Development of High-Resolution Simple Sequence Repeat Markers through Expression Profiling of Genes Associated with Pod Maturity of Soybean

Round 1

Reviewer 1 Report

The authors presented a paper describing the use of an efficient approach for designing SSR markers that detect genes associated with specific traits and improved accuracy in soybean.

This paper results clear and exhaustive. The introduction is exhaustive. Material and methods are clear and complete in all aspects, and the same appear the results, that are represented and described in a comprehensive way. The authors described and discussed well their results.

Just few changes are needed:

Introduction

Lines 32 to 34: Please add a reference;

Lines 46 to 48: Please add a reference;

Lines 53 to 57: I suggest adding more recent references to support better this section;

Results:

Lines 170: What the authors mean with “[…] were regulated in pods of more than one ecotype” (same in lines 230-231) Please clarify this aspect;

Line 194: Tm conventionally is the Melting Temperature and not Annealing;

Lines 201 to 211: Format mistake, please edit this part following the requested font.

Author Response

Response to Comments and Suggestions of Reviewer1

The authors presented a paper describing the use of an efficient approach for designing SSR markers that detect genes associated with specific traits and improved accuracy in soybean.

This paper results clear and exhaustive. The introduction is exhaustive. Material and methods are clear and complete in all aspects, and the same appear the results, that are represented and described in a comprehensive way. The authors described and discussed well their results.

Just few changes are needed:

Introduction

Lines 32 to 34: Please add a reference;

• We have cited two references [2,3]

2. Watanabe, S.; Harada, K.; Abe, J. Genetic and molecular bases of photoperiod responses of flowering in soybean. Breed Sci. 2012, 61, 531–543; DOI: 10.1270/jsbbs.61.531.
3. Cao, D.; Takeshima, R.; Zhao, C.; Liu, B.; Abe, J.; Kong, F. Molecular mechanisms of flowering under long days and stem growth habit in soybean. J. Exp. Bot. 2017, 68, 1873–1884; DOI: 10.1093/jxb/erw394.

 

Lines 46 to 48: Please add a reference;

• We have cited two references [9,10]

9. Sun, F.; Xu, M.; Park, C.; Dwiyanti, M.S.; Nagano, A.J.; Zhu, J.; Watanabe, S.; Kong, F.; Liu, B.; Yamada, T.; Abe, J. Characterization and quantitative trait locus mapping of late-flowering from a Thai soybean cultivar introduced into a photoperiod-insensitive genetic background. PLoS One 2019, 14, e0226116; DOI: 10.1371/journal.pone.0226116.
10. Jia, H.; Jiang, B.; Wu, C.; Lu, W.; Hou, W.; Sun, S.; Yan, H.; Han, T. Maturity Group Classification and Maturity Locus Genotyping of Early-Maturing Soybean Varieties from High-Latitude Cold Regions. PLoS One 2014, 9, e94139; DOI: 10.1371/journal.pone.0094139.

 

Lines 53 to 57: I suggest adding more recent references to support better this section;

• The references included here are the most recent available (Gao et al. 2012; Hu et al. 2014; Li et al. 2017).

 

Results:

Lines 170: What the authors mean with “[…] were regulated in pods of more than one ecotype” (same in lines 230-231) Please clarify this aspect;

• We have defined the up (relative expression ≥ 2.0) and downregulation (≤ 0.5) in the last paragraph (page 8, line 182-183) 
• “The six genes were either upregulated (relative expression ≥ 2.0) or downregulated (≤ 0.5) in pods of more than one ecotype (Table 3).”

Line 194: Tm conventionally is the Melting Temperature and not Annealing;

• We revised ‘Tm’ to ‘Ta’.

Lines 201 to 211: Format mistake, please edit this part following the requested font.

• We edited the section using the full alignment format.

Author Response File: Author Response.docx

Reviewer 2 Report

The paper on development of High-Resolution Simple Sequence Repeat Markers associated with pod maturity of soybean is a nice piece of research, with an appealing set-up. I have only minor comments and suggestions for improvement, expecially for the matherials and the discussion, that surely need to be reconsidered.

Here you can find my comments on a detailed list of points that you should improve (your text is in italic, my comments are in bold):

 

 

[53-56] However, some previously developed SSR markers were often found to be less polymorphic and failed to produce the expected PCR products. Further, these molecular markers can differ in their stability, quality of the amplification products, and equality of the amplification loci, and provide less information regarding the gene of interest.

I would say this is a weak paragraph... "some previously developed SSR.." suggests a citation. Also the last three lines should be justified with a citation.

 

[57-60] Macaulay et al. [12] developed a highly informative genotyping set of SSRs for genetic studies of barley that was not limited by these drawbacks. We conducted this study to develop SSR markers with high resolution as well as improved efficiency, precision, and reliability for the genetic study of soybean pod maturity trait.

I cannot see the link between the first sentence and the ones before, or the last one, in which you state the aim of the research

 

[63-68] We conducted experimental field tests at Suwon (37°16′N, 127°01′E), Republic of Korea in 2017 and 2018 to investigate DTM of the 100 soybean varieties (Table 1). The experimental plots were completely randomized. The soybean seeds were sown on Jun 12, 2017 and Jun 14, 2018 in a 70 cm by 15 cm spacing in plots of the experimental fields that were treated with a basic granular fertilizer N-P2O5-K2O = 30-30-34 kg/ha before sowing. DTM of the soybeans was defined as the number of days from sowing to full maturity (95% of the pods reached a mature pod color; stage R8).

Why only two years (and not three, making possible to statistically "remove" environment for the observed phenotype)? Were there replicates in the fields that provide statistical robustness?

 

[69-70] Table 1. Classification of 100 soybean varieties by maturity groups and ecotypes according to their day to maturity.

Linked to the previous comment: in this table, you reported the 100 varieties grouped for DTM. Having data from two enviroments, and probably replicates among them, I would expect to see statistics about the varietiesì behaviour under the two seasons. DTM column is in absolute number, is it the mean of the two years? If so, do you think this is a solid parameter?

 

[86] GRAMENE’s SSR tools [20]

Consider citing Temnykh et al. (2001), as requested by the developers...

 

[86 – 87] We analyzed the SSR in gene sequences using the GRAMENE’s SSR tools [20] and then selected genes with SSR regions.

With which parameters? Did you look for perfect, imperfect, compound... if perfect, min. parameters… or you just pick all the genetic ssr?

 

[87] We manually designed markers to amplify specific SSR regions.

How? Which parameters did you take into consideration? (i.e. min,max,optimal GC%, min,max,optimal Tm….)

 

[115-117] RT-qPCR was performed using the QuantStudio 5 Real-Time PCR System (Thermo Fisher Scientific) and the QuantiNova SYBR Green RT-PCR Kit (Qiagen) according to the specific annealing temperature for each primer pair (Table S2).

Did you design the primers? If yes, how?

Which PCR parameters? How many cycles?

 

[117-120] For relative expression profiling, we normalized the quantification cycle (Cq) values of each gene using TUB4 (Glyma.03G124400.1) and then calculated relative expression ratio of each gene based on UKN2 (Glyma.06G038500.1) expression.

