Combined Genome-Wide Association Studies (GWAS) and Linkage Mapping Identifies Genomic Regions Associated with Seedling Root System Architecture (RSA) under Different Nitrogen Conditions in Wheat (Triticum aestivum L.)
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsThe work presents the use of both GWAS and linkage mapping to decipher genomic regions associated with root traits under high and low N conditions in wheat. The work is interesting, addresses an important challenge of nitrogen use efficiency in wheat, and is well-presented; however, I have the following concerns:
Major comments:
The abstract does not display the outcome of a combined GWAS and linkage mapping. This section should be improved by highlighting the loci consistently mapped by both methods.
In addition, the authors did not clearly describe the experimental design. How were the seedling trays arranged in the lab? Was a specific experimental design followed to lay out the seedling trays? How many seedling trays per row? How many wheat seedlings were plated per tray? Was the experiment replicated? How many replications were there? Were all 243 GWAS panels and 123 RILs evaluated in a single experiment? Or were experiments conducted in batches? These details are important but missing.
Other comments:
In the method section, it would be nice also to indicate the generation of the RILs (F5?F6?....).
Section 2.4 is shallow; please, the PCA description should be removed and described under GWAS instead.
Line 133, what is the basis for setting a significant threshold at -log10 (P-value) > 5?
Under linkage mapping, what is the basis for setting LOD threshold at 2.5? To gain more confidence in these QTL, please run at least 1000 permutations test. The LOD threshold should be set based on the permutation test.
Line 86: what is it that was previously utilized to describe resistance to leaf rust and stripe rust? please, clarify the sentence.
It would be nice to see the heritability (repeatability) of these traits; why was it not computed?
In Figure 1, please indicate the position of the two parents used to create the RILs. It is important to show the contrast between the two parents for each trait in the frequency distribution.
Comments on the Quality of English LanguageThere are grammatical errors and sentence construction issues in this manuscript that should be improved by having an expert English expert read through the entire work.
Author Response
Response to the comments of reviewer #1
Major comments:
Comment 1: The abstract does not display the outcome of a combined GWAS and linkage mapping. This section should be improved by highlighting the loci consistently mapped by both methods. In addition, the authors did not clearly describe the experimental design. How were the seedling trays arranged in the lab?
Response: Sorry for our unclear expression. To express clearly, we revised as following:
1.We added the outcome of a combined GWAS and linkage mapping and highlight the loci consistently mapped by both methods with a description of “three colocalized intervals, AX-95160997/QRtrl.haust-3D, AX-109592379/QRnrt.haust-5A and AX-110924288/ QRtrl.haust-7D/QRtrs.haust-7D. According to the physical location of the colocalization of these two sites, it is between 39.61~43.74MB, 649.97~661.55 MB and 592.44~605.36 MB, which is called qRtrl-3D, qRnrt-5A and qRtrl-7D.” in line 27-30.
2.We also added “the experimental design. How were the seedling trays arranged in the lab?” as “A hydroponic seedling experiment using a 96-well tray was conducted in lab with two treatments including normal N (NN) and low N (LN). Five RSA traits include the relative number of root tips (RNRT), relative total root length (RTRL), relative total root surface area (RTRS), relative total root volume (RTRV), and relative average root diameter (RARD) were investigated.” in line 17-21.
Comments 2: Was a specific experimental design followed to lay out the seedling trays? How many seedling trays per row? How many wheat seedlings were plated per tray? Was the experiment replicated? How many replications were there? Were all 243 GWAS panels and 123 RILs evaluated in a single experiment? Or were experiments conducted in batches? These details are important but missing.
Response: Thank you for your thoughtful suggestions, We carried out a specific experimental design followed to lay out the seedling trays, but did not describe the details. In the revised manuscript, we added the information of the seedling trays arranged in the lab as “The nutrient solution culture method was used in a seedling stage test. The nutrient solution was referred to Hoagland et al. [28] nutrient solution and appropriately modified according to the nutritional characteristics of wheat (Table 1). Two treatments of normal N (NN, 4.0 mmol/L) and low N (LN, 0.8 mmol/L) were set in the laboratory. The LN nutrient solution contained the same nutrients concentration with NN nutrient solution except the concentration of Ca(NO3)2 with 0.4 mmol/L, CaCl2 with 2.1 mmol/L, and (NH4)2SO4 with 0 mmol/L. The pH of nutrition solution was adjusted to 6.0 with dilute HCl and NaOH before transferring. Wheat grains were soaked in a 10% H2O2 solution for 10 minutes, seeds were germinated in saturated CaSO4 solution for 7 days at 20℃. Once the seedlings reached the one-leaf stage and exhibited similar growth, and then the germinated seeds with residual endosperm removed were transferred to black 96-well seedling tray (12.7 cm long × 11.4 cm wide × 8.7 cm high, diameter of 6.3 mm and 12 holes per row), containing 800 mL of nutrient solution. The seedlings were randomly placed with five replications of each variety and grown in a intelligent artificial climate chamber (Zhejiang Top Clou-agri Technology Co. Ltd). The climate setting was 14 hours in light with a light intensity of 3000 Lx and a temperature of 25℃ and 10 hours in darkness with a light intensity of 0 Lx and a dark temperature of 18℃; and a humidity of 60%.). The solution was changed every 3 days. All experiments were completed in three batches.” in lines 101-118.
Comment 3: In the method section, it would be nice also to indicate the generation of the RILs (F5?F6?....)
Response: Thank you for your suggestion. The generation of the RILs is F6 in our experiment. We added this information as “The F6 RIL population of 123 lines” in line 96.
Comments 4: Section 2.4 is shallow; please, the PCA description should be removed and described under GWAS instead.
Response: We are agreement with your valuable advice. We removed the of description PCA and described it under GWAS instead. The modifications was in line 149-150 as “With the aid of Tassel v5.0 (https://www.maizegenetics.net/tassel), PCA and kinship analysis were carried out. The mixed linear model (MLM) in Tassel V5.0 software was used for GWAS.”
Comment 5: Line 133, what is the basis for setting a significant threshold at -log10 (P-value) > 5?
Response: Thank very much for your professional advice. The suggestive genome-wide significance threshold as - log (P) are also widely used in the relative researches, such as (He et al., 2017; Shi et al., 2022; Guo et al., 2023; Zhao et al., 2023), etc. In order to reduce false positives, this experiment also used the combination of FDR test and threshold for screening results in more reliable outcomes. However, due to the differences in genetic backgrounds between GWAS and RIL populations, the genetic backgrounds of GWAS panel used in GWAS analysis are generally more complex. Some significant loci in RIL populations may be masked in GWAS panel, and if the P-value set is too strict, the colocalization intervals cannot be found. When we determined the threshold, we tried several thresholds with - log (P) of 3, 4, 5, and 6, and found that there were 5176 SNP markers when - log (P) was 3 and 2659 SNP markers when - log (P) was 4. The standard was too loose, but when - log (P) was 6, there were only 135 SNP markers, which was too strict, resulting in many traits such as RTRL having no available SNP markers. Therefore, we chose - log (P) of 5. To reduce false positives, we also used the online software BioLadder (https://www.bioladder.cn/web/#/chart/58) Using the BH (Benjamini and Hochberg) method, adjust the P-value (corrected P-value) based on FDR, with - log10 (P-value)>5 and corrected P<0.05 as the cutoff criteria. For better expression the relative content, we added these contents to the text (line 159-167). Additionally, the FDR results are presented in Figure S2.
Guo, J., Guo, J., Li, L., Bai, X., Huo, X., and Shi, W., et al. (2023). Combined linkage analysis and association mapping identifies genomic regions associated with yield-related and drought-tolerance traits in wheat (Triticum aestivum L.). Theoretical and Applied Genetics 136(12).
He, Y., Wu, D., Wei, D., Fu, Y., Cui, Y., and Dong, H., et al. (2017). GWAS, QTL mapping and gene expression analyses in Brassica napus reveal genetic control of branching morphogenesis. Scientific reports 7(1), 15971.
