., Wieringa, J. G., Mahon, A. R. & Lamberti, G. A. Comparative cost and effort of fish distribution detection via environmental DNA analysis and electrofishing. Fisheries 42, 90–99 (2017).
Miya, M., Gotoh, R. O. & Sado, T. MiFish metabarcoding: A high-throughput approach for simultaneous detection of multiple fish species from environmental DNA and other samples. Fish. Sci. 86, 939–970 (2020).
Oka, S. et al. Environmental DNA metabarcoding for biodiversity monitoring of a highly diverse tropical fish community in a coral reef lagoon: Estimation of species richness and detection of habitat segregation. Environ. DNA 3, 55–69 (2021).
Thomsen, P. F. et al. Monitoring endangered freshwater biodiversity using environmental DNA. Mol. Ecol. 21, 2565–2573 (2012).
Pimm, S. L. et al. Emerging technologies to conserve biodiversity. Trends Ecol. Evol. 30, 685–696 (2015).
Rourke, M. L. et al. Environmental DNA (eDNA) as a tool for assessing fish biomass: A review of approaches and future considerations for resource surveys. Environ. DNA 4, 9–33 (2022).
Tsuji, S. et al. Real-time multiplex PCR for simultaneous detection of multiple species from environmental DNA: An application on two Japanese medaka species. Sci. Rep. 8, 1–8 (2018).
Kissling, W. D. et al. Building essential biodiversity variables (EBVs) of species distribution and abundance at a global scale. Biol. Rev. 93, 600–625 (2018).
Rodríguez-Ezpeleta, N. et al. Biodiversity monitoring using environmental DNA. Mol. Ecol. Resour. 21, 1405–1409 (2021).
Boivin-Delisle, D. et al. Using environmental DNA for biomonitoring of freshwater fish communities: Comparison with established gillnet surveys in a boreal hydroelectric impoundment. Environ. DNA 3, 105–120 (2021).
Deiner, K. et al. Environmental DNA metabarcoding: Transforming how we survey animal and plant communities. Mol. Ecol. 26, 5872–5895 (2017).
Doi, H. et al. Compilation of real-time PCR conditions toward the standardization of environmental DNA methods. Ecol. Res. 36, 379–388 (2021).
Kelly, R. P. Making environmental DNA count. Mol. Ecol. Resour. 16, 10–12 (2016).
Kumar, G., Eble, J. E. & Gaither, M. R. A practical guide to sample preservation and pre-PCR processing of aquatic environmental DNA. Mol. Ecol. Resour. 20, 29–39 (2020).
Ficetola, G. F., Miaud, C., Pompanon, F. & Taberlet, P. Species detection using environmental DNA from water samples. Biol. Let. 4, 423–425 (2008).
Kuwae, M. et al. Sedimentary DNA tracks decadal-centennial changes in fish abundance. Commun. Biol. 3, 1–12 (2020).
Lynggaard, C. et al. Airborne environmental DNA for terrestrial vertebrate community monitoring. Curr. Biol. 32, 701–707.e5 (2022).
Tsuji, S., Takahara, T., Doi, H., Shibata, N. & Yamanaka, H. The detection of aquatic macroorganisms using environmental DNA analysis—A review of methods for collection, extraction, and detection. Environ. DNA 1, 99–108 (2019).
Bylemans, J., Gleeson, D. M., Duncan, R. P., Hardy, C. M. & Furlan, E. M. A performance evaluation of targeted eDNA and eDNA metabarcoding analyses for freshwater fishes. Environ. DNA 1, 402–414 (2019).
Wozney, K. M. & Wilson, C. C. Quantitative PCR multiplexes for simultaneous multispecies detection of Asian carp eDNA. J. Great Lakes Res. 43, 771–776 (2017).
Evans, N. T. et al. Quantification of mesocosm fish and amphibian species diversity via environmental DNA metabarcoding. Mol. Ecol. Resour. 16, 29–41 (2016).
Fraija-Fernández, N. et al. Marine water environmental DNA metabarcoding provides a comprehensive fish diversity assessment and reveals spatial patterns in a large oceanic area. Ecol. Evol. 10, 7560–7584 (2020).
Kelly, R. P., Port, J. A., Yamahara, K. M. & Crowder, L. B. Using environmental DNA to census marine fishes in a large mesocosm. PLoS ONE 9, e86175 (2014).
