Study on Molecular Profiles of Staphylococcus aureus Strains: Spectrometric Approach

Round 1

Reviewer 1 Report

In this paper ,the authors’ work is very interesting and important,it is a promising method to identify S .aureus  strains via the MALDI TOF MS analysis and different computational methods. Before publication.the authors should answer the following questions and compliment some data:

1. in this manuscript, the authors think the different cultures can affect the identification. But authors didn’t explain what cause the phenomena clearly.the authors should give a reasonable explanation.
2. To identify bacteria strains via the MALDI TOF MS analysis,I think authors should establish a big database related to the Aureus if authors want MALDI TOF MS play a more import role in identification of bacteria.
3. The authors should compare MALDI TOF MS with traditional methods.

Author Response

Response to Reviewers’ Comments

 

1. Ref. No.:  molecules-956317

Study on molecular profiles of S. aureus strains – spectrometric approach

Michał Złoch, Paweł Pomastowski, Ewelina Maślak, Fernanda Monedeiro and Bogusław Buszewski

Firstly, the authors would like to thank the Editor and Reviewers for appreciation of our effort and secondly, for useful comments, remarks, and valuable suggestions that led to the increasing in the quality of the present work. Therefore, to better improve the article the authors have addressed all the comments as explained below.

In the manuscript file, all of the changes have been done using the "Track Changes" function in Microsoft Word.

Reviewers' comments:

Reviewer #1

In this paper ,the authors’ work is very interesting and important,it is a promising method to identify S .aureus  strains via the MALDI TOF MS analysis and different computational methods. Before publication.the authors should answer the following questions and compliment some data:

1. In this manuscript, the authors think the different cultures can affect the identification. But authors didn’t explain what cause the phenomena clearly. the authors should give a reasonable explanation.

Re: Thanks to the Reviewer for pointed out this important issue. In common opinion, the identification of the microorganisms via MALDI technique is considered as culture-independent since most of the signals in generated proteins profiles derived from highly abundant ribosomal proteins (metabolic independent). Therefore, utilization of the different culture media (enriched or poorer ones) it is considered in most cases in the context of obtaining a sufficiently rapid growth of biomass sufficient to collect the appropriate number of cells for analysis which is 5 × 10⁵ cells on a target spot. This phenomena was also noted in our study, where no significant impact of the culture medium type on the identification score was noted apart from mannitol salt agar medium (MAN). To clarify this finding, suitable information has been added to the manuscript – “However, this issue relates more to the effect of the type of culture medium on the growth rate of the bacteria than on the differences in the composition of the protein profile itself.” (P17.L.408-410). However, as we wrote in the Introduction “It is known that culture conditions such as the composition of the growth medium, incubation time, pH or temperature may lead to changes in the mass spectra profiles reflected in the variation of signals intensities, loss of signals, or the shift of signals probably caused by proteins expression differences” (P3.L.81-85) since in the proteomic profiles of the bacteria generated via MALDI are also peaks derived from ionization of non-ribosomal proteins, which are metabolic status dependent. This phenomenon is supposed to play an important role during the differentiation of close-related species or subspecies, where variation in the MS profiles compositions is subtle. Until now, this issue is poorly investigated. Therefore, the main goal of our study was to find which kind of culture media (universal, enriched, differential, selective) support discrimination of the S. aureus strains (and additionally which statistical method should be used to make the analysis more feasible) - such information will be useful for clinical laboratories. For this purpose, we used analysis in the linear mode as far most of the hospitals are equipped with linear devices. This kind of analysis is not sufficient for giving the full explanation of the mechanisms underlying the impact of the culture media type on the generated proteomic profiles. For this, analysis in the reflectron mode (MS/MS) should be applied what we plan to do for selected conditions and models in the next step of our study. To clarify this issue, the suitable information has been added to the manuscript (P.21-22.L.565-571)-“ “Since the presented studies used analysis in the linear mode (mostly applied in the clinical laboratories), it could be also promising to investigate the selected conditions using the reflectron mode (TOF/TOF MS) which could both help explain mechanisms underlying the revealed phenomena and give the opportunity to select specific peaks for improving the S. aureus differentiation. However, this represents a different approach with other criteria to be considered, like for the example lower intensity of the characteristic peaks (biomarkers), and will be the subject of further research.”

