Diallel Crosses of Cotton (Gossypium hirsutum L.)—Enhancement of Fiber Properties in Future Cultivars for Burkina Faso

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This study deals with an analysis of diallel crossing involving parents and F1's, of cotton genotypes from Bukina Faso and the USA. It aims to improve the fiber quality of the progenies and select the best hybrid combinations for the attributes under study. It has important results for the cotton genetic improvement program in Bukina Faso and important information for the scientific community, however the following points must be considered:

1)      The title is so big, so I suggest changing it to Diallel Crosses of Cotton for enhancement of Fiber Properties in Future Cultivars for Burkina Faso;

2)      To remove  the Keywords fiber Properties and Burkina Faso, once they are in the title;

3)      Line 73-74: The SCA refers to the effect of a specific hybrid combination;

4)      Line 85: to remove Bourgou et al.;

5)      Line 86: To remove "especially via SCA" as both CGA and SCA are important;      

6)      Line 121: “For the evaluation, and as a prelude to estimate heterosis....” it should be removed because there is no heterosis estimations and it was not the aim of this work;

7)      In data analysis, to say if the model is random or fixed. Also, how you got the GCA and SCA effects in the diallel Analysis. So, it should be fine if you put the diallel Analysis with all the expected Mean squares;

8)      Line 158-159: in the formula: 2GCA/(2GCA + SCA, describe if they are variance or effect;

9)      Replace all “p = 0.05” for p ≤ 0.05 and “p = 0.01” for p ≤0.01;

10)   Line 160-161: To show how you got σ2A=additive variance; σ2G=genotypic variance; σ2P=phenotypic variance;

11)   Line 170: Table 1. is not a table of Analysis of variance (ANOVA), it is a table of means and Skott and Knott test;

12)   The title of Table 2 should be Analysis of Variance of Combining Capacity….or simply diallel analysis…. This table is very confusing. the table must only contain what appears in the statistical model of “Model II” of Method I by Griffing (1956): Yij= u + gi + gj + Sij + Eij. It is not necessary to include a source of variation from the parents. Furthermore, its results are exactly the same as the GCA, which is very strange. Parental variation concerns phenotypic variance and GCA deals only with the variance of g effects. This may compromise all other analyses, results and interpretations;

13)   Because MAT does not show significance in table 2, it is better to remove all analyses and discussions about this trait. Also, if is zero, how can it be significant in this table and the others;

14)   Line 198: the authors say: “Mean squares analysis (Table 2) shows GCA effects were significant (p ≤0.01) for all the fiber property variables”, but for Mat all values are zero, then correct this;

15)   Lines 207-212: It is not correct to say that the results are not reliable, as they are normal for GCA analyses, as their values ​​depend on each other, and one contributes more or less for the characteristic. That's why their sum is always zero;

16)   Line 333:   +b was significant too;

17)   Line 444: To remove “colored”.

In conclusion, it is necessary for the analysis presented in Table 2 to be redone and all these corrections for this article to be approved.

Author Response

1)      The title is so big, so I suggest changing it to Diallel Crosses of Cotton for enhancement of Fiber Properties in Future Cultivars for Burkina Faso;

We removed 'from Burkina Faso and Texas A&M AgriLife Research' and left '-2' in to indicate a companion paper published earlier in Agronomy on agronomic properties. Thank you, the institutions involved do not need to be in the title.

2)      To remove  the Keywords fiber Properties and Burkina Faso, once they are in the title;

Removed.

3)      Line 73-74: The SCA refers to the effect of a specific hybrid combination;

Amended as noted.

4)      Line 85: to remove Bourgou et al.;

Removed.

5)      Line 86: To remove "especially via SCA" as both CGA and SCA are important;      

Removed.

6)      Line 121: “For the evaluation, and as a prelude to estimate heterosis....” it should be removed because there is no heterosis estimations and it was not the aim of this work;

Agreed and amended.