Thomsen, P. F. et al. Environmental DNA from seawater samples correlate with trawl catches of subarctic, deepwater fishes. PLoS ONE 11, e0165252 (2016).
Lamb, P. D. et al. How quantitative is metabarcoding: A meta-analytical approach. Mol. Ecol. 28, 420–430 (2019).
Lim, N. K. M. et al. Next-generation freshwater bioassessment: eDNA metabarcoding with a conserved metazoan primer reveals species-rich and reservoir-specific communities. R. Soc. Open Sci. 3, 160635 (2016).
Hoshino, T., Nakao, R., Doi, H. & Minamoto, T. Simultaneous absolute quantification and sequencing of fish environmental DNA in a mesocosm by quantitative sequencing technique. Sci. Rep. 11, 4372 (2021).
Smets, W. et al. A method for simultaneous measurement of soil bacterial abundances and community composition via 16S rRNA gene sequencing. Soil Biol. Biochem. 96, 145–151 (2016).
Ushio, M. et al. Quantitative monitoring of multispecies fish environmental DNA using high-throughput sequencing. Metabarcod. Metagenom. 2, e23297 (2018).
Miya, M. et al. MiFish, a set of universal PCR primers for metabarcoding environmental DNA from fishes: Detection of more than 230 subtropical marine species. R. Soc. Open Sci. 2, 150088 (2015).
Sato, M. et al. Quantitative assessment of multiple fish species around artificial reefs combining environmental DNA metabarcoding and acoustic survey. Sci. Rep. 11, 1–14 (2021).
Ushio, M. Interaction capacity as a potential driver of community diversity. Proc. R. Soc. B Biol. Sci. 289, 20212690 (2022).
Andruszkiewicz, E. A., Sassoubre, L. M. & Boehm, A. B. Persistence of marine fish environmental DNA and the influence of sunlight. PLoS ONE 12, e0185043 (2017).
Bylemans, J., Gleeson, D. M., Hardy, C. M. & Furlan, E. Toward an ecoregion scale evaluation of eDNA metabarcoding primers: A case study for the freshwater fish biodiversity of the Murray-Darling Basin (Australia). Ecol. Evol. 8, 8697–8712 (2018).
Civade, R. et al. Spatial representativeness of environmental DNA metabarcoding signal for fish biodiversity assessment in a natural freshwater system. PLoS ONE 11, e0157366 (2016).
Deiner, K., Fronhofer, E. A., Mächler, E., Walser, J.-C. & Altermatt, F. Environmental DNA reveals that rivers are conveyer belts of biodiversity information. Nat. Commun. 7, 12544 (2016).
Hänfling, B. et al. Environmental DNA metabarcoding of lake fish communities reflects long-term data from established survey methods. Mol. Ecol. 25, 3101–3119 (2016).
Nakagawa, H. et al. Comparing local-and regional-scale estimations of the diversity of stream fish using eDNA metabarcoding and conventional observation methods. Freshw. Biol. 63, 569–580 (2018).
Sato, H., Sogo, Y., Doi, H. & Yamanaka, H. Usefulness and limitations of sample pooling for environmental DNA metabarcoding of freshwater fish communities. Sci. Rep. 7, 14860 (2017).
Shaw, J. L. A. et al. Comparison of environmental DNA metabarcoding and conventional fish survey methods in a river system. Biol. Cons. 197, 131–138 (2016).
Valentini, A. et al. Next-generation monitoring of aquatic biodiversity using environmental DNA metabarcoding. Mol. Ecol. 25, 929–942 (2016).
Yamamoto, S. et al. Environmental DNA metabarcoding reveals local fish communities in a species-rich coastal sea. Sci. Rep. 7, 40368 (2017).
Jane, S. F. et al. Distance, flow and PCR inhibition: eDNA dynamics in two headwater streams. Mol. Ecol. Resour. 15, 216–227 (2015).
Harper, L. R. et al. Needle in a haystack? A comparison of eDNA metabarcoding and targeted qPCR for detection of the great crested newt (Triturus cristatus). Ecol. Evol. 8, 6330–6341 (2018).
Nichols, R. V. et al. Minimizing polymerase biases in metabarcoding. Mol. Ecol. Resour. 18, 927–939 (2018).
