Skip to main content

eDNA-based monitoring of Batrachochytrium dendrobatidis and Batrachochytrium salamandrivorans with ddPCR in Luxembourg ponds: taking signals below the Limit of Detection (LOD) into account



Batrachochytrium dendrobatidis (Bd) and Batrachochytrium salamandrivorans (Bsal) are two pathogenic fungi that are a significant threat to amphibian communities worldwide. European populations are strongly impacted and the monitoring of the presence and spread of these pathogens is crucial for efficient decision-making in conservation management.


Here we proposed an environmental DNA (eDNA) monitoring of these two pathogenic agents through droplet digital PCR (ddPCR) based on water samples from 24 ponds in Luxembourg. In addition, amphibians were swabbed in eight of the targeted ponds in order to compare the two approaches at site-level detection. This study allowed the development of a new method taking below-Limit of Detection (LOD) results into account thanks to the statistical comparison of the frequencies of false positives in no template controls (NTC) and below-LOD results in technical replicates. In the eDNA-based approach, the use of this method led to an increase in Bd and Bsal detection of 28 and 50% respectively. In swabbing, this resulted in 8% more positive results for Bd. In some samples, the use of technical replicates allowed to recover above-LOD signals and increase Bd detection by 35 and 33% respectively for eDNA and swabbing, and Bsal detection by 25% for eDNA.


These results confirmed the usefulness of technical replicates to overcome high levels of stochasticity in very low concentration samples even for a highly sensitive technique such as ddPCR. In addition, it showed that below-LOD signals could be consistently recovered and the corresponding amplification events assigned either to positive or negative detection via the method developed here. This methodology might be particularly worth pursuing in pathogenic agents’ detection as false negatives could have important adverse consequences. In total, 15 ponds were found positive for Bd and four for Bsal. This study reports the first record of Bsal in Luxembourg.

Peer Review reports


Two chytrid fungi, Batrachochytrium dendrobatidis (Bd) and Batrachochytrium salamandrivorans (Bsal), possibly originating from Asia and introduced in other continents through pet trade, were identified as major drivers responsible for the global decline of amphibians observed over the last decades [1,2,3,4]. The systematic monitoring of these skin pathogenic fungi lethal to amphibians is crucial as it allows investigating both the local and general virulence and dispersal patterns, while considering co-factors such as environmental conditions (e.g. temperature, humidity, cover) or human activities (e.g. trade entry points, pet trade, nitrous fertilization) [4, 5]. Moreover, accumulating and aggregating monitoring data could help in the fine-tuning of local niche models [6] and lead to enhanced safety in conservation actions such as reintroduction and translocation projects [7].

Initially, the detection of these chytrid fungi relied on histopathology [8] and immune-histochemical assays [9, 10]. In recent surveys, the detection shifted to molecular-based methods with several qPCR detection assays designed to detect the DNA of both chytrid fungi based on specific primers and probes [11,12,13]. These molecular methods enabled the detection of the pathogens at any stage of their life cycle and were first applied to DNA extracts from toe clips and skin swabbing [11, 12]. In the last decades, the development of approaches based on environmental DNA (eDNA) allowed the recovery of organisms’ DNA from environmental samples such as air, soil or water [14]. The Bd/Bsal detection benefited from these advances and an increasing number of monitoring studies were based on eDNA extracted from water samples [15,16,17,18,19,20,21,22].

Droplet digital PCR (ddPCR) is increasingly being used in the field of eDNA. Through emulsion partitioning, ddPCR produces thousands of independent PCR reactions allowing for high-sensitivity detection and absolute quantification of template concentration in samples [23, 24]. Partially overcoming inhibition effects and known to outcompete qPCR in low eDNA concentrations, this technique is gaining popularity in the field of eDNA-based detection (e.g. [25,26,27,28,29,30]). However, the use of ddPCR for the detection of the chytrid pathogenic fungi has remained limited and indirect to date. For instance, it has been employed to evaluate, in Bd positive samples, the copy number in different strains [31] or to quantify Bd-positive samples in a qPCR survey [32]. However, given its high sensitivity, the direct use of ddPCR has the potential to significantly increase the efficiency of Bd and Bsal detection.

In eDNA-based detection, the very low concentrations of target DNA often result in signals that are difficult to delineate from the background noise associated with the methodology. These very low template concentrations may lead to the generation of signals below the Limit of Detection (LOD i.e. the lowest concentration of target DNA that can be detected with sufficient confidence e.g. [33]), and also increase the stochasticity in amplifications [34]. Thus, below-LOD signals could originate from the actual amplification of the target DNA and could be worth sorting apart from mere background noise. Indeed, in either analytical chemistry or eDNA-based detection, below-LOD values were often proven to convey valuable information [35,36,37,38,39,40]. However, in eDNA studies, LOD can be used as a threshold below which amplification events are disregarded (e.g. [41,42,43]). Discarding these signals could bias detection and produce false negatives, which can be particularly perilous when the object of detection is a pathogenic agent. For this type of target, decreasing false negatives should be highly promoted [37, 44], as non-detection could have important impacts. The fact that below-LOD results could originate from the actual presence of the pathogen was already suggested in a previous Bsal swabbing survey, but these could still not be considered genuine positives [45].

The presence of Bd has been confirmed in Luxembourg in 2009 [46, 47] though the zoospore load was generally lower than the one observed during severe chytridiomycosis outbreaks [47]. Concerning Bsal, despite its presence in the border region of neighbouring countries, it was not detected in Luxembourg [48]. Concomitantly, the decline of amphibian populations in Luxembourg in recent decades has prompted various measures such as legal protection, road-crossing structures, habitat improvement, and breeding pond creation or restoration [49,50,51,52]. Translocation or reintroduction programs have also been employed to protect or restore the populations of some species [51]. However, given the risk posed by Bd and the high rate of spread of Bsal [53], such initiatives, without adequate monitoring, could result in the transfer of naïve amphibians into contaminated ponds or conversely the introduction of potential pathogen reservoir specimens into ponds originally devoid of pathogens.

In the present study, ddPCR was used for the detection of the fungal pathogens Bd and Bsal in both eDNA-based and swab-based monitoring of 24 and 8 ponds respectively in Luxembourg. This survey allowed to further investigate the use of below-LOD amplification events as valid detection events through a comparison with the background noise generated from the methodology and assays.


None of the negative controls exhibited amplification, thus positives can soundly be assumed to originate from genuine template amplification.


