Novel QTLs for salinity tolerance revealed by genome-wide association studies of biomass, chlorophyll and tissue ion content in 176 rice landraces from Bangladesh

Md Nafis Ul Alam; G M Nurnabi Azad Jewel; Tomalika Azim; Zeba I Seraj

doi:10.1371/journal.pone.0259456

. 2021 Nov 5;16(11):e0259456. doi: 10.1371/journal.pone.0259456

Novel QTLs for salinity tolerance revealed by genome-wide association studies of biomass, chlorophyll and tissue ion content in 176 rice landraces from Bangladesh

Md Nafis Ul Alam ¹, G M Nurnabi Azad Jewel ^1,², Tomalika Azim ¹, Zeba I Seraj ^1,^*

Editor: Prasanta K Subudhi³

PMCID: PMC8570475 PMID: 34739483

Abstract

Farmland is on the decline and worldwide food security is at risk. Rice is the staple of choice for over half the Earth’s people. To sustain current demands and ascertain a food secure future, substandard farmland affected by abiotic stresses must be utilized. For rapid crop improvement, a broader understanding of polygenic traits like stress tolerance and crop yield is indispensable. To this end, the hidden diversity of resilient and neglected wild varieties must be traced back to their genetic roots. In this study, we separately assayed 11 phenotypes in a panel of 176 diverse accessions predominantly comprised of local landraces from Bangladesh. We compiled high resolution sequence data for these accessions. We collectively studied the ties between the observed phenotypic differences and the examined additive genetic effects underlying these variations. We applied a fixed effect model to associate phenotypes with genotypes on a genomic scale. Discovered QTLs were mapped to known genes. Our explorations yielded 13 QTLs related to various traits in multiple trait classes. 10 identified QTLs were equivalent to findings from previous studies. Integrative analysis assumes potential novel functionality for a number of candidate genes. These findings will usher novel avenues for the bioengineering of high yielding crops of the future fortified with genetic defenses against abiotic stressors.

1. Introduction

10 million hectares of arable land is lost to urbanization every year [1]. An accelerated wave of climate change damages the quality of the remaining landmass [2]. About half of the world’s currently estimated farmland is affected by abiotic stresses, most notably, salinity [3, 4]. Among cereal crops, rice is the most sensitive to abiotic factors and yield losses [5] but has the largest contribution to global food production. Rice feeds more people across the world than any other crop [6], accounting for up to 80% of daily calorie intakes of half the world’s population [7]. It is surmised that global rice production has to double by the year 2050 to combat these burdens imposed on food security [8, 9]. The insufficiency in finite arable land has to be overcome by increases in crop productivity. This must involve the introgression and propagation of stress tolerance traits to enable farming in soil less ideal for present day elite varieties.

With the help of high-resolution genotyping technologies, it has become possible to elucidate the molecular signature of complex polygenic traits. Multitudes of small genetic effects and their interactions with the surroundings underlie the quantitative assertion of these phenotypes. This limits our ability to refine them through direct breeding or molecular approaches. Many statistical and genome-wide association based studies (GWAS) have been carried out for salinity tolerance [10, 11], chlorophyll [12] and yield [13] related traits in rice independently. Researchers often point out that complex traits are interlinked through their evolutionary origins and must be studied in conjunction to reach conclusive outcomes [14]. Few recent GWAS studies explore abiotic stress tolerance in relation to developmental and agronomic phenotypes in rice. Identification of major-effect QTLs for yield, tolerance and developmental attributes will continue to add layers of sophistication to the current understanding of plant biology. Characterization of causal polymorphisms will facilitate the introgression of salient features into elite varieties with the help of sophisticated molecular technologies such as CRISPR-Cas9 genome editing.

Bangladesh is the largest delta on the planet, endowed with fertile land, meandering rivers and a pronounced history of growing rice. To accelerate the improvements in cereal crop development, we phenotyped 176 local rice varieties from Bangladesh for 11 quantitative traits relating to plant biomass, chlorophyll content, tissue ion content and visual salt damage (SES). We incorporated next-generation sequencing data with high genomic coverage to estimate additive genetic effects and correlate the observed phenotypic values with genomic predictions. We tested over 4 million high-quality SNP markers for association signals separately for each trait. We cross-referenced our identified QTLs with functional annotations, publicly available gene expression profiles and allelic substitution effects. The compiled assessment of multiple traits belonging to several trait classes accentuated the genetic landscape influencing their effects. For identified QTLs, integrative analysis of auxiliary data enabled us to unravel potential novel gene roles.

2. Methods

2.1 Data and source code

All data, source code and shell commands are documented on GitHub at https://www.github.com/DeadlineWasYesterday/Cat-does-plant. The complete computational pipeline was executed in an HP Z840 workstation running on a 16-core Intel Xeon processor with 256GB of RAM.

2.2 Plant material and growth conditions

The 176 rice accessions were received from the IRRI seed bank to ascertain their identity with the 3,000-rice genome project [15]. Seeds were multiplied at BRRI (Bangladesh Rice Research Institute) fields in the months of March and April when the average day/night temperature was 32°/28°C with an average humidity of 72%. S1 Table lists the accession codes and metadata with local names and subpopulation information for all 176 varieties. Only 3 varieties belong to the USDA core (IRGC 31727, IRGC 27555 and IRGC 58736) and none are listed in the USDA mini-core collection. 106 accessions are traditional landraces that are cultivated locally in Bangladesh. The remaining 70 accessions are varieties that had been developed by BRRI from local landraces for local agronomy. With the exception of three (IRGC 49375, IRGC 126002 and IRGC 124432), all varieties had been originally collected from Bangladesh by the International Rice Research Institute. The starter template for the map of Bangladesh was collected from Vecteezy [16] and modified in Adobe Illustrator. Geolocations were labelled using the ggplot2 and ggrepel libraries in R.

The phenotype screening for salinity tolerance at the seedling stage took place in a nethouse enclosure at an average day/night temperature of 31°C/27°C and approximately 72% relative humidity. The screening for salinity tolerance at seedling stage was carried out following the methods described by Amin et al. 2012 [17]. Sprouted seeds were sown in netted Styrofoam and floated in 3×3 (3 for control and 3 for stress) replicated PVC trays containing 60L Yoshida solution [18]. The positions of the 176 accession seedlings were randomized in each of the 6 trays. The germinated seeds were allowed to grow for 14 days. Then, NaCl stress was applied gradually starting from 4dS/m up to 12dS/m by 2dS/m increments every 24 hours in all of the experiment trays. The solution level of the trays were maintained with water. Phenotypic measurements were taken after 16 days of stress application, when 90% of the leaves of the sensitive control (IR29) were damaged.

2.3 Phenotyping

After about 1 month of germination and 16 days from the first salt stress exposure for the stress condition plants, seedlings were systematically phenotyped. Fresh weight and length measurements were taken for whole roots and shoots. For root and shoot ion content data, plants were externally washed and oven dried at 70°C. Ground-up dried roots and leaves were treated with 1N HCl for 48 hours before being assayed in a flame photometer (Sherwood model 410, Sherwood, UK). Ion measurements are denoted in millimolar concentration. SES (Standard Evaluation System) scores were given to each plant on the basis of growth stunting, damage to leaves, leaf chlorosis, drying of biomass and plant vigor as described by Gregorio et al. 1997 [19, 20]. Chlorophyll content was measured in fully expanded third leaves. Leaf extracts of 1cm² size were dissolved in 80% acetone in the absence of light. Absorbance at 645nm and 663nm wavelengths were measured. The formulae for chlorophyll A and B estimation were adapted from Yoshida et al. 1976 [18] as follows:

C h l o r o p h y l l A c o n t e n t = (12.7 \times A_{663} - 2.69 \times A_{645}) \times \frac{V}{1000 \times W}

C h l o r o p h y l l B c o n t e n t = (22.9 \times A_{663} - 4.68 \times A_{645}) \times \frac{V}{1000 \times W}

Here A₆₆₃ and A₆₄₅ refer to the absorbance at 663nm and 645nm wavelengths respectively. V is volume in milliliters, W is the fresh weight of the leaf pieces in milligrams and chlorophyll content is expressed as milligrams per gram of fresh leaf tissue.

2.4 Phenotype statistics

We have catalogued 10 basic traits that we grossly assign to plant biomass, chlorophyll content and tissue ion content categories. The 11th basic trait is the visual SES score for evaluating salt damage which by definition is only measured in stress condition. Simple mathematical operations on the phenotypes gave rise to 12 additional derived phenotypes. For the 10 traits measured in both conditions, heritability was measured by the formula $h 2 = σ_{G}^{2} / (σ_{G}^{2} + \frac{σ_{e}^{2}}{n})$ where $σ_{G}^{2}$ is the variance of the genotypes, $σ_{e}^{2}$ is the residual variance and n is the number of replicates. The variance components were calculated by including the genotype, treatment condition and interaction between genotype and condition as random effects in the lmer function from the lme4 package in R. Marker-based heritability was calculated using the kinship matrix K (details in section 2.7) in the marker_h2 function from the heritability package in R. Broad-sense heritability was calculated by dividing the variance of the genotype means by the overall phenotypic variance. Correlation and GWAS studies involved phenotype means for three replicates. Linear correlation is expressed by the Pearson’s correlation coefficient. Genomic prediction for SNP effects on phenotype was calculated by ridge regression using the mixed.solve function from the rrBLUP package [21]. Assessment of data distributions was carried out by the SciPy module in Python. Phenotype transformation for GWAS was done by an ideal genotype derived warping function calculated by WarpedLMM [22].

2.5 Preprocessing genotype data

Next-generation sequencing data was compiled from the 3,000-rice genome project [15]. In the original study, libraries were created from young leaves and sequenced in the Illumina HiSeq2000 platform to generate 90bp paired-end reads. Raw sequence reads were aligned to the IRGSP 1.0 reference genome for Oryza sativa sub. japonica cultivar Nipponbare. The average sequencing depth was 14× with genome coverage and mapping rates over 90%. Publicly available VCF data from aligned reads were sourced for 176 varieties and compiled using VCFtools in the SAMtools suite. SNPs having phred-scaled quality scores below 30 were flagged as low-quality reads and considered missing data. Beagle 5.1 [23] was used for imputation of unphased genotypes. There were on average 2,930,739 missing genotypes out of 15,241,544 total marker calls per individual with a standard deviation of 492,558. For assessment of imputation accuracy, we prepared a test set having the original proportion of missing data by systematically removing high-quality reads from 1.85 million markers which had no missing reads. Unphased imputation accuracy was measured to be over 99.5% and the report can be found on the git repository. Working files were prepared using in-house python scripts. SNPs that originally had over 20% low quality reads were excluded from all studies. For GWAS, a minor allele frequency filter of 5% was applied to the whole population. Because of the smaller sample size, the minor allele frequency filter for the subpopulations was set at 10%.

2.6 Population structure estimation

A combination of methods was employed for deciphering population structure. Principle components were calculated using R and clustered in two and three dimensions respectively using a k-means clustering algorithm. The first three principal components showed 38% explained variation in scree plots. 2D and 3D scatter plots were also plotted in R using the first three principal components and can be seen in the git repository. Bayesian maximum likelihood-based population structure estimation software fastStructure [24] implemented in python was used for assessing admixture between groups. K (number of subpopulations) values from 2 to 15 were tested and the chooseK.py function was used to select k = 3. Distances for the neighbor-joining tree were calculated in TASSEL 5 [25] and visualized by the web-based application iTOL (interactive Tree Of Life) [26].