Hosoya, K. Yamakei Handy Illustrated Book 15: Freshwater fishes of Japan (Yama-Kei Publishers, 2019).
Nakabo, T. Fishes of Japan with Pictorial Keys to the Species (3-Volume Set). (Tokai University Press, 2013).
Goutte, A., Molbert, N., Guérin, S., Richoux, R. & Rocher, V. Monitoring freshwater fish communities in large rivers using environmental DNA metabarcoding and a long-term electrofishing survey. J. Fish Biol. 97, 444–452 (2020).
Barnes, M. A. & Turner, C. R. The ecology of environmental DNA and implications for conservation genetics. Conserv. Genet. 17, 1–17 (2016).
Collins, R. A. et al. Non-specific amplification compromises environmental DNA metabarcoding with COI. Methods Ecol. Evol. 10, 1985–2001 (2019).
Tsuji, S., Ushio, M., Sakurai, S., Minamoto, T. & Yamanaka, H. Water temperature-dependent degradation of environmental DNA and its relation to bacterial abundance. PLoS ONE 12, e0176608 (2017).
Elbrecht, V. & Leese, F. Can DNA-based ecosystem assessments quantify species abundance? Testing primer bias and biomass—sequence relationships with an innovative metabarcoding protocol. PLoS ONE 10, e0130324 (2015).
Nester, G. M. et al. Development and evaluation of fish eDNA metabarcoding assays facilitate the detection of cryptic seahorse taxa (family: Syngnathidae). Environ. DNA 2, 614–626 (2020).
Piñol, J., Mir, G., Gomez-Polo, P. & Agustí, N. Universal and blocking primer mismatches limit the use of high-throughput DNA sequencing for the quantitative metabarcoding of arthropods. Mol. Ecol. Resour. 15, 819–830 (2015).
Zhang, S., Zhao, J. & Yao, M. A comprehensive and comparative evaluation of primers for metabarcoding eDNA from fish. Methods Ecol. Evol. 11, 1609–1625 (2020).
Yamanaka, H. et al. A simple method for preserving environmental DNA in water samples at ambient temperature by addition of cationic surfactant. Limnology 18, 233–241 (2017).
Minamoto, T. et al. An illustrated manual for environmental DNA research: Water sampling guidelines and experimental protocols. Environ. DNA 3, 8–13 (2021).
Tsuji, S., Nakao, R., Saito, M., Minamoto, T. & Akamatsu, Y. Pre-centrifugation before DNA extraction mitigates extraction efficiency reduction of environmental DNA caused by the preservative solution (benzalkonium chloride) remaining in the filters. Limnology 23, 9–16 (2022).
R Core Team. R. A Language and Environment for Statistical Computing. (2021).
Venables, W. N. & Ripley, B. D. Modern Applied Statistics with S. (Springer, 2002).
Coulter, D. P. et al. Nonlinear relationship between Silver Carp density and their eDNA concentration in a large river. PLoS ONE 14, e0218823 (2019).
Doi, H. et al. Environmental DNA analysis for estimating the abundance and biomass of stream fish. Freshw. Biol. 62, 30–39 (2017).
Kanno, K., Onikura, N., Kurita, Y., Koyama, A. & Nakajima, J. Morphological, distributional, and genetic characteristics of Cottus pollux in the Kyushu Island, Japan: indication of fluvial and amphidromous life histories within a single lineage. Ichthyol. Res. 65, 462–470 (2018).
We thank laboratory members of Akamatsu laboratory, Yamaguchi University, and Inui Laboratory, Fukuoka Institute of Technology University, for help with the field sampling. We thank Mr. S. Kunimatsu and Mr. Y. Fuke (Kyoto University) for providing fish pictures. We thank Dr. M. Ushio (The Hong Kong University of Science and Technology) for his valuable comments and advices on the draft. This is supported by the River Fund of The River Foundation, Japan (2019-5211-030). This work was supported by YU Project for Formation of the Core Research Center.
The authors declare no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Tsuji, S., Inui, R., Nakao, R. et al. Quantitative environmental DNA metabarcoding shows high potential as a novel approach to quantitatively assess fish community. Sci Rep 12, 21524 (2022). https://doi.org/10.1038/s41598-022-25274-3
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.