The Impacts of Flowering Time and Tillering on Grain Yield of Sorghum Hybrids across Diverse Environments

Round 1

Reviewer 1 Report

Congratulations to the authors on the changes made to the manuscript. The manuscript merits publication with only a few modifications regarding points mentioned. Below.

Introduction: The description of the objectives is much improved. The inclusion of the current last sentence “The findings in this study can provide insight in to the design of breeding programs and selection of cultivars targeting diverse water deficit environments.” suggests that the authors may want to modify the second objective by adding “including diverse water deficit environments”, and then delete the current last sentence.

Also, photoperiod and photoperiod sensitivity should be mentioned as important determinants of time to flowering for much sorghum germplasm, particularly since the author indicated that this study was intended to be of global relevance.

Materials and Methods: The method for determining significance of genetic contribution of DTF or FTN to yield is not described.

Results:

Presentation of the number of tester/trial combinations with significant DTF/FTN associations with yield (is this based on significance of the regression coefficient? To clarify) and the number that are positive and negative before presenting the mean contributions (current lines 275-276 and 318-319) would be desirable.

Line 275-276 and 318-319: The method used for computing the mean contribution of DTF and FTN to yield can be clarified: is it based only on absolute values (disregarding positive or negative sign) and / or only considering those tester/trial combinations with significant regression coefficient or genetic contribution to yield?

Although Tables 4 and 5 indicate significance of % contribution to grain yield and regression coefficient (footnotes c and e), it is not clear what P level is taken for determining significance, and why there are cases with inconsistency in significance level between % contribution to grain yield and the regression coefficient within the same tester/trial combination.

Discussion:

Line 418-421: Difficult to understand with current wording.

Lines 418-421 for FTN and Lines 421-425 for DTF: Discussion of optimal levels can only be done for a specified context: a target population of genotypes and population of environments. This text is discussing optima without defining or discussing the target populations of genetic materials and environments. The context on which this studies results were obtained is the current Australian experimental hybrid pool in the 48 tester/trial combinations, and this should be mentioned. There is therefore need to briefly discuss (here or in Section 4.3) what the pool of Australian experimental hybrids represents (for example the genetic origin/background, expected level of photoperiod insensitivity or sensitivity, plant height) and what the population of trial environments represent (for example mean yield level, range of latitude, extent and type of water deficiency, soil and, or other parameters).

Making the conclusion that this study provides “reliable estimates” for the contribution of DTF and FTN to grain yield based on the number of experimental hybrids and trial environments (Line 462-465) does not take into consideration the important aspect of reliability for what kind(s) of condition(s) (noting that this studies’ estimates vary greatly in magnitude and direction (positive or negative) depending on the specific environment). An appropriate point to make is that this study provides an example of using results from a large number of independent trials (with differing genetic materials and environments) for examining the association of DTF and FTN with yield. Including discussion of the limitations of this study for examining the contribution of DTF and FTN to yield under water deficient conditions due to the small sample size, with only 3 of 21 trials characterized as experiencing more serious water deficiencies, would be appropriate here. Inclusion in the manuscript of your response to this reviewers’ prior question regarding yield levels of the set of 21 trial environments on which this study is based (text copied here below) would be useful and important for discussing the relevance of this studies estimates for this and other target populations of environments.

Authors’ response to Reviewer: “Part of the reason that high-yielding environments were over-sampled in our data set and only one trial has a mean yield less than the average yield in Australia is that our program tends to sample trials with reasonable stored soil moisture to reduce the chance of crop failure. In addition, many low yielding trials that would be harvested by farmers are failures as experiments and data is not collected or cannot be used (e.g. no genetic variation due to very high error, extreme lodging etc.). “

Author Response

Please find attached file for comments.

Author Response File: Author Response.pdf

Reviewer 2 Report

The present manuscript deals with the impact of days to flowering and tillering number on G x E interactions of sorghum yield in Australia. It comprises a large set of genotypes and single experiments, providing a robust data set for such a modelling approach. The limitations of this study (e. g. lack of weather data for several experiments) are adequately acknowledged, which is agreeable for readers (other manuscripts/publications are rather ambiguos in this regard). The results show that the impact of flowering time and tillering on genetic variation of yield is much lower than one could expect, which is valuable news for sorghum breeders. The discussion is adequate and the study as a whole has a high overall merit for sorghum breeding.

I am only somehow confused on Fig. 1, and generally about your heritability estimates: How can you have different heritabilities for one trait? Do the different heritabilites refer to different years, group of environments, set of testers etc.? This should be described more clearly. If heritability was calculated for each single trial/environment separately, from my point of view heritability would not be the right term, should be rather denoted "repeatability" in that case.

Author Response

“heritability” has been changed to “repeatability” throughout the manuscript.

 

Author Response File: Author Response.pdf

Reviewer 3 Report

Line 3: “across” should be capitalized.

Line 32: Make “Tillering” lowercase to be consistent.

Line 61: Consider changing “relatively simple traits with high heritability” to “traits with high heritability”. The inclusion of “simple” is undermined by the immediate subsequent text describing the confounding GxE aspect of that trait.

Line 108: Provide reference/link to what entails “local management practices”.

Line 118-122: seems like some of this could almost go into the discussion section.

Line 214: Include descriptions of ET type 1-5 in the methods.

Line 256: Figure 1. More journals are requiring the inclusion of each individual data point in boxplots, so for future graphs it would be a good idea to include the jitter method in ggplot2.