 

2. To identify bacteria strains via the MALDI TOF MS analysis,I think authors should establish a big database related to the Aureus if authors want MALDI TOF MS play a more import role in identification of bacteria.

 

Re: Thanks to the Reviewer for the remark. At this step, only a characterization of data acquired by a certain set of parameters was aimed. The referred variables comprised testing 7 bacterial strains, 8 culture media and 3 different MALDI matrices, resulting in the processing of 672 spectra. With this approach, optimal conditions for the cultivation and analysis of strain belonging to the same genus could be indicated. A further step of this research could in fact refer to the comprehensive analysis of a greater dataset, generated from the set of conditions pointed out by the present study. The use of a larger database can indeed be profitable, once machine learning models could be trained in a more proper manner and yield more reliable results, however, the quality of the generated data used for the learning must be assured, therefore the authors choose to be more attained to the study of influence of the main variables at this stage

 

3. The authors should compare MALDI TOF MS with traditional methods

 

Re: Thanks to the Reviewer for taking up this important issue. Although MALDI TOF MS analysis is considered as a relatively new microbial identification method compared to the e.g. sequencing of 16S rDNA PCR products or phenotypic biochemical tests, nevertheless since its introduction into clinical microbiology (over 10 years ago), has quickly become the standard routine identification tool for bacteria in most laboratories around the world. Besides improving the diagnosis of bacterial infections by accelerating analyses and improving the quality of results, MALDI TOF MS also dramatically changed the possibilities of differentiation of closely related bacteria (including both aerobic and anaerobic species, nonfermenters, HACEK group, Corynebacteria, Nocardia and Mycobacteria) which in many cases were difficult or even impossible to distinguish by traditional biochemical methods use. Currently, in the market there are also available the validated MALDI identification systems with IVD/CE certificates - VITEK^® MS IVD and MALDI Biotyper IVD-CE. Since the comparison of various method used for closed-related species differentiation was not the subject of our study, there was no need to use traditional method to prove result of S. aureus identification. In light of the mentioned information provided based on the latest literature reports, the result of the MALDI identification is considered as reliable. Nevertheless, the issue raised by the Reviewer should be a subject of further studies where selected the best analysis conditions and statistical models for S. aureus differentiation via MALDI approach should be compared with other methods, e.g. molecular subtyping methods.[ Kostrzewa et al, (2019): How MALDI-TOF mass spectrometry can aid the diagnosis of hard-to-identify pathogenic bacteria –the rare and the unknown, Expert Review of Molecular Diagnostics, DOI:10.1080/14737159.2019.1643238; Harris et al. 2010. Rapid identification of staphylococci from prosthetic joint infections using MALDI-TOF mass-spectrometry. Int J Artif Organs 9:568-574. doi: 10.1177/039139881003300902.; Sauget et al. 2017. Can MALDI-TOF Mass Spectrometry Reasonably Type Bacteria? Trends in Microbiology, 25(6): 447-455. https://doi.org/10.1016/j.tim.2016.12.006]

 

Reviewer 2 Report

the manuscript "Study on molecular profiles of S. aureus strains spectrometric approach" presents results from an interesting and required study for making MALDI a routine identification/classification tool in clinical labs. There are no major issues with the manuscript and most of my remarks are added in the attached document.

Four points I would like to emphasize.

Firstly the tables need to be made more readable, for some it is indicated how this can be done for others the authors should find better ways to present the data.

Secondly there is not one spectrum shown, this is a journal for the general (scientific) community, only a fraction of this public is able to form an idea of what it looks like and of the problems using this data. So a simple example showing two closely related strains would be very illustrative and make increase the impact of the work (for a non-specialist).

