Research Article |
Corresponding author: Trude Vrålstad ( trude.vralstad@vetinst.no ) Academic editor: Adam Petrusek
© 2022 Johannes C. Rusch, David A. Strand, Charlotte Laurendz, Tom Andersen, Stein I. Johnsen, Lennart Edsman, Trude Vrålstad.
This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Citation:
Rusch JC, Strand DA, Laurendz C, Andersen T, Johnsen SI, Edsman L, Vrålstad T (2022) Exploring the eDNA dynamics of the host-pathogen pair Pacifastacus leniusculus (Decapoda) and Aphanomyces astaci (Saprolegniales) under experimental conditions. NeoBiota 79: 1-29. https://doi.org/10.3897/neobiota.79.82793
|
The oomycete Aphanomyces astaci causes crayfish plague, a disease threatening native European crayfish. It is carried and transmitted by American crayfish species, which are the original hosts of A. astaci. In recent years, environmental DNA (eDNA) methods have been successfully implemented to monitor the spread of both A. astaci and its hosts. However, still little is known about how population density and other environmental factors influence the detectability of this host-pathogen complex. In a mesocosm experiment, we tested the influence of crayfish density, temperature and food availability on the detectability of eDNA for A. astaci and its host, signal crayfish Pacifastacus leniusculus. We also compared eDNA results with crayfish population density measured by catch per unit effort (CPUE) from two lakes with varying crayfish density and A. astaci prevalence. The mesocosm experiment revealed that a limited set of controlled factors can substantially change the detectable amount of eDNA, even though the physical presence of the target organisms remains the same. In cold, clear water, eDNA quantities of both targets increased far more than in a linear fashion with increased crayfish density. However, the presence of food decreased the detectability of crayfish eDNA, presumably through increased microbial-induced eDNA degradation. For A. astaci, where eDNA typically represents living spores, food did not affect the detectability. However, high water temperature strongly reduced it. The increased complexity and variability of factors influencing eDNA concentration under natural conditions, compared to a controlled experimental environment, suggests that establishing a reliable relationship between eDNA quantities and crayfish density is difficult to achieve. This was also supported by field data, where we found minimal correspondence between eDNA quantity and CPUE data. A comparison between quantitative real-time PCR (qPCR) analysis and droplet-digital PCR (ddPCR) analysis revealed higher detection success of the targets in field samples when using qPCR. Overall, our results support eDNA as an effective tool for presence-absence monitoring, but it seems less suited for biomass quantification and population density estimates. Detection of A. astaci and P. leniusculus is not influenced uniformly by respective environmental factors. Consequently, we recommend a strategy of monitoring both targets, where the detection of one may point towards the presence of the other.
crayfish plague, ddPCR, environmental biomonitoring, environmental DNA, freshwater crayfish, mesocosm experiment, occupancy modelling
The oomycete Aphanomyces astaci is a fungal-like water mould that causes crayfish plague, a disease lethal to crayfish indigenous to Europe (
Environmental DNA (eDNA) monitoring is increasingly used for biomonitoring of species, including both macroorganisms and microorganisms (
The host-pathogen pair Pacifastacus leniusculus and A. astaci are a particularly interesting model for studying eDNA dynamics as crayfish leave relatively low traces of eDNA in the water (
To link the experimental data to a real-life situation, we also included a small field survey where water samples were obtained in parallel with catch per unit effort (CPUE) data from two lakes with varying crayfish density and varying infection load with A. astaci. The results from the experiments and field survey will hopefully provide more detailed understanding of eDNA dynamics of the host-pathogen pair and provide knowledge that can help in designing better monitoring programmes involving A. astaci and freshwater crayfish.
In total, 125 P. leniusculus specimens (71 female, 54 male, average total length 109.6 mm ±16.8 mm) were obtained by trapping from two Norwegian lakes (Rødenessjøen and Øymarksjøen) within the Halden watercourse in south-eastern Norway. Crayfish in both lakes have a well-documented history of infection with A. astaci (
Capture, transport and husbandry of crayfish were conducted with permits from the Norwegian Food Safety Authority, the County Governor of Østfold and the Norwegian Environment Agency. This, along with euthanasia at the end of the experiment, was conducted in accordance with the Norwegian Animal Welfare Act (LOV-2009-06-19-97) and EU regulations (EU Directive; 2010/63/EU).