The investigation concerning the background artefact for the Bd assay yielded two amplifications out of 190 NTC. Limit of Blank (LOB) was 0.48 cp/μl (copies per μl), the derived LOD 2.15 cp/μl and Limit of Quantification (LOQ) 3.54 cp/μl. Specific correction factors for the delineation of positive amplification events were calculated for each pond sampled in order to fit the impact of inhibition on their fluorescence level (Table S1).

For eDNA amplifications, 14 out of the 24 sites sampled were flagged as Bd-positive (Fig. 1a, Table 1): ten with concentrations above LOD and four with measurements below LOD but for which the frequency of amplification events was found statistically different from artefacts frequency in NTC (Table 1). Globally, the concentrations measured ranged from 0 to 113.73 cp/μl (Table S2).

Fig. 1
figure 1

Map of the detection results for a) Bd and b) Bsal. Green dots = negatives; yellow dots = positives < LOD (frequency of < LOD events significantly higher than NTC false positives frequency); red dots = positive > LOD; red triangle = positives > LOQ. This map was generated using QGIS 3.2.1 [54]

Table 1 Results of Bd detection for eDNA and swabbing approaches. Chi2 test = Pearson’s Chi-squared test to compare the frequencies of artefacts in NTC and below-LOD amplifications

Among the 14 amphibian specimens found positive with swabbing, seven yielded signals above LOD and seven below LOD in the first pass (Table 1). After the amplification of additional technical replicates in below-LOD signals, four specimen extracts yielded signals above LOD and one a below-LOD signal (which was found positive when statistically compared to NTC background noise). Two specimen extracts produced no additional signals (Table 1). In total, four out of the eight sites surveyed in swabbing yielded a positive detection for Bd. Among all samples, concentrations ranged from 0 to 409.26 cp/μl (Table S2).

Out of 12 positive specimens, 11 were anurans (Hyla arborea and Pelophylax sp.). The remaining positive signal (below LOD, but yielding significantly more amplification events than NTC) was produced from the Urodela species Ichthyosaura alpestris. Thus, in spite of a strong taxonomic sampling bias, with most of the specimens sampled belonging to Urodela (Table S3–88% urodelans and 12% anurans), Bd detection was mostly achieved in Anura (Table S3–39.3% in anurans and 0.7% in urodelans).

Six out of the eight sites where amphibian specimens were swabbed were also found positive with the eDNA approach (Table 1). The two remaining Bd-positive sites showed disagreement between eDNA and swabbing approaches: the site H-Bert47, where eDNA yielded no amplification, and the site H-Bert50, where none of the extracts from specimen swabbing produced a signal (Table 1).


The background artefact assessment for this assay yielded two amplifications out of 182 NTC samples. LOB was 0.51 cp/μl, derived LOD 2.18 cp/μl and LOQ 3.35 cp/μl. Specific correction factors for the delineation of positive amplification events were calculated for each pond sampled (Table S1).

In eDNA-based detection, four ponds were found positive for Bsal, two above LOD and two below LOD (but statistically different from NTC artefact background noise) (Fig. 1b, Table 2). Globally, concentrations ranged from 0 to 7.04 cp/μl (Table S2).

Table 2 Results of Bsal detection for eDNA and swabbing approaches. Chi2 test = Pearson’s Chi-squared test to compare the frequencies of artefacts in NTC and below-LOD amplifications

None of the extracts from specimen swabbing yielded a detection for Bsal. However, as there was no overlap between the eDNA positive sites and the swabbing experiment no comparison or confirmation could be gained via the swabbing approach (Table 2).


Bd detection

Bd was successfully detected in 15 out of the 24 sites sampled either through the eDNA, the swabbing approach or both (Fig. 1a). Two previous studies based on specimen swabbing already detected Bd in Luxembourg. The first one detected the pathogenic fungus in the central part of the country in two out of eight sites surveyed; in addition, two other sites provided results below the LOD in southwestern Luxembourg and were considered as ‘possible’ positive sites [46] (Fig. 2). These uncertain detection events could have benefited from a higher repetition level and a comparison with NTC artefactual frequency as implemented here. The second study, undertaken a few years later, detected Bd in five out of 12 sites surveyed, mainly in the northern and eastern part of the country [47] (Fig. 2). Altogether the results of the three studies show that Bd is well implanted in Luxembourg and likely spreading: one locality surveyed and found negative in 2008 [46], Useldange, was found positive in the present study (Fig. 2).

Fig. 2
figure 2

Map summarizing Bd detection results for the sites sampled in [46] (green squares = negatives; red squares = positives), in [47] (green triangles = negatives; red triangles = positives) and in this study (green dots = negatives; red dots = positives). The black circle highlights the Useldange site that was tested negative in 2008 [46] and positive in this study. This map was generated using QGIS 3.2.1 [54]

While Bd infection is often lethal in anurans [3], urodelans usually recover from infection and were thus flagged as potential reservoirs [55]. However, low Bd prevalence was found in wild urodelans populations [56] and some species were proven resistant possibly due to skin defenses [57]. Thus, urodelans could actually be less susceptible to act as pathogen reservoirs as previously thought. This could explain the unbalanced detection ratio found in this study, with a much higher rate of positives in anurans (Table S3–39.3% in anurans and 0.7% in urodelans) despite the lower sampled number of specimens for this group (Table S3–88% urodelans and 12% anurans). It is notable that in the course of the swabbing not a single animal with visible signs of chytridiomycosis could be detected.

Bsal detection

Bsal was detected in four of the 24 sites sampled using eDNA (Fig. 1b). Although this pathogen was already detected in two of the neighboring countries of Luxembourg, Belgium and Germany [6, 58], this is the first detection event of Bsal for the country itself. The positive sites were located in the western part of the country, which could mean that Bsal may have spread already across several regions of Luxembourg (but see below in ‘Potential biological significance of low-level signals’ concerning the pathogen vectors). While this finding should be further confirmed through methods other than eDNA alone, such as swabbing or observation of chytridiomycosis symptoms in wild specimens, it is an early warning for Luxembourg but also for other European countries as it suggests that Bsal is currently spreading.

Retrieving positive signals below-LOD

In order to achieve a higher degree of confidence in below-LOD amplification events, background noise signal was defined in each assay through the amplification of numerous NTC replicates (in ddPCR, such a background noise i.e. positive amplification artefacts in NTC was previously described in several studies e.g. [59,60,61]). This allowed to assess the background signal inherent to each ddPCR assay, so it could be compared statistically with repetitions of samples that yielded below-LOD signals. Those samples had to be replicated sufficiently to make statistical comparisons possible. This increase in technical replicates in very low template samples was already known to help counter the stochasticity and the resulting sampling effect [34]. Here, the comparison of their amplifications results to background noise was added to further ascertain the detection.

