Pisum sativum L. ‘Eso’: Metabolic Profiling of Yellow Seeds to Define the Optimal Harvest Time

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The sections of Results, Discussion and Conclusions are not enough supported and clarified by data, analysis or support information:

1. Results. As part of the hypothesis/objective stated was evaluate or determine the harvest time for Eso variety with the major quantity of nutritional and bioactive compounds, in principal text there is not data or evidences of differences or similarities in compound profiles throughout time harvest. For example, in the main text incorporate the summary of the Figures S2, S3, S4 and S5 as well as the Table S2 just with compounds chosen (values with one decimal), including the respective ANOVA disaggregated in their components (factor harvest time, errors, etc.).

2. In terms of statistical analysis I recommend to do a principal component analysis based in the average values of compounds across harvest time in order to determine the main compounds that determine the difference between harvest time (i.e. 40-70 days to harvest), representing in a scatterplot the harvest time.

3. In the Discussion section is difficult to follow which are the causes from the differences between harvest time or is poorly discussed.

4. The Conclusions are not supported on results such as “All the analyses show how a harvest on the 40th day leads to a product richer in both free amino acids and secondary metabolites” or “The understanding of the dynamics and mechanisms of plant metabolome alterations and their biological functions is fundamental to support the choice of the best cultivar and culture conditions” because the results are referred just one cultivar.

Author Response

The sections of Results, Discussion and Conclusions are not enough supported and clarified by data, analysis or support information:

1. Results. As part of the hypothesis/objective stated was evaluate or determine the harvest time for Eso variety with the major quantity of nutritional and bioactive compounds, in principal text there is not data or evidences of differences or similarities in compound profiles throughout time harvest. For example, in the main text incorporate the summary of the Figures S2, S3, S4 and S5 as well as the Table S2 just with compounds chosen (values with one decimal), including the respective ANOVA disaggregated in their components (factor harvest time, errors, etc.).

We thank the reviewer for this observation. We added this sentence in the results: ”Comparing the spectra at the different harvesting times, only quantitative differences were observed, and not qualitative ones.” To denote that the reported molecules were detected and quantified in all the observed samples. Moreover, to better highlight the quantitative changes occurring in the seeds during the observation period, ANOVA table has been moved to the main text and modified the results accordingly.

2. In terms of statistical analysis I recommend to do a principal component analysis based in the average values of compounds across harvest time in order to determine the main compounds that determine the difference between harvest time (i.e. 40-70 days to harvest), representing in a scatterplot the harvest time.

We thank the reviewer for the observation, which allows us to explain in larger details the characteristics of ASCA. Being a multivariate generalization of the analysis of variance, in the case of a single factor, asca operates exactly as suggested by the reviewer. Indeed, it calculates an effect matrix that is made of the mean spectra corresponding to the different levels of the factor (i.e., the four different harvesting times) and then extracts principal component scores and loading from this matrix. With respect to a single PCA on the means, ASCA has the advantage of providing tools (e.g., permutation testing) to evaluate whether the differences between the profiles recorded at the 4 harvest times are statistically significant and, if so, to evaluate whether all the levels are different or just some of those. The latter option can be accomplished by backprojecting the data onto the PCA loadings calculated from the level means, so to have a graphical representation of the within-level scatter: accordingly, by comparing the difference between the means (between-levels scatter) with the within-level scatter it is possible to evaluate which harvesting times correspond to a different metabolic profile (See Figure 1). In this context ASCA combines dimension reductions and maximum covariance criterion of PCA with the partition of the variability in the data operated by ANOVA modeling (together with the possibility of “post hoc” analysis) with the dimensionality reduction typical of PCA. In that regard, for a better interpretability, at first we included in the text the 3D scores scatterplot showing the differences between factor level means and the within factor variability (Figure 1). Moreover, we reported on the main text the Longitudinal Scores Plot for each component, for observing differences for each harvesting group. The following sentences were also added to the main text:” The main observable difference between sample groups in the ASCA model is associated with the 40th day of collection compared to the rest (Figure 1). To better understand the variations between the samples of the different collections with respect to the metabolites, Longitudinal Scores Plots (i.e., the representation of the scores along each component as a function of harvesting time) are therefore reported in Figure 2. Figures 3-5 collect the loadings for the three SCs, together with their confidence intervals, which allow interpreting the observed difference in terms of the measured metabolites.”.

3. In the Discussion section is difficult to follow which are the causes from the differences between harvest time or is poorly discussed.