The experiment was designed as full-factorial to analyse the influence of crayfish density, availability of food and temperature on the detectability of eDNA from P. leniusculus and A. astaci (Fig.
Schematic drawing of the experimental setup with P. leniusculus and A. astaci. The numbers 2 and 20 represent the number of crayfish present in the respective tanks with 100 l of water. One trial consisted of two tanks with 2 crayfish (low density) and two tanks with 20 crayfish (high density). Crayfish from one tank of each density group were fed, while the crayfish in the parallel tanks got no food. Three replicate trials were run at high (20 °C) water temperature and another three replicate trials at low (10 °C) water temperatures, in total for six weeks.
For each week, crayfish were randomly picked from the communal tank, assigned to an experimental tank and their number-markings were recorded. All crayfish were kept in their respective tanks for one week. In all tanks, crayfish were provided with sufficient shelters made from PVC tubes. After seven days, triplicate water samples of 1 l (3 × 1 l) were taken from each tank using a peristaltic pump (Masterflex I/P, Cole-Parmer, Vermon Hills, USA), tygon tubing (Masterflex), an in-line filter holder (Millipore, Billerica, Massachusetts, USA) and glass fibre filters (47 mm AP25, Millipore) according to
At the end of the experimental period, the crayfish were euthanised by placing them in ice slush for anaesthesia, followed by piercing of the brain using a scalpel. Tissue samples were taken from the tail-fan of 45 crayfish used in the experiment and analysed with species-specific A. astaci quantitative real-time PCR (qPCR) assay for determining the A. astaci prevalence and semi-quantitative agent levels, as described in
Water samples were also obtained from two lakes with well-documented illegally introduced P. leniusculus populations (Table
Sampling points in Lake Øymarksjøen in Norway and Lake Stora Le in Sweden. The countries are indicated by their two-letter ISO codes: NO and SE. Sampling points in Øymarksjøen are numbered O1–O9, the sampling points in Lake Stora Le are S1–S3. The respective sampling points are depicted as red dots, the international border is represented by the black line. The location of the map is illustrated by the red area in the inset map in the top right corner.
List of sampling sites including location, sampling date and amount of water filtered.
Site code | Location | Date sampled | Sample volume (in l) | Coordinates |
---|---|---|---|---|
O1 | Øymarksjoen, west of Sandbøl | 08.06.2016 | 5 | 59.3522N, 11.6608E |
10.08.2016 | 5 | |||
O2 | Øymarksjoen, above Sandbøl | 08.06.2016 | 5 | 59.3501N, 11.6556E |
10.08.2016 | 4.5 | |||
O3 | Øymarksjoen, south of Sandbøl | 08.06.2016 | 5 | 59.3483N, 11.6472E |
10.08.2016 | 5 | |||
O4 | Øymarksjoen, Fossbekkbrua | 08.06.2016 | 4.5 | 59.3331N, 11.6364E |
10.08.2016 | 4 | |||
O5 | Øymarksjoen, hyttefelt | 08.06.2016 | 5 | 59.3283N, 11.6450E |
10.08.2016 | 4 | |||
O6 | Øymarksjoen, west of Bønesøya | 08.06.2016 | 5 | 56.3261N, 11.6528E |
10.08.2016 | 5 | |||
O7 | Øymarksjoen, Bønesøya | 08.06.2016 | 5 | 59.3294N, 11.6561E |
O8 | Øymarksjoen, Blåsnuppen | 08.06.2016 | 5 | 59.3242N, 11.6601E |
10.08.2016 | 2.5 | |||
10.08.2016 | 3.5 | |||
O9 | Mokallen, outlet to Strømselva | 08.06.2016 | 5 | 59.3117N, 11.6667E |
10.08.2016 | 3.5 | |||
S1 | Stora Le | 20.09.2016 | 5 (x5) | 59.1594N, 11.8625E |
S2 | Stora Le | 20.09.2016 | 5 (x5) | 59.2067N, 11.8231E |
S3 | Stora Le | 20.09.2016 | 5 (x5) | 59.2089N, 11.8261E |
In Lake Øymarksjøen, nine sampling sites were selected at which two water samples were collected per site (one in June 2016 and one in August 2016). In order to estimate P. leniusculus CPUE, a total of five foldable cylindrical crayfish traps (LiNi) with two entrances and a mesh size of 14 mm (
Three sampling sites with varying density of crayfish populations were chosen in Lake Stora Le based on previous monitoring (
Before DNA extraction, the filters were frozen at –80 °C and then freeze-dried for 24 h, using a vacuum freeze dryer (Heto drywinner, Thermo Fisher Scientific, Waltham, USA). DNA was extracted from the filters according to a cetyltrimethyl ammonium bromide (CTAB) protocol described in
All qPCR analyses were run on a Mx3005P qPCR thermocycler (Agilent, Santa Clara, USA), using the assay for A. astaci developed by
