In Silico Infrared Spectroscopy as a Benchmark for Identifying Seized Samples Suspected of Being N-Ethylpentylone

Round 1

Reviewer 1 Report

The manuscript discusses the application of chemometric evaluation/modelling of infrared analysis of experimental data in 68 samples seized by law enforcement.  The manuscript is well written and the data presented in a logical and clear manner.  There are some aspects that are neglected though and the reviewer would like to see these included.

[1] The specific application of chemometrics in forensic drug analysis (using NMR, GC-MS and LC-MS data) is very briefly discussed in the introduction and there are a number of key references that could have been included to show the work in this specific area to provide context to the work.

[2] The authors comment that "Identifying NPS is challenging. In the case of cathinones, homologous amphetamines can impair correct identification due to structural similarities" - this maybe true for infrared analysis, however, there are many papers which have utilised NMR, GC-MS (some of which have coupled these techniques with chemometrics) to discriminate between structural isomers - this section should be revised taking into account the prior work in the area by teams - as it implies that the authors are trying to suggest that this is something that hasn't been tackled before.

[3] Why has there been no confirmation of the content of the samples and comparison with a confirmatory technique (such as GC-MS or LC-MS?).  GC-MS and similar techniques are considered the "gold standard" in forensic analysis - however the authors have not provided any data, in the main manuscript or the SI, to confirm the content of the 68 samples.

[4] Building on point 3 - there have been at least four published studies in the past two years including: Green, Int. J. Drug Policy (2020); Wallace, Drug Testing Analysis (2021); Fregonese, Front. Psychiatry (2021) and Dixon, J. Pharm. Biomed. Analysis (2022) which describe the comparison of infrared analysis of seized forensic samples and each of them both indicate that infrared is the least discriminatory technique compared to other techniques.  How does the authors current data compare with these studies?  A comparison of the approach with a "gold standard" technique and discussion within the context of other studies is needed to clearly articulate the novelty of the approach described in the manuscript.  This is the papers biggest failing and must be addressed before publication is granted.

Overall the paper describes a good chemometric approach to the rapid analysis of seized samples - however the lack of a comparison of a gold standard technique and contextualisation of the method with previously published studies to benchmark the approach severely diminishes the work in terms of novelty.  Recommend MAJOR revision of the manuscript and inclusion of key data (GC-MS) as highlighted in this review.

Author Response

Response Letter

 

Manuscript ID: psychoactives-1982638

Type of manuscript: Article

Title: In silico infrared spectroscopy as a benchmark for identifying seized samples suspected of being N-ethylpentylone.

Prof. Dr. Ricardo Dinis-Oliveira, Editor-in-Chief, Psychoactives

  

Dear Prof. Dr. Dinis-Oliveira,

 

We have adjusted the manuscript on the basis of the recommendations and notes made by the reviewers. We have re-evaluated the introduction so that the section would provide more evidence of the existing problem and the solution proposed by us. We have introduced a more specific discussion of the relationship between this work and other published research. We agree with the reviewers about changing the title and the conclusion. Finally, we have adjusted the references as indicated by the reviewer and the journal's rules. Please find attached a description of the changes we have made to the manuscript, shown in blue. All the resubmission files have been modified and now include all necessary adjustments. Below are the reviewers' suggestions and the changes we have made accordingly.

We hope we have answered all your questions to your satisfaction, but do not hesitate to contact us for further clarification.

Thank you for your time and help.

Sincerely,

Aline Bruni

 

 

 

Reviewer 1

 

The manuscript discusses the application of chemometric evaluation/ modelling of infrared analysis of experimental data in 68 samples seized by law enforcement. The manuscript is well written and the data presented in a logical and clear manner. There are some aspects that are neglected though and their viewer would like to see these included.

 

[1] The specific application of chemometrics in forensic drug analysis (using NMR, GC-MS and LC-MS data) is very briefly discussed in the introduction and there are a number of key references that could have been included to show the work in this specific area to provide context to the work.

