The Application of eDNA for Monitoring Aquatic Non-Indigenous Species: Practical and Policy Considerations

Round 1

Reviewer 1 Report

The paper: The application of eDNA for monitoring aquatic non-native species: practical and policy considerations is a review paper that addresses the application of eDNA in the aquatic environment.
There are some considerations that I would like to bring to the authors' attention:
"DNA-based methods": the authors use this term throughout the text as a synonym for eDNA, but genomic DNA extraction from tissue is also a DNA-based method and not an eDNA application. I suggest changing the term, eDNA would be just fine.
"Passive monitoring": personally, I don't like this term that authors use as a synonym for metabarcoding. When I search Scopus for "passive method" AND "metabarcoding", I get 36 results out of 4907 articles dealing with "metabarcoding". The difference between barcoding and metabarcoding is that I use species-specific primers in the first case and supposedly universal primers in the second. Neither method is truly passive, although I agree that there are passive methods for collecting eDNA samples.
The authors provide a concise overview "of the latest developments in DNA-based methods" (line 46), but looking at the references, I see zero publications in 2023, 7 publications in 2022, 12 publications in 2021 (one of which was in a non-peer-reviewed repository, while published in a journal), and 12 publications in 2020 (again, one of them in a non-peer-reviewed repository while published in a journal), while there are many publications dealing with eDNA in the last four years (2006 documents).
Pathogen detection is a key topic, and I suggest adding more than a single reference to pathogens and eDNA, as there are many recent publications addressing this topic in both marine and freshwater.
The authors report in the abstract that presence/absence and relative abundances are possible with DNA metabarcoding, which is absolutely correct. In Figure 2, the authors report that DNA barcoding is suitable for quantifying abundance, while DNA metabarcoding (passive monitoring) is only suitable for presence/absence.
I personally disagree with the statement in Figure 2 and wonder why the authors believe that DNA metabarcoding isn't suitable for representing relative abundances. Moreover, it's very difficult to infer abundance from DNA barcoding because the amount of DNA depends on the species, their habits, their seasonal muteness, their external or internal release of gametes, etc. All these aspects aren't even mentioned in the study.
Figure S1: Among the optimizations of eDNA testing, the authors mention the inclusion of positive and negative controls and biological replicates, but there is no mention of technical replicates, which are also an important step for eDNA testing. In the context of this figure, it's not clear to me what the authors mean by "improving sample handling" to reduce type I errors and - II.
Another important aspect not mentioned in the paper is the handling of databases and all the in silico optimization required to design species-specific oligonucleotides.
While LOD is mentioned, the authors didn't mention LOQ, which is a tricky aspect of eDNA that deserves future attention and will be considered in the work of Klymus et al. (2020). Perhaps the authors have some considerations on LOQ in the context of eDNA.

I suggest a comprehensive review of this work as it has great potential that needs to be better developed.

Minor comments:
Line breaks: there are many spaces between words in the text that the authors should comprehensively revise (e.g., line 39, 83, 113, 116, 145, etc.)
The style of references needs to be corrected as not all references follow the same style.
The term "aka" in scientific papers is a bit awkward. I'm sure the authors are young scientists and this term sounds correct, but I suggest finding another term instead of "aka".
Line 77: The reference is inserted incorrectly in the text
Line 230: Please remove one parenthesis.

Author Response

REVIEWER 1

Comments and Suggestions for Authors

The paper: The application of eDNA for monitoring aquatic non-native species: practical and policy considerations is a review paper that addresses the application of eDNA in the aquatic environment. There are some considerations that I would like to bring to the authors' attention:

"DNA-based methods": the authors use this term throughout the text as a synonym for eDNA, but genomic DNA extraction from tissue is also a DNA-based method and not an eDNA application. I suggest changing the term, eDNA would be just fine.

AU: The authors used the term “DNA-based methods” in a broad sense, to refer to all molecular biology methods that involve nucleic acids, including both eDNA and genomic DNA extraction from tissues/animals. We have rectified the term in the ms in Pg2L69.

"Passive monitoring": personally, I don't like this term that authors use as a synonym for metabarcoding. When I search Scopus for "passive method" AND "metabarcoding", I get 36 results out of 4907 articles dealing with "metabarcoding". The difference between barcoding and metabarcoding is that I use species-specific primers in the first case and supposedly universal primers in the second. Neither method is truly passive, although I agree that there are passive methods for collecting eDNA samples.