All DNA isolates were analysed both undiluted and 10-fold diluted to account for potential inhibition, in total four replicates per filter sample. The level of inhibition was determined by calculating the difference in Ct-values between the undiluted and diluted samples (ΔCt) following
Genomic DNA from P. leniusculus and A. astaci with a known DNA copy number concentration was included in each run to create a standard curve for relative quantification of targeted DNA copies in each reaction (
Droplet digital PCR (ddPCR) was performed on a QX200 AutoDG Droplet Digital PCR System (Bio-Rad, Hercules, USA). For ddPCR analysis of the samples, we drew upon the qPCR assays developed for A. astaci (
We used linear regression on log(x + 1)-transformed variables to investigate the overall consistency between ddPCR- and qPCR-based copy numbers (per reaction) and assessed “goodness of fit” from the Pearson correlation coefficient (r) between the two. We used generalised linear models (GLMs) to estimate effect sizes of the treatments in the laboratory experiments. Since the positive droplet count in a ddPCR assay conforms better to statistical distributions of the exponential families than the non-integer copy number estimates derived from this statistic, we decided to model the logarithm of positive droplets using the logarithm of total droplets as offset (i.e. including a “+ offset(log(tot.drp))” term in the model formula). Using this model construct, we essentially modelled the fraction of droplets that are positive with maintaining a dependent variable that is an integer count. Since this type of data often exhibits more zero counts than expected from a Poisson distribution (so-called over-dispersion), we fitted models of both the Poisson and negative binomial families and compared their performances by Akaike’s Information Criterion (AIC). To investigate possible interactive effects between treatments, we fitted models with and without interactions and compared these also by AIC. To account for the pseudo-replication introduced by taking three samples from each tank at the end of each experimental run, we used Tank ID nested within experimental run as a random intercept effect (i.e. including a “+ (1 | Run / Tank)” term in the model formula). We also chose to sum the droplet counts from the A and B filter halves instead of having an additional hierarchical level in the models. We fitted the resulting generalised linear mixed models (GLMMs) with the glmmTMB package (
For the field data, we used 3-level hierarchical occupancy models to represent the variation between sites, between replicated filter samples from the same site and between assays on separate halves of the same filter. In this analysis, we focused on presence of P. leniusculus and A. astaci eDNA. Here also, a positive detection was defined as ≥ 3 positive droplets in a reaction with > 8000 total droplets (reactions with < 8000 total droplets were flagged as missing values). We fitted the resulting 3-level binomial models with a Bayesian approach using the msocc package for R (
While all DNA, PCR and environmental laboratory controls remained negative in the ddPCR analysis, we experienced low positive signals in some of the inlet water controls in weeks 5 and 6. To test if these weak positive detections influenced the results, we used the same GLMM analysis as described above. All samples collected in the same week as the positive inlet controls that were equal to or lower than the positive control for that week were set to zero. Thus, we used the droplet count of the positive inlet control as the threshold for scoring samples positive. The statistical GLMM tests for the effect of the contamination showed no difference in the significant factors when adjusting for the positive inlets controls (see Suppl. material
From the 45 analysed crayfish, representing 36% of the total amount of crayfish used in the experimental population, the prevalence of A. astaci was 78% and the agent level varied from A0 to A6. According to this classification, agent levels A0 and A1 are considered negative, while agent levels A2 to A6 indicate presence of the pathogen with exponentially increasing amounts of detectable pathogen DNA (
Suppl. material
Scatterplot of the estimated DNA copies per reaction of both qPCR and ddPCR analysis for A. astaci (A) and P. leniusculus (B) from the mesocosm trial. A significant positive correlation between the methods was observed. A) A. astaci: Pearson’s r = 0.98, p < 2.2×10–6. B) P. leniusculus: Pearsons’s r = 0.99, p < 2.2×10–16. Black line represents 1:1 correspondence between ddPCR and qPCR.