 

We appreciate the reviewer's observation. We have corrected the manuscript according to this recommendation. In the introduction, we detail the problem addressed by the work and include more references in the area, which will help the reader have more foundations if needed.

 

[2] The authors comment that "Identifying NPS is challenging. In the case of cathinones, homologous amphetamines can impair correct identification due to structural similarities" - this may be true for infrared analysis, however, there are many papers which have utilized NMR, GC-MS (some of which have coupled these techniques with chemometrics) to discriminate between structural isomers - this section should be revised taking into account the prior work in the area by teams - as it implies that the authors are trying to suggest that this is something that hasn't been tackled before.

 

We agree with the reviewer that there are papers on the identification of amphetamines and cathinones by techniques such as NMR and GC-MS. As a way of presenting these previous papers, in the introduction we have added some references that we deem relevant. This addition indicates to the reader that other publications have addressed other techniques with the same purpose. However, we emphasize that the approach presented in this work has never been addressed, so we highlight all the methodological steps and apply the models to the case at hand (identification of seized drugs).

 

 [3] Why has there been no confirmation of the content of the samples and comparison with a confirmatory technique (such as GC-MS or LC-MS?). GC-MS and similar techniques are considered the "gold standard" in forensic analysis - however the authors have not provided any data, in the main manuscript or the SI, to confirm the content of the 68 samples.

 

We are grateful for the reviewer's comment. When discussing forensic analysis, we understand that two types of examinations are needed, usually a preliminary and faster one and a second definitive and more robust one. In our work, we have constructed a database from in silico infrared spectroscopy data for identifying compounds. For the proposal to become feasible, we have used two more specific classes of drugs to apply these machine-learning methods. The next step will be to expand these spectral references so that forensic experts can use in silico data in their decision-making. Thus, we have not presented GC-MS or LC-MS data because the reality of the countries or regions that can count on this type of equipment is not known and because these data are not the focus of our work or other literature works, which do not show confirmation by these techniques, either.

 

 [4] Building on point 3 - there have been at least four published studies in the past two years including: Green, Int. J. Drug Policy (2020); Wallace, Drug Testing Analysis (2021); Fregonese, Front. Psychiatry (2021) and Dixon, J. Pharm. Biomed. Analysis (2022) which describe the comparison of infrared analysis of seized forensic samples and each of them both indicate that infrared is the least discriminatory technique compared to other techniques. How does the authors current data compare with these studies? A comparison of the approach with a "gold standard" technique and discussion within the context of other studies is needed to clearly articulate the novelty of the approach described in the manuscript. This is the papers biggest failing and must be addressed before publication is granted.

 

We thank the reviewer for the indications of published studies. We have included them in the revised manuscript. By studying these recommended studies, we understand that, compared with other techniques, point analysis of infrared spectra provides more general information, such as information about functional groups. We have reflected on these data and discussed them in the appropriate section of our work together with other data that we deem relevant.

 

Overall the paper describes a good chemometric approach to the rapid analysis of seized samples - however the lack of a comparison of a gold standard technique and contextualization of the method with previously published studies to benchmark the approach severely diminishes the work in terms of novelty. Recommend MAJOR revision of the manuscript and inclusion of key data (GC-MS) as highlighted in this review.

 

We appreciate the reviewer's recommendations and hope we have satisfactorily responded to all suggestions.

 

Reviewer 2

 

1. The main question posed in the title should be changed as it was not fully answered, but I guess what the authors meant was to show how theoretical methods could help with experimental data regarding synthetic cathinones.

 

We appreciate the reviewer’s observation, and we have changed the title as follows:

“In silico infrared spectroscopy as a benchmark for identifying seized samples suspected of being N-ethylpentylone”.