2.7 Genome-wide association studies

The primary algorithm for association testing was BLINK [27] in the GAPIT software package [28]. The more popular and statistically robust compressed mixed linear model (CMLM) applied both in GAPIT and TASSEL 5 were also run to validate the findings (data in git repository).

BLINK employs two successive fixed effects models. The first model tests for associations and calculates p values and the second model iteratively improves the first model by exporting markers as covariates using Bayesian Information Criterion (BIC).

The first model is denoted:

y_{i} = S_{i 1}^{*} b_{1} + S_{i 2}^{*} b_{2} + \dots + S_{i k}^{*} b_{k} + S_{i j} d_{j} + e_{i}

where y_i is the estimated observation for the i-th individual. $S_{i}^{*}$ terms represent covariates named pseudo QTNs that start off as an empty set and subsequently are iteratively selected by the second model. b terms are the corresponding effects for the pseudo QTNs. S_ij is the genotype for the testing marker j. d_j is the effect of the j-th marker and e_i is the residual error term having a distribution $e_{i} \sim N (0, σ_{e}^{2})$ .

The second model given below is similar but lacks the genotype term:

y_{i} = S_{i 1}^{*} b_{1} + S_{i 2}^{*} b_{2} + \dots + S_{i k}^{*} b_{k} + e_{i}

Markers are sorted by ascending order of p value and fitted in the second model one by one if considered significant after Bonferroni correction and found to not be in linkage disequilibrium (r² > 0.7) with any marker already included in the equation. Model selection by BIC finds the best fit second model from all combinations and the covariates from that model are exported to the first model.

This complete process is repeated in BLINK until the second model no longer selects new covariates for inclusion in the first model and the first model is established as the final testing model.

To correct for population structure and cryptic relatedness, the kinship matrix K, calculated using GAPIT and the Q matrix having the first three principal components were fit in the first model as covariates. For the compressed mixed linear models fit for validation tests, the optimum compression level was selected and the same covariates were included. Q-Q plots and Manhattan plots were visualized by the qqman package in R [29].

An adjusted Bonferroni correction was applied to set the genome wide significant and suggestive thresholds. The formula used for setting the significance threshold was −log (1/M) where M is the total number of testing markers. In the whole population, the value was 6.63 for 4,283,120 markers that remained after the MAF filter, which was rounded down to 6.5 for use. In Indica and Aus subpopulations, there were 2,735,794 and 2,562,596 markers respectively and the significant threshold was set at 6.4 accordingly. For suggestive thresholds, effective number of markers were calculated, defined as those being in approximate linkage equilibrium across the genome. A pairwise LD cutoff of r² = 0.2 with 50kb sliding windows and step size of 50kb in PLINK [30] was used to calculate the effective number of markers. The effective number was 384875, 255487 and 234071 for the whole, Indica and Aus sets respectively. The resulting suggestive threshold values were 5.37, 5.4 and 5.58, from which a common down-rounded value of 5.0 was used.

3. Results

3.1 Observation of population structure

The panel of 176 accessions predominantly consist of the Indica and Aus subgroups. Fig 1 shows 114 known geolocations from S1 Table where the labelled varieties have been historically cultivated. The appropriate number of subpopulations (K) was estimated to be 3. Three distinct clusters could also be visualized in a scatter plot of the first and second principal components. An admixture plot for 3 subpopulations and the NJ tree for the 176 individuals can be seen in Fig 2A and 2B respectively. The red individuals belong to the Aus subpopulation and the green shaded individuals are members of the Indica subpopulation. We analyzed 2D and 3D scatter plots for the first three principal components to infer population structure. Bayesian k-means clustering results can be seen in Fig 2C and 2D. The subpopulations assigned to the individuals by fastStructure, the groupings calculated by the k-means algorithm, and that inferred by manual inspection were found to be identical. Population structure inferences from all sources have been tabulated in S1 Table. In the NJ tree and the PCA scatter plots, the Aus subpopulation was found to cluster more tightly than the individuals belonging to the Indica subgroup. The third subpopulation lodges only a few individuals that are grouped as aromatic rice varieties and are locally known as Basmati or Sadri rice. Because of having a very small number of individuals, the aromatic subpopulation was excluded from subpopulation specific association testing along with three accessions that did not cluster with either subpopulation (S1 Table).

Fig 2 — Three distinct subpopulations can be inferred from these diagrams. (a) An admixture plot calculated from fastStructure. (b) Neighbor-joining tree drawn from genotype data. The green individuals belong to the Aus subgroup, red individuals are from the Indica subgroup and black lines denote plants distant from these groups. (c) K-means clustering results of the first two principal components. (d) K-means clustering results of a three-dimensional plot of the first three principal components.

3.2 Comprehensive analysis of phenotypes

Eight additional phenotypes could be derived from the biomass traits. Table 1 lists all basic and derived phenotypes with the formulations for each derivation from the basic traits. Biomass traits had lower variance across replicates and were more heritable than ion content traits. Within the biomass category, shoot traits were more heritable than root traits. Marker derived heritability for all phenotypes and broad-sense heritability for applicable phenotypes can be found in S2 Table. The means of phenotype values recorded for every genotype is stored in S3 Table. Narrow-sense heritability could not be calculated from additive marker effects because the grand mean of the phenotypes could explain more variance than the mixed.solve prediction model. This could be attributable to the relatively small sample size with diverse and heterozygous wildtype genotypes used for genomic prediction. The linear Pearson’s correlation coefficients between the observed phenotypes are shown in the bottom diagonal of the heatmap in Fig 3 and the correlation between the genotype effects for each phenotype calculated through genomic prediction occupies the upper diagonal of the heatmap. The correlation between the observed phenotypes and additive SNP effects are seen along the white diagonal of the map. All phenotype estimations made by genomic prediction can be found in S4 Table. An extended figure showing correlation coefficients between observed and estimated values of all 33 basic and derived phenotypes can be viewed in S1 Fig. A matrix of p values for the correlations shown in S1 Fig is provided in S5 Table.

Table 1. Basic and derived phenotypes.

Group	Basic traits	Heritability of basic traits	Derived traits	Formula for derivation	Heritability of derived traits
Biomass	Root weight	0.519935202	Lost root weight	Root weight in control condition—Root weight in stress condition	0.020508891
Biomass	Shoot weight	0.598616239	Lost shoot weight	Shoot weight in control condition—shoot weight in stress condition	0.063502564
Biomass	Root length	0.538476891	Lost root length	Root length in control condition—Root length in stress condition	0.186636646
Biomass	Shoot length	0.680380481	Lost shoot length	Shoot length in control condition—shoot length in stress condition	0.00847037
Biomass			Root weight per unit length^**	Root weight / Root length	0.441443378
Biomass			Shoot weight per unit length^**	Shoot weight / Shoot length	0.234151181
Biomass			Lost root weight per unit length	Root weight per unit length in control condition—Root weight per unit length in stress condition	0.128275455
Biomass			Lost shoot weight per unit length	Shoot weight per unit length in control condition—Shoot weight per unit length in stress condition	0.021422646
Ion content	Root sodium content	0
Ion content	Root potassium content	0.015590152
Ion content	Shoot sodium content	0.029297234
Ion content	Shoot potassium content	1.05E-09
Chlorophyll content	Chlorophyll A content	0.299207727	Lost chlorophyll A content	Chlorophyll A content in control condition—Chlorophyll A content in stress condition	0.177123042
Chlorophyll content	Chlorophyll B content	0.273868031	Lost chlorophyll B content	Chlorophyll B content in control condition—Chlorophyll B content in stress condition	0.181277873
SES	SES score^*	0.37256389

Open in a new tab

* traits measured in only one condition.

** abbreviated as ‘thickness’ throughout the manuscript and applicable to both control and stress conditions.

In Fig 3, we observe better trait correlations in salt stress conditions relative to control conditions. Rectangle C1 and S1 compare correlations between biomass traits in control and stress conditions, C2 and S2 compare correlations between chlorophyll, SES and biomass in control and stress conditions. C3 and S3 compare correlations between chlorophyll and SES in control and stress. The improved linear correlations between biomass, chlorophyll and SES traits imply that the encumbrance of salt stress on a plant prevents the disproportionate gain of biomass. Since plant biomass and chlorophyll content is markedly reduced under salt stress (S3 Table), we can conclude that salt stress does not affect all phenotypes for all genotypes at the same rate and the genetic advantages that any genotype has in terms of tissue growth (root/shoot) or chlorophyll accumulation will be compromised by the effect of abiotic stress at a greater magnitude than traits which have near-baseline values. On account of linear correlation, plant root length stands out to be the most independent phenotype, especially in control conditions.

We derived principal components from the biomass and chlorophyll phenotypes under salt stress to compare the salt tolerability of our accessions. Since the composite SES score is a visual score broadly based on stunting of growth, discoloration of leaves and loss of plant vigor, we expect the pattern of SES scores to mimic the major principal component derived from the stress phenotypes. The first principal component could explain 69.6% variation in the data and has been plotted from the top in Fig 4 in the order of ascending SES score. From the figure, we observe a visible correspondence between SES score and the related stress condition phenotypes. In terms of SES score, the best performing varieties are IRIS_313–10973, IRIS_313–11224 and IRIS_313–8244 and the most sensitive varieties are IRIS_313–11063, IRIS_313–11064, IRIS_313–11400, IRIS_313–12055 and IRIS_313–8641. Based on the axis of maximum variation, the most tolerant accessions are IRIS_313–11116, IRIS_313–11217 and IRIS_313–9594 and the least tolerant accessions are IRIS_313–12055, IRIS_313–10606 IRIS_313–10602 and IRIS_313–8641 (S3 Table).

Fig 4 — The principal component along the axis of maximum variation could explain 69.6% of the total variation. Metrics derived from simple numerical phenotypes could substitute for SES scores when a more objective index of salt injury is desired.

Density plots for all processed and derived phenotypes with subgroup specific curves for 82 Indica and 78 Aus varieties are tiled in S2 Fig. The biomass related traits we measured mostly followed a normal distribution with the exception of plant shoot length in both control and stress conditions. The ion content and chlorophyll content traits displayed large skews with few extreme values in both poles. Out of the ion content phenotypes, only shoot sodium content in stress condition exhibited a gaussian distribution (S2 Fig). Filtering out few extreme values mended the majority of the distributions to fit a normal curve. S6 Table lists the Shapiro-Wilk statistics for all phenotypes prior to filtering and the number of data points that needed to be omitted to improve the distribution. We did not use the filtered data in the primary association testing as it would compromise statistical power and risk spurious associations. Instead, data was transformed using an appropriate warping function and the changes in the distributions are shown in S3 Fig. Scatter plots of residuals from genomic estimates of phenotypes calculated before and after transformation can be seen in S4 Fig. The transformation method improved the centered clustering of residuals around zero and their uniformity in dispersion.