AU: We acknowledge Reviewer1 concerns but the authors used the term ‘passive monitoring’ to emphasize the non-targeted nature of metabarcoding, and the fact that it can detect a wide range of species without specifically targeting any particular group. This is not necessarily synonymous with metabarcoding, although metabarcoding can be used as a tool for passive monitoring. It should be noted, however, that "passive monitoring" can refer to a range of different monitoring methods, including those that do not involve environmental DNA analysis at all. The authors have clarified that "passive monitoring" specifically refers to the non-targeted approach of  eDNA analysis (Pg4L158-161).

 

The authors provide a concise overview "of the latest developments in DNA-based methods" (line 46), but looking at the references, I see zero publications in 2023, 7 publications in 2022, 12 publications in 2021 (one of which was in a non-peer-reviewed repository, while published in a journal), and 12 publications in 2020 (again, one of them in a non-peer-reviewed repository while published in a journal), while there are many publications dealing with eDNA in the last four years (2006 documents).

AU: Thank you for bringing this matter to our attention but in all honesty, there is not much literature on eDNA tools for NIS detection in 2023. Nonetheless, we have tried to include more recent literature on DNA-based tools for NIS detection, which can better reflect the current research in the field. For instance, there is now in the revised version seven publications from 2023 and seventeen new publications from 2022.

Pathogen detection is a key topic, and I suggest adding more than a single reference to pathogens and eDNA, as there are many recent publications addressing this topic in both marine and freshwater.

AU: Thank you, Reviewer1, for your suggestion. We agree that monitoring invasive and emergent disease species, including pathogens, is a critical aspect of environmental monitoring. However, the focus of our review is on non-pathogens since they represent a vast and complex area of research in their own right and apologize for any confusion on this point. We tried to clarify this aspect at the start of the introduction-objectives, Pg2L48-51 “In the interest of clarity and focus, we have chosen to limit our review to non-pathogenic invasive species, since the topic of invasive pathogens is complex and extensive, and deserves its own comprehensive review.’

The authors report in the abstract that presence/absence and relative abundances are possible with DNA metabarcoding, which is absolutely correct. In Figure 2, the authors report that DNA barcoding is suitable for quantifying abundance, while DNA metabarcoding (passive monitoring) is only suitable for presence/absence.
I personally disagree with the statement in Figure 2 and wonder why the authors believe that DNA metabarcoding isn't suitable for representing relative abundances. Moreover, it's very difficult to infer abundance from DNA barcoding because the amount of DNA depends on the species, their habits, their seasonal muteness, their external or internal release of gametes, etc. All these aspects aren't even mentioned in the study.

AU: We agree with Reviewer1 that metabarcoding is only suitable for representing relative abundances and the impossibility of estimating absolute abundances in eukaryotic multicellular organisms, particularly when using multicopy genes and also considering species biology. At the moment this can only be applied with confidence to unicellular organisms (e.g bacteria, protozoa, diatoms). The authors rectified Figure 2 accordingly, using relative abundances for both methods and discussed these aspects extensively in the manuscript (Pg4-5L197-261).

Figure S1: Among the optimizations of eDNA testing, the authors mention the inclusion of positive and negative controls and biological replicates, but there is no mention of technical replicates, which are also an important step for eDNA testing. In the context of this figure, it's not clear to me what the authors mean by "improving sample handling" to reduce type I errors and - II.

AU: We have included ‘technical replicates’ and clarified ‘sample handling’ in Figure S1 and Figure S1 legend. By improving sample handling we meant avoid cross-contamination between samples by taking simple steps as the use of gloves and minimize handling procedures as much as possible as illustrated in Figure 1.

Another important aspect not mentioned in the paper is the handling of databases and all the in-silico optimization required to design species-specific oligonucleotides.

AU: We agree with Reviewer1 and we have included in the manuscript ‘handling with databases’ and some ‘in-silico optimization considerations for species-specific primer design’. Pg8L339-344

While LOD is mentioned, the authors didn't mention LOQ, which is a tricky aspect of eDNA that deserves future attention and will be considered in the work of Klymus et al. (2020). Perhaps the authors have some considerations on LOQ in the context of eDNA.