 

2. The topic is relevant to the field. It helps to understand the drawbacks of some experimental data. The authors presented modeling study of synthetic cannabinoids in the past. Here they show interesting approach to synthetic cathinones. In my opinion it should be published, because it sheds some light on infrared data, which analysis is not trivial as one may expected (straightforward comparison of the spectra).

 

We thank the reviewer for the comment aimed at the relevance of the work.

 

4. The flaws that I can point out are: it is difficult to analyze mixtures using infrared spectroscopy; another technique could be used in several samples, as a control samples (e.g. powder X-ray diffraction), to show their real composition

 

We are grateful for the reviewer's observation. Indeed, challenges must still be overcome when it comes to analyzing mixtures by using infrared spectroscopy. One of these challenges concerns the complexities of spectral regions compared to the responses obtained by X-ray diffraction or NMR. As demonstrated in our work, the challenge of understanding all the details of a spectrum can be overcome by using chemometrics or machine learning methods. The professional must make their decision in the face of the response generated by the methodology.

 

5. The conclusions are consistent with the arguments presented in the manuscript; however, they do not address the main question posed in the title.

 

We are grateful for the reviewer's recommendation, and we have changed the conclusion as follows:

 

We have applied spectroscopy in the infrared region to obtain prompt information in the repressive context and for harm reduction. The proposed computational approach can identify amphetamines and/or cathinones in seized samples. The combination of computational simulation and statistical methods adequately creates a benchmark for new psychoactive substances. PLS-DA can predict the unknown samples in the modeled classes. However, infrared spectroscopy cannot differentiate between N-ethylpentylone and its homologous amphetamine. The reason for this behavior must be assessed so that the real influence of adulterants on the spectra can be realized. A more detailed study of the interferents may also be required. Other techniques, such as GC-MS, must provide more assertive identification in a forensic laboratory. Literature shows similar conclusions [89–93]. Nevertheless, this study can give presumptive information and be a first step toward using computer simulation to compose benchmarks for in-situ comparison. A more robust dataset can aid decision-making in a multidisciplinary and integrated context of NPS forensics.

Creating a comprehensive dataset to follow international surveillance trends can be meaningful. The methodology studied in this work provides essential and adaptable information to identify new substances or harmful contaminants. The main advantage is that these benchmarks dismiss the need for laboratory supplies and regulatory authorizations. Using computational standards associated with chemometrics can be helpful for law enforcement and human rights because these standards reduce the time needed for conducting the experimental tests.

 

6. The references are appropriate. However, references 32 and 37 are the same.

 

We agree with the reviewer and have made the necessary adaptations to the references.

 

7. The tables and figures are fine, but I recommend the text to be send for professional language check.

 

We appreciate the reviewer's recommendation. The version sent to the journal has been checked by a language professional.

Author Response File: Author Response.pdf

Reviewer 2 Report

1. The main question posed in the title should be changed as it was not fully answered, but I guess what the authors meant was to show how theoretical methods could help with experimental data regarding synthetic cathinones. 2. The topic is relevant to the field. It helps to understand the drawbacks of some experimental data. 3. The authors presented modeling study of synthetic cannabinoids in the past. Here they show interesting approach to synthetic cathinones. In my opinion it should be published, because it sheds some light on infrared data, which analysis is not trivial as one may expected (straightforward comparison of the spectra). 4. The flaws that I can point out are: it is difficult to analyze mixtures using infrared spectroscopy; another technique could be used in several samples, as a control samples (e.g. powder X-ray diffraction), to show their real composition 5. The conclusions are consistent with the arguments presented in the manuscript, however they do not address the main question posed in the title. 6. The references are appropriate. However, references 32 and 37 are the same. 7. The tables and figures are fine, but I recommend the text to be send for professional language check.

Author Response

Response Letter

 

Manuscript ID: psychoactives-1982638

Type of manuscript: Article

Title: In silico infrared spectroscopy as a benchmark for identifying seized samples suspected of being N-ethylpentylone.