Line 260: BLUP appears for first time as an acronym outside of the methods.

Line 312: I think “between trial” should be “between-trial”.

Line 355: Same for “all between”.

Line 375: Again, using the word “simple” to describe a trait is subjective.

Line 410: Should be “water stressed” not “waster stressed”.

Line 526: Get the conflict of interest statement filled in

Author Response

Please find the attached file for details responses to reviewer's comments.

Author Response File: Author Response.pdf

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

This manuscript is within the scope of the journal. The study looks at the correlation between flowering time, tillering capacity and sorghum grain yield in a rather innovative way. It shows that, although flowering time and tillering capacity sometimes contributed to explaining genetic differences in grain yield, some other factors than the time of flowering and tiller numbers needs to be considered to better understand genotype by environment interaction on sorghum grain yield.

The introduction is well written, nicely highlighting the novelty of the study. The methodology is straightforward though maybe some clarifications are still needed ( explicitly state i. the strengths and limitations of the dataset, ii. the added value of the BLUP analysis see detailed review), and the results are clearly described. The discussion could be a bit strengthened by explicitly relating environmental characteristics with the main conclusions. The manuscript is publishable with minor revisions.

Comments for author File: Comments.pdf

Reviewer 2 Report

This paper has potential to contribute to an important topic of research for sorghum breeding. Although this paper hopefully will reach high levels of “Significance of content”, “Interest to Readers”, and “Overall Merit” with revisions based on comments below, the levels for these categories are currently low for this manuscript in its’ currently state. The lack of clear, meaningful objective(s) and the lack of presentation and use of any meta data for characterizing the different test environments and examining relationships between the results obtained and these factors limits this paper in its current form to more of an academic exercise with little gain of knowledge (learning) for guiding or inspiring future research.

Introduction:

(1) The background and justification for this work are fairly well presented but could be somewhat strengthened. As the article focuses on flowering time, it may also be important to mention a) the potential large effect of photoperiod sensitivity on flowering time and on canopy formation for certain sorghum germplasm, and b) what is reported and known regarding the variation and degree of photoperiod sensitivity and its influence on flowering for sorghum germplasm at different sowing dates and latitudes in Australia. Similarly the role of temperature (accumulated degree days) for progress to flowering would be well to mention, and to what extent temperature differences are experienced during the vegetative stage of growth in the Australian production (and testing) areas.

(2) The last paragraph describes what this study aims to do but does not provide clarity on the objectives of this study. To “examine the impacts of flowering time (DTF) and FTN on grain yield using data from…” is what was done (ie. method) but does not describe what learning and new understandings are to be generated? This must be clarified as the subsequent analyses and presentation of results should be guided by what understandings/knowledge is/are to be gained.

Materials and Methods:

(3) There is no environmental meta-data presented to describe the key environmental parameters of the test environments. For example, what rainfall (total and distribution), what soils and measures/expectations of water holding capacity, what temperature – location? Have these test locations been characterized previously as belonging to a single or multiple mega-environments for sorghum production? Can the authors identify any more homogenous sub-groups of the test environments? The trials for mean flowering suggest that there are some considerable differences between test environments.

(4) Lines 145-150: The genetic correlations between pairs of environments averaging 0.66 for DTF (ie maybe about only 44% of total variation is accounted for), and 0.76 for FTN (ie maybe about 58% of total variation accounted for) appears to contradict the conclusion that there is low GxE for these traits and that overall BLUPs can be used for these traits across all trials. Can the usefulness of the BLUPS be examined, maybe by comparing them with the actual values in the different trials or by generating BLUPS from random half of the tests and then compared with the actual values in the other tests?

Results:

(5) Line 209: Do DTF and FTN means represent trial means or means for test hybrids?

(6) Lines 239-241: Not clear which environments used for correlations: all 21 environments?

(7) Lines 243-259: Specify for which environment(s) the DTF-Yield genetic correlation high, and what if any environmental factor(s) (or germplasm sampling) that was different for those trials relative to the others. Simply reporting counts of positive or negative signs of associations without looking at the agronomic/biological meaning of those associations and relationship with particular environmental factors does not help provide better understanding. What if any environmental factors are common or specific for the trials with negative associations (AYTM17PIN, AYTM17CAR, AYTM16DAL) in Table 4, and likewise for those with positive associations (AYTM15DAL, AYTM15GAT), and what if any environmental factors differ between these two groups?

(8) section starting line 308: Similar to (7) above. What if any environmental factors are common or specific for the trials with negative associations (AYTM16WAR, AYTM16JAN, AYTM15JIM) in Table 5, and likewise for those with positive associations (AYTM16BLA, AYTM16CRO, AYTM16SPR), and what if any environmental factors differ between these two groups?

(9) Lines 266-268: not clear what is “same”

Discussion:

Lines 356-358: Not appropriate as the paper has not (yet) reported environmental meta data nor examined relationship between water stress/sufficiency with type of associations.

Lines 359-362: Not justified, as simple counts of sign of associations, irregardless of the extent of contribution to yield has little meaning.

Lines 365-368: Important points! Need further discussion. What is the reason why only one of 21 trials had a mean yield below the average sorghum yields of Australia during this period? What consequences for breeding can be drawn?

Lines 381-382: Is association confused with causation?

Lines 428-429: An unreasonable expectation? What was the objective(s) pursued in this paper?