Thirdly, the authors emphasize the need for standardization and the use of elaborate informatic tools. Good, but the most simple thing to increase the specificity and accuracy of the approach is not even mentioned. The easiest way to to generate more/different data. This can be done in two easy ways not requiring any new sample and in seconds. Use different matrices and use both for classification, this asks nothing, just one microliter of matrix solution. Secondly the authors have a TOF/TOF, the statitical effect of not only having a peak (mass + intensity) but having fragments from this peak (mass2 + intensity, mass3 + intensity, ....) must be enormous. More difficult to apply but in my opinion more promising as well.

Lastly, when talking about intensity the authors fail to explain what it is. realtive compared to base peak, absolute, relative compared to internal standard, .... This must be clarified.

 

 

Comments for author File: Comments.pdf

Author Response

Response to Reviewers’ Comments

 

1. Ref. No.:  molecules-956317

Study on molecular profiles of S. aureus strains – spectrometric approach

Michał Złoch, Paweł Pomastowski, Ewelina Maślak, Fernanda Monedeiro and Bogusław Buszewski

Firstly, the authors would like to thank the Editor and Reviewers for appreciation of our effort and secondly, for useful comments, remarks, and valuable suggestions that led to the increasing in the quality of the present work. Therefore, to better improve the article the authors have addressed all the comments as explained below.

In the manuscript file, all of the changes have been done using the "Track Changes" function in Microsoft Word.

Reviewers' comments:

Reviewer #2

 

the manuscript "Study on molecular profiles of S. aureus strains spectrometric approach" presents results from an interesting and required study for making MALDI a routine identification/classification tool in clinical labs. There are no major issues with the manuscript and most of my remarks are added in the attached document.

 

Four points I would like to emphasize.

 

1. Firstly the tables need to be made more readable, for some it is indicated how this can be done for others the authors should find better ways to present the data.

 

Re: We are in full agreement with the Reviewer's opinion. The tables has been corrected (Table 3, Table 4, Table 5, Table 6) to be more readable. Moreover, all the tables has been supplemented with the description of the used abbreviations including matrices, culture media, methods and so one. 

 

2. Secondly there is not one spectrum shown, this is a journal for the general (scientific) community, only a fraction of this public is able to form an idea of what it looks like and of the problems using this data. So a simple example showing two closely related strains would be very illustrative and make increase the impact of the work (for a non-specialist).

 

Re: We agree with the Reviewer's opinion that an additional figure with the example of MS spectra obtained for investigated strains will help emphasize the significance of the presented work and will support understanding of presented findings, especially for a non-specialist. Therefore, an additional figure (Figure 1; P.6) which presents the exemplary MS spectra has been added to the manuscript.

 

3. Thirdly, the authors emphasize the need for standardization and the use of elaborate informatic tools. Good, but the most simple thing to increase the specificity and accuracy of the approach is not even mentioned. The easiest way to to generate more/different data. This can be done in two easy ways not requiring any new sample and in seconds. Use different matrices and use both for classification, this asks nothing, just one microliter of matrix solution. Secondly the authors have a TOF/TOF, the statitical effect of not only having a peak (mass + intensity) but having fragments from this peak (mass2 + intensity, mass3 + intensity, ....) must be enormous. More difficult to apply but in my opinion more promising as well. P.22.L.565: the authors have the availability of a TOF/TOF. What always strikes me is that the MS/MS capacities of MALDI is never used in this type of study. IF used the presence/absence of one peak could be sufficient for increasing the discrimination of related species. Some reflection on this? And why not using two matrices to increase the discrimination, I mean the extract is ready it is sufficient to put two spots and overlay it with a different matrix. On top of doing fancy informatics things just getting more data would be the easiest and most straightforward way to improve this approach. This is not a comment on the paper but something the authors must reflect on in the discussion.