In this study, for the sites that yielded below-LOD signals in the first pass (Fig. 3), the application of this approach allowed to recover signals higher than the LOD but also to gain positive detections from the comparison of below-LOD signals with the NTC background. This yielded additional detection events originating from very low template concentrations possibly produced by low-density fungus populations. Using this approach, and in comparison to considering below-LOD signals as negatives, the eDNA-based detection increased by 28 and 50% respectively for Bd and Bsal. In swabbing, it brought 8% more positive results for Bd. It is also worth noting that the repetition in itself allowed to gain higher level signals (i.e. above LOD) that also significantly increased the number of detection events: in Bd, by 35% for eDNA and 33% for swabbing and 25% for Bsal eDNA detection. This highlights the risk of false negatives due to high levels of stochasticity in very low concentration samples and thus the importance of repetitions to retrieve the signal even for a highly sensitive technique such as ddPCR.

Fig. 3
figure 3

Schematic representation of Bd/Bsal detection process and decision-making

Although this approach is replicate-intensive, it helps to objectively formalize the recovery of below-LOD amplifications in the detection process. However, it is worth noting that, due to low concentration template amplification stochasticity, the two steps of technical repetition (i.e. the two repetitions of all field replicates and the 12–13 repetitions of < LOD signals from the original amplification of the 10 field replicates, Fig. 3) should be operated separately: if the first pass processed yields any >LOD signal, the site could be considered as positive and the second part of the repetition saved. The H-BETM03 site exemplifies the utility of implementing the repetition amplification through several steps as four field replicates yielded < LOD signals thus increasing accordingly the replication number when the two repetitions of the field replicates sufficed to obtain >LOD signals and thus confirm Bd detection (Table 1).

If several scenarios concerning target concentration along with an increasing number of < LOD signals in the ten initial amplifications from the field replicates are to be taken into account, the best strategy to limit the number of replicates to be processed in order to achieve detection, beyond one < LOD replicate, would be to go for two repetitions of all field replicates (Table S4). This is confirmed by the results obtained in several sites (H-BETM03, T-RU27, T-US03 and T-US04 for Bd and T-RED09 for Bsal - Tables 1 and 2) where > LOD signals were obtained from repetitions for samples that produced only a single < LOD signal in the first place.

This method might be worth pursuing especially in pathogen detection where false negatives can have highly detrimental consequences. Applying it on a routine basis could allow for a higher confidence level in both decision-making and management.

Potential biological significance of low-level signals

Low-level signals might originate from biotic vectors with low vectorial capacity such as waterfowl transporting encysted spores on the scales of their feet [62,63,64] or even abiotic events, such as fog and rain, that could also be vectors for viable spores [65]. Their persistent free-living stage can last several years, thus allowing a considerable time lag between spore arrival and the presence of target or reservoir species [17, 66]. As such, even low frequency events that convey but a few viable spores could, in time, contribute to the actual colonization of ponds. This could explain how ponds might be infected over long distances as the pathogens could benefit from the broad dispersion range of waterfowls and/or abiotic vectors.

Thus, low-level detection signals could be worth documenting as they might indicate local introduction events, even if not successful (i.e. signals could originate from decaying spores material which failed to survive [67]) or the presence of a very small pioneer population. Such early warning signals could flag sites for further scrutiny in future management plans.

Disagreement between eDNA and swabbing results

A disagreement was found between eDNA and swabbing at the site H-Bert47, with no amplification from eDNA (the additional processing of two more technical replicates for each field replicate did not yield any amplification) when two of the DNA extracts from swabbing produced several over and below-LOD signals. This inconsistency might relate to one drawback of the sampling technique employed here (i.e. composite sample cf. ‘Methods’ paragraph), which was designed to produce dependent replicates that could be pooled for statistical tests. Indeed, this sampling design can produce high levels of dilution: one positive 1 L sample could be mixed with nine negative ones and as the ten 0.5 L replicates are sampled from the composite sample, this could lead to a drastic increase of both stochasticity and sampling bias. In turn, this could jeopardize further detection from the corresponding pool of replicates. Hence, the independent field replicates approach might prove more efficient than the composite sample strategy adopted in this study, as no signal dilution would hamper amplification. However, independent field replicates would require a higher number of total replicates for the statistical comparison of below-LOD signals with artefacts from NTC as technical replicates could not be pooled per site.

Concerning the site H-Bert50, eDNA yielded a positive result from the statistical comparison with artefactual yield when swabbing did not produce any positive amplification events. This low signal detection suggests a low template concentration and thus a limited presence of Bd. This could explain that swabbing missed the signal as only 30 specimens were processed for this pond; a higher number of swabbed specimens might have allowed for the detection of Bd on this site through swabbing as well.

Thus, for Bd detection at site level, eDNA and swabbing results showed only marginal differences (which might be improved by applying independent sampling for eDNA). However, eDNA monitoring, contrary to swabbing, allows disconnecting the pathogen detection from the presence of amphibian hosts. This can be crucial as chytrid fungi could persist for several years in a pond without either target or reservoir amphibian species [17, 66]. Moreover, the eDNA sampling is less time intensive, thus less costly [20] and is less taxing for the amphibian communities as they are not captured and manipulated as in swabbing. Also, the absence of specimen handling and the shorter sampling time on site in eDNA approach could reduce the risk of accidental cross-contaminations between sites. Nevertheless, beyond site-level detection, and contrary to swabbing, eDNA cannot bring information on the prevalence of the pathogenic agents either at the global amphibian population scale or in the different taxonomic compartments of the communities. Thus, swabbing remains a valuable tool to get crucial fine-scale information on the spreading of Bd and Bsal.

Conservation implications

This study shows that Bd is well implanted in Luxembourg and likely spreading. The newly reported presence of Bsal in the country is alarming as it could negatively affects populations of the fire salamander and the great crested newt [45]. This highlights the continued need for field workers to follow good practices in order to minimize the spread of the pathogenic fungi. A comprehensive national Bd/Bsal action plan is currently being developed for Luxembourg. It plans to implement quantitative monitoring of Bsal annually at some 20 fire salamander sites near the border with neighboring countries [48] and to perform an eDNA-based monitoring of some 20 different ponds each year. A further implication of these results is the critical need for great caution in amphibian reintroductions and translocations. Given the presence of both pathogenic agents, there are concrete risks of transmitting pathogenic agents either way between destination ponds and transferred animals. Therefore, any reintroduction or translocation initiative should be associated with a pathogen monitoring in both the source and recipient populations and/or ponds. Otherwise, these actions could be counterproductive and spread the pathogens.


ddPCR allowed to monitor Bd and Bsal in eDNA samples from pond water and DNA extracts from specimen swabs. The results indicated a strong presence of Bd in Luxembourg and allowed to detect Bsal for the first time in the western part of the country, implying that the pathogenic agent is gaining ground in Europe.