How did you calculate the relative expression? (i.e. using GeNorm, Vandesompele J, De Preter K, Pattyn F, Poppe B, Van Roy N, De Paepe A, et al. Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes. Genome Biol. 2002; pmid:12184808)

 

Why did you use UKN2? (Gao, M., Liu, Y., Ma, X., Shuai, Q., Gai, J., & Li, Y. (2017). Evaluation of reference genes for normalization of gene expression using quantitative RT-PCR under aluminum, cadmium, and heat stresses in soybean. PloS one, 12(1), e0168965)

 

[134-147] Analysis of variance (ANOVA) was analyzed using the R version 3.6.1, based on the package agricolae using three biological replicates and three technical replicates. Multiple comparisons on expression profiling of genes were performed by Duncan`s multiple range test (P < 0.05) […] The DTM of the 100 varieties varied extensively, from 79 to 117 days. The shortest DTM among the varieties was that of OT89-06 (79 days), while the longest was that of ‘Pungsannamul’ (117 days). We classified the 100 varieties into 8 MGs (0~VII) according to their DTM, with a 5-day range for each MG. The groups were: group 0, ≤ 79 DTM; group I, 80–84 DTM; group II, 85–89 DTM; group III, 90–94 DTM; group IV, 95–99 DTM; group V, 100–104 DTM; group VI, 105–109 DTM; and group VII, ≥ 110 DTM (Table 1). All the MGs were further classified into 3 maturity ecotypes: early (EM), middle (MM), and late maturity (LM). The EM ecotype consisted of MG 0 to II (30 varieties); the MM ecotype consisted of MG III and IV (36 varieties); and the LM ecotype comprised MG V to VII (34 varieties;

Did you perform analysis on the field trial data for DTM? If so, they should be present somewhere…

 

[150-151] We analyzed the gene annotations and identified 1,147 genes that were related to growth

Did you make differences between mono-, di-, tri-, tetra-, penta- and hexa-? Trinucleotides are for sure the most present in coding regions (Morgante et al., 2002; Subramanian et al., 2003; Portis et al., 2018)

 

[160-161] Analysis of the SSR regions in the 1,147 genes revealed 211 genes that harbored over 5 repeats of SSR motifs.

How? Based on GO annotation?

 

[165-167] Figure 2. PCR amplification products of three different soybean maturity ecotype varieties using 42 SSR markers. E, early maturity ecotype OT89-05; M, middle maturity ecotype IT213198; L, late maturity ecotype IT213177.

Are the markers scored on a 1% agarose gel? There are quite some differences between the alleles... Could you summarize this differences in product size somewhere/somehow?

 

[194 – 199] The number of alleles amplified by the selected markers varied from 1 to 5, with each showing a PIC value > 0.5 (Table 5). However, the SSR markers identified in this study exhibited lower PIC values than those of the previously established markers [26–31] (Table 5), although the sizes of the alleles amplified by the previous and new SSR markers were similar (Table 5).

It could be explained by the fact that the SSR identified in the two references were not focused only in genic regions. SSR in coding regions have been reported to be less polymorphic than the one in non-coding ones (Portis et al., 2016)… Furthermore, it is worth noting that “they may not provide an adequate level of polymorphism to distinguish between closely related ecotypes/varieties.” (Portis et al., 2018).

Consider speding few sentences, somewhere, evaluating the reasons why you observed lower polymorphism...

 

[202-211] The tree visualized for the previously established 6 SSR markers had 3 large agglomerative clusters, with the clusters A-I, A-II and A-III comprising 45, 31, and 24 varieties, respectively. Although the EM ecotype in cluster A-I and the MM ecotype in clusters A-II and A-III were the largest constituents, the distribution of these ecotypes was under 50% in each cluster (Fig. 5A). Another tree constructed based on the 6 newly developed SSR markers also had 3 clusters: B-I, B-II, and B-III. Clusters B-I, B-II, and B-III included 31, 34, and 35 varieties, respectively. In cluster B-I, 18 (58.1%) varieties belonged to the EM ecotype, while 70.6% of the varieties in the cluster B-II and 71.4% of the varieties in the cluster B-III belonged to the MM and LM ecotypes, respectively (Fig. 5B).

Consider rephrasing, making everything more clear and readable...

One question: are the six markers from literature identified within genic regions? Table 2 reports only the associated QTL and the reference.

 

[218 – 171] Discussion

Kind of a "wall of text"... this makes the discussion a little bit too heavy to read. I would consider splitting the chapter in sub-chapters and paragraphs.

Consider also rephrasing paying attention to verb tenses etc...

Author Response

Response to Comments and Suggestions of Reviewer 2

The paper on development of High-Resolution Simple Sequence Repeat Markers associated with pod maturity of soybean is a nice piece of research, with an appealing set-up. I have only minor comments and suggestions for improvement, expecially for the matherials and the discussion, that surely need to be reconsidered.

Here you can find my comments on a detailed list of points that you should improve (your text is in italic, my comments are in bold):

 

 

[53-56] However, some previously developed SSR markers were often found to be less polymorphic and failed to produce the expected PCR products. Further, these molecular markers can differ in their stability, quality of the amplification products, and equality of the amplification loci, and provide less information regarding the gene of interest.

I would say this is a weak paragraph... "some previously developed SSR.." suggests a citation. Also, the last three lines should be justified with a citation.

• We edited the text accordingly as ([9–11]à[13–15]), and have cited a new reference ([16]) as follows: 
• 16. Miah, G.; Rafii, M.Y.; Ismail, M.R.; Puteh, A.B.; Rahim, H.A.; Islam, K.; Latif, M.A. A review of microsatellite markers and their applications in rice breeding programs to improve blast disease resistance. Int. J. Mol. Sci., 2013, 14, 22499-22528; DOI:10.3390/ijms141122499.
• “However, certain previously developed SSR markers were often found to be less polymorphic and failed to produce the expected PCR products [13–15]. Further, these molecular markers can differ in their stability, amplification product quality, and equality of the amplification loci, while providing less information regarding the gene of interest [16]. These disadvantages of SSR markers limit their use and accuracy for genotyping or breeding programs [16].” (page 2, lines 53-57)

 

[57-60] Macaulay et al. [12] developed a highly informative genotyping set of SSRs for genetic studies of barley that was not limited by these drawbacks. We conducted this study to develop SSR markers with high resolution as well as improved efficiency, precision, and reliability for the genetic study of soybean pod maturity trait.

I cannot see the link between the first sentence and the ones before, or the last one, in which you state the aim of the research

• We have edited the text as follows: “These disadvantages of SSR markers limit their use and accuracy for genotyping or breeding programs [16]. Meanwhile, Macaulay et al. [17] developed a highly informative genotyping set of SSRs for genetic studies of barley that was not limited by these drawbacks. Therefore, we conducted this study to similarly develop SSR markers with high-resolution, as well as improved efficiency, precision, and reliability for the genetic study of soybean pod maturity trait.” (page 2, lines 56-61)

 

[63-68] We conducted experimental field tests at Suwon (37°16′N, 127°01′E), Republic of Korea in 2017 and 2018 to investigate DTM of the 100 soybean varieties (Table 1). The experimental plots were completely randomized. The soybean seeds were sown on Jun 12, 2017 and Jun 14, 2018 in a 70 cm by 15 cm spacing in plots of the experimental fields that were treated with a basic granular fertilizer N-P2O5-K2O = 30-30-34 kg/ha before sowing. DTM of the soybeans was defined as the number of days from sowing to full maturity (95% of the pods reached a mature pod color; stage R8).

Why only two years (and not three, making possible to statistically "remove" environment for the observed phenotype)? Were there replicates in the fields that provide statistical robustness?

• The days to maturity (DTM) is a distinct trait of soybean, which is primarily controlled by day length. Hence, two years is sufficient to analyze DTM of soybean in a specific region as the day length of a specific region dose not significantly change.
• Ha et al. [7] described a maturity grouping system (MG 0 to MG VII) for 100 Korean soybean cultivars, according to days to maturity (DTM) based on data obtained in 2005-2006 at Miryang, Korea.
• Lee et al. [8] used the maturity ecotype based on DTM to evaluate the effects of growth period and cumulative temperature on flowering, maturity, and yield of Korean soybean varieties at Suwon in 2018.

[69-70] Table 1. Classification of 100 soybean varieties by maturity groups and ecotypes according to their day to maturity.

Linked to the previous comment: in this table, you reported the 100 varieties grouped for DTM. Having data from two environments, and probably replicates among them, I would expect to see statistics about the varieties behaviour under the two seasons. DTM column is in absolute number, is it the mean of the two years? If so, do you think this is a solid parameter?