3.3 Genome-wide association studies

The BLINK algorithm [27] implemented in the GAPIT software package [28] is a recent innovation for powerful statistical association testing at considerably lower computational cost. BLINK achieves enhanced computational time by employing a linkage disequilibrium (LD) based iterative marker pruning strategy that fits two fixed effect models with a filtering step involving the second model. The algorithm is discussed in detail in the methods section. We also fit a more commonly used compressed mixed linear model (CMLM) to our phenotypes with optimum compression level to compare and confirm significant associations. Three independent studies on the whole population and Aus and Indica subpopulations were carried out respectively. 33 individual Manhattan and QQ plots each from three independent sets association tests are stored in the git repository under the ‘GWAS’ folder name. Table 2 lists 13 total QTLs identified and grouped into biomass or salt tolerance categories on the basis of their significant association to multiple traits belonging to the mentioned trait class. The first identified QTL, qCDP1.1, was observed in plant shoot length measured at stress conditions in the Indica subpopulation (Fig 5A). A suggestive peak on the same loci is found in the Manhattan plot for SES in the same subpopulation (S5 Fig). Two other associations noted for biomass traits were qCDP12.1 and qCDP12.3. Like qCDP1.1, qCDP12.3 can also be found associated with stress tolerance traits (S5 Fig). qCDP3.1 relates to potassium content in plant shoots (Fig 6A). Because of the central role of potassium and other cations in salinity tolerance, we observe qCDP3.1 as a stress tolerance trait. Four more signals, qCDP5.1, qCDP6.1, qCDP8.1 and qCDP9.1, relating to the effect of salinity on plant shoots shown in Fig 6B and 6C were discovered. qCDP7.1, qCDP9.2 and qCDP9.3 concerning root and biomass traits are observed in Fig 7. Fig 8 shows that apart from qCDP3.1, two more loci, qCDP10.1 and qCDP12.2 are associated with ion content phenotypes in stress conditions. 10 out of 13 QTLs in our findings were previously discovered by GWAS studies with similar phenotype connotations (Table 2). An extended version of Table 2 with peak marker locations, original phenotype names and references [10, 11, 31–44] can be found in S7 Table.

Table 2. Identified major QTLs.

Locus name	Trait class	Chromosome	Lead SNP position	P value	Traits where observed	Traits previously linked to by GWAS
qCDP1.1	Biomass	1	38423605	7.22E-07	Stress shoot length, SES	Shoot length, shoot to root ratio and potassium concentration in salt stress [Leon et al. 2016]; root weight under phosphate starvation [Mai et al. 2020]; shoot weight, plant weight [To et al 2019]; shoot length, shoot length under zinc stress [Zhang et al. 2016]; shoot height in control and under aluminium stress [Tao et al. 2018]; root length under salt treatment, relative dry shoot weight, relative biomass [Zhang et al. 2020]
qCDP3.1	Salt tolerance	3	1528621	2.65E-08	Shoot potassium
qCDP5.1	Salt tolerance	5	17632907	2.26E-07	Lost shoot length, lost shoot weight	Root length under phosphate deprivation [Mai et al. 2020]; total dry weight under salt stress [Yu et al. 2017]
qCDP6.1	Salt tolerance	6	20760852	2.23E-06	Stress shoot length, stress root thickness	Salt injury score, plant dry weight under salt stress [Leon et al. 2016]
qCDP7.1	Salt tolerance	7	21913935	1.05E-06	Stress root thickness, stress shoot thickness	Seed imbibition under salt stress [Cui et al. 2018]; relative maximum leaf length in salt stress [Frouin et al. 2018]; biomass, panicle dry weight, grain yield [Diop et al. 2020]
qCDP8.1	Salt tolerance	8	27323749	2.47E-06	Lost shoot weight	Root length in salt stress[Leon et al. 2016]; spikely sterility [Dingkuhn et al. 2017]; resistance to bacterial blight [Delorean et al. 2016], potassium content and sodium-potassium ratio in salt stress [Frouin et al. 2018]
qCDP9.1	Salt tolerance	9	9438999	7.54E-06	Lost shoot length	Low temperature seedling survivability [Schläppi et al. 2017]
qCDP9.2	Salt tolerance	9	11455731	1.91E-06	Shoot potassium, chlorophyll A, chlorophyll B, lost root length, shoot sodium content	Seed germination rate under salt stress [Cui et al. 2018]
qCDP9.3	Salt tolerance	9	20273741	1.46E-06	Stress shoot thickness, stress shoot weight, stress chlorophyll A
qCDP10.1	Salt tolerance	10	18641874	1.07E-06	Stress root potassium, root weight	SSI score, green leaves, shoot biomass, root biomass in salt stress [Rohila et al. 2019]
qCDP12.1	Biomass	12	2854998	6.51E-06	Shoot thickness	Shoot weight to length ratio [To et al. 2019]
qCDP12.2	Salt tolerance	12	16595682	5.50E-06	Stress root potassium, stress root sodium
qCDP12.3	Biomass	12	25769505	7.57E-06	Stress root length	Shoot length in salt stress [Leon et al. 2016]; low temperature seedling survivability and survival [Schläppi et al. 2017]

Open in a new tab

Fig 5 — Horizontal red and blue lines show the significant and suggestive thresholds respectively. Vertical black dotted lines show the locations of the three QTLs in the three Manhattan plots.

Fig 6 — Horizontal red and blue lines show the significant and suggestive thresholds respectively. Vertical black dotted lines show the locations of the five QTLs in the three Manhattan plots.

Fig 7 — Horizontal red and blue lines show the significant and suggestive thresholds respectively. Vertical black dotted lines show the locations of the three QTLs in the three Manhattan plots.

Fig 8 — Horizontal red and blue lines show the significant and suggestive thresholds respectively. Vertical black dotted lines show the locations of the two QTLs in the three Manhattan plots.

3.4 QTL mapping

The extent of linkage disequilibrium in rice has been found consistently in the range of 100kb to 200kb [44], although some authors suggest that it could extend over 500kb for some subpopulations [45]. In our experiments, LD decay to half maximal value of r² was observed at a distance of ~70kb in the whole population, ~100kb in the Aus subpopulation and ~50kb in the Indica subpopulation (S6 Fig). Based on these assumptions, we defined the range of our QTLs to be from 100kb upstream of the peak SNP up to 100kb downstream for association mapping. 313 markers showing significant and suggestive associations within the sequence range of known rice genes are described in S8 Table with metadata from the RAP-DB database [13]. All significant and suggestive markers from validation tests using the CMLM model can be found in S9 Table. We reviewed the annotations for genes situated in the vicinity of these significant associations for the elucidation of functional roles in regards to the implied phenotypes. From these annotations, 52 genes with notable functional roles corresponding to their associated phenotypes are included in the supplementary notes.

One significant marker from qCDP3.1 associated with shoot potassium content could be mapped to OsGSL10 (MSU ID: LOC_Os03g03610), a metabolic enzyme having 2-fold higher expression in shoots than any other tissue. In the same QTL, a cation antiporter associated to shoot potassium content, LOC_Os03g03590, could be mapped by a highly significant marker. This gene is also ubiquitously expressed with significantly higher expression in leaves and shoots. Based on its peptide sequence, the transporter protein is annotated to be localized in the chloroplast (Uniprot ID: Q10SB9). Within the same locus, two notable genes: OsCP29 (MSU ID: LOC_Os07g37240) and OsPetC (MSU ID: LOC_Os07g37030), are also localized in the chloroplast. Both these genes are expressed all throughout the plant in markedly higher expression levels and have almost 10-fold higher expression in leaf tissues. The presence of a chloroplast localized cation transporter adjacent to two particularly important chloroplast related genes in a gene locus identified through shoot potassium content suggests a role of this transporter protein and chloroplasts in the regulation of potassium ion content in plants. From the MSU7 brief loci annotations, genomic copies of the OsNAC6 gene (MSU ID: LOC_Os01g66120) which is known to have a functional role in plant development and stress tolerance, could be mapped to 4 identified QTLs: qCDP1.1 (LOC_Os01g66120), qCDP3.1 (LOC_Os03g03540), qCDP12.1 (LOC_Os12g05990) and qCDP12.3 (LOC_Os12g41680). qCDP1.1, qCDP12.1 and qCDP12.3 are QTLs belonging to our biomass category (Table 2). This is good evidence to suggest that OsNAC6 might have a prominent role in embryogenesis, germination and biomass gain in rice. The gene had also been implicated in stress response and abiotic stress tolerance which holds to its presence in the salt stress related QTLs: qCDP1.1, qCDP3.1 and qCDP12.3.

4. Discussion

4.1 Biomass traits and chlorophyll content could be valuable indices for the screening of salinity tolerance

There lies significant ambiguity in the extrapolation of salt tolerance mechanisms in rice [46]. Gregorio et al. 1997 [19] popularized visual injury scoring for the screening of salinity tolerance. The application of SES for assessment of salinity tolerance has gathered some criticism since then [14]. We found the SES score to have meaningful correlations with all observed traits except for ion content (S1 Fig). Our ion content traits are among the least heritable phenotypes (Table 1 and S2 Table) and their distributions deviate the most from a normal gaussian distribution (S2 Fig). They are also the most variable across genotypes and among replicates (S3 Table), least applicable for genomic prediction (S4 Fig) and the least correlated between two test conditions and with the other phenotypes, including the composite SES score (S1 Fig). There are however, significant near-binary differences in tissue sodium content for control and stress counterparts and the magnitude of this change greatly outweighs the variance observed across genotypes (S3 Table). This proved that in spite of being the most significant physiological alteration with the most deleterious metabolic consequences, changes in sodium content for individual genotypes between control and stress conditions could not be an informative variable for prediction models. This lack of correlation and lack of reliable difference across genotypes was previously observed by Pires et al. 2015 [14]. Pires et al. also found that potassium content in roots and leaves were not significantly altered between control and stress group plants and thus, could not be a reliable predictor of salt tolerance in statistical models. There has already been ample criticism against the traditional application of the K⁺/Na⁺ ratio for evaluation of stress tolerance [14, 47] and our observations incurred further scrutiny to the conventional practice of evaluating salinity tolerance across different genotypes using tissue ion content and K⁺/Na⁺ ratios. For these reasons, we did not derive any further phenotypes from sodium and potassium content traits.

Although ion content is linearly independent to salt injury, very high correlation between sodium and potassium content in roots under salt stress is observed (S1 Fig). The lack of this correlation in shoots under stress conditions suggests that sodium exclusion from plant leaves might be critical for survival in salt stress. Chlorophyll A and B are not independent variables and although relatively less heritable, show high degree of correlation in both control and stress conditions and correlate strongly to salt injury. This was reported in Pires et al. 2015 and our findings confirm their observations. The relationship observed between biomass traits and salinity tolerance in our studies is noteworthy. It is always the combination of a number of salt tolerance strategies that a plant employs to cope with stress. Intuitively, salinity tolerance benchmarks are depicted by the ability of plants to accumulate and retain biomass and color. We observed that proportional losses of shoot and root biomass is instilled by salt stress. Whereas plant genotypes that had more chlorophyll in control conditions, did not retain more chlorophyll in stress conditions (Fig 3). Hence for chlorophyll, how much a plant has under normal conditions is not much relevant to how much a plant of the same genotype could maintain under salt stress conditions. How much biomass a plant of certain genotype retains in salt stress is however, linearly correlated to how much biomass it can put on in control conditions. Greater biomass in stress conditions correlates strongly and negatively to damage scores. It was also found by other researchers that the effect of salinity on biomass is more predominant in the shoots compared to roots [48]. It has previously been shown that response to salt stress in the tolerant landrace Pokkali takes place swiftly at the transcriptional level, whereas there is a delayed response in sensitive IR29 [49]. The homeostatic adjustments in tolerant plants allows them to continue growth, photosynthesize and ultimately set grains [50]. In light of these observations, and on the basis of GWAS results, we place much emphasis on biomass traits and chlorophyll content for the assessment of salinity tolerance in rice.