Dealing with pathogen detection makes the prevention of false negatives crucial as these could interfere with monitoring attempts and thus jeopardize amphibian reintroduction or pathogen eradication attempts. Thus, it is crucial to recover as many detection signals as possible with a certain confidence level. In order to achieve this, a new method was proposed in this study through a thorough examination of the specific artefact levels in NTC for the two assays used and a higher level of replication for below-LOD signal samples. The statistical comparison of the frequencies of both items allowed for an objective and reproducible decision-making on the detection status of the targeted pathogens.

The application of this method allowed to recover a substantial amount of detection events which were likely originating from very low concentration templates. Those could be due to random and discrete propagule introductions through low intensity vectors such as wild birds, fog or rain. Even if some of these events may be non-viable introductions, they are worth documenting when the target monitored is a pathogenic agent such as Bd and Bsal, as even a mere signal of pathogen introduction is important to consider for management.


Water sampling

A network of 24 ponds located in Luxembourg was sampled during mid-April 2022 (Fig. 4, Table 3). At each site, a composite sample was established by mixing 10 1 L samples collected around the whole circumference of the pond. The sampling was conducted at a distance of 15–20 cm from the bank, sampling depth ranged from 5 to 10 cm from the surface. From this composite sample, ten replicates of 500 ml were subsampled using sterile Whirl Pak plastic bags. This sampling protocol was designed to generate dependent replicates for each pond, so these, along with their technical replicates, could be pooled in further statistical analyses.

Fig. 4
figure 4

Map of the sites sampled. This map was generated using QGIS 3.2.1 [54]

Table 3 Sampling sites (Pond areas were assessed from 2022 aerial photographs and includes the clearly discernible temporarily dry areas (

Each water sample, which had been stored at 4 °C before processing, was filtered through a 0.45 μm nitrocellulose membrane (Nalgene analytical funnel) using a peristaltic pump (Masterflex L/S Standard Pump Head) connected to a column drill (500 W). The resulting filter was then immediately stored in 800 μl ATL Qiagen lysis buffer at 4 °C in a 2 ml microfuge tube and frozen at − 20 °C back in the laboratory. For most ponds (76%), the filtration could not be performed for the entire 500 ml volume of each replicate due to the presence of high amounts of suspended sediments in the water column (Table S2- range 100–500 ml). For each site, a negative control was used to detect any possible cross-contamination: it consisted of a 500 ml bottle of distilled water that was taken into the field and subjected to the same storage and filtration process as the actual samples.

In order to prevent pathogens spreading between ponds, all the material used for the sampling of a pond was either disposed off (gloves), reused after a wash in 4% hypochlorite and a thorough water rinsing (buckets, bottles) or bleached with 4% hypochlorite (boots, plastic pants).

Specimen swabbing

For comparison with eDNA results, we conducted amphibian swabbing in eight of the surveyed ponds (Table 3). Up to 30 adult amphibians per pond were caught either in traps (3–4 aquatic funnel traps) or hand-netted for tree frogs. The animals were swabbed according to the following hierarchical order: tree frogs (five individuals in each of the three sites where the species occurred), any other frog or toad, great crested newts, all other newt species. Each animal was swabbed five times over the mouth, the inner thighs (anurans), the base of the tail (urodelans) and the webbing between the toes, resulting in 30 passages per animal. Swab tips were cut and immediately stored at 4 °C in 1.5 ml microfuge tubes containing 180 μl ATL Qiagen lysis buffer and then frozen at − 20 °C back in the laboratory.

To minimize contamination risk during swabbing, two persons were involved in the process: one holding the animal, and the other one swabbing and taking notes (species, sex). Both were wearing nitrile gloves that were discarded between animals. Nets were thoroughly cleaned and disinfected with ethanol (70%) between individuals and ponds. Traps were cleaned and dried for at least 24 hours. Specimen capture and swabbing always took place at a later date than water sampling for eDNA in order to avoid sediment resuspension before water collection. Animals were trapped using permits issued by the Luxembourg Ministry of the Environment, Climate and Biodiversity (References 98551GB/ne, 101,987, 99082GB/ne & 99087GB/ne). All specimens were released after swabbing.

DNA extraction


DNA was extracted from the nitrocellulose filter membranes using a Qiagen DNeasy Blood & Tissue Kit with a volume-adapted protocol: after using clean scissors to thoroughly shred the filter directly in its storage tube, 80 μl of Proteinase K were added to the lysis buffer used for filter preservation. After an overnight incubation at 56 °C, 600 μl of the lysis solution were recovered and mixed with 600 μl AL Qiagen lysis buffer (10 minutes incubation at 56 °C) and 600 μl 96% ethanol. After homogenization by mixing, the solution was transferred to Qiagen DNeasy 96 plates. The rest of the protocol followed the manufacturer’s instructions. DNA extracts were eluted in 100 μl of 56 °C-heated AE Qiagen elution buffer. To monitor any potential sample cross-contamination, negative controls (ATL buffer with proteinase K) were extracted alongside each series of samples.


DNA was extracted from swab tips using a Qiagen DNeasy Blood & Tissue Kit following the manufacturer’s protocol with an overnight lysis at 56 °C. DNA extracts were eluted in 100 μl of 56 °C-heated AE Qiagen elution buffer.


The DNA extracts were processed for ddPCR and read on a Bio–Rad QX200 suite according to the manufacturer’s instructions [68]. The ddPCR Supermix for Probes was used along with the specific primer sets previously designed for Bd [12] and Bsal [13] (with HEX and FAM fluorochromes respectively, quenchers were BHQ1 for both assays). The PCR cycling program followed the manufacturer’s instructions with an annealing temperature of 46 °C for Bd and 50 °C for Bsal for 50 cycles. The reaction mix was composed of 11 μl ddPCR Supermix for Probes, 760 nM primer and 430 nM probe completed with water to 21 μl and added. It was added with 1 μl template DNA extract for field/technical replicates and 1 μl water for NTC. In addition, several NTC reactions were processed for Bd and Bsal assays. In each assay, the amplification results from the negative controls and the additional NTC reactions were pooled to assess the frequency of false positives (for a total number of 190 and 182 respectively for Bd and Bsal). This pool is referred to as NTC in the rest of the text. All concentrations given in this study are concentrations assessed in the samples that were calculated by taking into account the sample dilution in the ddPCR mix (i.e. 1 μl of eDNA sample with 21 μl of ddPCR mix).