Shi, H., Chen, M., Gao, L., Wang, Y., Bai, Y., and Yan, H., et al. (2022). Genome-wide association study of agronomic traits related to nitrogen use efficiency in wheat. Theoretical and Applied Genetics 135(12), 4289-4302.
Zhao, Y., Yan, X., Zeng, Z., Zhao, D., Chen, P., and Wang, Y., et al. (2023). Integrated genome‐wide association study and QTL mapping reveals qSa‐3A associated with English grain aphid, Sitobion avenae (Fabricius) resistance in wheat. Pest Management Science.
Comment 6: Under linkage mapping, what is the basis for setting LOD threshold at 2.5? To gain more confidence in these QTL, please run at least 1000 permutations test. The LOD threshold should be set based on the permutation test.
Response: Thanks for your professional suggestion. The LOD threshold at 2.5 was set according to previous literatures (Ren et al., 2017; Yang et al., 2021; Guo et al., 2023; Zhao et al., 2023). Some literatures also use a threshold of 3.0. In our test, we found that there were almost no available QTLs for analysis using the 3.0 threshold. Therefore, we adopted the threshold of 2.5, which is also the default threshold for IciMapping V4.2. This threshold of 2.5 is quoted from Zhao (2023), who used the same population and analysis method. In order to descript accurately, we explained this in the revised manuscript (line177-178).
Guo, J., Guo, J., Li, L., Bai, X., Huo, X., and Shi, W., et al. (2023). Combined linkage analysis and association mapping identifies genomic regions associated with yield-related and drought-tolerance traits in wheat (Triticum aestivum L.). Theoretical and Applied Genetics 136(12).
Ren, Y., Qian, Y., Xu, Y., Zou, C., Liu, D., and Zhao, X., et al. (2017). Characterization of QTLs for root traits of wheat grown under different nitrogen and phosphorus supply levels. Frontiers in plant science 8, 2096.
Yang, M., Wang, C., Hassan, M.A., Wu, Y., Xia, X., and Shi, S., et al. (2021). QTL mapping of seedling biomass and root traits under different nitrogen conditions in bread wheat (Triticum aestivum L.). Journal of Integrative Agriculture 20(5), 1180-1192.
Zhao, Y., Yan, X., Zeng, Z., Zhao, D., Chen, P., and Wang, Y., et al. (2023). Integrated genome‐wide association study and QTL mapping reveals qSa‐3A associated with English grain aphid, Sitobion avenae (Fabricius) resistance in wheat. Pest Management Science.
Comment 7. Line 86: what is it that was previously utilized to describe resistance to leaf rust and stripe rust? please, clarify the sentence.
Response: Sorry for our unclear expression. In the original paper, we want to use "what is it that was previously utilized to describe resistance to leaf rust and stripe rust? " to indicate that the group (RIL population) in our test has used in before study about disease resistance, and the results have been very good, proving that the molecular data of this population is reliable. To express more clearly, we have made a modifications as “The CIMMYT RIL population used in our study was previously utilized to describe resistance to leaf rust and stripe rust” in lines 98-99.
Comments 8: It would be nice to see the heritability (repeatability) of these traits; why was it not computed?
Response: Thank you for your thoughtful suggestions, which have been instrumental in refining our manuscript. We calculated the heritability of 5 RSA traits, but not decribed it. In the revised manuscript, we added the heritability of 5 RSA traits as Supplementary Materials (Table S4) and modified the description as in line (137-143 and 194-205).
|
Table S4 Heritability analysis of 5 RSA traits |
|||||
|
Population |
Trait |
Genotype (G) |
Environment (E) |
G×E |
Heritability (h2) |
|
GWAS panel |
TRL |
2160.1 |
4370.3 |
3073.9 |
0.689 |
|
TRS |
382.74 |
77.85 |
331.96 |
0.688 |
|
|
ARD |
0.01037 |
0.10784 |
0.00039 |
0.486 |
|
|
TRV |
0.21467 |
0.08362 |
0.2708 |
0.599 |
|
|
NRT |
928.8 |
901.7 |
2295.4 |
0.429 |
|
|
RIL population |
TRL |
1106 |
221.2 |
1847 |
0.539 |
|
TRS |
342.9 |
23.66 |
196.9 |
0.773 |
|
|
ARD |
0.08794 |
0.01319 |
0.06923 |
0.867 |
|
|
TRV |
0.0522 |
0.00598 |
0.04171 |
0.710 |
|
|
NRT |
348.944 |
131.39 |
719.71 |
0.483 |
|
Comment 9: In Figure 1, please indicate the position of the two parents used to create the RILs. It is important to show the contrast between the two parents for each in the frequency distribution.
Response: We want to express my gratitude for your perceptive comments on my manuscript. Regrettably, in the process of trying it out, we found that it is technically difficult to label the position of parents in the frequency distribution histogram because the vertical axis represents the percentage of frequency, and parents cannot determine the percentage of the population they belong to. In order to better indicate the numerical value of parents, we have presented it in Figure 3.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsCertainly! Here is the translated article in academic English:
The article explores a scarcely investigated topic in wheat: the impact of root architecture on nitrogen use efficiency. To this end, researchers employ both association mapping and biparental mapping, successfully identifying three novel QTLs with potential impact on the genetic improvement of root architecture’s contribution to nitrogen use efficiency. The experiments are well-designed, with results consistent with the proposed methodologies. The discussion is well-developed, and the conclusions are well-formulated.
The main point for improvement is to have the English reviewed by a native speaker. Minor edits for consideration are included in the attached PDF.
Comments for author File: Comments.pdf
The main point for improvement is to have the English reviewed by a native speaker. Minor edits for consideration are included in the attached PDF.
Author Response
Response to the comments of reviewer #2
Page 1
Thank you for your thoughtful suggestions, which have been instrumental in refining our manuscript. We have made a modification based on each your comments.
Page 2
Comment 1: Please include reference.
Response: Thank you for your suggestion. We have added references in 46 as “[2,3]”.
Jin, Y.; Liu, J.; Liu, C.; Jia, D.; Liu, P.; Wang, Y. Genome-wide association study of nitrogen use efficiency related traits in common wheat (Triticum aestivum L.). Acta Agronomica Sinica 2021, 47, 394-404.
Maccaferri, M.; El-Feki, W.; Nazemi, G.; Salvi, S.; Canè, M.A.; Colalongo, M.C.; Stefanelli, S.; Tuberosa, R. Prioritizing quantitative trait loci for root system architecture in tetraploid wheat. J Exp Bot 2016, 67, 1161-1178.
Comment 2: Not clear this sentence, please improve
Response: Sorry for our unclear expression. We have made modifications to the language in lines 48-50 as “Improving the nitrogen fertilizer utilization efficiency of wheat can achieve high yields and promote environmental protection and sustainable development”.
Comment.3 I understand this verb should be in present tense, not past
Response: Thank you for your valuable feedback on our manuscript. we have made a modifications as “Integrating the significance of breeding wheat varieties with high NUE is important in China.” In lines 50-51.
Comments.4-6:
Response: Thank you for your valuable feedback on our manuscript. We have revised “map-ping” to “mapping”, “In the two latest studies” to “In two recent studies on”, and “gene regions” to “genomic regions” as your suggestion.
Comment 7: This sentence is not clear, please improve it
Response: We have revised the text as suggested. The revise was “Herein, we conducted a lab hydroponic seedling experiment using a GWAS population (243) and RIL population (123) to investigate the RSA traits associated with various genotypes during the seedling stage under the conditions of normal N (NN) and low N (LN) supply.” in lines 81-84.
Comment 8: I think that English can be improved in this paragraph: This study used two wheat panels.
Response: Sorry for our unclear expression. We have revised “This study utilized two panels made of wheat” to “This study used two wheat panels.” in line 94.