Re: We would like to thank the Reviewer for these remarks. Indeed, by this hypothesis, in tandem analysis could be promising for strain differentiation; However, some considerations may be outlined regarding the choice of the used approach. The higher complexity of the data may does not imply in a data with superior quality for the processing by classificatory algorithms (software native or self-implemented). Some important requirements are reproducibility of the obtained spectra and appropriate intensity of the signal. A very characteristic fragment likely will present a lower intensity, therefore the approach based on a set of “markers” stands as a reliable strategy. Size of input data is also important for the quality of its processing, a data with larger volume is more prone to not keep its integrity during computational treatment. Apart from this, although the used equipment allows this type of analysis, the development of a standard protocol may be useful for the application of the proposed methodology in different laboratories which does not necessarily afford this model of instrument and/or its native algorithms - this applies in particular to clinical laboratories, most of which are equipped with linear devices. Nevertheless, application of different matrices use both for classification and utilization the LIFT technologies for m/z at  4000 – 3500 (real mass rage for our Ultraflex Xtreme II, fragmentation of higher masses is – even by CID –  not efficient) is promising for evaluation of new statistical approach for molecular profiling of bacteria strain. In this study we would like to study impact of different culture media on S. aureus strain using different matrices to study their identification impact by standardized Biotyper platform, used in e.g. clinical laboratories. In further studies, for determination e.g. discrimination power of used statistical method of self -developed algorithms in the R environment, we will test the Reviewer’s suggestion.

The reflection about this approach has been added to the manuscript (P.21-22.L.565-571) – “Since the presented studies used analysis in the linear mode (mostly applied in the clinical laboratories), it could be also promising to investigate the selected conditions using the reflectron mode (TOF/TOF MS) which could both help explain mechanisms underlying the revealed phenomena and give the opportunity to select specific peaks for improving the S. aureus differentiation. However, this represents a different approach with other criteria to be considered, like for the example lower intensity of the characteristic peaks (biomarkers), and will be the subject of further research.”

4. Lastly, when talking about intensity the authors fail to explain what it is. realtive compared to base peak, absolute, relative compared to internal standard, .... This must be clarified.

Re: Thanks to the Reviewer for the important suggestion. The used values were the relative intensity of the peak in relation to the spectrum of the calibration standard (BTS, Bruker Bacterial Test Standard). This information was added to the manuscript (P.24.L.642) in order to clarify this issue.

 

5. Others:

- P.1.L.2: write the species full in the title

Re: According to Reviewer’s suggestion the full species in the title has been written

-P.1.L.27: working on this since 15 years, what is needed is standardization and the things mentioned here

Re: The sentence was corrected based on the Reviewer’s suggestion and term “standardization” was added as also crucial part for future studies on this technique

- P.2. L.48: remove “different”

Re: Correction has been applied

- P.3.L. 77-81: not very clear, no mention of bioinformatics above. Make two sentences

Re: According to the Reviewer’s suggestion, the sentence has been rewritten and separated into two sentences. Since bioinformatics methods were not mentioned – this term was removed from the final sentence.

-P.3.L. 89-90: influencing the presence/absence of peaks

Re: The sentence was corrected according to Reviewer’s suggestion – “influencing the presence/absence of peaks” replaced “enhancing the differential ionization of specific compounds”

-P.3.L.98-103: not clear rewrite

Re: According to the Reviewer’s suggestion, the sentence was rewritten as follow: “During the phyloproteomic analysis proteins expression levels (signal intensities) are as important as the presence or absence of specific signals, thus, the interpretation of big data sets generated during MALDI analysis is complicated. Therefore, the usage of appropriate bioinformatics tools for the clustering of the mass spectral data of proteins is recommended as a fast and efficient tool for the comparison of spectra”

- P.5.L. 144-157: explain what this is and how this is generated, a key issue. Furthermore since the Material is at the end a legend giving all abbreviations is needed.

Re: According to the Reviewer’s remarks the caption of Table 1 was supplemented with suitable information on how the results were generated, the full name of methods as well as a legend defining abbreviations used. In the same way the legends were applied for all others tables.

-P.7.L.184-185: I would say MAN performs worse independent of the matrix used.

Re: We agree with the Reviewer’s opinion and the sentence was changed to emphasized this finding – the sentence in the reviewed manuscript is: “Regarding the MSP data, the results for MAN medium performs worse independent of the matrix used.”