Of the 72 water samples taken during the aquarium experiment and analysed with ddPCR, 46 were positive for A. astaci and 60 were positive for P. leniusculus. A total of 21 DNA extraction subsamples were excluded from the analysis due to the total droplet count in the reaction being below 8000. The number of positive droplets per ddPCR reaction ranged from 3 to > 19433 (See Suppl. material
For A. astaci, the median eDNA copy number per litre was much lower at 20 °C than at 10 °C, irrespective of any other factor/influence (Table
Boxplot of detectable eDNA copies per litre for A. astaci (A) and P. leniusculus (B), as detected by ddPCR. For temperature, the blue box indicates the interquartile range at 10 °C, while the red box indicates the interquartile range at 20 °C. Density is indicated by 2 (crayfish per tank) and 20 (crayfish per tank) and the median is represented by the thick black horizontal bar within the boxes. A for A. astaci, the median copy number/l was generally very low at 20 °C, while high median copy numbers/l were observed at 10 °C and high crayfish density. Food had no apparent effect B for P. leniusculus, the highest median eDNA copy number/l was observed at high crayfish density at 10 °C, with no food. The median copy number/l was generally substantially lower at 20 °C and, in particular, in the tanks where crayfish were fed. Food had a negative effect on eDNA copy numbers both at 10 °C and 20 °C.
Summary of the median eDNA copies per litre with coefficient of variation in brackets of P. leniusculus eDNA and A. astaci eDNA for the combinations of test conditions: density, food availability and temperature. Fold change indicates the relative increase (x : 1) or decrease (1 : ×) in eDNA copy numbers per litre of water from low density (2 crayfish) to high density (20 crayfish).
Temp | Target | Food | No food | ||||
---|---|---|---|---|---|---|---|
2 crayfish | 20 crayfish | Fold change | 2 crayfish | 20 crayfish | Fold change | ||
10 °C | P. len | 5378 | 2533 | 1 : 2.1 | 8089 | 8488889 | 1049 : 1 |
(92%) | (21%) | (78%) | (93%) | ||||
20 °C | P. len | 844 | 1689 | 2 : 1 | 20667 | 17467 | 1 : 1.2 |
(139%) | (120%) | (94%) | (75%) | ||||
10 °C | A. ast | 262 | 44556 | 170 : 1 | 622 | 28622 | 46 : 1 |
(132%) | (107%) | (115%) | (108%) | ||||
20 °C | A. ast | 27 | 0 | NA | 0 | 53 | NA |
(103%) | (210%) | (147%) | (170%) |
These observations were reflected by the statistical modelling. For A. astaci, the two-way interaction model had the lowest AIC value. High crayfish density had a significant positive effect on eDNA quantity (positive droplets), whereas high temperature had a significant negative effect on eDNA quantity of A. astaci. The combination of high temperature and high density also had a significant negative effect on the amount of detectable A. astaci eDNA (See Suppl. material
Generalised mixed effect model analysis of the influence of temperature, density and food availability on the amount of detectable eDNA of A. astaci (A) and P. leniusculus (B) in the mesocosm experiment. The amount of detectable eDNA is represented as positive droplets per sample (log scale). A for A. astaci, the eDNA quantity (positive droplets) was significantly higher in tanks with high crayfish density (20 crayfish) at 10 °C, while high temperature (20 °C) had a significant negative effect on the eDNA quantity for all combinations B for P. leniusculus, the eDNA quantity (positive droplets) was significantly higher for the combination “no food” for 20 crayfish at 10 °C, while the combination 20 crayfish provided with food at 20 °C had a significant negative effect on the eDNA quantity.