Comment 9: If the GWAS panel included varieties I understand that you should not refer them as GWAS panel Names, just varieties. It would be interesting if you include same additional information of varieties included in the GWAS panel like at least grow habit (spring, winter, facultative), year of release
Response: We greatly appreciate the valuable feedback you have provided on my manuscript. Regrettably, this material was provided by another research group, and it is too difficult for us to show this information. Furthermore, this material includes some local varieties and special germplasm resources, which are not publicly available. If this information is made public in the future, we will present it in future articles. We hope you can appreciate the constraints.
Comments10-12:
Response: Thank you for your suggestion. We have made the modifications as each advice of you.
Comment 13: This might be moved to Results, unless it were previous data and you can include it as unpublished data.
Response: We are in agreement with your valuable input and have made the necessary adjustments as you suggested. We deleted the text based on your suggestion and provided details of the research findings in the results section in lines 220-221.
Comment 14. How did you do that? you should expand this idea
Response: Thank you for your insightful feedback. Indeed, we did not clearly convey the intended meaning of concept in original manuscript. This sentence might be redundant here, so we remove it. (line 99)
Page 3
Comments 1-8
Response: Thank you for your valuable feedback on our manuscript. We have revised the text as suggested.
Comment 6: please develop/explain the data quality check protocol
Response: Thank you for your suggestion. We have provided a brief explanation of data quality check protocol. as “The quality pretreatment of genotyping data was carried out for SNP call rate and MAF (minor allele frequency) with the PLINK software with threshold of maf 0.02 and geno 0.1. (http://zzz.bwh.harvard.edu/plink/tutorial.shtml) resulting in the retention of 395782 SNPs.” in line 146-149 of the revised manuscript.
Page 4
Comment 1: I would talk about GWAS panel, as I am not sure we can talk about natural population if the panel is composed of varieties.
Response: We have made the adjustments as you suggested. We have revised “natural population” to “GWAS panel.”.More detail please see the revised manuscript in lines 82,84,145,184,187,189,190,194,201,202,2065,210,214,335,336,345
Comments 2-10 and page 6
Response: Thank you for your suggestion. We have detailed made the modifications as your advice.
Page 8
Comments 1-2 I would use past tense in this sentence instead of present tense:
Response: Thank you for your suggestion. We have changed the tense in line 275-280.
Page10
Comments 1-3: what is unit point?
Response: Sorry for our unclear expression. We changed " the unit point might account for 8.16%–13.72% of the phenotypic
variation" to “explaining 81.6% to 13.72% of the phenotypic variances.” (lines 287-293)
Page 11
Comment 1: In which way? could you develop this idea?
Response: For example, using the method of correlation analysis, studies have shown that traits such as root length, diameter, and absorption area of the root system are significantly positively correlated with yield traits such as thousand grain weight, grain length, and grain width (Ruffel et al., 2014; Chen et al., 2016; Maccaferri et al., 2016). If we want to conduct in depth research in my group, we need to combine yield data, which is also what we need to do in my future work. However, we have not yet obtained these yield data. We will combine it with this RSA data in another article to explore their connection. We have added references in 46 as “Studies have shown that traits such as root length, diameter, and absorption area of the root system are significantly positively correlated with yield traits such as thousand grain weight, grain length, and grain width [2,3,7,8]”in line 314-316.
Chen, J., Zhang, Y., Tan, Y., Zhang, M., Zhu, L., and Xu, G., et al. (2016). Agronomic nitrogen‐use efficiency of rice can be increased by driving Os NRT 2.1 expression with the Os NAR 2.1 promoter. Plant Biotechnology Journal 14(8), 1705-1715.
Maccaferri, M., El-Feki, W., Nazemi, G., Salvi, S., Canè, M.A., and Colalongo, M.C., et al. (2016). Prioritizing quantitative trait loci for root system architecture in tetraploid wheat. Journal of experimental botany 67(4), 1161-1178..
Ruffel, S., Gojon, A., and Lejay, L. (2014). Signal interactions in the regulation of root nitrate uptake. Journal of Experimental Botany 65(19), 5509-5517.
Comment 2：Higher variation?
Response: Sorry for our unclear expression. It refers to the higher CV value of the GWAS panel. We explained it as “The coefficient of variation (CV) of GWAS panel is greater than that of RIL population,” in line 334-335.
Conments in Page 12, page13 and Page14:
Response: Thank you for your suggestion. We have detailed made the modifications as your advice.
Page 14
Comment 4 please delete the full stop, I would say ...were mapped according to the responses of root architecture related traits for first time in wheat.
Comment 6 you could say in the case of the linkage mapping which is the parental contributing positively in the three main NUE QTLs
Response: Thank you for your valuable feedback on our manuscript. We have revised the text as “were mapped according to the responses of root architecture related traits. In the case of the linkage mapping which is the parental contributing positively in the three main NUE QTLs.” in lines 421-423.
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for AuthorsThe manuscript entitled “Combined genome-wide association studies (GWAS) and link-2 age mapping identifies genomic regions associated with seed-3 ling root system architecture (RSA) under different nitrogen 4 conditions in wheat (Triticum aestivum L.)” is a nice approach to identify genomic regions controlling roots development through GWAS. Thus, providing a reference for future cloning of root-related genes. However, before it could be considered for publication some issues need to be addressed.
Regarding to material and methods section, the description of the procedures employed needs to be enhanced.
The section describes the nutrient solution, but not the conditions in which the plants were cultured, that is the type of growth chamber, temperature, photoperiod…
Also explain how many replicates by genotype were used, why only 5 were chosen…
Line 133, please explain why the threshold was set to -log10(P-value)>5 and if you performed any FDR correction. Also, section 2.5 seem to include duplicated information, please correct this.
Line 148 the authors refer to “compound interval mapping”, I believe it should be composite interval mapping, otherwise please explain in further detail the compound interval mapping procedure, as it is not something generally employed. Again, for the mapping explain why the threshold employed was chosen. And finally sentences like line 150 “MapChart V2.32 software was used to map the QTL linkage group” might create confusion, mapchart is used to represent the mapping results not to obtain them, please rephrase to clarify.
Regarding the results section:
Line 166 “The RSA traits of the natural population and the RIL population showed continuous distribution”, did authors mean data were normally distributed, if so, how do they test normality, what was the threshold…
Line 168, 169 “The correlation analysis of RSA traits showed that LN supply could increase TRL, TRS and NRT, and the increase was more obvious in the natural population” how did they test for correlation?, they used pearson, or spearmen correlation. Did they perform a test to asses the significance of the observed trends? Moreover, correlation is not the right analysis to asses if LN produces and increase, it will be more suitable to perform a paired t.test for example. Please give a better description of the analysis performed in the material and methods section and/or enhance those analysis. Later on, it is mentioned a t-test, but the procedure, and threshold of such test is not described. Also, be aware that the used of t-test requires a normal distribution and according to the distributions presented in figure 1 that might not be true for some traits, like TRS, TRL at least.
Based on figure 4 authors state that false discovery rate is controlled in their GWAs analysis, however according to the plots presented it does not seem the case. Please elaborate on this and verify the results. Also, Manhattan plot for RARD looks bizarre as many markers seem to have a 4 -log10(p-value), the images is strange, please verify the results.
Regarding discussion section, it could be enhanced to include more information about the detected QTLs, and the underling genes.
Comments on the Quality of English LanguageSome grammatical mistakes were found along the manuscript, for example:
Line 63 has mapping spelled “map-ping” even in the middle of the line. Lines 26, 72, 73, etc… have a similar mistake. Also, some sentences do not make full sense, line 86 “Using the single seed descent method [26].”, it stands alone and is difficult to understand the description about the material or to what it refers. The full text needs to be revised to correct grammatical mistakes.
Author Response
Response to the comments of reviewer #3
Comments 1: Regarding to material and methods section, the description of the procedures employed needs to be enhanced.