• As stated above, the days to maturity (DTM) is a distinct trait of soybean, which is primarily controlled by day length. Hence, two years is sufficient to analyze DTM of soybean in a specific region as the day length of a specific region dose not significantly change.
• The value in the DTM column represents the mean of two years.

[86] GRAMENE’s SSR tools [20]

Consider citing Temnykh et al. (2001), as requested by the developers...

• We have cited the reference (Temnykh et al. (2001)).

25. Temnykh, S.; DeClerck, G.; Lukashova, A.; Lipovich, L.; Cartinhour, S.; McCouch, S. Computational and experimental analysis of microsatellites in rice (Oryza sativa L.): frequency, length variation, transposon associations, and genetic marker potential. Genome Res. 2001, 11, 1441–1452; DOI: 10.1101/gr.184001.

 

[86 – 87] We analyzed the SSR in gene sequences using the GRAMENE’s SSR tools [20] and then selected genes with SSR regions.

With which parameters? Did you look for perfect, imperfect, compound... if perfect, min. parameters… or you just pick all the genetic ssr?

• We edited the text as follows: “We analyzed the SSR in gene sequences using the GRAMENE’s SSR tools [25] and then selected genes that harbored more than five repeats of SSR motifs.” (page 3, lines 88-89)

 

[87] We manually designed markers to amplify specific SSR regions.

How? Which parameters did you take into consideration? (i.e. min,max,optimal GC%, min,max,optimal Tm….)

• “The markers were manually designed to amplify specific SSR regions, using the following parameters: primer size from 18 to 28 with 23 as the optimum; product size from 80 to 500 bp; annealing temperature from 48 to 60 °C with an optimum of 56 °C; GC content from 45% to 55%, with 50% as optimum.” (page 3, lines 88-91)

 

[115-117] RT-qPCR was performed using the QuantStudio 5 Real-Time PCR System (Thermo Fisher Scientific) and the QuantiNova SYBR Green RT-PCR Kit (Qiagen) according to the specific annealing temperature for each primer pair (Table S2).

Did you design the primers? If yes, how?

• “We also manually designed primers, to profile expression of the genes, using the following parameters: primer size from 20 to 25 with 23 as the optimum; product size from 80 to 300 bp; annealing temperature from 52 to 58 °C with an optimum of 56 °C; GC content from 45% to 55%, with 50% as optimum.” (page 4, lines 115-118)

 

Which PCR parameters? How many cycles?

• “RT-qPCR was performed after 2 min at 95 °C for polymerase activation, and cycling between 95 °C (15 s) and the respective annealing temperature (15 s) for each primer pair (Table S2) followed by 40 cycles using the QuantStudio 5 Real-Time PCR System (Thermo Fisher Scientific) and the QuantiNova SYBR Green RT-PCR kit (Qiagen).” (page 5-6, lines 121-124)

 

 [117-120] For relative expression profiling, we normalized the quantification cycle (Cq) values of each gene using TUB4 (Glyma.03G124400.1) and then calculated relative expression ratio of each gene based on UKN2 (Glyma.06G038500.1) expression.

How did you calculate the relative expression? (i.e. using GeNorm, Vandesompele J, De Preter K, Pattyn F, Poppe B, Van Roy N, De Paepe A, et al. Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes. Genome Biol. 2002; pmid:12184808)

 

Why did you use UKN2? (Gao, M., Liu, Y., Ma, X., Shuai, Q., Gai, J., & Li, Y. (2017). Evaluation of reference genes for normalization of gene expression using quantitative RT-PCR under aluminum, cadmium, and heat stresses in soybean. PloS one, 12(1), e0168965)

 

• We did not use a GeNorm method for normalization but rather used two reference genes to improve the accuracy of normalization for expression profiling.
• Hu et al. (2009) validated that for six different cultivars in long day (LD) and short day (SD), their expression stability did not vary significantly with ACT11, UKN2 and TUB4 being the most stable genes.
• Among the genes, TUB4 (Glyma.03G124400.1) was the most highly expressed, with the associated Cq value in leaf, stem, and pod of three different ecotypes stable in the range of 20.24 to 21.45.
• Moreover, Li et al (2012) concluded that the reference genes identified as optimal for seed development were TUA5 and UKN2.
• Gao et al. (2017) also concluded that UKN2 is the best reference gene for analysis of roots and leaves together.

• We found out that Glyma.06G038500.1 (UKN2), for which the Cq values were 23.34 ± 0.17, 24.28 ± 0.23 and 23.87 ± 0.12 in leaf, stem, and pod of three different ecotypes, respectively, was the most stable when we analyzed expression stability of three UKN2s.“Two reference genes were also used, TUB4 (Glyma.03G124400.1), and UKN2 (Glyma.06G038500.1), to improve accuracy of normalization for expression profiling. For relative expression profiling, we modified the formula reported by Livak and Schmittgen (Livak and Schmittgen 2001): normalization (ΔCq) of the quantification cycle (Cq) values of each gene using TUB4, ΔCq = Cq (a target sample) − Cq (TUB4); ΔΔCq calculation based on ΔCq of UKN2, ΔΔCq = ΔCq (a target sample) − ΔCq (UKN2); relative expression ratio (R), R = (2^-ΔΔCq of a target sample)/(mean of 2^-ΔΔCq of all sample).” (page 5, lines 125-131)References 

• Accordingly, we have revised the Materials and Methods (2.4 Profiling of gene expression chapter) as follows:

1. Hu, R., Fan, C., Li, H., Zhang, Q., & Fu, Y. F. Evaluation of putative reference genes for gene expression normalization in soybean by quantitative real-time RT-PCR. BMC molecular biology, 2009, 10, 93; DOI: 10.1186/1471-2199-10-93.
2. Li, Q., Fan, C., Zhang, X. et al. Validation of reference genes for real-time quantitative PCR normalization in soybean developmental and germinating seeds. Plant Cell Rep 2012, 31, 1789–1798; DOI: 10.1007/s00299-012-1282-4.
3. Gao M, Liu Y, Ma X, Shuai Q, Gai J,et al. Evaluation of Reference Genes for Normalization of Gene Expression Using Quantitative RT-PCR under Aluminum, Cadmium, and Heat Stresses in Soybean. PLOS ONE 2017, 12(1): e0168965; DOI: 10.1371/journal.pone.0168965.
4. Livak, K.J.; Schmittgen, T.D. Analysis of relative gene expression data using realtime quantitative PCR and the 2^ΔΔC(T) Method. Methods, 2001, 25(4), 402–408; DOI: 10.1006/meth.2001.1262.

 

[134-147] Analysis of variance (ANOVA) was analyzed using the R version 3.6.1, based on the package agricolae using three biological replicates and three technical replicates. Multiple comparisons on expression profiling of genes were performed by Duncan`s multiple range test (P < 0.05) […] The DTM of the 100 varieties varied extensively, from 79 to 117 days. The shortest DTM among the varieties was that of OT89-06 (79 days), while the longest was that of ‘Pungsannamul’ (117 days). We classified the 100 varieties into 8 MGs (0~VII) according to their DTM, with a 5-day range for each MG. The groups were: group 0, ≤ 79 DTM; group I, 80–84 DTM; group II, 85–89 DTM; group III, 90–94 DTM; group IV, 95–99 DTM; group V, 100–104 DTM; group VI, 105–109 DTM; and group VII, ≥ 110 DTM (Table 1). All the MGs were further classified into 3 maturity ecotypes: early (EM), middle (MM), and late maturity (LM). The EM ecotype consisted of MG 0 to II (30 varieties); the MM ecotype consisted of MG III and IV (36 varieties); and the LM ecotype comprised MG V to VII (34 varieties;

Did you perform analysis on the field trial data for DTM? If so, they should be present somewhere…

• We had presented this information in the original manuscript on Line 63-64: “We conducted experimental field tests at Suwon (37°16′N, 127°01′E), Republic of Korea in 2017 and 2018 to investigate DTM of the 100 soybean varieties (Table 1).”