Double threshold approach as inhibition and artefact countermeasure

As described in [42], the inhibition level was specifically assessed in each pond for the two ddPCR assays. In short, for each pond, three positive control reactions (comprising 1 μl of gBlock gene fragments DNA described above) were respectively spiked with 1 μl eDNA extracted from three replicates. This allowed taking into account the impact of inhibition on the fluorescence level of positive amplification events for each pond [42]. In order to assess the fluorescence level in non-inhibited positive reactions for each of the two assays, 21 positive control replicates were generated with 1 μl gBlocks alone.

From raw fluorescence amplitude measurements and approximating their distribution as normal either for negative and positive amplification events, two correction factors were calculated for each pond: an Upper and a Lower Threshold Correction Factor (UTCF and LTCF [42]). These correction factors were then used to define an upper threshold (to sieve out the high fluorescence artefacts produced by ddPCR i.e. ‘stars’ [42]) and a lower threshold (to sieve out intermediate fluorescence amplifications between positive and negative amplification events i.e. ‘rain’ cf. Biorad documentation). The use of these two thresholds produced a consistent and specific delineation of positive amplification events in replicates processed from each pond. It allowed, in each sample, the positive amplification events delineation process to fit both the baseline fluorescence shifting and the fluorescence dropping due to the various inhibition level in ponds [42]. Through this process, absolute concentrations of the samples were recovered (Table S2).

Limit of blank, limit of detection and limit of quantification

For both assays, several metrics were calculated based on NTC and a decimal dilution series of gBlock fragments (ranging from 5.5 ng/μl to 5.5.10−10 ng/μl and 6 ng/μl to 6.10−10 ng/μl for Bd and Bsal respectively, with 6 replicates for each concentration) [69]: 1) the Limit of Blank (LOB), defined as the highest concentration that can be found in NTC, was determined from 190 NTC for Bd and 182 NTC for Bsal (LOB = mean NTC + 3 SD NTC); 2) the Limit of Detection (LOD), defined as the lowest concentration of target DNA that can be detected with sufficient confidence, was derived from LOB using the standard deviation from the lowest detectable concentration of positive control (LOD = LOB + 3 (SDlow concentration sample)). To strongly minimize false positives, a 99.73% confidence interval was chosen for LOB and LOD calculations. The LOQ (Limit of Quantification) i.e. the lowest concentration that can be assessed in 90% of the replicates, was established from the decimal dilution series of gBlock fragments.

Decision making in Bd/Bsal detection (Fig. 3)

In a first pass, the 10 field replicates from each pond were analyzed for the presence of Bd and Bsal. If at least one of the replicates exhibited a concentration higher than the LOD, the pathogen was considered as detected on the site. If no replicate yielded any signal, the site was considered negative for the targeted pathogen. If one or several of the field replicates produced a signal below LOD, these replicates were repeated 12–13 times. In addition, all the other replicates from the site were replicated two times in order to compensate for the sampling bias generated by the composite sampling method. If no signal over the LOD was obtained from these additional technical replicates, the frequency of the below-LOD positive amplification events for the whole site was tested against the frequency of false positives in NTC assessed for each assay (with a total of 30 + 12-13n replications per site, n being the number of < LOD signals obtained in the first pass; 43 replicates total for sites when n = 1 and 82 when n = 4 (Tables 1 and 2)). The intrinsic dependency of replicates from the composite sample collected in each site allowed this pooling of amplification results at site level for statistical analyses. The statistical comparison between the frequencies of false positives in NTC and below-LOD signals in the targeted sites allows the recovery of pathogen detection information with a higher level of confidence for signals that would have been otherwise discarded. The same process was applied to the results obtained from amphibian swabbing with the difference that the statistical test of below-LOD results was used at the specimen level.

Statistics and data analysis

The test of the frequency deviation between artefacts in NTC and below-LOD amplifications with Pearson’s Chi-squared (with simulated p-values based on 2000 Monte Carlo simulation replicates) along with the calculation of the correction factors were conducted in the RStudio environment 2022.02.3 (Build 492) [70] with R version 4.2.0 [71]. Maps were generated using QGIS 3.2.1 [54].

Availability of data and materials

Raw data and materials are available on request.


  1. Martel A, Blooi M, Adriaensen C, Van Rooij P, Beukema W, Fisher MC, Farrer RA, Schmidt BR, Tobler U, Goka K, et al. Recent introduction of a chytrid fungus endangers Western Palearctic salamanders. Science. 2014;346(6209):630–1.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  2. Yuan ZY, Martel A, Wu J, Van Praet S, Canessa S, Pasmans F. Widespread occurrence of an emerging fungal pathogen in heavily traded Chinese urodelan species. Conserv Lett. 2018;11(4):e12436.

    Article  Google Scholar 

  3. Scheele B, Pasmans F, Skerratt LF, Berger L, Martel A, Beukema W, Acevedo AA, Burrowes PA, Carvalho T, Catenazzi A, et al. Amphibian fungal panzootic causes catastrophic and ongoing loss of biodiversity. Science. 2019;363(6434):1459.

    Article  PubMed  CAS  Google Scholar 

  4. Fisher MC, Garner TWJ. Chytrid fungi and global amphibian declines. Nat Rev Microbiol. 2020;18(6):332–43.

    Article  PubMed  CAS  Google Scholar 

  5. Fisher MC, Henk DA, Briggs CJ, Brownstein JS, Madoff LC, McCraw SL, Gurr SJ. Emerging fungal threats to animal, plant and ecosystem health. Nature. 2012;484(7393):186–94.

    Article  PubMed  CAS  Google Scholar 

  6. Beukema W, Erens J, Schulz V, Stegen G, Spitzen-van der Sluijs A, Stark T, Laudelout A, Kinet T, Kirschey T, Poulain M, et al. Landscape epidemiology of Batrachochytrium salamandrivorans: reconciling data limitations and conservation urgency. Ecol Appl. 2021;31(5):e02342.