7)      In data analysis, to say if the model is random or fixed. Also, how you got the GCA and SCA effects in the diallel Analysis. So, it should be fine if you put the diallel Analysis with all the expected Mean squares;

A statement has been added to address that a fix model was used.

Analysis with the DIAL Win98 software according to Griffing's [12] Method II Model I" generated combining abilities effects by trait studied (analysis 1). For each trait Table 3 is mean squares, 4 is GCA effects, 5 is SCA effects.

Edits were made and hopefully this has been clarified in the text.

8)      Line 158-159: in the formula: 2GCA/(2GCA + SCA, describe if they are variance or effect; 

this ratio is variance component, it was removed from the table

9)      Replace all “p = 0.05” for p ≤ 0.05 and “p = 0.01” for p ≤0.01;

Replaced.

10)   Line 160-161: To show how you got σ2A=additive variance; σ2G=genotypic variance; σ2P=phenotypic variance;

a statement with reference was added. 

11)   Line 170: Table 1. is not a table of Analysis of variance (ANOVA), it is a table of means and Skott and Knott test;

Table title is amended to reflect contents.

12)   The title of Table 2 should be Analysis of Variance of Combining Capacity….or simply diallel analysis…. This table is very confusing. the table must only contain what appears in the statistical model of “Model II” of Method I by Griffing (1956): Yij= u + gi + gj + Sij + Eij. It is not necessary to include a source of variation from the parents. Furthermore, its results are exactly the same as the GCA, which is very strange. Parental variation concerns phenotypic variance and GCA deals only with the variance of g effects. This may compromise all other analyses, results and interpretations;

table has been corrected, and parental line and GCA/SCA ratio removed

13)   Because MAT does not show significance in table 2, it is better to remove all analyses and discussions about this trait. Also, if is zero, how can it be significant in this table and the others;

This is correct. It is an unfortunate miscommunication with fiber testing lab in that maturity cannot be directly measured, only calculated, from HVI tests, and is already estimated by micronaire. Therefore, maturity is removed from all analyses and discussion.

14)   Line 198: the authors say: “Mean squares analysis (Table 2) shows GCA effects were significant (p ≤0.01) for all the fiber property variables”, but for Mat all values are zero, then correct this;

Mat has been removed from the analysis and discussion.

15)   Lines 207-212: It is not correct to say that the results are not reliable, as they are normal for GCA analyses, as their values ​​depend on each other, and one contributes more or less for the characteristic. That's why their sum is always zero;

This has been changed to indicate it would have been more desirable to include more genotypes, but, but as suggested, the results are normal for this project.

16)   Line 333:   +b was significant too;

Yes! Thank you.

17)   Line 444: To remove “colored”.

'good colored' was replaced by 'high grade' since the +b and Rd refer to color grade.

In conclusion, it is necessary for the analysis presented in Table 2 to be redone and all these corrections for this article to be approved.

Thank you for your comments, all points have been addressed with a reply here and in the manuscript. 

Reviewer 2 Report

Comments and Suggestions for Authors

Dear Editor,

The manuscript by Bourgou et al. is very interesting manuscript and authors have discussed the breeding of cotton and Enhancement of Fiber Properties in Future Cultivars for Burkina Faso. The manuscript is up-to-date and discussed important aspects including fiber properties relating to breeding process. I really enjoyed while reading it, however I have few suggestions which need to be incorporated to improve quality of manuscript. I recommend acceptance of this manuscript after revision.

Comments:

1.      There are some typing errors in manuscript, which needs to be fixed.

2.      Abstract re write and clear the main concept and add some results.

3.      Introduction add the more citation in paragraph from line 35-42

4.      Introduction line 43-59 there is no citation in the paragraphs please add relevant citations

5.      In material and methods section add planting material in tabular form in given form much confusing

6.      In Crosses section LREC mean?

7.      The study is four year old in current scenario any study have done on same please add

8.      Improve the discussion section co relate with results in same sequence

9.      Authors can consult following literature;

          https://doi.org/10.3390/agronomy10010112

                https://doi.org/10.1186/s12870-023-04131-z

10.  If authors can incorporate these changes in manuscript, I am sure this article will enhance our understanding of fiber properties and breeding processes of cotton crop.