AU: We have included LOQ and Klymus et al. (2020) in the manuscript. Pg7-8L357-367

I suggest a comprehensive review of this work as it has great potential that needs to be better developed.

AU: We would like to thank Reviewer1 for the insightful feedback on our manuscript. We have taken comments and suggestions into careful consideration and made revisions accordingly. We believe that these revisions have significantly improved the quality and accuracy of the manuscript.

Minor comments:

Line breaks: there are many spaces between words in the text that the authors should comprehensively revise (e.g., line 39, 83, 113, 116, 145, etc.)

AU: The manuscript was revised and modified accordingly

The style of references needs to be corrected as not all references follow the same style.

AU: The manuscript was revised and modified accordingly

The term "aka" in scientific papers is a bit awkward. I'm sure the authors are young scientists and this term sounds correct, but I suggest finding another term instead of "aka".

AU: The authors have modified aka by ‘also known as’

Line 77: The reference is inserted incorrectly in the text

AU: Modified accordingly

Line 230: Please remove one parenthesis.

AU: Modified accordingly

Reviewer 2 Report

The review is well written and focused and deserves to be published after some minor revisions.

Perhaps due to its focused writing, I feel that some possible options for DNA-based NIS surveillance have been left out from the review. I try to give examples of them here, but this is not an exhaustive listing, and the authors may look for more details before reviewing their text. Further, I try to discuss some points that the authors may take into consideration in their reviewed version.

Viability PCR (Emerson et al. 2017) is an option to disentangle living (organisms present at the time point of sampling) from dead material (organisms absent at the time point or long-distance drifted DNA). It is well-established in prokaryotic fields but has been tested for multicellular organisms as well (HIrohara et al 2021) and maybe it deserves to be mentioned in the review. Another well-established prokaryotic method that could be mentioned if monitoring unicellular eukaryotic NIS is CARD-FISH (Piwosz et al. 2021). It is a targeted method to quantify abundance of chosen species. Thirdly, multiplexing qPCR/ddPCR assays may allow for simultaneous testing for numerous species, reducing processing and handling times and lowering costs (e.g., Wozney & Wilson 2017).

Although eDNA-based methods “minimizes the onus of taxonomic expertise”, taxonomic expertise is still desperately needed in the analyses and reporting of the results. Taxonomic experts can validate reference sequences and point gross mistakes in the reference sequences or identifications based on the reference sequences that will be quite easily missed by a bioinformatician that trusts one’s codes and quality of the references. Should perhaps be mentioned in the review that eDNA-based methods do not replace taxonomic experts but complements them?

In the chapter “eDNA fate and impact on NIS detection”, the authors mention that DNA may remain detectable in sediments for several years. In fact, DNA may remain detectable in sediments at least several tens of years – there is a complete field of research on this topic, on ancient DNA.

As the authors mention in their review, long-distance spreading of traces of DNA causes confusion in the results. This spreading can in fact be very long-distance, see Gottschling et al. (2021). Due to this fact, eDNA-based NIS surveillance should perhaps include some kind of modelling/evaluation approach to give an estimate of the probability of long-distance carryover of DNA. Of course, this might be difficult to achieve but at least it should be discussed. Validation of the eDNA-based presence observations should be conducted with traditional surveys, to get physical evidence of viable organisms. This might be important for legal aspects of the monitoring. Unfortunately, burden of proofing is on the eDNA methods although traditional surveys have their own biases. eDNA-based NIS monitoring is expected to be perfect before applying it in “real-life” cases. Therefore, a validation scheme that Thalinger et al. (2020) proposed, is a minimum requirement for the case studies showcasing applicability of eDNA NIS monitoring. The authors could give more pronounced advice on this topic. The case reports should be written directly towards the management and policy makers, not fellow researchers.

One source of false positives that I have encountered frequently is not mentioned in the review. This is false positive due to errors in reference library – very common phenomenon at least in insects. Although BOLD is well annotated, it still includes several reference sequences that originate in fact from Ricketsia, Wolbachia or other endosymbionts/prey/ectosymbionts, and which causes false positives. I can envisage that the same holds for other organism groups and reference libraries as well. This is a big concern that needs to be tackled before DNA-based NIS monitoring can be implemented – difficult to apply legally binding monitoring if such errors are frequent. This is where an array of controls (lab and computer) and taxonomic experts become important. And standardization. When standardized, stakeholders are ready for the method. For false negatives, documenting and preventing PCR inhibition is indeed important. The best way of doing so might be performing the PCR inhibition test with adding an artificial sequence into the samples because it elucidates the effect of PCR inhibition on eDNA detection and concentration estimations (Doi et al. 2021).