4.2 Some Bangladeshi landraces exhibit high degrees of salt tolerance

Two local endemic landraces, Kala Digha (IRIS_313–10973) and Kala Muna (IRIS_313–11115) are known to be highly salt tolerant. Similar degrees of tolerance were shown by Lal Digha (IRIS_313–11487) and Patnai 31–679 (IRIS_313–11224). The salinity tolerance of these accessions was in par with that of the tolerant landrace Pokkali (IRIS_313–8244) which is routinely cultivated in the coastal areas of Sri Lanka and Southern India. All of these accessions belong to the Ind2 subpopulation in the 3KRG database. The majority of our Ind2 plants are moderately or highly salt tolerant whereas the sensitive plants mostly belong to the Aus subpopulation (S3 Table). The varieties Chitraj (IRIS_313–11220), Kola Muchi (IRIS_313–10975) and Patnai 31–675 (IRIS_313–9049) stand out by having relatively high shoot biomass in spite of being highly salt sensitive. Kortik Kaika (IRIS_313–11116), Kal Shuli (IRIS_313–11113) and Pokkali (IRIS_313–8244) had the highest shoot biomass in stress conditions. The traditional varieties were overall more tolerant than the developed varieties. This is expected because developed varieties are generally bred and selected for yield traits rather than their wildtype characteristics that confer tolerance to environmental stress. In many places across the country where drought is common and along the southern coast where soil is saline, these local landraces are predominantly farmed as they perform better than the sensitive elite varieties.

4.3 Potential genes of interest were found in the identified QTLs

Studies on previously published rice gene expression data and functional annotations highlight a number of genes in the identified QTLs (S1 Text). Expression data of annotated genes within the QTLs taken from MSU7 are provided in S10 Table and processed RNA-seq data from the roots and shoots of a representative tolerant rice variety, Baldo and a sensitive variety, Vialone Nano adapted from Formentin et al. 2018 [51] are given in S11 Table. A heatmap of differentially expressed genes is shown in S7 Fig (see S1 Text for details). Effects of the natural polymorphisms at the protein level can be found in S12 Table. Significant differences in phenotypes between the variant alleles are observed by ANOVA in S13 Table and by the Student’s t-test in S14 Table.

Variant alleles of the chloroplast transporter gene discussed in the QTL mapping section (section 3.4), LOC_Os03g03590 found in qCDP3.1 did not exhibit significant differences in phenotype. OsPRX86 is a peroxidase precursor protein belonging to the QTL qCDP6.1. The locus has been previously found to be associated with salt injury scores in rice [10]. A polymorphism causing a nonsense consequence in the gene was found to affect multiple salinity tolerance traits, namely, SES score, shoot length under stress and normal conditions and shoot weight. Another unannotated gene, LOC_Os09g15389.2, belonging to another previously identified locus, qCDP9.1 [52], was observed to affect salt injury score, shoot weight and shoot length when the protein is truncated. OsPIL (MSU ID: LOC_Os12g41650.4) is a phytochrome interacting helix-loop-helix protein gene present in qCDP12.3 that has 5 to 10-fold greater expression levels in plant leaves, and is known to play a part in abiotic stress response [53]. Our observed SNP only affects transcript 4 out of 5 known transcripts of this gene. LOC_Os09g19160 is an uncharacterized serine/threonine kinase from qCDP9.2, that belongs to a cluster housing another serine/threonine kinase gene, LOC_Os09g19229. These kinases were found to affect root length in normal and stress conditions and also show significant differential expression under control and stress conditions. Similar kinases have been observed to play an important part in root development and nodulation [54]. An SNP variation in the 6-phosphogluconolactonase gene from qCDP8.1, LOC_Os08g43370.1 disrupted root length and shoot weight in stress condition. This gene also exhibits significant differential expression under control and stress conditions. These genes appear to be promising subjects for discrete functional studies to find any contributions they might have in abiotic stress tolerance.

5. Conclusion

We have mapped 13 QTLs for various traits including 3 novel QTLs. The novel QTLs qCDP3.1 and qCDP12.2 are associated with tissue ion content and qCDP9.3 is associated with plant biomass and chlorophyll content under stress conditions. The 10 other identified QTLs had previously been reported in literature against identical quantitative traits. When mapped to the genome, these QTLs were found to harbor many genes implicated in abiotic stress response and tolerance. Supplementary analysis of gene expression and functional annotations assumes potential roles for a number of genes within the identified QTLs. A number of accessions from Bangladesh display high degree of salt tolerance. Some resilient and notable landraces include Kala Digha (IRIS_313–10973), Kala Muna (IRIS_313–11115) Lal Digha (IRIS_313–11487) and Patnai 31–679 (IRIS_313–11224). Although it is well established that biomass and chlorophyll related traits alongside tissue ion content aids the assessment of salinity tolerance, our studies show that the numerical values of plant biomass and chlorophyll content could manifest themselves into an objective index of salt injury unabated by human error.

GWAS studies are convenient protocols for novel genetic discoveries. They have become even more popular with the subsistence and growth of publicly available genomic data. The automation of genomic studies is more relevant today because of recent advancements in mechanical and drone-based phenotype imaging technologies. Public documentation of our experimental design on GitHub will assist future researchers and facilitate GWAS and other studies in plant genomics.

Supporting information

S1 Text. Gene expression and functional studies within the identified QTLs.

(DOCX)

Click here for additional data file.^{(40.4KB, docx)}

S1 Fig. Extended correlation matrix of observed traits.

(JPG)

Click here for additional data file.^{(5.3MB, jpg)}

S2 Fig. Density plots for basic and derived phenotypes.

(JPG)

Click here for additional data file.^{(809.5KB, jpg)}

S3 Fig. Changes in phenotype distributions after applying an ideal warping function.

(JPG)

Click here for additional data file.^{(324.5KB, jpg)}

S4 Fig. Residual plots for basic and derived phenotypes.

(JPG)

Click here for additional data file.^{(1.1MB, jpg)}

S5 Fig. Collage of Manhattan plots for traits with biomass associations.

(JPG)

Click here for additional data file.^{(7.4MB, jpg)}

S6 Fig. Observed extent of linkage disequilibrium.

(JPG)

Click here for additional data file.^{(300.1KB, jpg)}

S7 Fig. Gene expression heatmaps from S1 Text.

(JPG)

Click here for additional data file.^{(4.3MB, jpg)}

S1 Table. Data on vegetation and population structure of 176 plants from the study.

(XLSX)

Click here for additional data file.^{(24KB, xlsx)}

S2 Table. Heritability of traits calculated from genomic kinship and broad-sense heritability calculated from variance components.

(XLSX)

Click here for additional data file.^{(11.1KB, xlsx)}

S3 Table. Observed phenotype means for each genotype.

(XLSX)

Click here for additional data file.^{(69.6KB, xlsx)}

S4 Table. Phenotype values for genotypes estimated by genomic prediction.

(XLSX)

Click here for additional data file.^{(91.5KB, xlsx)}

S5 Table. P values for entire correlation matrix in S1 Fig.

(XLSX)

Click here for additional data file.^{(27.6KB, xlsx)}

S6 Table. Details on phenotype distributions before and after filtering.

(XLSX)

Click here for additional data file.^{(12.4KB, xlsx)}

S7 Table. Details on QTLs from previous studies that were found to have similar connotations as our findings.

(XLSX)

Click here for additional data file.^{(11.9KB, xlsx)}

S8 Table. All significant and suggestive marker associations that fall within the sequence range of known genes.

(XLSX)

Click here for additional data file.^{(59.3KB, xlsx)}

S9 Table. All significant and suggestive marker association within known gene regions from validation tests by the CMLM model in GAPIT.

(XLSX)

Click here for additional data file.^{(58.6KB, xlsx)}

S10 Table. Gene expression data for candidate genes.

(XLSX)

Click here for additional data file.^{(85KB, xlsx)}

S11 Table. Gene expression data from leaves and roots of a sensitive and a tolerant variety.

(XLSX)

Click here for additional data file.^{(172.6KB, xlsx)}

S12 Table. Effects of all 28188 possible substitution events within candidate gene regions.

(XLSX)

Click here for additional data file.^{(2MB, xlsx)}

S13 Table. Results of One-way ANOVA carried out for all markers within candidate gene regions.

(XLSX)

Click here for additional data file.^{(223.6KB, xlsx)}

S14 Table. Results of student’s t test carried out for all markers within candidate gene regions.

(XLSX)

Click here for additional data file.^{(972KB, xlsx)}

Acknowledgments

The authors would like to acknowledge Dr. Sabrina M. Elias from Independent University, Bangladesh and MU Sharif Shohan from the University of Dhaka for their valuable insights and guidance throughout the project. We acknowledge the Center for Bioinformatics Learning Advancement and Systematic Training (cBLAST) from the University of Dhaka and the University of Dhaka itself for enabling the necessary computational and experimental resources. We acknowledge Dr. Md. Sazzadur Rahman from Bangladesh Rice Research Institute (BRRI) for his assistance in multiplying seeds and preparing the germplasm. We also acknowledge the members of the Plant Biotechnology Lab at the University of Dhaka with special mention to Rabin Sarker and Raju Ahmed for help in handling plant material, screening for salinity tolerance and in enumerating and tabulating wet lab data. Finally, we acknowledge Anika Tahsin from the University of Dhaka for help in retrieving geolocational origins of the studied plant varieties and being a constant source of inspiration.

Data Availability

All data and source code can be found at https://github.com/DeadlineWasYesterday/Cat-Does-Plant/.

Funding Statement

The author(s) received no specific funding for this work.