 

[150-151] We analyzed the gene annotations and identified 1,147 genes that were related to growth

Did you make differences between mono-, di-, tri-, tetra-, penta- and hexa-? Trinucleotides are for sure the most present in coding regions (Morgante et al., 2002; Subramanian et al., 2003; Portis et al., 2018)

• We only used the open annotation information for the genes.
• We supplied characteristics of the SSR markers, including SSR pattern and repeat number of SSR, in the Supplementary Table S1.
• We have also added information regarding the SSR motifs in the revised Discussion as follows:  
• “As shown in supplementary Table S1, we developed 211 genic SSR markers. Among the 211 SSR repeat motifs, tri-nucleotide repeats (TNRs) were found to be the most abundant, accounting for 75.4% (159), followed by 16.6% of di-nucleotide repeats (DNRs), and 8.0% tetra-nucleotide repeats (TtNRs). Meanwhile, longer than penta-nucleotide repeats were not found. These results were similar to previously published data with TNRs reported as the most frequent (54–78%), followed by DNRs (17.1–40.4%), and TTNRs (3–6%) among cereal species [50].” (page 12-13, lines 239-244)

[160-161] Analysis of the SSR regions in the 1,147 genes revealed 211 genes that harbored over 5 repeats of SSR motifs.

How? Based on GO annotation?

• We analyzed the SSR regions in the full sequence of each gene transcript using the SSR tool.

[165-167] Figure 2. PCR amplification products of three different soybean maturity ecotype varieties using 42 SSR markers. E, early maturity ecotype OT89-05; M, middle maturity ecotype IT213198; L, late maturity ecotype IT213177.

Are the markers scored on a 1% agarose gel? There are quite some differences between the alleles... Could you summarize this differences in product size somewhere/somehow?

• When we selected SSR markers with over two alleles on the three different ecotypes, we used a 1% agarose gel. Although we used a capillary electrophoresis equipment QIAxcel Advanced system (Qiagen) when we scored the alleles, which were formed by each SSR marker for agglomerative clustering.

[194 – 199] The number of alleles amplified by the selected markers varied from 1 to 5, with each showing a PIC value > 0.5 (Table 5). However, the SSR markers identified in this study exhibited lower PIC values than those of the previously established markers [26–31] (Table 5), although the sizes of the alleles amplified by the previous and new SSR markers were similar (Table 5).

It could be explained by the fact that the SSR identified in the two references were not focused only in genic regions. SSR in coding regions have been reported to be less polymorphic than the one in non-coding ones (Portis et al., 2016). Furthermore, it is worth noting that “they may not provide an adequate level of polymorphism to distinguish between closely related ecotypes/varieties.” (Portis et al., 2018).

Consider speding few sentences, somewhere, evaluating the reasons why you observed lower polymorphism...

• We have discussed the lower level of polymorphisms in the revised Discussion as follows:
• “Genic SSR markers have been reported to not only be less polymorphic than non-genic markers [67], but to also have 3.5 alleles per locus, with a higher PIC of 0.824 among the mulberry species [68]. Similarly, previous studies reported that the genic SSR markers for jute (2.7 alleles and a PIC of 0.34) [69] and flax (2.3 alleles and PIC of 0.35) [70] show high polymorphism and are expected to be of use for the characterization of germplasm, as well as for variety identification, and marker assisted breeding. Therefore, although the average number of alleles and PIC values of the six newly developed genic SSR markers were low (3.1 alleles and a PIC of 0.61) compared to those of the previously established six SSR markers (3.8 alleles and PIC of 0.70), the allelic variations exhibited by the new SSR markers were sufficient to generate a clustering and distinguish the soybean varieties by ecotype group. Finally, these alleles can serve as a marker for characterization of populations, varieties, and germplasm, as previously reported in wheat [71], jatropha [72], and oil palm [73].” (page 14, lines 282-292)

 

[202-211] The tree visualized for the previously established 6 SSR markers had 3 large agglomerative clusters, with the clusters A-I, A-II and A-III comprising 45, 31, and 24 varieties, respectively. Although the EM ecotype in cluster A-I and the MM ecotype in clusters A-II and A-III were the largest constituents, the distribution of these ecotypes was under 50% in each cluster (Fig. 5A). Another tree constructed based on the 6 newly developed SSR markers also had 3 clusters: B-I, B-II, and B-III. Clusters B-I, B-II, and B-III included 31, 34, and 35 varieties, respectively. In cluster B-I, 18 (58.1%) varieties belonged to the EM ecotype, while 70.6% of the varieties in the cluster B-II and 71.4% of the varieties in the cluster B-III belonged to the MM and LM ecotypes, respectively (Fig. 5B).

Consider rephrasing, making everything more clear and readable...

One question: are the six markers from literature identified within genic regions? Table 2 reports only the associated QTL and the reference.

• We have revised the text as follows:
• “The tree generated for the previously established six SSR markers [26–31], had three large agglomerative clusters, with clusters A-I, A-II and A-III comprising 45, 31, and 24 varieties, respectively. Although the EM ecotype in cluster A-I, and the MM ecotype in clusters A-II and A-III, were the largest ecotypes, their distribution was under 50% in each cluster (Fig. 5A). Another tree that was constructed based on the six newly developed SSR markers, also had three clusters: B-I, B-II, and B-III, which included 31, 34, and 35 varieties, respectively. In cluster B-I, 18 (58.1%) varieties belonged to the EM ecotype, while 24 (70.6%) varieties in cluster B-II, and 25 (71.4%) varieties in cluster B-III belonged to the MM and LM ecotypes, respectively (Fig. 5B).” (page 11, lines 215-222)
• The six markers in Table 2 are not genic SSR markers. These markers have been applied for QTL mapping related to pod maturity of soybean in previous studies [31-36].

[218 – 171] Discussion

Kind of a "wall of text"... this makes the discussion a little bit too heavy to read. I would consider splitting the chapter in sub-chapters and paragraphs.

Consider also rephrasing paying attention to verb tenses etc...

• We have revised the Discussion accordingly and have divided it into four subsections to improve the overall readability.

Author Response File: Author Response.docx

Reviewer 3 Report

The paper “Development of High-Resolution Simple Sequence Repeat Markers through Expression Profiling of Genes Associated with Pod Maturity of Soybean” aims at developing some SSR markers as a high-resolution tool for genotyping/QTL mapping of soybean pod maturity. Such markers would be valuable to categorize soybean accessions into maturity groups, better suited to be grown in different regions. The authors take advantage of existing genetic resources to obtain gene sequences, annotated as having a potential role in pod maturity date. They then filter them for the presence of  SSR motifs, and validate SSR markers lying in 211 of them, detecting polymorphism in 42. They then profile the expression of these polymorphic genes, finding that 6 of them show some differential expression in pods of different ecotypes (categorized based on early, mid and late pod maturity date). Agglomerative clustering trees appear to suggest that these newly developed SSR marker are better at predicting the three ecotypes than a set of 6 makers found in the literature.

The paper contains a lot of work, both for SSR marker development and gene expression profiling, but in my opinion there are some issues in its conceptualization. The article is written in a similar way to projects involving QTL mapping or GWAS, where tight genotype/phenotype association pinpoints possible candidate genes responsible for a character. In this paper instead the list of genes of interest was made by functional prediction and further reduced to those containing probable SSR in their sequence.

The six most interesting genes were then chosen according to differences in expression in pod tissues. I am of the opinion that there is no real evidence of any involvement of these genes in pod maturity and different levels of expression in pod between ecotypes could easily be due to the action of trans-acting factors encoded by genes lying on totally unrelated portions of the genome.

The clustering, although suggestive of a possible association, does not convincingly prove, in my opinion, that the six markers show  “strong relationship” with each maturity ecotype, as stated in the discussion.