Comments on the Quality of English Language

Need to improve English quality

Author Response

1. There are some typing errors in manuscript, which needs to be fixed.

These errors are addressed as the manuscript is revised according to all the suggestions. Thank you for the comment.

2. Abstract re write and clear the main concept and add some results.

'Lack of variation among genetic resources available in Burkina Faso hinders breeding progress to meet fiber quality demands in future cultivars.' was added between the first and second sentence and specific numbers of hybrids replaced 'most' or other ambiguous results.

3. Introduction add the more citation in paragraph from line 35-42

The information is noted as being from [1]

Estur, G. Organisation et performances des filières cotonnières africaines: qualité et commercialisation du coton fibre en Afrique. Document pour la Banque Mondiale 2008, 75.

4. Introduction line 43-59 there is no citation in the paragraphs please add relevant citations

The information is updated as being from reference [1]

5. In material and methods section add planting material in tabular form in given form much confusing

A table (Table 1) is added showing line designation, code, origin, and type if the lines used in the crossing scheme. 

6. In Crosses section LREC mean?

'Texas A&M AgriLife Research at Lubbock' replaced its abbreviation, LREC (as noted in previous section). 

7. The study is four year old in current scenario any study have done on same please add

No further study is done since additional seed was not exported and progress was slowed by pandemic restrictions. 

8. Improve the discussion section co relate with results in same sequence

The discussion section has been reorganized to more closely match the results section.

9. Authors can consult following literature;

          https://doi.org/10.3390/agronomy10010112

                https://doi.org/10.1186/s12870-023-04131-z

10. If authors can incorporate these changes in manuscript, I am sure this article will enhance our understanding of fiber properties and breeding processes of cotton crop.

Thank you for your comments, all points have been addressed with a reply here and in the manuscript. 

Reviewer 3 Report

Comments and Suggestions for Authors

L39: Should “silk” be replaced with “fiber”? Or is silk the traditional term in Burkina Faso?

L123: I have a major concern that the experiment was only carried out in one environment. Due to environmental differences, a combining ability experiment is typically carried out in more than one environment. Can the authors add a justification in the manuscript for how the results can be correctly interpreted when only carried out in one environment?

L147-149: Please provide definitions and units of measure for each of these fiber properties. Also, I am familiar with HVI fiber measurements but have never received a maturity measure from HVI. How was maturity measured? Also, does SFI measure fibers less than 12,7 mm—please clarify in your definition.

Table 1: This table is not the analysis of variance. Rather mean values are presented. I suggest renaming something like “Means of fiber quality traits…”

L199-200: I suggest rewording as follows “…additive gene action is prevailing over dominant or epistatic expression…”

L209-210: The authors indicate that “the GCA effects calculated are not reliable due to the paucity of genotypes.” Then why do the authors present tables of results and discussion of the GCA effects? If the authors know that they cannot trust the GCA data, why is all the GCA data a focus of the paper and shared with readers? Please clarify if I have misinterpreted this sentence.

Author Response

L39: Should “silk” be replaced with “fiber”? Or is silk the traditional term in Burkina Faso?

The intention may have been to compare requirements to silk as to polyester, but it does make much more sense to replace silk with fiber.

L123: I have a major concern that the experiment was only carried out in one environment. Due to environmental differences, a combining ability experiment is typically carried out in more than one environment. Can the authors add a justification in the manuscript for how the results can be correctly interpreted when only carried out in one environment?

Crossing and F1 multiplication was done in the greenhouse in Texas under limited time constraints (and with fewer than desired genotypes due to import/export restrictions and space requirements). Since the objective was specifically to evaluate genetic diversity for improvement of future cultivars from Burkina Faso, the decision was taken to evaluate the populations in the target environment, Burkina Faso. This rationale is added.