References

1.German Environment Agency, F.M.f.E.C.a.D., Ten million hectares of arable land worldwide are ’lost’ every year: less and less fertile and healthy soil. 2015, Umweltbundesamt: Germany https://www.umweltbundesamt.de/en/press/pressinformation/ten-million-hectares-of-arable-land-worldwide-are.
2.Olesen J.E., et al., Impacts and adaptation of European crop production systems to climate change. European Journal of Agronomy, 2011. 34(2): p. 96–112. [Google Scholar]
3.Gupta G.S.J.R.E.S., Land degradation and challenges of food security. 2019. 11: p. 63. [Google Scholar]
4.Hossain M.S.J.I.R.J.B.S., Present scenario of global salt affected soils, its management and importance of salinity research. 2019. 1: p. 1–3. [Google Scholar]
5.Eynard A., Lal R., and Wiebe K., Crop Response in Salt-Affected Soils. Journal of Sustainable Agriculture, 2005. 27(1): p. 5–50. [Google Scholar]
6.Tsioumani E., The State of the World’s Biodiversity for Food and Agriculture: a Call to Action? Environmental Policy and Law, 2019. 49: p. 110–112. [Google Scholar]
7.Solis C.A., et al., Back to the Wild: On a Quest for Donors Toward Salinity Tolerant Rice. 2020. 11(323). [DOI] [PMC free article] [PubMed] [Google Scholar]
8.McCouch S., et al., Feeding the future. 2013. 499(7456): p. 23–24. [DOI] [PubMed] [Google Scholar]
9.Kromdijk J. and Long S., One crop breeding cycle from starvation? How engineering crop photosynthesis for rising CO 2 and temperature could be one important route to alleviation. Proceedings of the Royal Society B: Biological Sciences, 2016. 283: p. 20152578. doi: 10.1098/rspb.2015.2578 [DOI] [PMC free article] [PubMed] [Google Scholar]
10.De Leon T.B., Linscombe S., and Subudhi P.K., Molecular Dissection of Seedling Salinity Tolerance in Rice (Oryza sativa L.) Using a High-Density GBS-Based SNP Linkage Map. Rice, 2016. 9(1): p. 52. doi: 10.1186/s12284-016-0125-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Yu J., et al., Genome-wide association study and gene set analysis for understanding candidate genes involved in salt tolerance at the rice seedling stage. Molecular Genetics and Genomics, 2017. 292(6): p. 1391–1403. doi: 10.1007/s00438-017-1354-9 [DOI] [PubMed] [Google Scholar]
12.Zhao Y., et al., New alleles for chlorophyll content and stay-green traits revealed by a genome wide association study in rice (Oryza sativa). Scientific Reports, 2019. 9(1): p. 2541. doi: 10.1038/s41598-019-39280-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
13.Kawahara Y., et al., Improvement of the Oryza sativa Nipponbare reference genome using next generation sequence and optical map data. Rice, 2013. 6(1): p. 4. doi: 10.1186/1939-8433-6-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Pires I.S., et al., Comprehensive phenotypic analysis of rice (Oryza sativa) response to salinity stress. 2015. 155(1): p. 43–54. [DOI] [PubMed] [Google Scholar]
15.Mansueto L., et al., Rice SNP-seek database update: new SNPs, indels, and queries. Nucleic acids research, 2017. 45(D1): p. D1075–D1081. doi: 10.1093/nar/gkw1135 [DOI] [PMC free article] [PubMed] [Google Scholar]
16.annieart0, Free Bangladesh Map Illustration. https://www.vecteezy.com/vector-art/122269-free-bangladesh-map-illustration.
17.Amin M., et al., Over-expression of a DEAD-box helicase, PDH45, confers both seedling and reproductive stage salinity tolerance to rice (Oryza sativa L.). Molecular Breeding, 2012. 30(1): p. 345–354. [Google Scholar]
18.Cock J., Yoshida S., and Forno D.A., Laboratory manual for physiological studies of rice. 1976: Int. Rice Res. Inst. [Google Scholar]
19.Gregorio G., Senadhira D., and Mendoza R., Screening rice for salinity tolerance, vol 22, IRRI discussion paper series. International Rice Research Institute, 1997. [Google Scholar]
20.IRRI I.J.I.R.R.I., Philippine, Standard evaluation system for rice. 2002: p. 1–45. [Google Scholar]
21.Endelman J.B., Ridge Regression and Other Kernels for Genomic Selection with R Package rrBLUP. 2011. 4(3): p. 250–255. [Google Scholar]
22.Fusi N., et al., Warped linear mixed models for the genetic analysis of transformed phenotypes. Nature Communications, 2014. 5(1): p. 4890. doi: 10.1038/ncomms5890 [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Browning B.L., Zhou Y., and Browning S.R., A One-Penny Imputed Genome from Next-Generation Reference Panels. The American Journal of Human Genetics, 2018. 103(3): p. 338–348. doi: 10.1016/j.ajhg.2018.07.015 [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Raj A., Stephens M., and Pritchard J.K., fastSTRUCTURE: Variational Inference of Population Structure in Large SNP Data Sets. 2014. 197(2): p. 573–589. [DOI] [PMC free article] [PubMed] [Google Scholar]
25.Bradbury P.J., et al., TASSEL: software for association mapping of complex traits in diverse samples. Bioinformatics, 2007. 23(19): p. 2633–2635. doi: 10.1093/bioinformatics/btm308 [DOI] [PubMed] [Google Scholar]
26.Letunic I. and Bork P., Interactive Tree Of Life (iTOL): an online tool for phylogenetic tree display and annotation. Bioinformatics, 2007. 23(1): p. 127–128. doi: 10.1093/bioinformatics/btl529 [DOI] [PubMed] [Google Scholar]
27.Huang M., et al., BLINK: a package for the next level of genome-wide association studies with both individuals and markers in the millions. GigaScience, 2019. 8(2). doi: 10.1093/gigascience/giy154 [DOI] [PMC free article] [PubMed] [Google Scholar]
28.Tang Y., et al., GAPIT version 2: an enhanced integrated tool for genomic association and prediction. 2016. 9(2): p. 1–9. [DOI] [PubMed] [Google Scholar]
29.Turner S.D.J.B., qqman: an R package for visualizing GWAS results using QQ and manhattan plots. 2014: p. 005165. [Google Scholar]
30.Chang C.C., et al., Second-generation PLINK: rising to the challenge of larger and richer datasets. GigaScience, 2015. 4(1). doi: 10.1186/s13742-014-0042-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
31.Mai N.T.P., et al., Discovery of new genetic determinants controlling the morphological plasticity in rice root and shoot under phosphate starvation using GWAS. 2020: p. 2020.10.31.363556. [DOI] [PubMed] [Google Scholar]
32.To H.T.M., et al., Unraveling the Genetic Elements Involved in Shoot and Root Growth Regulation by Jasmonate in Rice Using a Genome-Wide Association Study. Rice, 2019. 12(1): p. 69. doi: 10.1186/s12284-019-0327-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
33.Zhang J., et al., QTL mapping and candidate gene analysis of ferrous iron and zinc toxicity tolerance at seedling stage in rice by genome-wide association study. BMC Genomics, 2017. 18(1): p. 828. doi: 10.1186/s12864-017-4221-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
34.Tao Y., et al., Genome-wide association mapping of aluminum toxicity tolerance and fine mapping of a candidate gene for Nrat1 in rice. PLOS ONE, 2018. 13(6): p. e0198589. doi: 10.1371/journal.pone.0198589 [DOI] [PMC free article] [PubMed] [Google Scholar]
35.Zhang Y., et al., QTL identification for salt tolerance related traits at the seedling stage in indica rice using a multi-parent advanced generation intercross (MAGIC) population. Plant Growth Regulation, 2020. 92(2): p. 365–373. [Google Scholar]
36.Cui Y., Zhang F., and Zhou Y., The Application of Multi-Locus GWAS for the Detection of Salt-Tolerance Loci in Rice. 2018. 9(1464). [DOI] [PMC free article] [PubMed] [Google Scholar]
37.Frouin J., et al., Tolerance to mild salinity stress in japonica rice: A genome-wide association mapping study highlights calcium signaling and metabolism genes. PLOS ONE, 2018. 13(1): p. e0190964. doi: 10.1371/journal.pone.0190964 [DOI] [PMC free article] [PubMed] [Google Scholar]
38.Diop B., et al., Bridging old and new: diversity and evaluation of high iron-associated stress response of rice cultivated in West Africa. Journal of Experimental Botany, 2020. 71(14): p. 4188–4200. doi: 10.1093/jxb/eraa182 [DOI] [PubMed] [Google Scholar]
39.Dingkuhn M., et al., Crop-model assisted phenomics and genome-wide association study for climate adaptation of indica rice. 2. Thermal stress and spikelet sterility. Journal of Experimental Botany, 2017. 68(15): p. 4389–4406. doi: 10.1093/jxb/erx250 [DOI] [PubMed] [Google Scholar]
40.Delorean, E.E.J.-.-C.T. and Dissertations, Detecting durable resistance to rice bacterial blight. 2016.
41.Schläppi M.R., et al., Assessment of Five Chilling Tolerance Traits and GWAS Mapping in Rice Using the USDA Mini-Core Collection. 2017. 8(957). [DOI] [PMC free article] [PubMed] [Google Scholar]
42.Rohila J.S., et al., Identification of Superior Alleles for Seedling Stage Salt Tolerance in the USDA Rice Mini-Core Collection. 2019. 8(11): p. 472. [DOI] [PMC free article] [PubMed] [Google Scholar]
43.Misra G., et al., Whole genome sequencing-based association study to unravel genetic architecture of cooked grain width and length traits in rice. Scientific Reports, 2017. 7(1): p. 12478. doi: 10.1038/s41598-017-12778-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
44.Huang X., et al., Genome-wide association studies of 14 agronomic traits in rice landraces. Nature Genetics, 2010. 42(11): p. 961–967. doi: 10.1038/ng.695 [DOI] [PubMed] [Google Scholar]
45.Mather K.A., et al., The extent of linkage disequilibrium in rice (Oryza sativa L.). 2007. 177(4): p. 2223–2232. [DOI] [PMC free article] [PubMed] [Google Scholar]
46.Bhowmik S., et al., Identification of salt tolerant rice cultivars via phenotypic and marker-assisted procedures. 2007. 10(24): p. 4449–4454. [DOI] [PubMed] [Google Scholar]
47.GARCIA A., et al., Sodium and potassium transport to the xylem are inherited independently in rice, and the mechanism of sodium: potassium selectivity differs between rice and wheat. 1997. 20(9): p. 1167–1174. [Google Scholar]
48.Negrão S., et al., Recent updates on salinity stress in rice: from physiological to molecular responses. 2011. 30(4): p. 329–377. [Google Scholar]
49.Golldack D., et al., Salinity stress-tolerant and -sensitive rice (Oryza sativa L.) regulate AKT1-type potassium channel transcripts differently. Plant Mol Biol, 2003. 51(1): p. 71–81. doi: 10.1023/a:1020763218045 [DOI] [PubMed] [Google Scholar]
50.Kawasaki S., et al., Gene expression profiles during the initial phase of salt stress in rice. Plant Cell, 2001. 13(4): p. 889–905. doi: 10.1105/tpc.13.4.889 [DOI] [PMC free article] [PubMed] [Google Scholar]
51.Formentin E., et al., Transcriptome and Cell Physiological Analyses in Different Rice Cultivars Provide New Insights Into Adaptive and Salinity Stress Responses. Frontiers in Plant Science, 2018. 9(204). doi: 10.3389/fpls.2018.00204 [DOI] [PMC free article] [PubMed] [Google Scholar]
52.Schläppi M.R., et al., Assessment of Five Chilling Tolerance Traits and GWAS Mapping in Rice Using the USDA Mini-Core Collection. Frontiers in Plant Science, 2017. 8(957). doi: 10.3389/fpls.2017.00957 [DOI] [PMC free article] [PubMed] [Google Scholar]
53.Jeong J. and Choi G., Phytochrome-interacting factors have both shared and distinct biological roles. Molecules and cells, 2013. 35(5): p. 371–380. doi: 10.1007/s10059-013-0135-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
54.Krusell L., et al., Shoot control of root development and nodulation is mediated by a receptor-like kinase. Nature, 2002. 420(6914): p. 422–6. doi: 10.1038/nature01207 [DOI] [PubMed] [Google Scholar]

PLoS One. doi: 10.1371/journal.pone.0259456.r001

Decision Letter 0

Prasanta K Subudhi

1 Jul 2021

PONE-D-21-14535

Comprehensive analysis and genome-wide association studies of biomass, chlorophyll, seed and salinity tolerance related traits in rice highlight genetic hotspots for crop improvement

PLOS ONE

Dear Dr. Seraj,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Specifically, both reviewers expressed concern regarding English language, redundancy, and some technical issues in this manuscript. Validation of few candidate genes is also suggested. I suggest to make a thorough revision with reorganization of contents without redundancy to make the manuscript reader-friendly.