Response: As your suggestion, we enhanced the description of the procedures employed as “The pH of nutrition solution was adjusted to 6.0 with dilute HCl and NaOH before transferring. Wheat grains were soaked in a 10% H2O2 solution for 10 minutes, seeds were germinated in saturated CaSO4 solution for 7 days at 20℃. Once the seedlings reached the one-leaf stage and exhibited similar growth, and then the germinated seeds with residual endosperm removed were transferred to black 96-well seedling tray (12.7 cm long×11.4 cm wide×8.7 cm high, diameter of 6.3 mm and 12 holes per row), containing 800 mL of nutrient solution. The seedlings were randomly placed with five replications of each variety and grown in a intelligent artificial climate chamber (Zhejiang Top Clou-agri Technology Co. Ltd). The climate setting was 14 hours in light with a light intensity of 3000 Lx and a temperature of 25℃ and 10 hours in darkness with a light intensity of 0 Lx and a dark temperature of 18℃; and a humidity of 60%.). The solution was changed every 3 days. All experiments were completed in three batches.” in Line 107-118.
Comments 2: The section describes the nutrient solution, but not the conditions in which the plants were cultured, that is the type of growth chamber, temperature, photoperiod…
Response: Thanks for our professional comments. We carried out a lab experiment using seedling trays, but we did not describe this details. In the revised manuscript, we added the information of the seedling trays arranged in the lab as “The nutrient solution culture method was used in a seedling stage test. The nutrient solution was referred to Hoagland et al. [28] nutrient solution and appropriately modified according to the nutritional characteristics of wheat (Table 1). Two treatments of normal N (NN, 4mmol/L) and low N (LN, 0.8mmol/L) were set in the laboratory. The LN nutrient solution contained the same nutrients concentration with NN nutrient solution except the concentration of Ca(NO3)2 with 0.4 mmol/L, CaCl2 with 2.1 mmol/L, and (NH4)2SO4 with 0 mmol/L. The pH of nutrition solution was adjusted to 6.0 with dilute HCl and NaOH before transferring. Wheat grains were soaked in a 10% H2O2 solution for 10 minutes, seeds were germinated in saturated CaSO4 solution for 7 days at 20℃. Once the seedlings reached the one-leaf stage and exhibited similar growth, and then the germinated seeds with residual endosperm removed were transferred to black 96-well seedling tray (12.7 cm long×11.4 cm wide×8.7 cm high, diameter of 6.3 mm and 12 holes per row), containing 800 mL of nutrient solution. The seedlings were randomly placed with five replications of each variety and grown in a intelligent artificial climate chamber (Zhejiang Top Clou-agri Technology Co. Ltd). The climate setting was 14 hours in light with a light intensity of 3000 Lx and a temperature of 25℃ and 10 hours in darkness with a light intensity of 0 Lx and a dark temperature of 18℃; and a humidity of 60%.). The solution was changed every 3 days. All experiments were completed in three batches.” in (Line 101-118).
Comments 3: Also explain how many replicates by genotype were used, why only 5 were chosen.
Response: We established five replicates, referencing the studies conducted by [1-3]. In these studies, they utilized a range of 3 to 5 replicates, and therefore, I opted for five replicatesRen, Y.; Qian, Y.; Xu, Y.; Zou, C.; Liu, D.; Zhao, X.; Zhang, A.; Tong, Y. Characterization of QTLs for root traits of wheat grown under different nitrogen and phosphorus supply levels. Front Plant Sci 2017, 8, 2096.
Fan, X.; Zhang, W.; Zhang, N.; Chen, M.; Zheng, S.; Zhao, C.; Han, J.; Liu, J.; Zhang, X.; Song, L. Identification of QTL regions for seedling root traits and their effect on nitrogen use efficiency in wheat (Triticum aestivum L.). Theor Appl Genet 2018, 131, 2677-2698, doi:10.1007/s00122-018-3183-6.
Xiong, H.; Guo, H.; Zhou, C.; Guo, X.; Xie, Y.; Zhao, L.; Gu, J.; Zhao, S.; Ding, Y.; Liu, L. A combined association mapping and t-test analysis of SNP loci and candidate genes involving in resistance to low nitrogen traits by a wheat mutant population. Plos One 2019, 14, e211492, doi:10.1371/journal.pone.0211492.
Comment 4: Line 133, please explain why the threshold was set to -log10(P-value)>5 and if you performed any FDR correction. Also, section 2.5 seem to include duplicated information, please correct this.
Response: Thank very much for your professional advice. Indeed, other reviewer (#1&3) also mentioned “why the threshold was set to -log10(P-value)>5”. Suggested thresholds are also widely used in research, such as (Zhao et.al, LI et.al, He et.al, Guo et.al), etc. When we determined the threshold, we tried several thresholds with - log (P) of 3, 4, 5, and 6, and found that there were 5176 SNP markers when - log (P) was 3 and 2659 SNP markers when - log (P) was 4. The standard was too loose, but when - log (P) was 6, there were only 135 SNP markers, which was too strict, resulting in many traits such as RTRL having no available SNP markers. Therefore, we chose - log (P) of 5. In order to reduce false positives, this experiment also used The combination of FDR test and threshold for screening results in more reliable outcomes. Due to the differences in genetic backgrounds between GWAS and RIL populations, the genetic backgrounds of GWAS panel used in GWAS analysis are generally more complex. Some significant loci in RIL populations may be masked in GWAS panel, and if the P-value set is too strict, the colocalization intervals cannot be found. In order to better search for co-localization intervals, a threshold of - log10 (P-value)>5 was used. To reduce false positives, use the online software BioLadder（https://www.bioladder.cn/web/#/chart/58）Using the BH (Benjamini and Hochberg) method, adjust the P-value (corrected P-value) based on FDR, with - log10 (P-value)>5 and corrected P<0.05 as the cutoff criteria, We added this content in (line 159-166,). The FDR results are presented in Figure S2.
Zhao, Y.; Yan, X.; Zeng, Z.; Zhao, D.; Chen, P.; Wang, Y.; Chen, F.; Wang, C. Integrated genome‐wide association study and QTL mapping reveals qSa‐3A associated with English grain aphid, Sitobion avenae (Fabricius) resistance in wheat. Pest Manag Sci 2023, doi:10.1002/ps.7598.
Li, X.; Guo, T.; Wang, J.; Bekele, W.A.; Sukumaran, S.; Vanous, A.E.; McNellie, J.P.; Tibbs-Cortes, L.E.; Lopes, M.S.; Lamkey, K.R. An integrated framework reinstating the environmental dimension for GWAS and genomic selection in crops. Mol Plant 2021, 14, 874-887, doi:10.1016/j.molp.2021.03.010.
He, Y.; Wu, D.; Wei, D.; Fu, Y.; Cui, Y.; Dong, H.; Tan, C.; Qian, W. GWAS, QTL mapping and gene expression analyses in Brassica napus reveal genetic control of branching morphogenesis. Sci Rep-Uk 2017, 7, 15971, doi:10.1038/s41598-017-15976-4.
Guo, J.; Guo, J.; Li, L.; Bai, X.; Huo, X.; Shi, W.; Gao, L.; Dai, K.; Jing, R.; Hao, C. Combined linkage analysis and association mapping identifies genomic regions associated with yield-related and drought-tolerance traits in wheat (Triticum aestivum L.). Theor Appl Genet 2023, 136, doi:10.1007/s00122-023-04494-9.
- Sorry, the statement in section 2.5 was incorrect, we have already deleted the duplicate information in section 2.5.
Comments 5: Line 148 the authors refer to “compound interval mapping”, I believe it should be composite interval mapping, otherwise please explain in further detail the compound interval mapping procedure, as it is not something generally employed. Again, for the mapping explain why the threshold employed was chosen. And finally sentences like line 150 “MapChart V2.32 software was used to map the QTL linkage group” might create confusion, mapchart is used to represent the mapping results not to obtain them, please rephrase to clarify.