We thank the reviewer for the opportunity to improve this paragraph. The discussion was revised according your comment (lines 287-413).

4. The Conclusions are not supported on results such as “All the analyses show how a harvest on the 40th day leads to a product richer in both free amino acids and secondary metabolites” or “The understanding of the dynamics and mechanisms of plant metabolome alterations and their biological functions is fundamental to support the choice of the best cultivar and culture conditions” because the results are referred just one cultivar.

We thank the reviewer for this comment, and we rewrote the conclusions as follows: “The present study showed the changes occurring in the nutritional profile of seeds already ripe and left to dry in the field, therefore suggesting how the choice of the harvest time could be employed to define their application. Indeed, seeds harvested at 40 days could be regarded as a source of free amino acids, more easily digested than proteins. Moreover, this harvest time is also the one with the lowest content of antinutrients. The combination of these factors makes this time and this product optimal also for possible human food use, especially for the growing plant-based meat market [60]. At the same time, seeds harvested at later times could be employed as a source of prebiotics, mainly oligosaccharides. The understanding of the dynamics and mechanisms of plant metabolome alterations and their biological functions is a fundamental step to define the optimal agricultural practices or how the culture conditions could be employed to diversify the production.”

Reviewer 2 Report

Comments and Suggestions for Authors

This paper was interesting and meaningful, metabolic profile of yellow pea during growth can indicate the nutritional formation and agronomic characteristics. There are some details need improve.

1 Is the meterial 'Eso' can represent the species yellow peas, why this cultivar?

2 the maturity of yellow pea was calculated from sowing, mang species was  calculated as 'after anthesis', which one is more appropriate?

3 The unit 'ppm' seem no longer used, please change it. 

4 the conclusion indicated the best harvest time was 40d after sowing, the sampling time was from 40 to 70d, is any chance that the day before 40 was more appropriate?

Comments on the Quality of English Language

The english was fine.

Author Response

This paper was interesting and meaningful, metabolic profile of yellow pea during growth can indicate the nutritional formation and agronomic characteristics. There are some details need improve.

1 Is the meterial 'Eso' can represent the species yellow peas, why this cultivar?

We thank the reviewer for this comment. ‘Eso’ is a yellow pea cultivar, which belongs to the same specie (Pisum sativum L.) of green peas. Therefore, we specifically indicate this cultivar to avoid misunderstandings. Regarding the choice of this cultivar, ‘Eso’ is one of the most promising varieties to be introduced in the Italian agricultural practices due to its high yield. This aspect has been evidenced in the introduction (lines 72-73).

2 the maturity of yellow pea was calculated from sowing, mang species was calculated as 'after anthesis', which one is more appropriate?

We thank the reviewer for this comment. Botanically it should be more appropriate ‘after anthesis’, but due to the scalar nature of flowering and fructification, for agronomists the ‘sowing’ is a more practical choice.

3 The unit 'ppm' seem no longer used, please change it. 

We thank the reviewer for this comment, but we feel that there is a misunderstanding since in NMR spectroscopy ‘ppm’ is not a quantitative unit, but a frequency unit (it indicates the resonance frequency of the protons of the molecules in solution) that is necessary to be employed to confront experiments carried out with different instrumentations.

4 the conclusion indicated the best harvest time was 40d after sowing, the sampling time was from 40 to 70d, is any chance that the day before 40 was more appropriate?

We thank the reviewer for this comment. 40 days is the minimum time for the seeds to complete their development, therefore it was even not possible to separate them from their epicarps. In any case, that could be an interesting matrix to be analyzed in further studies.

To better clarify the concept, we changed the conclusion as follows: “The present study showed the changes occurring in the nutritional profile of seeds already ripe and left to dry in the field, therefore suggesting how the choice of the harvest time could be employed to define their application. Indeed, seeds harvested at 40 days could be regarded as a source of free amino acids, more easily digested than proteins. Moreover, this harvest time is also the one with the lowest content of antinutrients. The combination of these factors makes this time and this product optimal also for possible human food use, especially for the growing plant-based meat market [60]. At the same time, seeds harvested at later times could be employed as a source of prebiotics, mainly oligosaccharides. The understanding of the dynamics and mechanisms of plant metabolome alterations and their biological functions is a fundamental step to define the optimal agricultural practices or how the culture conditions could be employed to diversify the production.”

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

As general comments:

In strict practice sense on information relevant in the Results section, the scores presented in the Figure 2 seems not so relevant as those related with the Figures S2, S3 and S4.