For P. leniusculus, the highest median number of eDNA copies per litre (> 8.4×106) was observed in the treatment group with high crayfish density and no food at 10 °C. However, the treatment group with high crayfish density and no food at 20 °C had a median of 17467 eDNA copies per litre, lower in fact than the treatment group with low crayfish density and no food at 20 °C (median eDNA copies per litre = 20667) (Table
The results of the model matched the results of the detected eDNA copy numbers per litre of P. leniusculus. Here, the three-way interaction model had the lowest AIC value. High density and the combination of no food at low temperature and high density were determined to have a significant positive effect on the amount of eDNA quantity (positive droplets) by the GLMM model. The two combinations of high temperature with food and high temperature with high density had a significant negative effect (Fig.
The effect of crayfish density on the amount of detected eDNA copies per litre, both for P. leniusculus and A. astaci eDNA, varied considerably. At 10 °C, we observed a 170-fold increase of the median eDNA quantity (represented by DNA copies per litre) of A. astaci from tanks with 2 crayfish to tanks with 20 crayfish provided with food. In the absence of food, a 46-fold increase was observed. At 20 °C, almost no A. astaci eDNA was detected in any of the tanks, only trace levels close to or below LOD (3 positive droplets) were observed (Table
For P. leniusculus at 20 °C, we found only a two-fold increase of the median eDNA quantity between the tanks with 2 and 20 crayfish provided with food and even a minor (1.2-fold) decrease when food was missing. At 10 °C, the median eDNA quantity was 2.1 fold lower in the tanks with 20 crayfish compared to 2 crayfish, when food was provided. However, in the absence of food, the median eDNA quantity was as much as 1049-fold higher in the tanks with 20 crayfish compared to 2 crayfish (Table
Of the 15 samples analysed from Lake Stora Le, 10 (66.7%) were positive for A. astaci eDNA and 7 (46.7%) were positive for P. leniusculus eDNA using ddPCR. Of the 18 samples analysed from Øymarksjøen, 11 (61.1%) were positive for A. astaci eDNA, while none was positive for P. leniusculus eDNA with ddPCR. For qPCR, 13 (72.2%) samples were positive for A. astaci and 12 (66.7%) were positive for P. leniusculus eDNA (Table
Summary of results from field samples at Lake Stora Le and Lake Øymarksjøen for eDNA detection of A. astaci and P. leniusculus. The dates of sampling are provided together with the location and sample replicate in Table
Lake | Location | # samples | Volume (l) | CPUE | ω ddPCR / qPCR | detection probability (ψ) | ||
---|---|---|---|---|---|---|---|---|
A. ast | P. len | A. ast | P. len | |||||
Stora Le | ||||||||
S1 | 5 | 25 | 20 | 0.8 / NA | 0.8 / NA | 1 | 1 | |
S2 | 5 | 25 | 3.6 | 1 / NA | 0.6 / NA | 0.99 | 0.96 | |
S3 | 5 | 25 | 0.6 | 0.2 / NA | 0 / NA | 0.94 | 0.29 | |
Øymarksjøen | ||||||||
O1 | 2 | 10 | 4 | 0.5 / 0 | 0 / 0.5 | 0.99 | 0.00 | |
O2 | 2 | 9.5 | 9.6 | 1 / 1 | 0 / 0 | 1 | 0.00 | |
O3 | 2 | 10 | 13.2 | 0.5 / 0.5 | 0 / 0.5 | 1 | 0.00 | |
O4 | 2 | 8.5 | 17.6 | 0 / 0.5 | 0 / 0.5 | 1 | 0.00 | |
O5 | 2 | 9 | 25.4 | 1 / 1 | 0 / 1 | 1 | 0.00 | |
O6 | 2 | 10 | 12.2 | 1 / 1 | 0 / 1 | 1 | 0.00 | |
O7 | 1 | 5 | 25.8 | 0 / 0 | 0 / 0 | 1 | 0.00 | |
O8 | 3 | 11 | 13.2 | 1 / 1 | 0 / 1 | 1 | 0.00 | |
O9 | 2 | 8.5 | 6 | 0 / 1 | 0 / 1 | 0.99 | 0.00 |
While there was relatively good correlation between the qPCR and ddPCR results from Lake Øymarksjøen for A. astaci (Fig.