-P.7-8.L.190-193: please explain why here only on MSP? From table 1 we see that MSP is slightly less than RAW. This makes it all the more important to detail the way these MSP are generated

Re: Thanks to the Reviewer for the important remarks. For identification of microorganisms via MALDI Biotyper platform both raw (whole) spectra as well as generated main spectra (MSPs) can be used. Nevertheless, for creating phyloproteomic trees only MSPs can be used via this software – this is an explanation why we created dendrograms only for MSPs. For comparison raw spectra another tool can be used – Composite Correlation Index (CCI) Matrix analysis – which is also included in the software. Thus, CCI analysis we used for raw spectra analysis (see Table 4). In general, an MSP is considered by Bruker as the basis of classification using the MALDI Biotyper. Therefore, in our work we focused on the comparison MSP analysis recommended by Bruker with analysis of the whole spectra using own protocols knowing that number of peaks in MSPs are reduced to the most relevant ones according to software (for this we used recommended preprocessing method called “Biotyper Preprocessing Standard Method”).

The part mentioned visual inspection was preceded by an appropriate description explaining the basis of the method -  “Based on the obtained MSPs (Main Spectra), created by the transformation of raw spectra into peak lists by extracting information on peak mass, peak frequency, and peak intensity distribution using Biotyper Preprocessing Standard Method, phyloproteomic trees were generated”.

- P.8.L.199: again since Material is at the back giving the total number of tested strains is essential

Re: We agree with the Reviewer’s remark. The number of investigated strains was added.

- P.9L.223: and BLA

Re: We thank the Reviewer for his insightful attention. This information was added to the mentioned sentence.

- P.9L.227: (composite correlation index method)

Re: According to the Reviewer’s suggestion the abbreviated name of the method was supplemented with its full name.

 

-P.12.L.291-295: the data is so clear that PCA is not the most suited method for doing this, is it needed to discuss the details of these results? One sentence would be sufficient, there is not one point at which PCA does equally good or better than MSP or HCA. Thsi would make figure 2 (which is not learning us anything) also obsolete.

Re: Thanks to the Reviewer for this valuable remark. Indeed, the findings are evident, so there is no need go into details. Therefore, second sentence “The best results, in terms of the percentage of the correct classification, were achieved using the MAN and BLA + HCCA (71%, in both cases – Fig. 3A) or VRE and MAN + DHB (57 and 43%, respectively, Fig. 3B) or VRE and BLA + SDHB (43 and 29%, respectively – Fig. 3C” ) was replaced by the sentence “Considering all matrices, 3 culture media were characterized by the best strains discrimination – BLA, MAN, and VRE” since such information, in our opinion, is the most important findings – these media appeared to be the best among all investigated for S. aureus differentiation. For supporting this finding, Figure 3 (formerly Fig. 2), in our opinion should be left in the manuscript.

-P.14.L.332-340: make the table better readable, avoid the splitting of content on two lines (for instance the % sign), use only 3 significant digits for the error, and if not better use cell delimitation. as it is now this is not clear at all.

Re: Thanks to the Reviewer for valuable advice. The table has been rearranged according to suggestion given. 

-P.14.L.340: no coloring is observed

Re: Thanks to the Reviewer’s insightful attention. Information about colors was added accidentally – the mistake was corrected and such information was removed from the table caption.

- P.14.L.342: for the other analytical methods

Re: the sentence was corrected according to the Reviewer’s suggestion.

- P.17.L.412-417: make two sentences

Re: Thanks to the Reviewer for the valuable remark. The sentence was separated into two sentences.

-P.17.L.423: this seems rather strange, basically the authors say that more than 50% of all proteins found in a cell are ribosomal. I propose to cut the part "and can comprise up to 50% of all cellular proteins"

Re: Thanks to the Reviewer for the valuable remark. The part "and can comprise up to 50% of all cellular proteins" has been removed.

-P.17.L.428-430: ?????

Re: We agree with the Reviewer that second part of the sentence is confusing, therefore, we decided to remove this part from the final sentence which in the revised version is “It emphasizes the need for the standardization of the sample preparation like the choice of matrix or culture medium, especially in the view of subspecies differentiation.” 

- P.18.L.460-465: these are proteins of more than 250 amino acids, did the authors observe any peaks around 27.5kDa? In my experience this is not the case so at best fragments of these can be observed. In general when combining MS with MS/MS on a fourrier transform most peaks in this approach are ribosomal proteins with fragments of highly abundant metabolic proteins (GAPDH for instance) making up for most of the non-ribosomal

Re: Thanks to the Reviewer for the important remark. Indeed, used conditions prevented the detection of such large proteins as 27.5 kDa – we investigated proteins up to 20 kDa. Therefore, according to Reviewer’s suggestion, we removed this sentence from the manuscript.

- P.21.L.543: write full name

Re: The full name has been provided – Enterococcus faecium

-P.23.L.623-626: details need to be provided, this is again a "manipulation" that will have a huge impact on the result.

Re: Thanks for the important remarks. For the smoothing the Savistsky-Golay method was applied (width 2 m/z, 10 cycles) while for Baseline substraction the TopHat algorithm (signal to noise threshold 2; peak detection algorithm – centroid) recommended by the software provider. Details has been added to the manuscript. 

-P.23.L.634-637: describe what "main spectra" is

Re: Thanks to the Reviewer for the suggestion. The description of “main spectra” has been added to the paragraph - Main Spectra (MSP) - peaks lists created by the transformation of raw spectra by extracting information on peak mass, peak frequency, and peak intensity distribution using Biotyper Preprocessing Standard Method recommended and provided by manufacturer (Bruker Daltonik GmbH, Bremen, Germany).

-P.25L.677-679: rewrite, this is an important sentence and it is not clear 

Re: Thanks to the Reviewer for the important suggestion. The sentence has been rewritten “Currently, due to advancements in dedicated software, computational algorithms, as well as the greater availability of the devices (lower prices), the application of the MALDI approach for microbial identification at the subspecies level is under the insightful debate.”

-P.30-31.L.872-875: correct this reference

Re: The reference has been corrected according to Reviewer’s suggestion.

Reviewer 3 Report

This manuscript is an interesting and well written study. The authors have evaluated the effects of culture conditions and matrix type on the differentiation of molecular profiles of Staphylococcus aureus strains via MALDI TOF MS analysis and different computational analysis.

The abstract and the aim are clear and are supported by the conclusions. In the introduction, previous studies on the subject are well presented and correctly supported by references.

The results are well described, and the figures and tables correctly provide the information. The materials and methods are explained clearly and in detail.

Two small comments:

-the introduction should be enriched with the main characteristics of Staphylococcus aureus

-In addition, the captions of the figures would be better placed under the figures themselves, in order to facilitate reading.

Author Response

Response to Reviewers’ Comments

 

1. Ref. No.:  molecules-956317

Study on molecular profiles of S. aureus strains – spectrometric approach

Michał Złoch, Paweł Pomastowski, Ewelina Maślak, Fernanda Monedeiro and Bogusław Buszewski

Firstly, the authors would like to thank the Editor and Reviewers for appreciation of our effort and secondly, for useful comments, remarks, and valuable suggestions that led to the increasing in the quality of the present work. Therefore, to better improve the article the authors have addressed all the comments as explained below.

In the manuscript file, all of the changes have been done using the "Track Changes" function in Microsoft Word.

Reviewers' comments:

Reviewer #3

 

 This manuscript is an interesting and well written study. The authors have evaluated the effects of culture conditions and matrix type on the differentiation of molecular profiles of Staphylococcus aureus strains via MALDI TOF MS analysis and different computational analysis.

 

The abstract and the aim are clear and are supported by the conclusions. In the introduction, previous studies on the subject are well presented and correctly supported by references.

 

The results are well described, and the figures and tables correctly provide the information. The materials and methods are explained clearly and in detail.

 

Two small comments:

-P.2.L.33-36   -the introduction should be enriched with the main characteristics of Staphylococcus aureus

Re: According to the Reviewer’s suggestion the Introduction part has been supplemented with the main characteristic of S. aureus.

2. -In addition, the captions of the figures would be better placed under the figures themselves, in order to facilitate reading.

Re: Thanks to Reviewer for the important remark. The correction of the Figures and Tables captions has been changed.