Response: Thank very much for your professional advice. Based on your suggestion, we have made the following modifications:
1.The “compound interval mapping” should be “composite interval mapping” and we have made a revisions in line 177 of the revised manuscript.
2.Indeed, reviewer #1 also mentioned “why the threshold employed was chosen”. The suggestive genome-wide significance threshold as - log (P) are also widely used in the relative researches, such as (He et al., 2017; Shi et al., 2022; Guo et al., 2023; Zhao et al., 2023), etc. Due to the differences in genetic backgrounds between GWAS and RIL populations, the genetic backgrounds of GWAS panel used in GWAS analysis are generally more complex. Some significant loci in RIL populations may be masked in GWAS panel, and if the P-value set is too strict, the colocalization intervals cannot be found. In order to better search for co-localization intervals, a threshold of - log10 (P-value)>5 was used. When we selected the threshold, we also referred to many literature and found that many studies also used a threshold of 2.5, such as (Zhao et al, Ren.et.al, Yang.et.al）Some literature also uses a threshold of 3.0, but we found that there were almost no available QTLs for analysis using the 3.0 threshold. Therefore, we adopted the threshold of 2.5, which is also the default threshold for IciMapping V4.2. This 2.5 is the threshold we quoted from Zhao's article. We used the same population and analysis method, and based on your suggestion, I explained this in the article.
Guo, J., Guo, J., Li, L., Bai, X., Huo, X., and Shi, W., et al. (2023). Combined linkage analysis and association mapping identifies genomic regions associated with yield-related and drought-tolerance traits in wheat (Triticum aestivum L.). Theoretical and Applied Genetics 136(12).
He, Y., Wu, D., Wei, D., Fu, Y., Cui, Y., and Dong, H., et al. (2017). GWAS, QTL mapping and gene expression analyses in Brassica napus reveal genetic control of branching morphogenesis. Scientific reports 7(1), 15971.
Shi, H., Chen, M., Gao, L., Wang, Y., Bai, Y., and Yan, H., et al. (2022). Genome-wide association study of agronomic traits related to nitrogen use efficiency in wheat. Theoretical and Applied Genetics 135(12), 4289-4302.
Zhao, Y., Yan, X., Zeng, Z., Zhao, D., Chen, P., and Wang, Y., et al. (2023). Integrated genome‐wide association study and QTL mapping reveals qSa‐3A associated with English grain aphid, Sitobion avenae (Fabricius) resistance in wheat. Pest Management Science.
Ren, Y., Qian, Y., Xu, Y., Zou, C., Liu, D., and Zhao, X., et al. (2017). Characterization of QTLs for root traits of wheat grown under different nitrogen and phosphorus supply levels. Frontiers in plant science 8, 2096.
Yang, M., Wang, C., Hassan, M.A., Wu, Y., Xia, X., and Shi, S., et al. (2021). QTL mapping of seedling biomass and root traits under different nitrogen conditions in bread wheat (Triticum aestivum L.). Journal of Integrative Agriculture 20(5), 1180-1192.
3.Based on your suggestion, we revised “MapChart V2.32 software was used to map the QTL linkage group” to “According to the results obtained from IciMapping V4.2, MapChart V2.32 software was used to map the QTL linkage group.” (line 177-179)
Comments 6: Line 166 “The RSA traits of the natural population and the RIL population showed continuous distribution”, did authors mean data were normally distributed, if so, how do they test normality, what was the threshold…
Response: Thank very much for your professional advice. We carried out a normality test when prepared the manuscript, but we did not show it. The results of the normality test are：“Except for TRL (NN), TRS (LN), NRT in GWAS panel, and TRL and TRS in RIL populations that do not follow a normal distribution, all other traits follow a normal distribution (P>0.05)”in line 206-213. The reason may be that RSA influenced by multiple genes, each with potentially small effects, and that the trait's expression can be highly dependent on environmental conditions. The combination of genetic and environmental factors can lead to a distribution that does not conform to a normal pattern. For instance, in our study mixed linear models (MLM) can be used to account for random effects, which can reduce the risk of false positives and improve the accuracy of the analysis. We also revised the expression as “Using SPSS 22.0 and Kolmogorov Smirnov (K-S test) for normality testing, if P>0.05, it indicates that the sample data is normally distributed” in line 134-137.
Comments 7: Line 168, 169 “The correlation analysis of RSA traits showed that LN supply could increase TRL, TRS and NRT, and the increase was more obvious in the GWAS panel” how did they test for correlation?, they used pearson, or spearmen correlation. Did they perform a test to asses the significance of the observed trends? Moreover, correlation is not the right analysis to asses if LN produces and increase, it will be more suitable to perform a paired t.test for example. Please give a better description of the analysis performed in the material and methods section and/or enhance those analysis. Later on, it is mentioned a t-test, but the procedure, and threshold of such test is not described. Also, be aware that the used of t-test requires a normal distribution and according to the distributions presented in figure 1 that might not be true for some traits, like TRS, TRL at least.
Response: We are grateful for the constructive feedback you provided on our work. As your suggestion, we have made the following modifications:
1.We are very sorry for our incorrect expression. The statement should not be “the correlation analysis”. Therefore, we conducted a T-test in Table 2 and revised the text as “The normality test results indicated that the TRL and TRS in RIL populations, the RSA traits followed a normal distribution (P>0.05) except for the TRL under NN, TRS under LN, and NRT in GWAS panel. However, when applied a natural logarithm transformation, all the traits followed to a normal distribution (P>0.05). Thus, we proceeded to conduct a paired T-test using the natural logarithm values of these values. The paired t-test results showed that LN supply could increase TRL, TRS and NRT in GWAS panel, and TRL in RIL population, and the increase was greater in the GWAS panel than RIL population (Table 2).” (line 206-213)
Table 2. Summary of RSA traits in different populations under different nitrogen levels
|
Treatment |
Trait |
GWAS panel |
|
RIL population |
||||||||
|
Min. |
Max. |
Average |
Standard deviation |
CV (%) |
|
Min. |
Max. |
Average |
Standard deviation |
CV (%) |
||
|
NN nutrient solution |
TRL (cm) |
60.07 |
372.1 |
143.72** |
57.09 |
39.72 |
|
58.19 |
279.43 |
132.28* |
53.46 |
40.41 |
|
TRS (cm2) |
14.13 |
135.1 |
51.73** |
19.59 |
37.87 |
|
19.29 |
209.41 |
45.27 |
22.16 |
48.95 |
|
|
ARD (mm) |
0.67 |
1.62 |
1.15* |
0.13 |
11.39 |
|
0.84 |
2.3 |
1.08** |
0.15 |
14.29 |
|
|
TRV (cm3) |
0.25 |
3.9 |
1.5** |
0.58 |
38.95 |
|
0.45 |
2.52 |
1.18 |
1.08 |
37.14 |
|
|
NRT |
38 |
293 |
91** |
35.6 |
39.2 |
|
23 |
214 |
81 |
33.16 |
41.15 |
|
|
LN nutrient solution |
TRL (cm) |
80.12 |
474.4 |
237.38 |
89.54 |
37.72 |
|
66.91 |
290.3 |
146.09 |
55.6 |
44.09 |
|
TRS (cm2) |
20.3 |
159.3 |
77.71 |
27.42 |
35.29 |
|
18 |
232.67 |
44.35 |
24.43 |
55.09 |
|
|
ARD (mm) |
0.23 |
2.31 |
1.11 |
0.29 |
26.36 |
|
0.84 |
2.55 |
1.13 |
0.19 |
16.39 |
|
|
TRV(cm3) |
0.09 |
4.54 |
1.91 |
0.85 |
44.62 |
|
0.38 |
2.48 |
1.17 |
1.32 |
40 |
|
|
NRT |
43 |
457 |
134 |
72 |
54.19 |
|
29 |
171 |
82 |
33 |
39.88 |
|
|
Relative ratio |
RTRL |
0.93 |
5.68 |
1.8 |
0.83 |
45.88 |
|
0.31 |
2.45 |
1.05 |
0.46 |
44.34 |
|
RTRS |
0.79 |
5.68 |
1.62 |
0.69 |
42.8 |
|
0.34 |
2.42 |
1.06 |
0.46 |
42.95 |
|
|
RARD |
0.2 |
2.22 |
0.98 |
0.29 |
29.24 |
|
0.69 |
1.64 |
1.06 |
0.16 |
14.78 |
|
|
RTRV |
0.06 |
5.61 |
1.4 |
0.78 |
55.72 |
|
0.36 |
2.66 |
1.1 |
0.5 |
45.57 |
|
|
RNRT |
0.73 |
5.58 |
1.56 |
0.82 |
52.26 |
|
0.27 |
3.02 |
1.13 |
0.53 |
46.62 |
|
Note: ＊＊ indicates significantly defferent level of NN and LN paired T test at P＜0.01 level．＊indicates significantly defferent level of NN and LN paired T test at P＜0.05 level Convert non normally distributed samples into natural logarithmic values (ln(x)) for comparison. Total root length (TRL), total root surface area (TRS), total root volume (TRV), number of root tips (NRT), average root diameter (ARD), relative number of root tips (RNRT), relative total root length (RTRL), relative total root surface area (RTRS), relative total root volume (RTRV), relative average root diameter (RARD).