The paper has the merit of identifying around forty new, inexpensive soybean SSR markers, but their usefulness can only be determined in association studies.

Small general comments:

Line 86: ref 20 should be “Temnykh et al. 2001 Genome Res. 11(8):1441-1452” rather than the Graeme SSRtool link

Line 106: I cannot find a description of the PIC formula in the reference given - Wang et al. [21]

Table 3 is rather large and of little utility for the paper itself, its information being only used in table 4. I feel Table 3 could easily be moved to Supplementary information

Author Response

Response to Comments and Suggestions of Reviewer 3

The paper “Development of High-Resolution Simple Sequence Repeat Markers through Expression Profiling of Genes Associated with Pod Maturity of Soybean” aims at developing some SSR markers as a high-resolution tool for genotyping/QTL mapping of soybean pod maturity. Such markers would be valuable to categorize soybean accessions into maturity groups, better suited to be grown in different regions. The authors take advantage of existing genetic resources to obtain gene sequences, annotated as having a potential role in pod maturity date. They then filter them for the presence of SSR motifs, and validate SSR markers lying in 211 of them, detecting polymorphism in 42. They then profile the expression of these polymorphic genes, finding that 6 of them show some differential expression in pods of different ecotypes (categorized based on early, mid and late pod maturity date). Agglomerative clustering trees appear to suggest that these newly developed SSR markers are better at predicting the three ecotypes than a set of 6 makers found in the literature.

The paper contains a lot of work, both for SSR marker development and gene expression profiling, but in my opinions there are some issues in its conceptualization. The article is written in a similar way to projects involving QTL mapping or GWAS, where tight genotype/phenotype association pinpoints possible candidate genes responsible for a character. In this paper instead the list of genes of interest was made by functional prediction and further reduced to those containing probable SSR in their sequence.

The six most interesting genes were then chosen according to differences in expression in pod tissues. I am of the opinion that there is no real evidence of any involvement of these genes in pod maturity and different levels of expression in pod between ecotypes could easily be due to the action of trans-acting factors encoded by genes lying on totally unrelated portions of the genome.

• Thank you for these comments. We believe that the gene expression profiling results also serve as evidence to verify that the genes were related to growth and development of pod maturity.
• Although we did not specifically define that the genes control growth and development of pod maturity, we putatively concluded as follows: 
• “These results imply that all six genes are related to growth and development of soybean pods. Therefore, we predicted that the newly developed SSR markers derived from the six genes can serve for high-resolution detection and characterization of genetic diversity based on the DTM of soybean varieties.” (page 13, lines 268-271)

The clustering, although suggestive of a possible association, does not convincingly prove, in my opinion, that the six markers show “strong relationship” with each maturity ecotype, as stated in the discussion.

• Although the clustering was not a definite prove for that the six markers show strong relationship with each maturity ecotype, it is reasonable that we should use strong, moderate or weak based on the statistical results when describing relationships between variables.
• This paper has the merit of identifying around forty new, inexpensive soybean SSR markers, but their usefulness can only be determined in association studies.

 

Small general comments:

Line 86: ref 20 should be “Temnykh et al. 2001 Genome Res. 11(8):1441-1452” rather than the Graeme SSRtool link

• We cited the reference (Temnykh et al. (2001)) as follows:

25. Temnykh, S.; DeClerck, G.; Lukashova, A.; Lipovich, L.; Cartinhour, S.; McCouch, S. Computational and experimental analysis of microsatellites in rice (Oryza sativa L.): frequency, length variation, transposon associations, and genetic marker potential. Genome Res. 2001, 11, 1441–1452; DOI: 10.1101/gr.184001.

 

Line 106: I cannot find a description of the PIC formula in the reference given - Wang et al. [21]

• The reference was revised to a more appropriate one as follows: 
• 26. Botstein, D.; White, R.L.; Skolnick, M.; Davis, R.W. Construction of a genetic linkage map in man using restriction fragment length polymorphisms. Am. J. Hum. Genet. 1980, 32, 314–331.

Table 3 is rather large and of little utility for the paper itself, its information being only used in table 4. I feel Table 3 could easily be moved to Supplementary information

• It is required to be in main text because Table 3 is important for understanding how to classify the genes related to pod maturity.

 

Author Response File: Author Response.docx

Reviewer 4 Report

The manuscript entitled with “Development of High-Resolution Simple Sequence Repeat Markers through Expression Profiling of Genes Associated with Pod Maturity of Soybean” by Park et al. (Manuscript ID: applsci-894790) described on the identification of high-resolution gene based markers for pod maturity in soybean. In general, the experiments and manuscript were well conceived and conducted. The results compared with appropriate controls and important finding is presented in results and figures. Appropriate literature cited and discussed all relevant results. This is the start of a useful contribution. The technique is translatable to other systems and this increases the value of the work. However, a few concerns have to be addressed before publication.

Therefore, I recommend this manuscript for the publication in Applied Sciences journal after incorporating some corrections suggested below.

 

Major comments

1. I think the Discussion section seems to be a little bit of short and low information. Authors have described only identified 6 genes and their expression. The manuscript contains valuable information for marker development and new constructed agglomerative clusters. Authors have to explain that more detail in Discussion section.

 

2. Figure 3 was shown the expression profiling of selected genes for maturity. I think authors might compare qRT-PCR expressions with leaf, stem and pod to identify pod maturity specific genes. However, authors have to distinguish plant sample strictly to compare the maturity in same tissue. The maturity level should be different according to tissue position. For example, the maturity levels in leaf tip, leaf blade, leaf base, and leaf sheath are different. The leaf sheath is the oldest sample but leaf tip is the youngest sample. There are several references which demonstrated distinct expression according to maturity level or leaf position. In this manuscript, authors had performed qRT-PCR for maturity related genes in leaf, stem, and pod. I think that the expression profiling may be ignored the sample maturity or the difference in sample position. Authors have to describe about experimental disregard in Discussion and detailed sample harvest in Materials and Methods.

 

Minor comments

1. Line 38: What is MG 000? I think it is simple mistypo.

Author Response

Response to Comments and Suggestions of Reviewer 4
The manuscript entitled with “Development of High-Resolution Simple Sequence Repeat Markers through Expression Profiling of Genes Associated with Pod Maturity of Soybean” by Park et al. (Manuscript ID: applsci-894790) described on the identification of high-resolution gene based markers for pod maturity in soybean. In general, the experiments and manuscript were well conceived and conducted. The results compared with appropriate controls and important finding is presented in results and figures. Appropriate literature cited and discussed all relevant results. This is the start of a useful contribution. The technique is translatable to other systems and this increases the value of the work. However, a few concerns have to be addressed before publication.
Therefore, I recommend this manuscript for the publication in Applied Sciences journal after incorporating some corrections suggested below.
 
Major comments
1. I think the Discussion section seems to be a little bit of short and low information. Authors have described only identified 6 genes and their expression. The manuscript contains valuable information for marker development and new constructed agglomerative clusters. Authors have to explain that more detail in Discussion section.
→  We newly described the valuable information for marker development (page 14, lines 287-292) and new constructed agglomerative clusters (page 14, lines 295-298) in the Discussion as follows:
“although the average number of alleles and PIC values of the six newly developed genic SSR markers were low (3.1 alleles and a PIC of 0.61) compared to those of the previously established six SSR markers (3.8 alleles and PIC of 0.70), the allelic variations exhibited by the new SSR markers were sufficient to generate a clustering and distinguish the soybean varieties by ecotype group. Finally, these alleles can serve as a marker for characterization of populations, varieties, and germplasm, as previously reported in wheat [71], jatropha [72], and oil palm [73].” (page 14, lines 287-292)
“We used agglomerative hierarchical clustering, a common bottom-up clustering method that uses the neighbor-joining method, for creating the phylogenetic trees [30]. This method provides a snapshot of the data that can facilitate more detailed analysis, while rapidly producing well-scaled informative networks for several hundred taxa [74].” (page 14, lines 295-298)
→  We have revised the Discussion accordingly and have divided it into four subsections to improve the overall readability.