Please submit your revised manuscript by Aug 14 2021 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.
A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.
An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Prasanta K. Subudhi, Ph.D.

Academic Editor

PLOS ONE

Additional Editor Comments:

Major revision

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. We note that Figure 1 in your submission contain map images which may be copyrighted. All PLOS content is published under the Creative Commons Attribution License (CC BY 4.0), which means that the manuscript, images, and Supporting Information files will be freely available online, and any third party is permitted to access, download, copy, distribute, and use these materials in any way, even commercially, with proper attribution. For these reasons, we cannot publish previously copyrighted maps or satellite images created using proprietary data, such as Google software (Google Maps, Street View, and Earth). For more information, see our copyright guidelines: http://journals.plos.org/plosone/s/licenses-and-copyright.

We require you to either (1) present written permission from the copyright holder to publish these figures specifically under the CC BY 4.0 license, or (2) remove the figures from your submission:

2.1. You may seek permission from the original copyright holder of Figure 1 to publish the content specifically under the CC BY 4.0 license.

We recommend that you contact the original copyright holder with the Content Permission Form (http://journals.plos.org/plosone/s/file?id=7c09/content-permission-form.pdf) and the following text:

“I request permission for the open-access journal PLOS ONE to publish XXX under the Creative Commons Attribution License (CCAL) CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). Please be aware that this license allows unrestricted use and distribution, even commercially, by third parties. Please reply and provide explicit written permission to publish XXX under a CC BY license and complete the attached form.”

Please upload the completed Content Permission Form or other proof of granted permissions as an "Other" file with your submission.

In the figure caption of the copyrighted figure, please include the following text: “Reprinted from [ref] under a CC BY license, with permission from [name of publisher], original copyright [original copyright year].”

2.2. If you are unable to obtain permission from the original copyright holder to publish these figures under the CC BY 4.0 license or if the copyright holder’s requirements are incompatible with the CC BY 4.0 license, please either i) remove the figure or ii) supply a replacement figure that complies with the CC BY 4.0 license. Please check copyright information on all replacement figures and update the figure caption with source information. If applicable, please specify in the figure caption text when a figure is similar but not identical to the original image and is therefore for illustrative purposes only.

The following resources for replacing copyrighted map figures may be helpful:

USGS National Map Viewer (public domain): http://viewer.nationalmap.gov/viewer/

The Gateway to Astronaut Photography of Earth (public domain): http://eol.jsc.nasa.gov/sseop/clickmap/

Maps at the CIA (public domain): https://www.cia.gov/library/publications/the-world-factbook/index.html and https://www.cia.gov/library/publications/cia-maps-publications/index.html

NASA Earth Observatory (public domain): http://earthobservatory.nasa.gov/

Landsat: http://landsat.visibleearth.nasa.gov/

USGS EROS (Earth Resources Observatory and Science (EROS) Center) (public domain): http://eros.usgs.gov/#

Natural Earth (public domain): http://www.naturalearthdata.com/

3. Please include captions for your Supporting Information files at the end of your manuscript, and update any in-text citations to match accordingly. Please see our Supporting Information guidelines for more information: http://journals.plos.org/plosone/s/supporting-information.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

Reviewer #2: Partly

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: I Don't Know

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: No

Reviewer #2: No

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: The manuscript by Alam et al. titled “Comprehensive analysis and genome-wide association studies of biomass, chlorophyll, seed and salinity tolerance related traits in rice highlight genetic hotspots for crop improvement” describes a GWAS mapping analysis of 15 phenotypes yielding 17 QTL and list of 21 associated candidate genes using 176 mostly land race accessions from Bangladesh.

The general information provided by this manuscript and the QTL and candidate gene list are potentially useful for crop improvement of elite rice varieties challenged by abiotic stress such as high salinity. However, the manuscript is too long, needs to be streamlined to make it more reader friendly, and authors have not validated any of the 21 candidate genes. Below are specific comments that need to be addressed.

Line 58: “No recent GWAS study..” is too absolute as a statement; modify to “Few recent GWAS studies…”

Line 122: Change “Seed husks” to “Seed hulls”? Hulls is more commonly used than husks.

Lines 244-245 & Fig. 2a: what is the “blue” subpopulation? Is it aromatic?

Lines 285-286: what do you mean by “fortification of trait correlations in salt stress conditions relative to control conditions”? It is not clear how e.g. C1 compared to S1 in Fig. 3 allows you to come to this conclusion: what exactly is compered here? Fig. 3 is confusing in other ways: what do you mean by “Linear correlations between observed and predicted phenotypes”? How do you “predict” phenotypes? A genomic prediction method is mentioned on lines 149-141, but it is not clear how this SNP effect on phenotype is used for the Pearson’s correlation shown in Fig. 3.

Lines 262: since seed traits were only measured under control conditions, it is not clear how this data set is useful for abiotic stress tolerance improvement. To streamline the paper, please remove those data unless you can provide a good explanation in the manuscript for having them.

Lines 295-325: Some of the results section should be moved to the methods section or only briefly mentioned. The data transformation part to achieve normal distributions before GWAS analysis should be described in methods and Fig. 5 be included as a supplementary figure to streamline the manuscript.

Other result sections also contain quite a bit of methods descriptions that should be moved from Results to Methods.

Section 3.3., GWAS analysis, should be streamlined (again, do not describe methods here). The naming of QTL could be improved: instead of giving CDP consecutive numbers that seem to random, it would be useful to label a QTL on chromosome 1 as pCDP1-x etc (-x if there are more than one QTL on the same chromosome). Also, what does “CDP” stand for?

Fig. 10: heat map colors need to be quantified.

Lines 453-458: Should not refer to SNPs as “mutations”, b/c these are naturally or artificially selected variants

Section 4.1 (lines 479-494) does not contribute much to the discussion and could be removed or modified. What is the main message the authors what to convey?

Lines 612-613: Os09g19160 is mentioned twice but should be different genes.

Gene expression data were from publicly available RNAseq data only and not validated by the authors. Gene expression should be determined under both control and salt stress conditions for selecting the best candidate genes for crop improvement. Moreover, just focusing on genes with synonymous aa substitutions or frameshifts in some accessions might not be the best way to select candidate genes because they could be differentially expressed between stress tolerant and stress sensitive accessions. To improve the manuscript, validate a few candidate genes by qPCR in a selected stress tolerant and a selected stress sensitive accession.

This reviewer is also a bit confused with some aspects of the manuscript: wasn’t the main objective to identify salt tolerance QTL by looking at biomass and chlorophyll content under salt stress and perhaps looking at the relative reduction of both compared to control? So why the long discussion in section 4.3 about the chr. 3 chlorophyll QTL found under control conditions?

Lastly, the section 5 (Conclusion) could be improved: it is too long and should summarize the main points of the manuscript and should not read like an additional discussion.

Reviewer #2: The manuscript explains a GWAS study coupled with an extensive candidate gene analysis. The amount of effort made by authors is appreciable. However, there are serious concerns about the study and presentation. Some of my comments and suggestions are as below:

General comments:

1. Nowhere in result section authors have mentioned about novel QTLs identified in the study (they discussed about this in abstract & conclusion). To me this should be their key finding which is not explained properly.

2. Authors claimed a number of times about functional role of genes without experimental evidences which is not appropriate.

3. Results have not been explained, instead just the tables were referred in most of cases.

4. There is too much redundancy in write up, result section is full of material-methods,

5. Complex language is big problem in this MS. Authors should write simple language for science but they rather chosen very complex and idiomatic language. Some of the examples are: “handful of individuals”, “environment embellish immense diversity”, “scientists endeavor to elucidate”, “interactions with the surroundings underlie the quantitative assertion of these phenotypes”, “But such is customary to plant breeding experiments and many known and unknown factors are in play here.”, “contra-intuitive for practical purposes”, “an elegant-demonstration”, “Intuitively, salinity tolerance benchmarks are depicted by the ability of plants to accumulate and retain biomass and color.”, “It did not escape our notice”, “inspires us to suggest a novel role for the gene in seed dimension regulation”, “thrive in substandard land without any compromise in yield”, “we attempt to rapidly sieve these desirable attributes into our modern cultivars”, “These observations are however, not inconceivable.”, “firm belief” and many more.

Specific comments:

Line 58-59: “No recent GWAS study explores abiotic stress tolerance in relation to developmental and agronomic phenotypes in rice.” – there are plenty of GWAS studies on agronomic traits, presume that agronomic phenotypes relates to agronomic traits. What author means developmental phenotype is probably evolutionary basis but how does this MS linked to GWAS for the evolutionary traits? Didn’t find any evolutionary in-depth analysis in this MS except grouping which is very common.

Line 62 – 64: CRISPR is very much functional but with targeted gene insertion or knock-outs. Is there any example where QTL that carries multiple genes spanning very long DNA fragment substituted the alternate allele through CRISPR and worked? Or it is just hypothesis?

Line 83: What do you mean by ‘176 rice accessions ordered from IRRI seed bank’? Author can say got the accessions or received from IRRI gene bank.

Line 84-85: “BRRI (Bangladesh Rice Research Institute) fields where the day/night temperature was 32°/28°C”. Is it possible to have the same temperature in field throughout cropping season?

Line 97-98: Screening for salinity tolerance at seedling stage was carried out as per Amin et al (2012). This article is not an open access article hence could not go through the details but as mentioned in M & M, germinated seeds grown for 14 days the salt stress applied in 5 increments from 4 to 12 dS/m. Then phenotypic measurements were made after 16 days from 4 dS/m treatment (30 days). SES scores and other parameters were taken as described by Gregorio et al 1997. The screening as per Amin et al and measurement as per Gregorio et al are two different things and will produce two different scenarios. Gregorio used to treat the plant when they are really at seedling and single tiller stage (first treatment of 6 dS/m at 6th days from germination and raised to 12 dS/m on 9th day). Once the seedling becomes big enough (started multiple tillering), the response to salinity becomes different, so not sure why authors used old seedlings to screen when they already start attaining the tolerance to salinity.

Line 117 and 118: What are the details of the formulae to measure the chlorophyll? Volume in ml or l? Weight (fresh or dry) in, mg or g ? Absorbance etc. Explain each formula.

Line 153: 2,930,739 missing ‘genotypes’? – it is typo

Line 159-161: Why MAF was chosen to 10% for subpopulations : Should be explained.

Line 197-199: Authors have discussed a lot about the FDR (through Bonferroni), however, more important is the effect of population structure on associations. The way authors have explained above about the model for testing marker effects (in two equations), authors can explain OR simplify the explanation about population structure correction.

Results: For presenting the results, authors have just referred the tables/ figures and in text there is too much redundancy with material-methods. Result section needs to be re-organized in way it befits for a research paper, explaining the crystal-clear results ONLY. Say for example line 226 to almost 239 (except 236 and 237), is superfluous (“largest delta on Earth with bountiful rivers draining into the northeastern part of the Indian Ocean by way of the Bay of Bengal”) or part of M & M. There are plenty of instances where information fit for ‘M & M’ are written in ‘Results’ like 245-247, 254-257, 259-265, 327-337 and so on).