    Article  PubMed  Google Scholar 

  7. Scheele BC, Hollanders M, Hoffmann EP, Newell DA, Lindenmayer DB, McFadden M, Gilbert DJ, Grogan LF. Conservation translocations for amphibian species threatened by chytrid fungus: a review, conceptual framework, and recommendations. Conserv Sci Pract. 2021;3(11):e524.

    Article  Google Scholar 

  8. Berger L, Speare R, Kent A. Diagnosis of chytridiomycosis of amphibians by histological examination. Zoos Print J. 1999;15:184–90.

    Article  Google Scholar 

  9. Hyatt AD, Boyle DG, Olsen V, Boyle DB, Berger L, Obendorf D, Dalton A, Kriger K, Heros M, Hines H, et al. Diagnostic assays and sampling protocols for the detection of Batrachochytrium dendrobatidis. Dis Aquat Org. 2007;73(3):175–92.

    Article  CAS  Google Scholar 

  10. Skerratt LF, Mendez D, McDonald KR, Garland S, Livingstone J, Berger L, Speare R. Validation of diagnostic tests in wildlife: the case of Chytridiomycosis in wild amphibians. J Herpetol. 2011;45(4):444–50.

    Article  Google Scholar 

  11. Blooi M, Pasmans F, Longcore JE, Spitzen-van der Sluijs A, Vercammen F, Martel A. Duplex real-time PCR for rapid simultaneous detection of Batrachochytrium dendrobatidis and Batrachochytrium salamandrivorans in amphibian samples. J Clin Microbiol. 2013a;51:4173–7.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  12. Boyle DG, Boyle DB, Olsen V, Morgan JAT, Hyatt AD. Rapid quantitative detection of chytridiomycosis (Batrachochytrium dendrobatidis) in amphibian samples using real-time Taqman PCR assay. Dis Aquat Org. 2004;60(2):141–8.

    Article  CAS  Google Scholar 

  13. Spitzen-van der Sluijs A, Stark T, DeJean T, Verbrugghe E, Herder J, Gilbert M, Janse J, Martel A, Pasmans F, Valentini A. Using environmental DNA for detection of Batrachochytrium salamandrivorans in natural water. Environmental DNA. 2020;2(4):565–71.

    Article  Google Scholar 

  14. Barnes MA, Turner CR. The ecology of environmental DNA and implications for conservation genetics. Conserv Genet. 2016;17(1):1–17.

    Article  CAS  Google Scholar 

  15. Shin J, Bataille A, Kosch TA, Waldman B. Swabbing often fails to detect amphibian Chytridiomycosis under conditions of low infection load. PLoS One. 2014;9(10):e111091.

    Article  PubMed  PubMed Central  Google Scholar 

  16. Kamoroff C, Goldberg CS. Using environmental DNA for early detection of amphibian chytrid fungus Batrachochytrium dendrobatidis prior to a ranid die-off. Dis Aquat Org. 2017;127(1):75–9.

    Article  Google Scholar 

  17. Mosher BA, Huyvaert KP, Bailey LL. Beyond the swab: ecosystem sampling to understand the persistence of an amphibian pathogen. Oecologia. 2018;188(1):319–30.

    Article  PubMed  Google Scholar 

  18. Julian JT, Glenney GW, Rees C. Evaluating observer bias and seasonal detection rates in amphibian pathogen eDNA collections by citizen scientists. Dis Aquat Org. 2019;134(1):15–24.

    Article  CAS  Google Scholar 

  19. Barnes MA, Brown AD, Daum MN, de la Garza KA, Driskill J, Garrett K, Goldstein MS, Luk A, Maguire JI, Moke R, et al. Detection of the amphibian pathogens Chytrid fungus (Batrachochytrium dendrobatidis) and Ranavirus in West Texas, USA, using environmental DNA. J Wildl Dis. 2020;56(3):702–6.

    Article  PubMed  CAS  Google Scholar 

  20. Gonzalez DL, Balaz V, Vojar J, Chajma P. Dual detection of the Chytrid Fungi Batrachochytrium spp. with an Enhanced Environmental DNA Approach. J Fungus. 2021;7(4):258.

    Article  CAS  Google Scholar 

  21. Hossack BR, Oja EB, Owens AK, Hall D, Cobos C, Crawford CL, Goldberg CS, Hedwall S, Howell PE, Lemos-Espinal JA, et al. Empirical evidence for effects of invasive American bullfrogs on occurrence of native amphibians and emerging pathogens. Ecol Appl. 2022;33(2):e2785.

  22. Kamoroff C, Goldberg CS, Grasso RL. Rapid detection of amphibian chytrid fungus Batrachochytrium dendrobatidis using in situ DNA extraction and a handheld mobile thermocycler. Dis Aquat Org. 2023;152:99–108.

    Article  Google Scholar 

  23. Vogelstein B, Kinzler KW. Digital PCR. Proc Natl Acad Sci U S A. 1999;96(16):9236–41.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  24. White RA, Blainey PC, Fan HC, Quake SR. Digital PCR provides sensitive and absolute calibration for high throughput sequencing. BMC Genomics. 2009;10:1–2.

    Google Scholar 

  25. Doi H, Takahara T, Minamoto T, Matsuhashi S, Uchii K, Yamanaka H. Droplet Digital polymerase chain reaction (PCR) outperforms real-time PCR in the detection of environmental DNA from an invasive fish species. Environ Sci Technol. 2015a;49(9):5601–8.

    Article  PubMed  CAS  Google Scholar 

  26. Doi H, Uchii K, Takahara T, Matsuhashi S, Yamanaka H, Minamoto T. Use of Droplet Digital PCR for estimation of fish abundance and biomass in environmental DNA surveys. PLoS One. 2015b;10(3):e0122763.

    Article  PubMed  PubMed Central  Google Scholar 

  27. Hamaguchi M, Shimabukuro H, Hori M, Yoshida G, Terada T, Miyajima T. Quantitative real-time polymerase chain reaction (PCR) and droplet digital PCR duplex assays for detecting Zostera marina DNA in coastal sediments. LIMNOL OCEANOGR-METH. 2018;16(4):253–64.

    Article  CAS  Google Scholar 

  28. Baker CS, Steel D, Nieukirk S, Klinck H. Environmental DNA (eDNA) from the wake of the whales: Droplet Digital PCR for detection and species identification. Front Mar Sci. 2018;5:11.

    Article  Google Scholar 

  29. Tsang HH, Domingos JA, Westaway JAF, Kam MHY, Huerlimann R, Gomes GB. Digital Droplet PCR-based environmental DNA tool for monitoring Cryptocaryon irritans in a marine fish farm from Hong Kong. Diversity-Basel. 2021;13(8):350.