2. Figure 3 was shown the expression profiling of selected genes for maturity. I think authors might compare qRT-PCR expressions with leaf, stem and pod to identify pod maturity specific genes. However, authors have to distinguish plant sample strictly to compare the maturity in same tissue. The maturity level should be different according to tissue position. For example, the maturity levels in leaf tip, leaf blade, leaf base, and leaf sheath are different. The leaf sheath is the oldest sample but leaf tip is the youngest sample. There are several references which demonstrated distinct expression according to maturity level or leaf position. In this manuscript, authors had performed qRT-PCR for maturity related genes in leaf, stem, and pod. I think that the expression profiling may be ignored the sample maturity or the difference in sample position. Authors have to describe about experimental disregard in Discussion and detailed sample harvest in Materials and Methods.
→  We used two reference genes, TUB4 (Glyma.03G124400.1) and UKN2 (Glyma.06G038500.1), to improve accuracy of normalization for expression profiling.
→  Hu et al. (2009) validated that the expression stability did not significantly vary for six different cultivars in long day (LD) and short day (SD), with ACT11, UKN2 and TUB4 being the most stable genes.
→  Among the genes, TUB4 (Glyma.03G124400.1) was the most highly expressed, with the associated Cq value in leaf, stem, and pod of three different ecotypes stable in the range of 20.24 to 21.45.
→  Moreover, Li et al (2012) concluded that the reference genes identified as optimal for seed development were TUA5 and UKN2.
→  Gao et al. (2017) also concluded that UKN2 is the best reference gene for analysis of roots and leaves together.
→  We found out that Glyma.06G038500.1 (UKN2), for which the Cq values were 23.34 ± 0.17, 24.28 ± 0.23 and 23.87 ± 0.12 in leaf, stem, and pod of three different ecotypes, respectively, was the most stable when we analyzed expression stability of three UKN2s.
→  Accordingly, we have revised the Materials and Methods (2.4 Profiling of gene expression chapter) as follows:
“Two reference genes were also used, TUB4 (Glyma.03G124400.1), and UKN2 (Glyma.06G038500.1), to improve accuracy of normalization for expression profiling. For relative expression profiling, we modified the formula reported by Livak and Schmittgen (Livak and Schmittgen 2001): normalization (ΔCq) of the quantification cycle (Cq) values of each gene using TUB4, ΔCq = Cq (a target sample) − Cq (TUB4); ΔΔCq calculation based on ΔCq of UKN2, ΔΔCq = ΔCq (a target sample) − ΔCq (UKN2); relative expression ratio (R), R = (2-ΔΔCq of a target sample)/(mean of 2-ΔΔCq of all sample).” (page 5, lines 125-131)

References
1. Hu, R., Fan, C., Li, H., Zhang, Q., & Fu, Y. F. Evaluation of putative reference genes for gene expression normalization in soybean by quantitative real-time RT-PCR. BMC molecular biology, 2009, 10, 93; DOI: 10.1186/1471-2199-10-93.
2. Li, Q., Fan, C., Zhang, X. et al. Validation of reference genes for real-time quantitative PCR normalization in soybean developmental and germinating seeds. Plant Cell Rep 2012, 31, 1789–1798; DOI: 10.1007/s00299-012-1282-4.
3. Gao M, Liu Y, Ma X, Shuai Q, Gai J,et al. Evaluation of Reference Genes for Normalization of Gene Expression Using Quantitative RT-PCR under Aluminum, Cadmium, and Heat Stresses in Soybean. PLOS ONE 2017, 12(1): e0168965; DOI: 10.1371/journal.pone.0168965.
4. Livak, K.J.; Schmittgen, T.D. Analysis of relative gene expression data using realtime quantitative PCR and the 2ΔΔC(T) Method. Methods, 2001, 25(4), 402–408; DOI: 10.1006/meth.2001.1262.

Minor comments
1. Line 38: What is MG 000? I think it is simple mistypo.

→  The MG 000 was used by Boerma et al. (2004) to classify maturity group (MG) for the extremely early maturing varieties.

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

In my opinion, thanks to the adjustments that the authors made after the first round of review, the manuscript has been substantially improved.

 

However, please consider few more things. Here I clarify few points that the authors missed from my first review report:

 

[65-67] We conducted experimental field tests at Suwon (37°16′N, 127°01′E), Republic of Korea in 2017 and 2018 to investigate DTM of the 100 soybean varieties (Table 1).

I read your answer. Still, we are not to the point... I will try to reformulate my comment: here you say that you “conduced experimental field tests… to investigate DTM”. If so, this is the point in the manuscript in which you identify and classify your plant matherial. In your answer to my previous comment, you said that you used the average of the two years. This sounds to me like "weak"… I would like to see some descriptive statistics of the cultivars somewhere… if not in the table, as supplemental. In my opinion, you cannot classify them without statistic.

I will imagine a possible scenario for the experimental design: you had a field with 100 cultivars. I expect you to have blocks containing a certain number of plants (sowing density could help you extimate them), not only one soy plant for cultivar. Then you score your blocks for DTM taking 10-20 plants for cultivar... and here you made an average. You did that for two years, and then you made an average of the two years and you classified your cultivars according to table 1.

I think you will agree with me if I suggest oyu to provide at least the standard deviation of the cultivars inside the year and between the two years. Then, I would say that you could classify your cultivars grouping the ones the dod not significantly differ from the DTM groups. In this way you are not using the average, that has no statistical power, but you provide robustness for your cultivar classification.

 

[164-167] A total of 55,589 genes were downloaded from the SoyBase database; among them, 49,638 were annotated. We analyzed the gene annotations and identified 1,147 genes that were related to growth and development of soybean pods; these were classified into four groups: flower, hormone synthesis, seed, and senescence (Table 3).

 

I will reformulate my previous question here: how did you decided which genes were related with growth and developments, flower, hormone synthesis etc.? Did you followed GO annotation?

I would like to see few lines here explaining your process.

 

4.2 Expression profiling shows that six genes were linked to growth and development of soybean pod

I would suggest to split it in subparagraphs, according to the different genes families, this would make it more readable

 

 

I would also suggest to review the manuscript for small english adjustments, here some examplese:

 

[50] the breeding program

Into breeding programs

 

[51] must first identified

Must first be

 

[57] These disadvantages of SSR markers limit their use

These disadvantages limit the use of SSR markers

 

[58-60] Meanwhile, Macaulay et al. [17] developed a highly informative genotyping set of SSRs for genetic studies of barley that was not limited by these drawbacks.

The use of meanwhile seems to me to be uncorrected here. To link the two concepts, I suggest to use something like “In 20xx, Mecaylay et al [17] developed…”, explaining how… then you state that you want to do something similar.

 

In general, I would suggest a strong recheck for english in the discussion and in the conclusions. Furthermore, the authors should consider some partial rephrasing in these sections: the content of the paper is nice and there is a lot of work in there. However, I would be aware of not claiming too much; the authors developed a nice marker-assisted system that seems to work in their matherial... If these markers could or could not be significant for breeding selection or QTL mapping, it need to be verified in the future.

Author Response

Response to Comments and Suggestions of Reviewer 2 (Round 2)

In my opinion, thanks to the adjustments that the authors made after the first round of review, the manuscript has been substantially improved.

 

However, please consider few more things. Here I clarify few points that the authors missed from my first review report:

 

[65-67] We conducted experimental field tests at Suwon (37°16′N, 127°01′E), Republic of Korea in 2017 and 2018 to investigate DTM of the 100 soybean varieties (Table 1).