Scatterplot of the estimated DNA copies per reaction of both qPCR and ddPCR analysis for A. astaci (A) and P. leniusculus (B) from Lake Øymarksjøen. A for A. astaci, the correlation between qPCR and ddPCR results is relatively good (Pearson’s r = 0.81, p = 2.4×10–10) B for P. leniusculus, the correlation between qPCR and ddPCR results is poor (Pearson’s r = 0.53, p = 0.0011). Black line represents 1:1 correspondence between ddPCR and qPCR.
Using the msocc package, we calculated the statistical probability of detecting A. astaci and P. leniusculus at crayfish densities ranging from 0 to 20 CPUE, based on the detection rates from field samples (Fig.
Modelling of probability of detection for A. astaci (A) and P. leniusculus (B) with respect to catch per unit effort (CPUE) in lakes Øymarksjøen (green line) and Stora Le (purple line) using msocc, based on ddPCR results. The thick lines represent the median detection probability, while the thin lines represent the upper and lower quantile. The figures are based on 11000 iterations, the first 1000 as warm-up and the rest thinned by 10. Figure
The probability of detecting eDNA of A. astaci using the sampling method described above reached 100% at a crayfish density of 2 CPUE in both Stora Le and Øymarksjøen. For P. leniusculus, we calculated a 100% eDNA detection probability above a crayfish density of 5 CPUE in Stora Le. The lack of positive detections in Lake Øymarksjøen using ddPCR provided us with insufficient data points to calculate the detection probability for Lake Øymarksjøen accurately. In the subsequent analysis using qPCR data, we calculated a 100% eDNA detection probability above 3 CPUE in Øymarksjøen. The eDNA concentration in the samples obtained from the field was consistently lower than in the aquarium samples, even in locations with high CPUE.
The mesocosm experiment conducted in our study demonstrates that environmental factors might drastically change the detectable amount of eDNA from A. astaci and P. leniusculus. In the cold and clear water in the experimental tanks, i.e. in the absence of food supplies, eDNA quantities of P. leniusculus and A. astaci increased far more than in a linear fashion with crayfish density. However, food availability seemed to contribute to a faster degradation of P. leniusculus eDNA. A. astaci, on the other hand, was unaffected by the presence of food in the cold water, while a water temperature of 20 °C had a surprisingly huge negative impact on A. astaci detectability from eDNA, regardless of food availability.
We found little support for our hypothesis that eDNA emitted from P. leniusculus scales directly with the number of individuals. Instead, we observed that small changes to the experimental environment led to large changes – both positive and negative – in the quantity of detectable eDNA. This indicates that the complexity and variability of influencing factors under field conditions obstructs predictable correlations between eDNA quantities and crayfish density. This is supported by our field data with no clear correlation between eDNA detectability and crayfish population density (as estimated by CPUE). A study on another crustacean, the green crab (Carcinus maenas), recently concluded that eDNA cannot be used to rigorously predict the biomass of the target species under controlled conditions (
For surveillance purposes, our study supports a strategy of detecting both the host and the pathogen. As the eDNA detectability of this alien host-pathogen couple seems to be affected differently, eDNA surveillance of both targets will increase the total detection probability, since detection of one may also suggest the presence of the other. This will, of course, only apply in habitats or regions where A. astaci is prevalent in alien crayfish hosts and not for American crayfish populations with very low or even missing pathogen prevalence (
Under field conditions, eDNA itself and the detectability of eDNA is subjected to a multitude of factors, such as UV radiation, dilution, inhibition through humic acids, retention in substrate and transport that expedite its degradation or disappearance from the system (
Even though sporulation of A. astaci has been described as most efficient below 20 °C (