L147-149: Please provide definitions and units of measure for each of these fiber properties. Also, I am familiar with HVI fiber measurements but have never received a maturity measure from HVI. How was maturity measured? Also, does SFI measure fibers less than 12,7 mm—please clarify in your definition.

Units of measure are added and SFI clarified. 

Table 1: This table is not the analysis of variance. Rather mean values are presented. I suggest renaming something like “Means of fiber quality traits…”

Thank you. The title has been changed to reflect the contents.

L199-200: I suggest rewording as follows “…additive gene action is prevailing over dominant or epistatic expression…”

Excellent catch - change has been made. 

L209-210: The authors indicate that “the GCA effects calculated are not reliable due to the paucity of genotypes.” Then why do the authors present tables of results and discussion of the GCA effects? If the authors know that they cannot trust the GCA data, why is all the GCA data a focus of the paper and shared with readers? Please clarify if I have misinterpreted this sentence.

The sentence has been changed to reflect results are normal for GCA analyses, as their values ​​depend on each other, and one contributes more or less for the characteristic.

Thank you for your comments, all points have been addressed with a reply here and in the manuscript. 

Reviewer 4 Report

Comments and Suggestions for Authors

Bourgou et al., had created F1 populations by half diallel crosses between six germplasms from Texas A&M AgriLife Research and 6 six parents from Burkina Faso. They have evaluated the fiber quality traits of both parents and hybrids, measured the general and special combining ability and select some parents with higher GCA and some hybrids with high SCA. This study will help researchers to find germplasms which could be used in the cultivar development process. This study has detailed data and analysis, but i still have some concerns which should be explain.

1.     Diallel cross and half diallel cross was different. In this study, authors have created 60 hybrids with 12 parents by half diallel cross, so they should express it more clearly.

2.     In Table 1, ** = significantly different at p = 0.01, this difference was compared by what?

3.     In the result part of Table 3, authors just have analyzed the GCA of one parent in one specific trait was positive or negative, but we most want to known was which parent was better in all fiber quality traits. For example, TX4 have both higher GCA in UHML and Str, authors should made more detailed analysis.

4.     Same as Q3, in table 4, hybrid (BK1×BK4) have both higher SCA in UHML, UI, and Str, but have negative SCA in SFI, authors should analysis the performance of hybrid based on both traits, don’t just focus on only one trait.

5.     In table 5, the narrow and broad heritability of Mic, UHML, UI, SFI, and Rd was identical, authors should explain this in the result and discussion part.

6.     Page 4, Line 166, ‘p’ should be italic.

7.     Page 5, Line 189, TX6 (6.54) should be TX6 (6.54%)

8.     Table 4, the symbol of cross should be ‘×’

9.     Table 5, σ²A, σ²G, σ²P, h­²n, h²b, et. Should be italic.

Comments on the Quality of English Language

No.

Author Response

1. Diallel cross and half diallel cross was different. In this study, authors have created 60 hybrids with 12 parents by half diallel cross, so they should express it more clearly.

The material and methods (line 83) states a half-diallel mating design was used, and some table titles have been changed to clarify that the data are from a half-diallel design. 

2. In Table 1, ** = significantly different at p = 0.01, this difference was compared by what?

The table title has been changed to clarify that means are presented in the table, and the footnote has been edited to clarify significance levels.

3. In the result part of Table 3, authors just have analyzed the GCA of one parent in one specific trait was positive or negative, but we most want to known was which parent was better in all fiber quality traits. For example, TX4 have both higher GCA in UHML and Str, authors should made more detailed analysis.

This is true, and an indication of how difficult it is to improve numerous fiber quality traits simultaneously. The results indicate which parents contributed to improvement of only one or two or more traits and which ones. There is some differentiation discussion about traits that are considered more economically important than others.

4. Same as Q3, in table 4, hybrid (BK1×BK4) have both higher SCA in UHML, UI, and Str, but have negative SCA in SFI, authors should analysis the performance of hybrid based on both traits, don’t just focus on only one trait.