- We used pearson and revise the expression as “Pearson correlation map”. We also added the not as “ * refer to the different significance at P<0.05 level respectively .” (Figure 2)
- We carried out a paired sample t-test, some traits do not conform to a normal distribution due to some extreme individuals. To address this issue, we have applied a natural logarithm transformation to the values. This transformation has effectively normalized the data, and the transformed data fit to a normal distribution (P>0.05). Therefore, we proceeded to conduct a paired t-test using the natural logarithm values of the values. The use of the natural logarithm transformation is a common practice in the field when dealing with data that do not meet the assumptions of normality, and it has been shown to be effective in stabilizing variance and making the data more amenable to parametric tests (Altman and Bland, 1995; Mishra et al., 2019; Liaw et al., 2021; West, 2022).We have included the results of the Kolmogorov-Smirnov test for normality before and after the transformation in our revised manuscript, along with a detailed explanation of the rationale behind our choice of statistical tests in line 134-136, 206-213
Altman, D.G., and Bland, J.M. (1995). Statistics notes: the normal distribution. Bmj 310(6975), 298.
Liaw, K.L., Khomik, M., and Arain, M.A. (2021). Explaining the Shortcomings of Log‐Transforming the Dependent Variable in Regression Models and Recommending a Better Alternative: Evidence From Soil CO2 Emission Studies. Journal of Geophysical Research: Biogeosciences 126(5), e2021J-e6238J.
Mishra, P., Pandey, C.M., Singh, U., Gupta, A., Sahu, C., and Keshri, A. (2019). Descriptive statistics and normality tests for statistical data. Annals of cardiac anaesthesia 22(1), 67-72.
West, R.M. (2022). Best practice in statistics: The use of log transformation. Annals of Clinical Biochemistry 59(3), 162-165.
Comments 8: Based on figure 4 authors state that false discovery rate is controlled in their GWAs analysis, however according to the plots presented it does not seem the case. Please elaborate on this and verify the results. Also, Manhattan plot for RARD looks bizarre as many markers seem to have a 4 -log10(p-value), the images is strange, please verify the results.
Response: It is really true as you suggested that Manhattan plot for RARD looks bizarre. As your advice, we verify the data and found some values were mistake in data format conversion. We are very Sorry for these mistakes. We have revised Figure4 to correct the errors. Some loci have also changed, and we have made modifications in the article and Table S2.
Comment 9: Regarding discussion section, it could be enhanced to include more information about the detected QTLs, and the underling genes.
Response: We greatly appreciate you valuable feed back. In the revised manuscript, we enhanced the discussion to include more information about the detected QTLs as “Some of these loci overlapped or coincided with multiple RSA related loci discovered by Jin et al. [2], Xiong et al. [43], Fan et al [47], Yang et al. [36], and Ren et al. [56] (Table S3). In which, the mapped area contains numerous known NUE related genes such as NRT2s genes on chromosome 6A, which can be associated with RTRL, RTRS, RNRT, and other traits. In addition, in our genetic mapping interval, we have found that several genes related to RSA may be associated with NUE in wheat, including genes NPF6.2, FD-GOGAT, and NR1.2. Within the overlapping intervals identified by both GWAS and QTL mapping, we found the GS2 genes [57-59]. In the genetic mapping interval, we also found auxin responsive gene. Despite the discovery of genes within the genetic mapping interval, the large size of the candidate gene interval currently prevent us from definitively identifying these genes as the specific candidates of interest. Further refinement of the interval through additional genetic and molecular analyses will be necessary to narrow down the list of potential genes and pinpoint those related to NUE.” (line 362-375). However, due to the fact that most studies use traditional RFLP, SSR, or DArT markers, reliable comparisons cannot be made based on existing wheat genetic or physical maps, and it is currently uncertain whether they are consistent with the loci found in this study.
Comment 10：Line 63 has mapping spelled “map-ping” even in the middle of the line. Lines 26, 72, 73, etc… have a similar mistake. Also, some sentences do not make full sense, line 86 “Using the single seed descent method [26].”, it stands alone and is difficult to understand the description about the material or to what it refers. The full text needs to be revised to correct grammatical mistakes.
Response:
1.Thank you for your suggestion. The misplaced dashes are stemming from line shifts during formatting. We have deleted the dashes“-“.
2.We apologize for the incomplete expression. We have removed unclear statements. The manuscript has been revised as “The CIMMYT RIL population used in our study was previously utilized to describe resistance to leaf rust and stripe rust” in line 98-99.
Author Response File: Author Response.pdf
Reviewer 4 Report
Comments and Suggestions for AuthorsI think that, from a standpoint of materials and methods used in the paper by Jia et al., it would be interesting to the readers of Agriculture. However, the paper is marred by linguistic issues, some of which may have been caused by edits done during writing of the manuscript. The writing has to be corrected to the point I'm going to recommend a major revision.
The sentences I'm referring to using line ranges below are wholly contained in the range to avoid ambiguity in which sentence I'm referring to.
82: "panels made of wheat" is probably not a correct description
86: What was achieved using the method? The sentence seems incomplete.
86-87: What was utilized?
87-88: If talking about results of this study, the results do not belong to the Materials & Methods chapter. If results of some previous work are meant by this, a more clear statement and some details on the study and its results are needed.
96-97: The sentence about N supply is unclear with poor English. Maybe it would be better if LN treatment concentrations were included in Table 1.
105-117: Explanations of the trait shortcuts, preferrably in the form of a table, should be added here as well (they are currently only in the abstract/introduction) so people can reference them quickly. Descriptions of how the traits were measured are needed as well.
118-120: This section is too short, it should either be expanded or incorporated somewhere else. Also poor English.
123-124: I don't understand the meaning of the sentence.
124-125: Quality check procedure should be detailed. If the reference details the procedure this should be stated, but some short description should still be in this paper so people do not have to search in another paper.
133: While the cutoff for significant values of -log10 p > 5 seems reasonable to me, I would recommend using a false discovery rate measure in the future, for example the q-values provided by R package qvalue (DOI: 10.18129/B9.bioc.qvalue). A change to using q-values in this paper would probably be an unreasonable amount of work, but I would ask the authors to provide the q-value cutoff that corresponds to the chosen p-value cutoff for their results. I.e. to calculate the q-values for the Tassel results and provide the highest q-value among the p-values that were deemed significant by the paper's methods in the text of the paper.
134-136: the cited paper doesn't seem to support the choice of interval size
141: "the marker" invokes a one certain SNP in my view, but none is mentioned (as it sholdn't in the Materials & Methods chapter). The sentence should be corrected.
144: The sentence is probably missing some vital part, it should be fixed.
144: What did the screening and de-redundancy entail?
161-162: The repetition of the word "relative" is unnecessary.
155-165: Dashes showing range (e.g. 11.39%-39.72%) were probably mistakenly retained during changes to the manuscript as minus signs (e.g. 11.39% (ARD) -39.72% (TRL)), wich causes a lot of confusion. Reference to Table 2 would be appreciated as well.
199-200: There seem to be no "lowercase letters" indicating significant difference; perhaps this was forgotten from an earlier version of the figure after "ns", "*" and "**" started to be used to show significance of the difference. Also, I'm missing information what do the A-E subfigures display, I presume different traits.
238: More datailed explanation of the figure is needed, for example the y-axis, which I assume is -log10(p).
274: According to Figure 1, it seems that reduced N supply lowered TRS in the RIL population
279-280, 296-300: I don't think that merely a high variability of a trait proves high genetic control in a trait. I would at least need to see variability within some clonal populations and an appropriate statistical test.
286-291: These sentences say the same things as the sentences on lines 281-285, probably an editing error.
295-296: Citation is needed.
Comments on the Quality of English Language
As per the reviewer guidelines, I din't focus on particular language issues and I've noticed the report form having an English quality comment section only after I've written the comments in an external editing software. The language-related notes crucial to the scientific aspect of the paper are in the general comments section.
Here are lines of few of the biggest issues I've noted: 93, 107, 136-137, 141-142, 306-308, 355-356. There were also misplaced dashes, probably stemming from line shifts during formatting.
Author Response
Response to the comments of reviewer #4
Comment 1 in line 82: "panels made of wheat" is probably not a correct description
Response: Sorry for our unclear expression. We have modified the expression as “This study used two wheat panels” in line 94.
Comment 2 in line 86-87: what was achieved using the method? The sentence seems incomplete. What was utilized?
Response: We apologize for the incomplete expression. We have removed unclear statements. The manuscript has been revised as “The CIMMYT RIL population used in our study was previously utilized to describe resistance to leaf rust and stripe rust” in line 98-99.
Comments 3 in line 87-88: If talking about results of this study, the results do not belong to the Materials & Methods chapter. If results of some previous work are meant by this, a more clear statement and some details on the study and its results are needed.
Response: We deleted the text based on your suggestion and provided the detail findings in the results section in lines 220-221.
Comment 4: 96-97: The sentence about N supply is unclear with poor English. Maybe it would be better if LN treatment concentrations were included in Table 1.
Response: Thank you for your thoughtful suggestions. Your insights have prompted us to revise the section starting from line 104-106 and Table1.The revised text as “Two treatments of normal N (NN, 4mmol/L) and low N (LN, 0.8mmol/L) were set in the laboratory. The LN nutrient solution contained the same nutrients concentration with NN nutrient solution except the concentration of Ca(NO3)2 with 0.4 mmol/L, CaCl2 with 2.1 mmol/L, and (NH4)2SO4 with 0 mmol/L..”
Comments 5: 105-117: Explanations of the trait shortcuts, preferrably in the form of a table, should be added here as well (they are currently only in the abstract/introduction) so people can reference them quickly. Descriptions of how the traits were measured are needed as well.
Response: We are grateful for the constructive feedback you provided on our work. We have made the revisions as “such as total root length (TRL), total root surface area (TRS), total root volume (TRV), number of root tips (NRT), and average root diameter (ARD)” in lines 129-131. We also added trait shortcuts to the Figures1,2,4.
Comment 6: 118-120: This section is too short, it should either be expanded or incorporated somewhere else. Also poor English.
Response: Thank very much for your professional advice. As your and the reviewer#1 advises, this section has been merged into the GWAS section in lines 149-150.
Comment 7:123-124: I don't understand the meaning of the sentence.
Response: Sorry for our unclear expression. In the original manuscript, we want to express that the GWAS panel is usable, because previous study (Yang, et.al 2019) had done the preliminary analysis of wheat stem rot disease based on this GWAS panel. But upon consideration, we decided this sentence might not be appropriate here, so we removed it and added references (Yang, et.al 2019)
Yang, X.; Pan, Y.; Singh, P.K.; He, X.; Ren, Y.; Zhao, L.; Zhang, N.; Cheng, S.; Chen, F. Investigation and genome-wide association study for Fusarium crown rot resistance in Chinese common wheat. Bmc Plant Biol 2019, 19, 1-14, doi:10.1186/s12870-019-1758-2.
Comments 8:124-125: Quality check procedure should be detailed. If the reference details the procedure this should be stated, but some short description should still be in this paper so people do not have to search in another paper.
Response: Thank you for your suggestion. We provided a brief explanation of the quality inspection process as “The quality pretreatment of genotyping data was carried out for SNP call rate and MAF (minor allele frequency) with the PLINK software with threshold of maf 0.02 and geno 0.1. (http://zzz.bwh.harvard.edu/plink/tutorial.shtml) resulting in the retention of 395782 SNPs.”, and this change is reflected in line 146-149.
Comment 9 in 133: While the cutoff for significant values of -log10 p > 5 seems reasonable to me, I would recommend using a false discovery rate measure in the future, for example the q-values provided by R package qvalue (DOI: 10.18129/B9.bioc.qvalue). A change to using q-values in this paper would probably be an unreasonable amount of work, but I would ask the authors to provide the q-value cutoff that corresponds to the chosen p-value cutoff for their results. I.e. to calculate the q-values for the Tassel results and provide the highest q-value among the p-values that were deemed significant by the paper's methods in the text of the paper.
Response: We are grateful for the constructive feedback you provided on our work. We have calculated the q value, and revised the text as “We also used R package qvalue to calculate the q value corresponding to the P value (Table S2)” When the P value is greater than 5, the q value of RTRL is range from 0.0151 to 0.0351; the q value of RTRS is range from 8.59E-7 to 5.34; the q value of RARD is range from 0.234 to 0.558; the q value of RTRV is range from 0.00163 to 0.319; the q value of RNRT is range from 0.0146 to0.0208. Because suggestive genome-wide significance threshold as - log (P) are also widely used in the relative researches, such as (He et al., 2017; Shi et al., 2022; Guo et al., 2023; Zhao et al., 2023), etc. When we determined the threshold, we tried several thresholds with - log (P) of 3, 4, 5, and 6, and found that there were 5176 SNP markers when - log (P) was 3 and 2659 SNP markers when - log (P) was 4. The standard was too loose, but when - log (P) was 6, there were only 135 SNP markers, which was too strict, resulting in many traits such as RTRL having no available SNP markers. Therefore, we chose - log (P) of 5. In order to reduce false positives, this experiment also used. the combination of FDR test and threshold for screening results in more reliable outcomes. Due to the differences in genetic backgrounds between GWAS and RIL populations, the genetic backgrounds of GWAS panel used in GWAS analysis are generally more complex. Some significant loci in RIL populations may be masked in GWAS panel, and if the P-value set is too strict, the colocalization intervals cannot be found. In order to better search for co-localization intervals, a threshold of - log10 (P-value)>5 was used. To reduce false positives, we use the online software BioLadder (https://www.bioladder.cn/web/#/chart/58) Using the BH (Benjamini and Hochberg) method, adjust the P-value (corrected P-value) based on FDR, with - log10 (P-value)>5 and corrected P<0.05 as the cutoff criteria. We added these contents in line 159-165. The FDR results are presented in Figure S2. Please find the revised manuscript attached for your review.
Guo, J., Guo, J., Li, L., Bai, X., Huo, X., and Shi, W., et al. (2023). Combined linkage analysis and association mapping identifies genomic regions associated with yield-related and drought-tolerance traits in wheat (Triticum aestivum L.). Theoretical and Applied Genetics 136(12).
He, Y., Wu, D., Wei, D., Fu, Y., Cui, Y., and Dong, H., et al. (2017). GWAS, QTL mapping and gene expression analyses in Brassica napus reveal genetic control of branching morphogenesis. Scientific reports 7(1), 15971.
Shi, H., Chen, M., Gao, L., Wang, Y., Bai, Y., and Yan, H., et al. (2022). Genome-wide association study of agronomic traits related to nitrogen use efficiency in wheat. Theoretical and Applied Genetics 135(12), 4289-4302.
Zhao, Y., Yan, X., Zeng, Z., Zhao, D., Chen, P., and Wang, Y., et al. (2023). Integrated genome‐wide association study and QTL mapping reveals qSa‐3A associated with English grain aphid, Sitobion avenae (Fabricius) resistance in wheat. Pest Management Science.
Comments 10:134-136: the cited paper doesn't seem to support the choice of interval size
Response: Thank you for your thoughtful suggestions, which have been instrumental in refining our manuscript. We added the whole genome LD decay map (Figure S1) with a description as “Linkage disequilibrium (LD) among markers was calculated for the whole genomes in PLINK software.” in line 168-169. The LD analysis of the linkage disequilibrium average attenuation distance (LD) group shows that the r2 value sharply decreases when the physical distance increases. When the intercept of r2 decreases to half, the average r2 value of the population is about 10MB (Figure S1) (line 266-269)
Comments 11: 141: "the marker" invokes a one certain SNP in my view, but none is mentioned (as it sholdn't in the Materials & Methods chapter). The sentence should be corrected.
Response: Sorry for our unclear expression. We have revised it as “To combine the GWAS results from all traits, a uniform suggestive genome-wide significance threshold- log10 (P-value) >5 was used” in lines 159-160.
Comment 12: 144: The sentence is probably missing some vital part, it should be fixed.144: What did the screening and de-redundancy entail?
Response: Sorry for our careless. This genetic map was created by Zhao et al. we used their genetic map for QTL mapping work. We have revised the text as “According to Zhao's genetic linkage map data. A high-density genetic linkage map was created from the whole genomes of the RIL population and its two parents using a wheat DArT array.” (line173-175)
Comment 13: 161-162: The repetition of the word "relative" is unnecessary.
Response: Thank you for your suggestion. We removed the redundant “relative”. And we change the “relative values” to “relative ratios“.(190-191)
Comment 14:155-165: Dashes showing range (e.g. 11.39%-39.72%) were probably mistakenly retained during changes to the manuscript as minus signs (e.g. 11.39% (ARD) -39.72% (TRL)), which causes a lot of confusion. Reference to Table 2 would be appreciated as well.
Response: We are in agreement with your valuable advice. We changed the text as “ranged from 11.39% (ARD) to 39.72% (TRL) and ranged from 14.29% (ARD) to 48.98% (TRS), while under the LN condition, the variation coefficient of each character of the GWAS panel and the RIL population ranged from 26.36% (ARD) to 54.19% (NRT) and ranged from 16.39% (ARD) to 55.09% (TRS) under LN conditions. Comparing the relative values, it was found that the values of all RSA traits were greater than 1 except RARD (0.98) in the GWAS population. The variation coefficients of each trait in the GWAS panel and the RIL population ranged from 29.24% (RARD) to 52.26% (NRT) and ranged from 14.78% (RARD) to 46.62% (RNRT).”in line 187-193.
Comments 15: 199-200: There seem to be no "lowercase letters" indicating significant difference; perhaps this was forgotten from an earlier version of the figure after "ns", "*" and "**" started to be used to show significance of the difference. Also, I'm missing information what do the A-E subfigures display, I presume different traits.
Response: We are grateful for the constructive feedback you provided on our work. Firstly, we supplemented the information of images A-E in (Figure 3). Meanwhile, as your advice, we recalculated the difference among two parents and N treatments, and add the "lowercase letters" in the figure 3.
Comments 16: 238: More datailed explanation of the figure is needed, for example the y-axis, which I assume is -log10(p).
Response: Thank very much for your professional advice. We have modified and add the noted Figure 4,5, as well as other Tables and Figures including y-axis annotation.
Comments 17:274: According to Figure 1, it seems that reduced N supply lowered TRS in the RIL population
Response: Yes, the average TRS of LN is slightly lower than that of NN, which is also reflected in Table 2. Sorry for the mistake expression, we have corrected it as “LN supply could increase TRL, TRS and NRT in GWAS panel, and TRL in RIL population” in line 333-334 of the revised manuscript.
Comments 18: 279-280, 296-300: I don't think that merely a high variability of a trait proves high genetic control in a trait. I would at least need to see variability within some clonal populations and an appropriate statistical test.
Response: Thank you for your thoughtful suggestions, which have been instrumental in refining our manuscript. We calculated the heritability of 5 RSA traits, but not decribed it. In the revised manuscript, we added the heritability of 5 RSA traits as Supplementary Materials (Table S4) and modified the description in line (137-143 and 194-205).
|
Table S4 Heritability analysis of 5 RSA traits |
|||||
|
Population |
Trait |
Genotype (G) |
Environment (E) |
G×E |
Heritability (h2) |
|
GWAS panel |
TRL |
2160.1 |
4370.3 |
3073.9 |
0.689 |
|
TRS |
382.74 |
77.85 |
331.96 |
0.688 |
|
|
ARD |
0.01037 |
0.10784 |
0.00039 |
0.486 |
|
|
TRV |
0.21467 |
0.08362 |
0.2708 |
0.599 |
|
|
NRT |
928.8 |
901.7 |
2295.4 |
0.429 |
|
|
RIL population |
TRL |
1106 |
221.2 |
1847 |
0.539 |
|
TRS |
342.9 |
23.66 |
196.9 |
0.773 |
|
|
ARD |
0.08794 |
0.01319 |
0.06923 |
0.867 |
|
|
TRV |
0.0522 |
0.00598 |
0.04171 |
0.710 |
|
|
NRT |
348.944 |
131.39 |
719.71 |
0.483 |
|
Comments 19: 286-291: These sentences say the same things as the sentences on lines 281-285, probably an editing error.
Response: We are very sorry for our incorrect writing. We have removed duplicate sentences in line 337-339.
Comment 20: 295-296: Citation is needed.
Response: Thank you for your suggestion. We added the references of（Fan et.al 2018; Jin et.al 2021; Ren et.al 2017）
Fan, X.; Zhang, W.; Zhang, N.; Chen, M.; Zheng, S.; Zhao, C.; Han, J.; Liu, J.; Zhang, X.; Song, L. Identification of QTL regions for seedling root traits and their effect on nitrogen use efficiency in wheat (Triticum aestivum L.). Theor Appl Genet 2018, 131, 2677-2698,
Jin, Y.; Liu, J.; Liu, C.; Jia, D.; Liu, P.; Wang, Y. Genome-wide association study of nitrogen use efficiency related traits in common wheat (Triticum aestivum L.). Acta Agronomica Sinica 2021, 47, 394-404,.
Ren, Y.; Qian, Y.; Xu, Y.; Zou, C.; Liu, D.; Zhao, X.; Zhang, A.; Tong, Y. Characterization of QTLs for root traits of wheat grown under different nitrogen and phosphorus supply levels. Front Plant Sci 2017, 8, 2096.
Comments 21: Here are lines of few of the biggest issues I've noted: 93, 107, 136-137, 141-142, 306-308, 355-356. There were also misplaced dashes, probably stemming from line shifts during formatting.
Response: Thank you for your suggestion. The misplaced dashes are stemming from line shifts during formatting. We have deleted the dashes“-“.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
Comments and Suggestions for AuthorsThe authors have improved the revised version of the manuscript and addressed the concerns raised earlier. Regarding the usage of subjective LOD threshold to declare QTL significance, authors should be aware that QTL signals that disappear after applying permutation test lack confidence and may never be useful in breeding.
Reviewer 4 Report
Comments and Suggestions for AuthorsI am content with the changes to the manuscript and have not found anything specific to note. Therefore, I recommend to accept the paper.