References

Doi, H., Minamoto, T., Takahara, T., Tsuji, S., Uchii, K., Yamamoto, S., Katano, I., & Yamanaka, H. (2021). Compilation of real-time PCR conditions toward the standardization of environmental DNA methods. Ecological Research, 36, 379– 388.

Emerson, J.B., Adams, R.I., Román, C.M.B. et al. Schrödinger’s microbes: Tools for distinguishing the living from the dead in microbial ecosystems. Microbiome 5, 86 (2017).

Gottschling, M., Czech, L., Mahé, F., Adl, S. & Dunthorn, M. The windblown: possible explanations for dinophyte DNA in forest soils. J. Eukaryot. Microbiol. 68, e12833 (2021).

Hirohara T, Tsuri K, Miyagawa K, Paine RTR and Yamanaka H (2021) The Application of PMA (Propidium Monoazide) to Different Target Sequence Lengths of Zebrafish eDNA: A New Approach Aimed Toward Improving Environmental DNA Ecology and Biological Surveillance. Front. Ecol. Evol. 9:632973.

Piwosz K, Mukherjee I, Salcher MM, Grujčić V and Šimek K (2021) CARD-FISH in the Sequencing Era: Opening a New Universe of Protistan Ecology. Front. Microbiol. 12:640066.

Wozney, K. M., & Wilson, C. C. (2017). Quantitative PCR multiplexes for simultaneous multispecies detection of Asian carp eDNA. Journal of Great Lakes Research, 43(4), 771– 776.

Author Response

REVIEWER 2

Comments and Suggestions for Authors

The review is well written and focused and deserves to be published after some minor revisions.  Perhaps due to its focused writing, I feel that some possible options for DNA-based NIS surveillance have been left out from the review. I try to give examples of them here, but this is not an exhaustive listing, and the authors may look for more details before reviewing their text. Further, I try to discuss some points that the authors may take into consideration in their reviewed version.

AU: We appreciate Reviewer2's feedback and constructive suggestions, and we are grateful for the time and effort in helping to improve our manuscript.

Viability PCR (Emerson et al. 2017) is an option to disentangle living (organisms present at the time point of sampling) from dead material (organisms absent at the time point or long-distance drifted DNA). It is well-established in prokaryotic fields but has been tested for multicellular organisms as well (HIrohara et al 2021) and maybe it deserves to be mentioned in the review. Another well-established prokaryotic method that could be mentioned if monitoring unicellular eukaryotic NIS is CARD-FISH (Piwosz et al. 2021). It is a targeted method to quantify abundance of chosen species. Thirdly, multiplexing qPCR/ddPCR assays may allow for simultaneous testing for numerous species, reducing processing and handling times and lowering costs (e.g., Wozney & Wilson 2017).

AU: As mentioned previously to Reviewer1, the focus of our review is mainly on aquatic eukaryotes (e.g.meio-macrofauna), and we apologize for any confusion on this point. We have clarified this aspect in the revised version of the manuscript to better reflect the focus and scope of the review. Thank you for suggesting these methods, the first methods are more targeted towards microbes and despite the card FISH being applied to bacteria or viruses infecting NIS species the authors feel this might be out of scope. Nonetheless, we have included in the revised version the multiplexing qPCR/ddPCR and PMA techniques as means to detect NIS in marine environments. Pg3L131-158

 

Although eDNA-based methods “minimizes the onus of taxonomic expertise”, taxonomic expertise is still desperately needed in the analyses and reporting of the results. Taxonomic experts can validate reference sequences and point gross mistakes in the reference sequences or identifications based on the reference sequences that will be quite easily missed by a bioinformatician that trusts one’s codes and quality of the references. Should perhaps be mentioned in the review that eDNA-based methods do not replace taxonomic experts but complements them?

AU: We agree that both approaches are complementary and both have their respective strengths and limitations, and we have emphasized this in the revised manuscript. Under the section ‘Detecting aquatic NIS: morphology, eDNA and eRNA’ Pg2L62

 

In the chapter “eDNA fate and impact on NIS detection”, the authors mention that DNA may remain detectable in sediments for several years. In fact, DNA may remain detectable in sediments at least several tens of years – there is a complete field of research on this topic, on ancient DNA. 

AU: We recognize that ancient DNA is a complex and extensive subject, and we have touched upon some aspects on this topic as an example of eDNA fate in sediments. Pg6L315-321

 

As the authors mention in their review, long-distance spreading of traces of DNA causes confusion in the results. This spreading can in fact be very long-distance, see Gottschling et al. (2021). Due to this fact, eDNA-based NIS surveillance should perhaps include some kind of modelling/evaluation approach to give an estimate of the probability of long-distance carryover of DNA. Of course, this might be difficult to achieve but at least it should be discussed.

AU: The authors have tried to include these aspects of active and passive dispersal in the manuscript, Pg6L264-304

Validation of the eDNA-based presence observations should be conducted with traditional surveys, to get physical evidence of viable organisms. This might be important for legal aspects of the monitoring. Unfortunately, burden of proofing is on the eDNA methods although traditional surveys have their own biases. eDNA-based NIS monitoring is expected to be perfect before applying it in “real-life” cases. Therefore, a validation scheme that Thalinger et al. (2020) proposed, is a minimum requirement for the case studies showcasing applicability of eDNA NIS monitoring. The authors could give more pronounced advice on this topic. The case reports should be written directly towards the management and policy makers, not fellow researchers. 

AU: The authors have tried to emphasize these aspects in the manuscript (Pg8-9L404-413

 One source of false positives that I have encountered frequently is not mentioned in the review. This is false positive due to errors in reference library – very common phenomenon at least in insects. Although BOLD is well annotated, it still includes several reference sequences that originate in fact from Ricketsia, Wolbachia or other endosymbionts/prey/ectosymbionts, and which causes false positives. I can envisage that the same holds for other organism groups and reference libraries as well. This is a big concern that needs to be tackled before DNA-based NIS monitoring can be implemented – difficult to apply legally binding monitoring if such errors are frequent. This is where an array of controls (lab and computer) and taxonomic experts become important. And standardization. When standardized, stakeholders are ready for the method.

AU: Thank you Reviewer2 for raising such pertinent point. The authors have included these aspects in the revised version of the manuscript. (Pg7L341-345, Pg8-9L405-449 and Pg10L495-516)

For false negatives, documenting and preventing PCR inhibition is indeed important. The best way of doing so might be performing the PCR inhibition test with adding an artificial sequence into the samples because it elucidates the effect of PCR inhibition on eDNA detection and concentration estimations (Doi et al. 2021).

AU: The authors have now mentioned this as suggested and incorporated Doi et al. 2021. Pg7L351-357

 

References

Doi, H., Minamoto, T., Takahara, T., Tsuji, S., Uchii, K., Yamamoto, S., Katano, I., & Yamanaka, H. (2021). Compilation of real-time PCR conditions toward the standardization of environmental DNA methods. Ecological Research, 36, 379– 388.

 

Emerson, J.B., Adams, R.I., Román, C.M.B. et al. Schrödinger’s microbes: Tools for distinguishing the living from the dead in microbial ecosystems. Microbiome 5, 86 (2017).

Gottschling, M., Czech, L., Mahé, F., Adl, S. & Dunthorn, M. The windblown: possible explanations for dinophyte DNA in forest soils. J. Eukaryot. Microbiol. 68, e12833 (2021).

Hirohara T, Tsuri K, Miyagawa K, Paine RTR and Yamanaka H (2021) The Application of PMA (Propidium Monoazide) to Different Target Sequence Lengths of Zebrafish eDNA: A New Approach Aimed Toward Improving Environmental DNA Ecology and Biological Surveillance. Front. Ecol. Evol. 9:632973.

 

Piwosz K, Mukherjee I, Salcher MM, Grujčić V and Šimek K (2021) CARD-FISH in the Sequencing Era: Opening a New Universe of Protistan Ecology. Front. Microbiol. 12:640066.

Wozney, K. M., & Wilson, C. C. (2017). Quantitative PCR multiplexes for simultaneous multispecies detection of Asian carp eDNA. Journal of Great Lakes Research, 43(4), 771– 776.

Round 2

Reviewer 1 Report

Dear Editor and authors,

I believe that the manuscript in its present form is suitable for publication.