I read your answer. Still, we are not to the point... I will try to reformulate my comment: here you say that you “conducted experimental field tests… to investigate DTM”. If so, this is the point in the manuscript in which you identify and classify your plant materials. In your answer to my previous comment, you said that you used the average of the two years. This sounds to me like "weak"… I would like to see some descriptive statistics of the cultivars somewhere… if not in the table, as supplemental. In my opinion, you cannot classify them without statistic.

• According to your comment, we have conducted analysis of variance (ANOVA) to compare the DTM of the varieties; while multiple comparisons between the DTMs were performed by the least significance difference (LSD) method. We have indicated the LSD value (2.51 days) in the Table 1 legend.
•  

I will imagine a possible scenario for the experimental design: you had a field with 100 cultivars. I expect you to have blocks containing a certain number of plants (sowing density could help you estimate them), not only one soy plant for cultivar. Then you score your blocks for DTM taking 10-20 plants for cultivar... and here you made an average. You did that for two years, and then you made an average of the two years and you classified your cultivars according to table 1.

• To more precisely define the plot design, we have edited the Materials and methods _2.1 Experimental field tests as follows: 
• “The three experimental plots were completely randomized with three replicates. The soybean seeds were sown on Jun 12, 2017 and Jun 14, 2018 in 4 m (row length) and 70 cm (row width) plots with 15 cm spacing, in the experimental fields that were treated with a basic granular fertilizer N-P₂O₅-K₂O = 30-30-34 kg/ha before sowing. This design allowed for 27 plants in a row, 4 rows per one replicate of each variety.” (page 2, lines 64-71)

I think you will agree with me if I suggest you to provide at least the standard deviation of the cultivars inside the year and between the two years. Then, I would say that you could classify your cultivars grouping the ones the dod not significantly differ from the DTM groups. In this way you are not using the average, that has no statistical power, but you provide robustness for your cultivar classification.

• As stated above, Analysis of variance has been conducted to compare the DTM means of the varieties. Multiple comparisons between the DTMs were performed by the least significance difference (LSD) method.
• We have indicated the LSD value in the Table 1 legend and edited the Materials and methods_2.6 Statistical analysis as follows: 
• “Multiple comparisons for gene expression profiling and DTM of varieties were performed by Duncan`s multiple range test and least significance difference (LSD), respectively.” (page 6, lines 152-154)

[page 6, lines 167-170] A total of 55,589 genes were downloaded from the SoyBase database; among them, 49,638 were annotated. We analyzed the gene annotations and identified 1,147 genes related to growth and development of soybean pods; these were classified into four groups: flower, hormone synthesis, seed, and senescence (Table 3).

 

I will reformulate my previous question here: how did you decided which genes were related with growth and developments, flower, hormone synthesis etc.? Did you follow GO annotation?

I would like to see few lines here explaining your process.

• I have firstly followed the GO annotation for deciding functional classification of genes related to pod maturity and then followed the information of other annotations. We have also referred to the expression patterns of each gene published in database of the Soybase website.
• For example: Glyma.03g201600.1 is annotated with GO:0006952 (GO-bp_defense response), GO:0007165 (GO-bp signal transduction), GO:0009506 (GO-cc plasmodesma), GO:0016020 (GO-cc membrane), GO:0016021 (GO-cc integral component of membrane), GO:0046658 (GO-cc anchored component of plasma membrane), GO:0004871 (GO-mf signal transducer activity) but with PF03168 (PFAM_Late embryogenesis abundant protein). Moreover, the expression pattern of Glyma.03g201600 showed that it is highly expressed in the pod and seed at all reproductive stages. Therefore, we concluded that the gene is related to seed development, particularly in the embryo.

 

4.2 Expression profiling shows that six genes were linked to growth and development of soybean pod

I would suggest to split it in subparagraphs, according to the different gene families, this would make it more readable

• We have revised the chapter so that it is broken down into three paragraphs according to the different gene families.

 

I would also suggest to review the manuscript for small English adjustments, here some examples:

 

[50] the breeding program

Into breeding programs

• We have revised the text according to your suggestion.
•  

[51] must first identified

Must first be

• We have revised the text accordingly.

 

[57] These disadvantages of SSR markers limit their use

These disadvantages limit the use of SSR markers

• We have revised the text accordingly.

 

[58-60] Meanwhile, Macaulay et al. [17] developed a highly informative genotyping set of SSRs for genetic studies of barley that was not limited by these drawbacks.

The use of meanwhile seems to me to be uncorrected here. To link the two concepts, I suggest to use something like “In 20xx, Mecaylay et al [17] developed…”, explaining how… then you state that you want to do something similar.

• We have revised the text to, “In 2001, Macaulay et al…”

 

In general, I would suggest a strong recheck for English in the discussion and in the conclusions. Furthermore, the authors should consider some partial rephrasing in these sections: the content of the paper is nice and there is a lot of work in there. However, I would be aware of not claiming too much; the authors developed a nice marker-assisted system that seems to work in their materials... If these markers could or could not be significant for breeding selection or QTL mapping, it need to be verified in the future.

• Thank you for these comments. We will verify whether or not the new developed markers are significant for breeding selection or QTL mapping in our future work.

Author Response File: Author Response.docx

Reviewer 3 Report

The paper contents have not changed from the previous version and I remain unconvinced the approach used is a viable one to develop markers representing genetic divesity linked to the DTM phenotype.

The  gene expression profiling results may well serve as evidence that the selected genes could have an involvment in pod growth and development and ultimately maturity, but it should not be assumed that variants in their sequence can be used as genetic markers to predict the phenotype of an individual.

If informative markers could be develped that easily, the trascriptome data avalable for countless plant species would have been put to a better use already, and nobody would bother with expensive and time consuming QTL or association mapping studies.

 

Author Response

Response to Comments and Suggestions of Reviewer 3 (Round 2)

The paper contents have not changed from the previous version and I remain unconvinced the approach used is a viable one to develop markers representing genetic diversity linked to the DTM phenotype.

• We believe that our revised manuscript contains several significant changes compared to the original version of our manuscript. For instance:We have also discussed the lower level of polymorphisms and merits of agglomerative clustering in the revised Discussion as follows:  
• “We used agglomerative hierarchical clustering, a common bottom-up clustering method that uses the neighbor-joining method, for creating the phylogenetic trees [30]. This method provides a snapshot of the data that can facilitate more detailed analysis, while rapidly producing well-scaled informative networks for several hundred taxa [74].” (page 14, lines 306-314)
• “Genic SSR markers have been reported to not only be less polymorphic than non-genic markers [67], but to also have 3.5 alleles per locus, with a higher PIC of 0.824 among the mulberry species [68]. Similarly, previous studies reported that the genic SSR markers for jute (2.7 alleles and a PIC of 0.34) [69] and flax (2.3 alleles and PIC of 0.35) [70] show high polymorphism and are expected to be of use for the characterization of germplasm, as well as for variety identification, and marker assisted breeding. Therefore, although the average number of alleles and PIC values of the six newly developed genic SSR markers were low (3.1 alleles and a PIC of 0.61) compared to those of the previously established six SSR markers (3.8 alleles and PIC of 0.70), the allelic variations exhibited by the new SSR markers were sufficient to generate a clustering and distinguish the soybean varieties by ecotype group. Finally, these alleles can serve as a marker for characterization of populations, varieties, and germplasm, as previously reported in wheat [71], jatropha [72], and oil palm [73].” (page 14, lines 293-303)
• We have split the Discussion into four sub-chapters and have explained the Discussion section in more detail to improve the overall readability.
• We have also added information regarding the SSR motifs in the revised Discussion as follows:
• “As shown in supplementary Table S1, we developed 211 genic SSR markers. Among the 211 SSR repeat motifs, tri-nucleotide repeats (TNRs) were found to be the most abundant, accounting for 75.4% (159), followed by 16.6% di-nucleotide repeats (DNRs), and 8.0% tetra-nucleotide repeats (TtNRs). Meanwhile, longer than penta-nucleotide repeats were not found. These results were similar to previously published data with TNRs reported as the most frequent (54–78%), followed by DNRs (17.1–40.4%), and TTNRs (3–6%) among cereal species [50].” (page 12-13, lines 246-251)
• As stated in the Discussion, to overcome these disadvantages of SSR markers, over 1,000 genic SSR markers have been developed for soybean genetic research [43,44]. Genic SSR markers are used as an effective tool to identify genetic resources, since SSRs are located in transcribed genes that are annotated a putative function based on biological information and homology search [45], and as a selection tool for specific traits in breeding populations of diverse crops [46–49]. (page 12, lines 241-246)
•  

The gene expression profiling results may well serve as evidence that the selected genes could have an involvement in pod growth and development and ultimately maturity, but it should not be assumed that variants in their sequence can be used as genetic markers to predict the phenotype of an individual.

• We partially disagree with your opinion that it should not be assumed that variants in their sequence can be used as genetic markers to predict the phenotype of an individual.
• The results demonstrating that allele diversity was identified by the newly developed genic SSR markers imply that the genes have variants in their sequence harboring SSR motifs. Therefore, we concluded that the newly developed SSR markers derived from the six genes can serve as effective genetic markers for genotyping and fine QTL mapping of soybean pod maturity but not for predicting the phenotype of an INDIVIDUAL. However, we acknowledge that it will be necessary to verify whether or not the newly developed markers are significant for breeding selection or QTL mapping in future studies.

 

If informative markers could be developed that easily, the transcriptome data available for countless plant species would have been put to a better use already, and nobody would bother with expensive and time consuming QTL or association mapping studies.

• We respectfully disagree with your opinion that our informative markers were easily developed.
• We have conducted a great deal of work consisting of difficult processes to develop the previously estimated SSR markers, including experimental field tests, gene analysis and selection (a total of 55,589 genes), development and selection of SSR markers, profiling of gene expression, and agglomerative clustering.
• As stated above, to overcome the disadvantages of SSR markers, many scientists and breeders have developed over 1,000 genic SSR markers for soybean genetic research.

Author Response File: Author Response.docx

Reviewer 4 Report

Comments to Author:

Thank you for your revision (Manuscript ID: applsci-894790), which addressed most of the comments and has substantially, improved the manuscript. Updated paragraphs improve the quality of Discussion section and may provide impressive ideas to the readers.

However, I could not find the description of detail sample position for qRT-PCR. I believe author easily find references for the expression distinction in same tissue according to maturity. For example, PMB (2017) 93:465-478 reported the different expression at different position in accordance with senescence. I think authors easily update their manuscript about this comment. Therefore, I recommend the manuscript for the publication in Applied Science after major revision.

Author Response

Response to Comments and Suggestions of Reviewer 4 (Round 2)

Thank you for your revision (Manuscript ID: applsci-894790), which addressed most of the comments and has substantially, improved the manuscript. Updated paragraphs improve the quality of Discussion section and may provide impressive ideas to the readers.

However, I could not find the description of detail sample position for qRT-PCR. I believe author easily find references for the expression distinction in same tissue according to maturity. For example, PMB (2017) 93:465-478 reported the different expression at different position in accordance with senescence. I think authors easily update their manuscript about this comment. Therefore, I recommend the manuscript for the publication in Applied Science after major revision.

• We have added information regarding the sampling process in the Materials and methods_2.4 Profiling of gene expression section, as follows:
• “We conducted expression profiling of genes harboring the selected SSRs using tissues (leaf, stem, pod) of the three maturity ecotypes to further select genes closely associated with growth and development of the soybean pod. For profile gene expression according to reproductive stages, soybean plants of three distinct stages, based on maturity ecotypes, were collected from stage R6 (full seed, LM ecotype) to R8 (EM ecotype) as shown in Fig. 1. The leaves, stem and pod were then simultaneously collected. The tissue samples were immediately placed in liquid nitrogen, and stored at −80 °C. All samples were collected in three biological replicates.” (page 4, lines 116-122)
• Kim et al (Plant Mol Biol 2017, 93, 465–478) collected four different leaf sections of transgenic or non-transgenic sugarcane for comparing Xyl10B activity in detail. However, we collected three tissues to select genes related to growth and development of soybean pod.
• We have also edited the Results section as follows:
• “With advanced stage (R6 to R8), five of the genes were downregulated in pod while Glyma.19g198600.1 was upregulated (Fig. 3).” (page 8, lines 190-191)
• We have also revised the Discussion as follows: “Glyma.19g158400.1 was upregulated in the pods of all three ecotypes although the highest expression was observed in the pods of the LM ecotype, indicating that the gene is downregulated with stage advancement.” (page 13, line 273-275)
• “These results demonstrate that although different isoforms belong to the same gene family in soybean, the genes are differentially regulated with stage advancement [55] and by tissue [56]”. (page 13, line 264-265)
• “Glyma.19g158400.1 was upregulated in the pods of all three ecotypes although the highest expression was observed in the pods of the LM ecotype, indicating that the gene is downregulated with stage advancement.” (page 13, line 273-275)

References

Kim, J.Y., Nong, G., Rice, J.D. et al. In planta production and characterization of a hyperthermostable GH10 xylanase in transgenic sugarcane. Plant Mol Biol 93, 465–478 (2017). https://doi.org/10.1007/s11103-016-0573-5

55. Teixeira, R.N.; Ligterink, W.; França-Neto, J.B.; Hilhorst, H.W.; da Silva, E.A. (2016). Gene expression profiling of the green seed problem in soybean. BMC Plant Biol. 16:37. doi: 10.1186/s12870-016-0729-0.
56. Gu, Y.; Li, W.; Jiang, H.; Wang, Y.; Gao, H.; Liu, M.; Chen, Q.; Lai, Y.; He, C. Differential expression of a WRKY gene between wild and cultivated soybeans correlates to seed size. J. Exp. Bot. 2017. 68, 2717–2729, DOI: 10.1093/jxb/erx147.

Author Response File: Author Response.docx

Round 3

Reviewer 2 Report

Dear authors,

thanks for all your answers and for accepting my comments and suggesitons.

In my opinion, the manuscript has been extremely improved in the last weeks and it is now ready for pubblication on this journal.

 

I only spotted one small error in the text, please fix it:

 

[84-89] annotation through evolutionary relationship) classification system [21]; 6) the first versions of Soybean Metabolic Pathway (SoyCyc) of the primary proteins described and assigned using Soybean Metabolic Pathway Database [22]; 6) the seventh versions of Soybean Metabolic Pathway (SoyCyc7) of the primary proteins described and assigned using PMN: Plant Metabolic Network [23]; and 7) the soycyc_enzymes of Soybean Metabolic Pathway (SoyCyc7-rxn) for the primary proteins determined after downloading from the TAIR database [24]

 

6) compares two times, review the list

 

 

For this reasons, in this last review, I evaluated this paper as "Accept after minor revision". 

Author Response

Response to Comments and Suggestions of Reviewer 2 (Round 3)

Thanks for all your answers and for accepting my comments and suggesitons.

In my opinion, the manuscript has been extremely improved in the last weeks and it is now ready for pubblication on this journal.

I only spotted one small error in the text, please fix it:

[84-89] annotation through evolutionary relationship) classification system [21]; 6) the first versions of Soybean Metabolic Pathway (SoyCyc) of the primary proteins described and assigned using Soybean Metabolic Pathway Database [22]; 6) the seventh versions of Soybean Metabolic Pathway (SoyCyc7) of the primary proteins described and assigned using PMN: Plant Metabolic Network [23]; and 7) the soycyc_enzymes of Soybean Metabolic Pathway (SoyCyc7-rxn) for the primary proteins determined after downloading from the TAIR database [24]

6) compares two times, review the list

→ Thanks for all your comments and suggestions. We have revised the list.

Author Response File: Author Response.docx