The huge increase (> 1000 fold) in eDNA concentrations in the high-density tanks with non-fed crayfish at 10 °C might be explained by injuries from aggressive interactions (Sint et al. 2021) combined with the relatively clean water with assumingly low microbiological activity. In a similar tank experiment,
When using ddPCR, we observed a relatively good detectability of eDNA from both targets in the field samples in Lake Stora Le and also good detectability of A. astaci in Lake Øymarksjøen. Surprisingly, we did not detect P. leniusculus in any of the samples from Lake Øymarksjøen with ddPCR, but in 66.7% of the samples when using qPCR. It is unlikely that this was caused by insufficient assay specificity as we obtained satisfactory results from the mesocosm experiment using the same assay on P. leniusculus originating from the interconnected lakes Øymarksjøen and Rødenessjøen. However, these results are similar to those in the study by
The overall detection rate for both organisms was higher in Lake Stora Le than in Lake Øymarksjøen. A speculative explanation is that this may result from trapping (for logistical reasons) prior to sampling in Lake Stora Le. Ideally, eDNA sampling should be carried out before trapping, as crayfish are drawn to the bait from their shelters and feeding activity combined with increased interactions may lead to higher rates of eDNA shedding. Nonetheless, we also observed higher turbidity in Lake Øymarksjøen than in Lake Stora Le. In Stora Le, non-detection of both A. astaci and P. leniusculus occurred only at locations with low CPUE (0.6 and 3.6) and the detected eDNA quantity corresponded well to the crayfish density. This also may be attributed to trapping prior to sampling as the data suggest from Øymarksjøen and other recent studies where no clear or only weak correlations were found between crayfish density and eDNA concentration (
Through the mesocosm experiment and the comparison with additional field data, we demonstrated that the detectability of both P. leniusculus and A. astaci eDNA is influenced by much more than mere population density. When sampling to monitor the presence of A. astaci, it is advisable to analyse the samples for eDNA of both the host and the pathogen for optimal detection efficiency. The crayfish plague agent A. astaci requires a crayfish host (or another freshwater decapod crustacean, see
This work was financially supported from several sources: 1) J.C. Rusch’s PhD project “Environmental DNA (eDNA) monitoring of two different freshwater pathogen-host complexes in the interface between nature and aquaculture” (eDNAqua-Fresh; 13076) funded by the Norwegian Veterinary Institute, 2) the project “Targeted strategies for safeguarding the noble crayfish against alien and emerging threats” (TARGET) NFR – 242907 funded by the Research Council of Norway and the Swedish Agency for Marine and Water Management (for the monitoring of Lake Stora Le). We are grateful to the Food Safety Authority and the Norwegian Environmental Agency for permits allowing the capture of live crayfish for aquarium trials. We are also grateful to Christian Stratton for his help and troubleshooting with the msocc calculations.
Primers and probes used in the present study
Data type: table (pdf file)
Explanation note: Primers and probes for Aphanomyces astaci and signal crayfish (Pacifastacus leniusculus) used in the present study.
ddPCR and qPCR data from the mesocosm experiments
Data type: table (excel file)
Explanation note: ddPCR and qPCR data from the mesocosm experiments at 20° C ("warm") (EXP1-EXP3). ddPCR and qPCR data from the mesocosm experiments at 10° C ("low temp") (EXP4-EXP6).
R-Script of GLM analysis
Data type: statistics
Explanation note: R-Script of GLM analysis of data from the mesocosm experiments.
R-script of MCOCC occupancy analysis
Data type: statistics
Explanation note: R-script of the MCOCC occupancy analysis for field data.
Results of the GLMM model
Data type: statistics
Explanation note: Results of the GLMM model for A. astaci (A. ast) and signal crayfish (P. len) determining statistical significance of three factors (food, temperature, density) on quantity of eDNA represented by positive droplets.
Agent levels of Aphanomyces astaci in individuals of Pacifastacus leniusculus used in the experiment
Data type: table (pdf file)
Explanation note: Agent levels of Aphanomyces astaci in individuals of Pacifastacus leniusculus used in the experiment. The agent level categories (A0-A5;