    Article  CAS  Google Scholar 

  30. Vautiera M, Chardon C, Goulona C, Guillarda J, Domaizon I. A quantitative eDNA-based approach to monitor fish spawning in lakes: application to European perch and whitefish. Fish Res. 2023;264:106708.

    Article  Google Scholar 

  31. Rebollar EA, Woodhams DC, LaBumbard B, Kielgast J, Harris RN. Prevalence and pathogen load estimates for the fungus Batrachochytrium dendrobatidis are impacted by ITS DNA copy number variation. Dis Aquat Org. 2017;123(3):213–26.

    Article  Google Scholar 

  32. Kostanjsek R, Turk M, Vek M, Gutierrez-Aguirre I, Cimerman NG. First screening for Batrachochytrium dendrobatidis, B. Salamandrivorans and Ranavirus infections in wild and captive amphibians in Slovenia. Salamandra. 2021;57(1):162–6.

    Google Scholar 

  33. Forootan A, Sjöback R, Björkman J, Sjögreen B, Linz L, Kubista M. Methods to determine limit of detection and limit of quantification in quantitative real-time PCR (qPCR). Biomol Detect Quantif. 2017;12:1–6.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  34. Kay S, Van den Eede G. The limits of GMO detection. Nat Biotechnol. 2001;19(5):405–5.

    Article  PubMed  CAS  Google Scholar 

  35. Gilliom RJ, Hirsch RM, Gilroy EJ. Effect of censoring trace-level water-quality data on trend-detection capability. Environ Sci Technol. 1984;18(7):530–5.

    Article  CAS  Google Scholar 

  36. Coleman D, Auses J, Grams N. Regulation - from an industry perspective or relationships between detection limits, quantitation limits, and significant digits. Chemom Intell Lab Syst. 1997;37(1):71–80.

    Article  CAS  Google Scholar 

  37. Trujillo-Gonzalez A, Becker JA, Huerlimann R, Saunders RJ, Hutson KS. Can environmental DNA be used for aquatic biosecurity in the aquarium fish trade? Biol Invasions. 2019;22(3):1011–25.

    Article  Google Scholar 

  38. Abbott C, Coulson M, Gagné N, Lacoursière-Roussel A, Parent GJ, Bajno R, Dietrich C, May-McNally S. Guidance on the use of targeted environmental DNA (eDNA) analysis for the Management of Aquatic Invasive Species and Species at risk. Canadian Science Advisory Research Document; 2021. 2021/019. iv: 1–42.

    Google Scholar 

  39. Klymus KE, Merkes CM, Allison MJ, Goldberg CS, Helbing CC, Hunter ME, Jackson CA, Lance RF, Mangan AM, Monroe EM, et al. Reporting the limits of detection and quantification for environmental DNA assays. Environmental DNA. 2019;2(3):271–82.

    Article  Google Scholar 

  40. Kralik P, Ricchi M. A basic guide to real time PCR in microbial diagnostics: definitions, parameters, and everything. Front Microbiol. 2017;8:108.

    Article  PubMed  PubMed Central  Google Scholar 

  41. Strand DA, Johnsen SI, Rusch JC, Agersnap S, Larsen WB, Knudsen SW, Moller PR, Vralstad T. Monitoring a Norwegian freshwater crayfish tragedy: eDNA snapshots of invasion, infection and extinction. J Appl Ecol. 2018;56(7):1661–73.

    Article  Google Scholar 

  42. Porco D, Hermant S, Purnomo CA, Horn M, Marson G, Colling G. Getting rid of ‘rain’ and ‘stars’: mitigating inhibition effects on ddPCR data analysis, the case study of the invasive crayfish Pacifastacus leniusculus in the streams of Luxembourg. PLoS One. 2022;17(11):e0275363.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  43. Uthicke S, Robson B, Doyle JR, Logan M, Pratchett MS, Lamare M. Developing an effective marine eDNA monitoring: eDNA detection at pre-outbreak densities of corallivorous seastar (Acanthaster cf. solaris). Sci Total Environ. 2022;851:158143.

    Article  PubMed  CAS  Google Scholar 

  44. Farrell JA, Whitmore L, Duffy DJ. The promise and pitfalls of environmental DNA and RNA approaches for the monitoring of human and animal pathogens from aquatic sources. Bioscience. 2021;71(6):609–25.

    Article  Google Scholar 

  45. Lötters S, Wagner N, Albaladejo G, Boning P, Dalbeck L, Dussel H, Feldmeier S, Guschal M, Kirst K, Ohlhoff D, et al. The amphibian pathogen Batrachochytrium salamandrivorans in the hotspot of its European invasive range: past - present - future. Salamandra. 2020;56(3):173–88.

    Google Scholar 

  46. Wood LR, Griffiths RA, Schley L. Amphibian chytridiomycosis in Luxembourg. Bull Soc Nat Luxemb. 2009;110:109–14.

    Google Scholar 

  47. Proess P, Ohst T, Plötner J, Engel E. Untersuchungen zum Vorkommen der Geburtshelferkröte (Alytes obstetricans) und zur Verbreitung des Chytrid-Pilzes (Batrachochytrium dendrobatidis) in Luxemburg. Bull Soc Nat luxemb. 2015;117:63–76.

    Google Scholar 

  48. Muller J, Plewnia A, Böning P, Frantz AC, Stassen R. Quantitative larval monitoring of Salamandra salamandra as an early warning system for Batrachochytrium salamandrivorans detection in Luxembourg. Bull Soc Nat Luxemb. 2022;124:167–78.

    Google Scholar 

  49. Junck C, Schoos F. Neuausbreitung des Laubfrosches (Hyla arborea) in der Folge von Biotopverbesserungsmassnahmen im Zentrum Luxemburgs. Bull Soc Nat luxemb. 2000;100:97–101.

    Google Scholar 

  50. Proess R. Verbreitungsatlas der Amphibien des Großherzogtums Luxemburg, vol. 75. Luxembourg: Musée national d’histoire naturelle; 2016.

    Google Scholar 

  51. Glesener L, Gräser P, Schneider S. Conservation and development of great crested newt (Triturus cristatus Laurenti, 1768) populations in the west and south-west of Luxembourg. Bull Soc Nat luxemb. 2022;124:107–24.

    Google Scholar 

  52. Schneider S, Glesener L. In situ-Ansiedlung der Gelbbauchunke in Luxemburg. Artenschutzmaßnahmen und deren Abhängigkeit von Witterungs- und Klimabedingungen. Digitalmagazin Expertenbrief Landschaftspflege. 2023;3.