Prof. Dr. Ricardo Dinis-Oliveira, Editor-in-Chief, Psychoactives

  

Dear Prof. Dr. Dinis-Oliveira,

 

We have adjusted the manuscript on the basis of the recommendations and notes made by the reviewers. We have re-evaluated the introduction so that the section would provide more evidence of the existing problem and the solution proposed by us. We have introduced a more specific discussion of the relationship between this work and other published research. We agree with the reviewers about changing the title and the conclusion. Finally, we have adjusted the references as indicated by the reviewer and the journal's rules. Please find attached a description of the changes we have made to the manuscript, shown in blue. All the resubmission files have been modified and now include all necessary adjustments. Below are the reviewers' suggestions and the changes we have made accordingly.

We hope we have answered all your questions to your satisfaction, but do not hesitate to contact us for further clarification.

Thank you for your time and help.

Sincerely,

Aline Bruni

 

 

 

Reviewer 1

 

The manuscript discusses the application of chemometric evaluation/ modelling of infrared analysis of experimental data in 68 samples seized by law enforcement. The manuscript is well written and the data presented in a logical and clear manner. There are some aspects that are neglected though and their viewer would like to see these included.

 

[1] The specific application of chemometrics in forensic drug analysis (using NMR, GC-MS and LC-MS data) is very briefly discussed in the introduction and there are a number of key references that could have been included to show the work in this specific area to provide context to the work.

 

We appreciate the reviewer's observation. We have corrected the manuscript according to this recommendation. In the introduction, we detail the problem addressed by the work and include more references in the area, which will help the reader have more foundations if needed.

 

[2] The authors comment that "Identifying NPS is challenging. In the case of cathinones, homologous amphetamines can impair correct identification due to structural similarities" - this may be true for infrared analysis, however, there are many papers which have utilized NMR, GC-MS (some of which have coupled these techniques with chemometrics) to discriminate between structural isomers - this section should be revised taking into account the prior work in the area by teams - as it implies that the authors are trying to suggest that this is something that hasn't been tackled before.

 

We agree with the reviewer that there are papers on the identification of amphetamines and cathinones by techniques such as NMR and GC-MS. As a way of presenting these previous papers, in the introduction we have added some references that we deem relevant. This addition indicates to the reader that other publications have addressed other techniques with the same purpose. However, we emphasize that the approach presented in this work has never been addressed, so we highlight all the methodological steps and apply the models to the case at hand (identification of seized drugs).

 

 [3] Why has there been no confirmation of the content of the samples and comparison with a confirmatory technique (such as GC-MS or LC-MS?). GC-MS and similar techniques are considered the "gold standard" in forensic analysis - however the authors have not provided any data, in the main manuscript or the SI, to confirm the content of the 68 samples.

 

We are grateful for the reviewer's comment. When discussing forensic analysis, we understand that two types of examinations are needed, usually a preliminary and faster one and a second definitive and more robust one. In our work, we have constructed a database from in silico infrared spectroscopy data for identifying compounds. For the proposal to become feasible, we have used two more specific classes of drugs to apply these machine-learning methods. The next step will be to expand these spectral references so that forensic experts can use in silico data in their decision-making. Thus, we have not presented GC-MS or LC-MS data because the reality of the countries or regions that can count on this type of equipment is not known and because these data are not the focus of our work or other literature works, which do not show confirmation by these techniques, either.

 

 [4] Building on point 3 - there have been at least four published studies in the past two years including: Green, Int. J. Drug Policy (2020); Wallace, Drug Testing Analysis (2021); Fregonese, Front. Psychiatry (2021) and Dixon, J. Pharm. Biomed. Analysis (2022) which describe the comparison of infrared analysis of seized forensic samples and each of them both indicate that infrared is the least discriminatory technique compared to other techniques. How does the authors current data compare with these studies? A comparison of the approach with a "gold standard" technique and discussion within the context of other studies is needed to clearly articulate the novelty of the approach described in the manuscript. This is the papers biggest failing and must be addressed before publication is granted.

 

We thank the reviewer for the indications of published studies. We have included them in the revised manuscript. By studying these recommended studies, we understand that, compared with other techniques, point analysis of infrared spectra provides more general information, such as information about functional groups. We have reflected on these data and discussed them in the appropriate section of our work together with other data that we deem relevant.

 

Overall the paper describes a good chemometric approach to the rapid analysis of seized samples - however the lack of a comparison of a gold standard technique and contextualization of the method with previously published studies to benchmark the approach severely diminishes the work in terms of novelty. Recommend MAJOR revision of the manuscript and inclusion of key data (GC-MS) as highlighted in this review.

 

We appreciate the reviewer's recommendations and hope we have satisfactorily responded to all suggestions.

 

Reviewer 2

 

1. The main question posed in the title should be changed as it was not fully answered, but I guess what the authors meant was to show how theoretical methods could help with experimental data regarding synthetic cathinones.

 

We appreciate the reviewer’s observation, and we have changed the title as follows:

“In silico infrared spectroscopy as a benchmark for identifying seized samples suspected of being N-ethylpentylone”.

 

2. The topic is relevant to the field. It helps to understand the drawbacks of some experimental data. The authors presented modeling study of synthetic cannabinoids in the past. Here they show interesting approach to synthetic cathinones. In my opinion it should be published, because it sheds some light on infrared data, which analysis is not trivial as one may expected (straightforward comparison of the spectra).

 

We thank the reviewer for the comment aimed at the relevance of the work.

 

4. The flaws that I can point out are: it is difficult to analyze mixtures using infrared spectroscopy; another technique could be used in several samples, as a control samples (e.g. powder X-ray diffraction), to show their real composition

 

We are grateful for the reviewer's observation. Indeed, challenges must still be overcome when it comes to analyzing mixtures by using infrared spectroscopy. One of these challenges concerns the complexities of spectral regions compared to the responses obtained by X-ray diffraction or NMR. As demonstrated in our work, the challenge of understanding all the details of a spectrum can be overcome by using chemometrics or machine learning methods. The professional must make their decision in the face of the response generated by the methodology.

 

5. The conclusions are consistent with the arguments presented in the manuscript; however, they do not address the main question posed in the title.

 

We are grateful for the reviewer's recommendation, and we have changed the conclusion as follows:

 

We have applied spectroscopy in the infrared region to obtain prompt information in the repressive context and for harm reduction. The proposed computational approach can identify amphetamines and/or cathinones in seized samples. The combination of computational simulation and statistical methods adequately creates a benchmark for new psychoactive substances. PLS-DA can predict the unknown samples in the modeled classes. However, infrared spectroscopy cannot differentiate between N-ethylpentylone and its homologous amphetamine. The reason for this behavior must be assessed so that the real influence of adulterants on the spectra can be realized. A more detailed study of the interferents may also be required. Other techniques, such as GC-MS, must provide more assertive identification in a forensic laboratory. Literature shows similar conclusions [89–93]. Nevertheless, this study can give presumptive information and be a first step toward using computer simulation to compose benchmarks for in-situ comparison. A more robust dataset can aid decision-making in a multidisciplinary and integrated context of NPS forensics.

Creating a comprehensive dataset to follow international surveillance trends can be meaningful. The methodology studied in this work provides essential and adaptable information to identify new substances or harmful contaminants. The main advantage is that these benchmarks dismiss the need for laboratory supplies and regulatory authorizations. Using computational standards associated with chemometrics can be helpful for law enforcement and human rights because these standards reduce the time needed for conducting the experimental tests.

 

6. The references are appropriate. However, references 32 and 37 are the same.

 

We agree with the reviewer and have made the necessary adaptations to the references.

 

7. The tables and figures are fine, but I recommend the text to be send for professional language check.

 

We appreciate the reviewer's recommendation. The version sent to the journal has been checked by a language professional.

Author Response File: Author Response.pdf