Agree that it would be best to identify parents or hybrids that improve the most economically important traits simultaneously, but in this study with limited genotypes, no hybrid was a magic bullet improving all the desirable traits.

5. In table 5, the narrow and broad heritability of Mic, UHML, UI, SFI, and Rd was identical, authors should explain this in the result and discussion part.

In the results section this was addressed : "Because additive variance was close to genotypic variance, narrow-sense heritability was near equal to broad-sense heritability." ( σ2G = σ2A + σ2D ; because σ2D is near zero, σ2G is closely equal to σ2A )

A similar explanation has been added in the discussion section.  

A statement relating the impact of reported GCA and SCA on additive gene action has also been added in the results section. "This study found GCA effects to be higher than SCA effects which indicates large additive gene action and little non-additive gene action."

6. Page 4, Line 166, ‘p’ should be italic.

Format is corrected.

7. Page 5, Line 189, TX6 (6.54) should be TX6 (6.54%)

Percent sign is added.

8. Table 4, the symbol of cross should be ‘×’

This is corrected. Thank you.

9. Table 5, σ²A, σ²G, σ²P, h­²n, h²b, et. Should be italic.

This formatting has been corrected.

Thank you for your comments, all points have been addressed with a reply here and in the manuscript. 

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

As the authors left Gossipium in the title, it must be removed from the key words.

In table 3, old 2, there are still errors, I suggest redoing it as suggested in the text, attached, or

removing the sources of repetition, cross and PxL variation.

Comments for author File: Comments.pdf

Author Response

As the authors left Gossipium in the title, it must be removed from the key words.

Thank you. Gossypium has been removed from the key words.

In table 3, old 2, there are still errors, I suggest redoing it as suggested in the text, attached, or

In the revision, table 2 was still mistakenly referred to as ANOVA instead of means in the discussion, which is corrected.

removing the sources of repetition, cross and PxL variation.

Table 3 and its title was amended to include analysis of variance for GCA and SCA.

Reviewer 3 Report

Comments and Suggestions for Authors

I would like to thank the authors for addressing previous comments. However, I want to follow up on my comment regarding the number of environments where the half-diallel entries were grown.  I may not have been clear in my previous comment. The field experiment was grown in Burkina Faso in 2019. That's one environment. I agree with the authors rationale that it was best to grow the experiment in Burkina Faso rather than the US. However, why was the experiment not grown in another location in Burkina Faso in 2019 or grown at the same location in 2020 to generate more data for analysis from at least two environments? Can the authors justify that weather conditions do not vary from year to year in this cotton-growing region of Burkina Faso? In my experience, you need data from multiple environments (year-location combinations) to have better accuracy in your analysis and interpretations. 

Author Response

I would like to thank the authors for addressing previous comments. However, I want to follow up on my comment regarding the number of environments where the half-diallel entries were grown.  I may not have been clear in my previous comment. The field experiment was grown in Burkina Faso in 2019. That's one environment. I agree with the authors rationale that it was best to grow the experiment in Burkina Faso rather than the US. However, why was the experiment not grown in another location in Burkina Faso in 2019 or grown at the same location in 2020 to generate more data for analysis from at least two environments? Can the authors justify that weather conditions do not vary from year to year in this cotton-growing region of Burkina Faso? In my experience, you need data from multiple environments (year-location combinations) to have better accuracy in your analysis and interpretations. 

This concern is completely understandable. Unfortunately, it may not be one we can properly address. Crosses to create the F1 seed were made in the US during a Borlaug Fellowship visit with limited seed of the African parents brought in under strict import permit. The F1 seed and very limited seed of American parents were approved for export to Burkina Faso and remnant African parent seed was returned. Therefore, the study was conducted with the F1 seed created in US, with no more F1 seed created, and the seed quantity was limited. Further research is possible on subsequent generations (but not on F1 since more F1 seed production using the same parents is restricted); however, this paper is a companion paper to one previously published in Agronomy (reference #8) on agronomic traits of this same population, so we wanted to publish the fiber quality traits using the same population in the same environment.