  53. Schmidt BR, Bozzuto C, Lotters S, Steinfartz S. Dynamics of host populations affected by the emerging fungal pathogen Batrachochytrium salamandrivorans. R Soc Open Sci. 2017;4(3):160801.

    Article  PubMed  PubMed Central  Google Scholar 

  54. QGIS Geographic Information System. QGIS Association; 2022.

    Google Scholar 

  55. Davidson EW, Parris M, Collins JP, Longcore JE, Pessier AP, Brunner J. Pathogenicity and transmission of chytridiomycosis in tiger salamanders (Ambystoma tigrinum). Copeia. 2003;3:601–7.

    Article  Google Scholar 

  56. Keitzer SC, Goforth R, Pessier AP, Johnson AJ. Survey for the pathogenic Chytrid fungus Batrachochytrium dendrobatidis in southwestern North Carolina salamander populations. J Wildl Dis. 2011;47(2):455–8.

    Article  PubMed  Google Scholar 

  57. Pasmans F, Van Rooij P, Blooi M, Tessa G, Bogaerts S, Sotgiu G, Garner TWJ, Fisher MC, Schmidt BR, Woeltjes T, et al. Resistance to Chytridiomycosis in European plethodontid salamanders of the genus Speleomantes. PLoS One. 2013;8(5):e63639.

  58. More S, Miranda MA, Bicout D, Botner A, Butterworth A, Calistri P, Depner K, Edwards S, Garin-Bastuji B, Good M, et al. Risk of survival, establishment and spread of Batrachochytrium salamandrivorans (Bsal) in the EU. EFSA J. 2018;16(4):e05259.

    PubMed  PubMed Central  Google Scholar 

  59. Witte AK, Fister S, Mester P, Schoder D, Rossmanith P. Evaluation of the performance of quantitative detection of the listeria monocytogenes prfA locus with droplet digital PCR. Anal Bioanal Chem. 2016;408(27):7583–93.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  60. Kiselinova M, Pasternak AO, De Spiegelaere W, Vogelaers D, Berkhout B, Vandekerckhove L. Comparison of Droplet Digital PCR and Seminested real-time PCR for quantification of cell-associated HIV-1 RNA. PLoS One. 2014;9(1):8.

    Article  Google Scholar 

  61. Strain MC, Lada SM, Luong T, Rought SE, Gianella S, Terry VH, Spina CA, Woelk CH, Richman DD. Highly precise measurement of HIV DNA by Droplet Digital PCR. PLoS One. 2013;8(4):8.

    Article  Google Scholar 

  62. Garmyn A, Van Rooij P, Pasmans F, Hellebuyck T, Van den Broeck W, Haesebrouck F, Martel A. Waterfowl: potential environmental reservoirs of the Chytrid fungus Batrachochytrium dendrobatidis. PLoS One. 2012;7(4):e35038.

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  63. Stegen G, Pasmans F, Schmidt BR, Rouffaer LO, Van Praet S, Schaub M, Canessa S, Laudelout A, Kinet T, Adriaensen C, et al. Drivers of salamander extirpation mediated by Batrachochytrium salamandrivorans. Nature. 2017;544(7650):353.

    Article  PubMed  CAS  Google Scholar 

  64. Hanlon SM, Henson JR, Kerby JL. Detection of amphibian chytrid fungus on waterfowl integument in natural settings. Dis Aquat Org. 2017;126(1):71–4.

    Article  Google Scholar 

  65. Prado JS, Ernetti JR, Pontes MR, Toledo LF. Chytrid in the clouds: an alternative passive transport of a lethal pathogen for amphibians. Hydrobiologia. 2023;850:2061–73.

  66. Chestnut T, Anderson C, Popa R, Blaustein AR, Voytek M, Olson DH, Kirshtein J. Heterogeneous occupancy and density estimates of the pathogenic fungus Batrachochytrium dendrobatidis in waters of North America. PLoS One. 2014;9(9):e106790.

    Article  PubMed  PubMed Central  Google Scholar 

  67. Blooi M, Martel A, Vercammen F, Pasmans F. Combining ethidium monoazide treatment with real-time PCR selectively quantifies viable Batrachochytrium dendrobatidis cells. Fungal Biol. 2013b;117(2):156–62.

    Article  PubMed  CAS  Google Scholar 

  68. Droplet Digital PCR Applications Guide [].

  69. Armbruster DA, Pry T. Limit of blank, limit of detection and limit of quantitation. Clin Biochem Rev. 2008;29:S49–52.

    PubMed  PubMed Central  Google Scholar 

  70. RStudio_Team. RStudio: integrated development for R. Boston, MA: RStudio, PBC; 2020.

    Google Scholar 

  71. R_Core_Team. R: a language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing; 2022.

    Google Scholar 

Download references


We want to thank Tiago de Sousa and Louis Lestang from the Administration de la Nature et des Forêts (ANF) as well as Nora Elvinger from the Ministère de l’Environnement, du Climat et de la Biodiversité (MECB) from Luxembourg for their support.


The project was funded by the MECB, the ANF from Luxembourg, and an internal grant from National Natural History Museum of Luxembourg.

Author information

Authors and Affiliations



Conceptualisation: DP. Field work: CAP; RP; LG; DP; KJ. Laboratory work: SL; DP. Data curation: DP. Formal analysis: DP. Funding acquisition: AF; GC; SS; RS; DP. Figures/tables: DP. Writing original draft: DP. All authors reviewed the manuscript. All authors read and approved the final manuscript.

Corresponding author

Correspondence to David Porco.

Ethics declarations

Ethics approval and consent to participate

All experiments were approved by the committee of Administration de la Nature et des Forêts (ANF) and the Ministère de l’Environnement, du Climat et de la Biodiversité (MECB) from Luxembourg (committee for the implementation of the national plan for emerging infectious diseases of amphibians), and performed in accordance with relevant guidelines and regulations. Animals were trapped using permits issued by the MECB (References 98551GB/ne, 101987, 99082GB/ne & 99087GB/ne).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

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 The Creative Commons Public Domain Dedication waiver ( applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Porco, D., Purnomo, C.A., Glesener, L. et al. eDNA-based monitoring of Batrachochytrium dendrobatidis and Batrachochytrium salamandrivorans with ddPCR in Luxembourg ponds: taking signals below the Limit of Detection (LOD) into account. BMC Ecol Evo 24, 4 (2024).

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: