Attenuated Total Reflectance-Fourier Transform Infrared Spectroscopy (ATR-FTIR) Coupled with Chemometrics, to Control the Botanical Authenticity and Quality of Cold-Pressed Functional Oils Commercialized in Romania

Round 1

Reviewer 1 Report

The authors measured the Raman spectra of the commercial oils in Romania and analyzed the spectra by chemometrics to control the oils' authenticity and quality. The reviewer felt that this study provided an important contribution to food inspection and this method can be applied for the classification of the basic ingredient and quality of oils.

Although the reviewer that their results and analysis are acceptable, the authors should consider the following points.

Comments

1. Although the UV and IR spectra are shown in the supplementary files, they should be  shown in the manuscript because the readers want to compare the spectra and the analytical results.

2. The FT-IR spectra in Figure 2S are too small to find out in detail. The author should make these spectra larger. Although the spectra show between 700 and 1800 cm–1, the spectra should be shown from 700 to 3600 cm–1. Because the authors measured the IR spectra in the region 700–3600 cm–1 and the PCA scores were obtained using the whole spectra ( 700 to 3600 cm–1) in Figure 1b.

3. p11 line 5: Authors suggested that the vibrational levels 2852 and 2922 cm–1 are good predictions for discrimination from Figure 5. According to Table 2, these peaks' assignments are stretching of -C-H (CH3, CH2). Why are these vibrational levels related to the discrimination of each oil?

4. Figure 6: What are the numbers (3 to –3) shown on the right of the Figure? How did the authors calculate these values? The authors mentioned that if the correlation matrix of the sample oil differs from that of the authentic one, the sample oil is mislabeled or adulterated. However, since the oils are produced from the plant may include some other molecules in the basic ingredient, it may affect the correlation matrix.  The reviewer thinks that there are other reasons when the correlation matrix shows different from the authentic ones.

Careless misses

P11 line 4: wrong II = IV, correct III + IV.

P12 line 4: wrong SF02, correct SFO2 

P12 line 5: wrong SFO1, correct SFO4 (according to Table 1, SOF4 is an authentic sample.)

P12 line 5: wrong PO4, correct PO1 (according to Figure 6, the correlation matrix of PO1 much differs from PO4 that is an authentic sample.)

P12 line 5: wrong different)2, correct ) and 2 should be removed.

Author Response

Dear reviewer,

First of all, many thanks for your detailed comments and suggestions.

My answers are below and the corrected form of the manuscript also, as an attachment. I do hope you can accept our comments and modifications.

all the best.

1. Although the UV and IR spectra are shown in the supplementary files, they should be  shown in the manuscript because the readers want to compare the spectra and the analytical results.

We moved Figure S1 (comparative UV-Vis spectra) to the main text, as Figure 1.

We introduced a Figure 2 representing the whole ATR-FTIR spectra (3600-650 cm-1) and  the fingerprint region 1800-650 cm-1) for seabuckthorn oil as an example.

2. The FT-IR spectra in Figure 2S are too small to find out in detail. The author should make these spectra larger. Although the spectra show between 700 and 1800 cm–1, the spectra should be shown from 700 to 3600 cm–1. Because the authors measured the IR spectra in the region 700–3600 cm–1 and the PCA scores were obtained using the whole spectra ( 700 to 3600 cm–1) in Figure 1b.

We kept the figure S2 as supplementary file since they represent the comparative fingerprint regions. We avoided to introduce in the whole text the charged spectra for the whole region, especially considering the resolution was not high.

Figure 1b  was  eliminated. We kept the Figure 1a which is now  Figure 3, since the PCS scores for the fingerprint region are more relevant for the aunthenticity. Any way the PCs were very similar for these 2 regions.

3. p11 line 5: Authors suggested that the vibrational levels 2852 and 2922 cm–1 are good predictions for discrimination from Figure 5. According to Table 2, these peaks' assignments are stretching of -C-H (CH3, CH2). Why are these vibrational levels related to the discrimination of each oil?

According to literature (ref. 22) for some oils, the IR regions of 2956–2853, 1740 cm⁻¹ are attributed to the fatty acid chain length.

In our measurements, the statistical data obtained by metaboanalyst showed that especially for highly unsaturated oils ( linseed, sea buckthorn and olive) can be discriminated by these wavenumbers. That’s why we made this comments.

4. Figure 6: What are the numbers (3 to –3) shown on the right of the Figure? How did the authors calculate these values?

These values are considered by the Metaboanalyst software. The scale 3 to -3 represents the intensity of prediction, from higher positive (3) to lower negative (-3). It is not a formula calculation by authors, but a Random forest algorithm. Details in Metaboanalyst 4.0. tutorial.

The authors mentioned that if the correlation matrix of the sample oil differs from that of the authentic one, the sample oil is mislabeled or adulterated. However, since the oils are produced from the plant may include some other molecules in the basic ingredient, it may affect the correlation matrix. 

Excepting the oils produced by in the plant, considered as authentic, the others we analysed are commercial, so possibly mislabelled or adulterated. That’s why we assumed that this correlation matrix could identify in the commercialized samples signs of mislabeling or adulteration. Of course the assumption of the reviewer that there are other reasons when the correlation matrix shows different from the authentic ones is accepted.

Careless misses

P11 line 4: wrong II = IV, correct III + IV. It is corrected, thank you

P12 line 4: wrong SF02, correct SFO2 It is corrected, thank you

P12 line 5: wrong SFO1, correct SFO4 (according to Table 1, SOF4 is an authentic sample.) It is corrected, thank you

P12 line 5: wrong PO4, correct PO1 (according to Figure 6, the correlation matrix of PO1 much differs from PO4 that is an authentic sample.) It is corrected, thank you

P12 line 5: wrong different)2, correct ) and 2 should be removed. It is corrected, thank you

Author Response File: Author Response.docx

Reviewer 2 Report

Manuscript ID: applsci-1026445 reports on the combined use of Attenuated Total Reflectance-Fourier Transform Infrared Spectroscopy (ATR-FTIR) and chemometrics as a reliable, rapid and easy-to-use methodology to evaluate vegetable oils quality and authenticity. The article is well-written and provides the literature with new knowledge. Data are handled properly and exhaustively with chemometrics. I have some recommendations for authors to improve further their study. These follow the text sequence:

-Introduction

Line 47.’’ …extra virgin’’.

Line 76. ‘’Linear Discriminant Analysis’’. In addition, LDA is a supervised statistical technique. Kindly rephrase in line 76.

-Lines 92-109 should be removed to the results and discussion section. The introduction section should be reduced. It is too long.

-Line 350. There is a problem near the parenthesis.               

-Figures

The size and the quality of figures, which are indeed very informative, must be improved. Kindly modify them using at least 300 dpi resolution.

Based on the aforementioned, I suggest a minor revision of the present article prior its publication.

Author Response

Dear reviewer,

First of all, many thanks for your detailed comments and suggestions.

My answers are below and the corrected form of the manuscript also, as an attachment. I do hope you can accept our comments and modifications.

1. Introduction

Line 47.’’ …extra virgin’’. It was corrected

Line 76. ‘’Linear Discriminant Analysis’’. In addition, LDA is a supervised statistical technique. Kindly rephrase in line 76. It was corrected

-Lines 92-109 should be removed to the results and discussion section. The introduction section should be reduced. It is too long. It was removed and shortened. Replaced at lines 204-223. Concomitantly I changed the numerotation of references in the text and in the final reference list.

-Line 350. There is a problem near the parenthesis.     It was corrected          

Figures

The size and the quality of figures, which are indeed very informative, must be improved. Kindly modify them using at least 300 dpi resolution.

We made improvements for the figure S1 which was moved in the main text as Figure 1 with an improved resolution. Anyway, the released image from the spectrometer were not of enogh high resolution. Also for Figure 2 introduced in the main text FTIR spectra of SBO oils.

Thank you, all the best

Author Response File: Author Response.docx