Line 274 – 276: Why it is so, please explain: “Narrow-sense heritability could not be calculated from additive marker effects because the grand mean of the phenotypes could explain more variance than the mixed.solve prediction model”

Line 306: What is bulky seed in term of their shape and size? Heard for the first time.

Line 361: Authors need to explain simply that what novel QTLs/ association they found and what was their effect. Their relation with candidate genes can be explained in next section. Could not find the information about the novel QTL in this section (QTL mapping) which could be the most important basis for acceptance of article for publication.

Line 377: What is trait like seed height? I have heard about seed length but not seed height? Are they same? If they are same, then why mentioned as “We previously found seed length to have

no correlation with seed height” (line 607-698).

Line 468-475: This can go as the legend or the footnote of the table.

Line 481-487: “Individual differences in plants of identical genotype are disturbingly disorderly. We found that variance within genotypes in our growth experiments were distinctly large for most traits other than seed properties. But such is customary to plant breeding experiments and many known and unknown factors are in play here. The question here is not how we can remove these confounding variables, since it is not only impossible but also contra-intuitive for practical purposes, but rather how we can derive informative features from observable phenotypes.”. What authors want to convey is not clear. The message should be simple and clear. ‘Individual differences in plants of identical genotype’ : It is not new when you are working with landraces. I am fail to understand the message out of this.

Line 493 – 494: This is well establish fact, what is new in this?

Line 574: Heading indicate the table number (“More candidate genes in table 4”)? Can’t be some better heading?

Line 627-628: Concluding a gene/QTL/marker effect, should require concrete experimental results. Just annotation based on literature does not make proper sense. Yes, you can relate but concluding is different and should be based on sound demonstrated results. Same for the functional role for the protein (where are results ? – line 630-631).

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

PLoS One. 2021 Nov 5;16(11):e0259456. doi: 10.1371/journal.pone.0259456.r002

Author response to Decision Letter 0

22 Jul 2021

We appreciate the concerns raised by the reviewers. Their scrutiny is by and large logical and reasonable. We have made extensive changes throughout the manuscript to address most of their comments and voiced our rebuttals for only a few out of all the points they have raised.

In this revision, we have made an effort to improve the language and organization of the manuscript and particularly refine the contents of the results section. We have incorporated gene expression data collected from a previous transcriptome analysis of a sensitive and a tolerant variety of rice where the plants were subjected to salt stress to assess differential expression [1]. All seed related data and results including 4 QTLs and all candidate genes and SNP effects from those QTLs have been removed with hopes of streamlining the manuscript to better fit a salinity stress tolerance related category.

We acknowledge that the weakest aspect of our experiments is the absence of thorough functional studies on the QTLs for validation at the gene and protein level. Statistical and GWAS studies have been carried out at length for a large number of phenotypes and this has led to a very long manuscript. There are many GWAS papers being published today that do not include functional studies [2-10]. This started out as a student project. As such the work not only had time constraints but was not budgeted for either. Therefore, we wish to avoid going into functional studies for this manuscript and rather build upon the current body of work in the near future by undertaking new projects.

Response to reviewer #1

We are thankful that the reviewer finds merit in our work. We address their concerns in detail below.

Comment: Line 58: “No recent GWAS study..” is too absolute as a statement; modify to “Few recent GWAS studies…”

Response: Modified in line 55.

Comment: Change “Seed husks” to “Seed hulls”? Hulls is more commonly used than husks.

Response: Seed related traits and results were removed as per the reviewer’s suggestion.

Comment: Lines 244-245 & Fig. 2a: what is the “blue” subpopulation? Is it aromatic?

Response: Yes. We have labelled it in figure 2 to make it more explicit.

Comment: what do you mean by “fortification of trait correlations in salt stress conditions relative to control conditions”? It is not clear how e.g. C1 compared to S1 in Fig. 3 allows you to come to this conclusion: what exactly is compered here? Fig. 3 is confusing in other ways: what do you mean by “Linear correlations between observed and predicted phenotypes”? How do you “predict” phenotypes? A genomic prediction method is mentioned on lines 149-141, but it is not clear how this SNP effect on phenotype is used for the Pearson’s correlation shown in Fig. 3.

Response: Figure 3 was inspired by Figure 2 (a) of Crowell et al. 2015 [11]. A correlation matrix is essentially a triangle, since you have one square in the heatmap for every one-to-one comparison. The correlations between the means of observed phenotype values given in supplementary table 3 are seen in the lower triangle of the heatmap and the correlation coefficients between the estimated phenotype values given in supplementary table 4 are given in the upper triangle. The diagonal given in white is simply one-to-one comparisons between the observed value and the predicted value of a given phenotype.

Phenotypes can be estimated from SNP and observed phenotype data using the mixed.solve() function. The function calculates a base value for each phenotype and then estimates a phenotype for a variant by adding up the marker effects of each SNP marker in that individual. Thus, the upper triangle of the figure is more representative of the correlations between the phenotypes when the difference in values between the variants (supplementary table 4) is entirely based on differences in their genotype.

Fortification of trait correlation simply means that stressed plants have a more linear relationship (heatmap colors are deeper) between certain phenotypes (marked by green rectangles in figure 3) than plants under control conditions.

We have added the following lines (255-258) in the manuscript to make the exposition of this figure more reader-friendly: “Rectangle C1 and S1 compare correlations between biomass traits in control and stress conditions, C2 and S2 compare correlations between chlorophyll, SES and biomass in control and stress conditions. C3 and S3 compare correlations between chlorophyll and SES in control and stress.”

Comment: Lines 262: since seed traits were only measured under control conditions, it is not clear how this data set is useful for abiotic stress tolerance improvement. To streamline the paper, please remove those data unless you can provide a good explanation in the manuscript for having them.

Response: All data, results and discussions about seed traits have been omitted in this revised version.

Comment: Lines 295-325: Some of the results section should be moved to the methods section or only briefly mentioned. The data transformation part to achieve normal distributions before GWAS analysis should be described in methods and Fig. 5 be included as a supplementary figure to streamline the manuscript.

Response: Extensive changes to the results section has been made. Data distribution image was moved to supplementary as suggested.

Comment: Other result sections also contain quite a bit of methods descriptions that should be moved from Results to Methods.

Response: We have omitted the redundant methodological details from the results.

Comment: Section 3.3., GWAS analysis, should be streamlined (again, do not describe methods here). The naming of QTL could be improved: instead of giving CDP consecutive numbers that seem to random, it would be useful to label a QTL on chromosome 1 as pCDP1-x etc (-x if there are more than one QTL on the same chromosome). Also, what does “CDP” stand for?

Response: Results were streamlined and naming convention for the QTLs were changed as per the author’s suggestion. CDP is simply the initials for the original project title: Comprehensive Data for Plants.

Comment: Fig. 10: heat map colors need to be quantified.

Response: The image (Figure 8 (a) now) was modified to add the color key.

Comment: Lines 453-458: Should not refer to SNPs as “mutations”, b/c these are naturally or artificially selected variants.

Response: The authors are embarrassed about this mishap and it has been corrected throughout.

Comment: Section 4.1 (lines 479-494) does not contribute much to the discussion and could be removed or modified. What is the main message the authors what to convey?

Response: The section has been removed entirely.

Comment: Lines 612-613: Os09g19160 is mentioned twice but should be different genes.

Response: Fixed.

Comment: Gene expression data were from publicly available RNAseq data only and not validated by the authors. Gene expression should be determined under both control and salt stress conditions for selecting the best candidate genes for crop improvement. Moreover, just focusing on genes with synonymous aa substitutions or frameshifts in some accessions might not be the best way to select candidate genes because they could be differentially expressed between stress tolerant and stress sensitive accessions. To improve the manuscript, validate a few candidate genes by qPCR in a selected stress tolerant and a selected stress sensitive accession.

Response: We have included gene expression data from a previous transcriptome analysis [1] of a sensitive and a tolerant variety of rice. Formentin et al. 2018 [1] had carried out deep RNA-seq studies in both control and stress conditions. Their data was acquired from supplementary table 2 of their study and further processed to facilitate necessary comparisons between our candidate genes in control vs stress conditions and in a sensitive and a tolerant check.

Comment: This reviewer is also a bit confused with some aspects of the manuscript: wasn’t the main objective to identify salt tolerance QTL by looking at biomass and chlorophyll content under salt stress and perhaps looking at the relative reduction of both compared to control? So why the long discussion in section 4.3 about the chr. 3 chlorophyll QTL found under control conditions?

Response: The section has been removed.

Comment: Lastly, the section 5 (Conclusion) could be improved: it is too long and should summarize the main points of the manuscript and should not read like an additional discussion.

Response: The conclusions section has been edited to only include the bare essentials.

Response to reviewer #2:

The authors are grateful for the critical comments and expert evaluation of the reviewer. We have made an effort to address all concerns raised and improve the manuscript in light of their suggestions.

Comment: 1. Nowhere in result section authors have mentioned about novel QTLs identified in the study (they discussed about this in abstract & conclusion). To me this should be their key finding which is not explained properly.

Response: We have made a lot of changes to the results section in this revision and hope that it does a better job of highlighting the key findings of our study. The GWAS results in section 3.3 and table 2 lead directly to the genes mapped to the QTLs in section 3.4 and table 3. Section 3.5 and 3.6 explore additional analysis which are brought together in tables 3 and 4. The QTLs have been discussed in the text in section 3.3 lines 295-308. There is further discussion about the genes within these QTLs in lines 323-344, lines 367-374, and also in section 4.2.

Comment: 2. Authors claimed a number of times about functional role of genes without experimental evidences which is not appropriate.

Response: The authors completely agree that is extremely inappropriate to make any claims without concrete experimental evidence. We strive to not make any unsubstantiated claims about our findings. We have noted the gene roles as ‘potential’ gene roles in the introduction, result and conclusions.

We acknowledge the absence of functional validation to be the main shortcoming of our manuscript, but as we have cited at the very beginning of this response letter, there have been many GWAS studies that were considered complete and suitable for publication without specific functional studies on the QTLs or genes.

Comment: 3. Results have not been explained, instead just the tables were referred in most of cases.

Response: Extensive edits and rewrites have been carried out on the results section to correct the flaws of the previous version. We have labored to include an exposition for every table and figure that has been cited in the text. For example, figure 1 and 2 stem directly from supplementary table 1, lines 236-238 observe supplementary table 2, data from supplementary tables 3, 4 and 5 give rise to figure 3, and so on.

Comment: 4. There is too much redundancy in write up, result section is full of material-methods.

Response: Major changes have been made to fix these issues.

Comment: 5. 5. Complex language is big problem in this MS. Authors should write simple language for science but they rather chosen very complex and idiomatic language. Some of the examples are: “handful of individuals”, “environment embellish immense diversity”, “scientists endeavor to elucidate”, “interactions with the surroundings underlie the quantitative assertion of these phenotypes”, “But such is customary to plant breeding experiments and many known and unknown factors are in play here.”, “contra-intuitive for practical purposes”, “an elegant-demonstration”, “Intuitively, salinity tolerance benchmarks are depicted by the ability of plants to accumulate and retain biomass and color.”, “It did not escape our notice”, “inspires us to suggest a novel role for the gene in seed dimension regulation”, “thrive in substandard land without any compromise in yield”, “we attempt to rapidly sieve these desirable attributes into our modern cultivars”, “These observations are however, not inconceivable.”, “firm belief” and many more.

Response: We feel that the reviewer is perhaps a little too scrupulous here. The phrase “It did not escape our notice” has been adapted from the last paragraph of James Watson and Francis Crick’s landmark paper on the discovery of the double helix where they hint at the implications of their discovery in explaining the molecular basis of Mendelian inheritance.

Nonetheless, many of these quoted sections have been removed and with many others, we have made changes to accommodate the reviewer’s preferences as follows:

Line 229: “handful of individuals” was changed to “few individuals”.

“environment embellish immense diversity” was removed.

Line 48 “scientists endeavor to elucidate” was changed to “it has become possible to elucidate”.

Line 49: “interactions with the surroundings underlie the quantitative assertion of these phenotypes” was left unchanged.

“But such is customary to plant breeding experiments and many known and unknown factors are in play here.”, section was removed.

“contra-intuitive for practical purposes”, section was removed.

Line 432: “an elegant-demonstration”, the authors feel that it truly is an elegant demonstration, so left unchanged.

Line 438: “Intuitively, salinity tolerance benchmarks are depicted by the ability of plants to accumulate and retain biomass and color.”, the authors feel that it conveys an important message and so left unchanged.

“It did not escape our notice”, section was removed.

“inspires us to suggest a novel role for the gene in seed dimension regulation”, section was removed.

“thrive in substandard land without any compromise in yield”, was removed.

“we attempt to rapidly sieve these desirable attributes into our modern cultivars”, was removed.

“These observations are however, not inconceivable.”, was removed.

Line 500: “firm belief”, left unchanged because the authors feel that an individual’s beliefs must always be firm and infallible.

Specific comments by reviewer #2

Comment: Line 58-59: “No recent GWAS study explores abiotic stress tolerance in relation to developmental and agronomic phenotypes in rice.” – there are plenty of GWAS studies on agronomic traits, presume that agronomic phenotypes relates to agronomic traits. What author means developmental phenotype is probably evolutionary basis but how does this MS linked to GWAS for the evolutionary traits? Didn’t find any evolutionary in-depth analysis in this MS except grouping which is very common.

Response: The authors have yet to come across any GWAS study that explores biomass, ion content, chlorophyll and seed traits in conjunction and this had been the basis of this statement. But as per the suggestions of reviewer #1 (comment 1) and reviewer #2 here, we have modified the sentence to “few recent GWAS studies…” in line 55.

Comment: Line 62 – 64: CRISPR is very much functional but with targeted gene insertion or knock-outs. Is there any example where QTL that carries multiple genes spanning very long DNA fragment substituted the alternate allele through CRISPR and worked? Or it is just hypothesis?

Response: We completely agree with the reviewer. We merely suggest that the genes within these QTLs are amenable to CRISPR-Cas9 editing. We have changed ‘salient QTLs’ to ‘salient features’ in line 59 to clarify the statement.

Comment: Line 83: What do you mean by ‘176 rice accessions ordered from IRRI seed bank’? Author can say got the accessions or received from IRRI gene bank.

Response: ‘ordered’ was changed to ‘received’ in line 79.

Comment: Line 84-85: “BRRI (Bangladesh Rice Research Institute) fields where the day/night temperature was 32°/28°C”. Is it possible to have the same temperature in field throughout cropping season?

Response: These numbers denote the average ambient temperature of the cropping season. We have edited the section to include the months of the cropping season. Quoting lines 80 to 82: “Seeds were multiplied at BRRI (Bangladesh Rice Research Institute) fields in the months of March and April when the average day/night temperature was 32°/28°C with an average humidity of 72%.”

Comment: Line 97-98: Screening for salinity tolerance at seedling stage was carried out as per Amin et al (2012). This article is not an open access article hence could not go

through the details but as mentioned in M & M, germinated seeds grown for 14 days the salt stress applied in 5 increments from 4 to 12 dS/m. Then phenotypic measurements were made after 16 days from 4 dS/m treatment (30 days). SES scores and other parameters were taken as described by Gregorio et al 1997. The screening as per Amin et al and measurement as per Gregorio et al are two different things and will produce two different scenarios. Gregorio used to treat the plant when they are really at seedling and single tiller stage (first treatment of 6 dS/m at 6th days from germination and raised to 12 dS/m on 9th day). Once the seedling becomes big enough (started multiple tillering), the response to salinity becomes different, so not sure why authors used old seedlings to screen when they already start attaining the tolerance to salinity.

Response: The authors have tried to convey that the screening protocol in its entirety mimics the works of Amin et al. 2012 [12]. Gregorio et al. 1997 [13] is only referenced in relation to the criteria for salt injury scoring in terms of SES. Rice plants are still seedlings at 14-30 days of age and at the time of harvest for our analysis, they had not started producing any tillers at all. These methods have been devised and followed as appropriate for screening of salinity tolerance at the seedling stage. Another point to note here is that our studies were in a net house, where the growth of the seedlings is dependent on the weather. Therefore we observe the size of the plant rather than the days. The variation in the days when the plants are deemed ready is usually between 12-16 days. The plants in this case were of different genotypes and the sizes varied between 6-8 inches, when the salt exposure regime was started. Earlier salt exposure would have been appropriate for an exclusive post-germination/early-seedling stage tolerance study.

Comment: Line 117 and 118: What are the details of the formulae to measure the chlorophyll? Volume in ml or l? Weight (fresh or dry) in, mg or g ? Absorbance etc. Explain each formula.

Response: The recommended explanations have been appended. Quoting lines 114 to 116: “Here A_663 and A_645 refer to the absorbance at 663nm and 645nm wavelengths respectively. V is volume in milliliters, W is the fresh weight of the leaf pieces in milligrams and chlorophyll content is expressed as milligrams per gram of fresh leaf tissue.”

Comment: Line 153: 2,930,739 missing ‘genotypes’? – it is typo.

Response: These refer to individual SNP marker genotypes not the whole genome/genotype of the individual. The authors affirm that this terminology is common in computational genomics to refer to the ‘genomic configuration’ of the SNP markers.

Comment: Line 159-161: Why MAF was chosen to 10% for subpopulations : Should be explained.

Response: The purpose of the minor allele filter is to filter out rare alleles that occur in very low frequency because the inclusion of these rare alleles in the study compromises statistical power and leads to spurious associations. MAF of 5% for 176 individuals means there has to be at least 9 instances of the minor allele, which is a suitable number for GWAS, but 5% of 82 or 78 is only ~4 which could yield false positive results with very low p-values. That is why a 10% MAF is more suitable for our subpopulation-based GWAS tests. We have added the following in lines 149 to 151 in the methods section to explain this briefly: “For GWAS, a minor allele frequency filter of 5% was applied to the whole population. Because of the smaller sample size, the minor allele frequency filter for the subpopulations was set at 10%.”

Comment: Line 197-199: Authors have discussed a lot about the FDR (through Bonferroni), however, more important is the effect of population structure on associations. The way authors have explained above about the model for testing marker effects (in two equations), authors can explain OR simplify the explanation about population structure correction.

Response: The authors agree that the BLINK algorithm can seem quite complex. But population structure correction is a relatively simple part of the model and is devoid of the complexity present in the exposition of the BLINK algorithm in section 2.7. A kinship matrix for the study sample and principal components from the genotype data are calculated and they are simply included in the final model as covariates after the BLINK algorithm has been executed. This has been explained in lines 184 to 187 of section 2.7: “To correct for population structure and cryptic relatedness, the kinship matrix K, calculated using GAPIT and the Q matrix having the first three principal components were fit in the first model as covariates. For the compressed mixed linear models fit for validation tests, the optimum compression level was selected and the same covariates were included.”

The authors feel that further explanation would needlessly inflate the methods section even further. Additionally, readers can always refer to the original article on the BLINK algorithm [14].

Comment: Results: For presenting the results, authors have just referred the tables/ figures and in text there is too much redundancy with material-methods. Result section needs to be re-organized in way it befits for a research paper, explaining the crystal-clear results ONLY. Say for example line 226 to almost 239 (except 236 and 237), is superfluous (“largest delta on Earth with bountiful rivers draining into the northeastern part of the Indian Ocean by way of the Bay of Bengal”) or part of M & M. There are plenty of instances where information fit for ‘M & M’ are written in ‘Results’ like 245-247, 254-257, 259-265, 327-337 and so on).

Response: Previous lines 226 to 239 have been mostly omitted. Previous lines 245-247 were changed. Previous lines 254-257 were kept as is since these are direct rationales that stem from the population study leading into the GWAS study. Previous lines 259-265 have been modified. Previous lines 327-337 have also been modified.

Extensive rewrites have been carried out on the results section and we hope that the reviewer will have a better impression of the section in this manuscript.

Comment: Line 274 – 276: Why it is so, please explain: “Narrow-sense heritability could not be calculated from additive marker effects because the grand mean of the phenotypes could explain more variance than the mixed.solve prediction model”

Response: The output of the prediction model explains less variation than the grand mean. We generally compare any model to the mean to find out how good it is. Since the difference in variation in this case is negative, it could not be done (or would not make statistical sense to do). We have added the following in lines 242-244 in section 3.2 assuming why this might be the case: “This could be attributable to the relatively small sample size with diverse and heterozygous wildtype genotypes used for genomic prediction.”

Comment: Line 306: What is bulky seed in term of their shape and size? Heard for the first time.

Response: All seed related data and results were omitted as per the suggestions of reviewer #1.

Comment: Line 361: Authors need to explain simply that what novel QTLs/ association they found and what was their effect. Their relation with candidate genes can be explained in next section. Could not find the information about the novel QTL in this section (QTL mapping) which could be the most important basis for acceptance of article for publication.

Response: Discovered QTLs have been summarized in table 2. The authors feel that in a properly formatted, published version where the table is viewed right below the text, it would be easy and much more comprehensible.

Comment: Line 377: What is trait like seed height? I have heard about seed length but not seed height? Are they same? If they are same, then why mentioned as “We previously found seed length to have no correlation with seed height” (line 607-698).

Response: We acknowledge that seed height is an unusual term. It had been explained previously in lines 125 and 126 of the first version. But it is no longer relevant since these all have been removed.

Comment: Line 468-475: This can go as the legend or the footnote of the table.

Response: The table has been removed on the basis of the feedback from the reviewers.

Comment: Line 481-487: “Individual differences in plants of identical genotype are disturbingly disorderly. We found that variance within genotypes in our growth experiments were distinctly large for most traits other than seed properties. But such is customary to plant breeding experiments and many known and unknown factors are in play here. The question here is not how we can remove these confounding variables, since it is not only impossible but also contra-intuitive for practical purposes, but rather how we can derive informative features from observable phenotypes.”. What authors want to convey is not clear. The message should be simple and clear. ‘Individual differences in plants of identical genotype’ : It is not new when you are working with landraces. I am fail to understand the message out of this.

Response: Section was removed.

Comment: Line 493 – 494: This is well establish fact, what is new in this?

Response: Section was removed.

Comment: Line 574: Heading indicate the table number (“More candidate genes in table 4”)? Can’t be some better heading?

Response: The authors are embarrassed about the unintelligent selection of the section title. It has been modified in line 451 to “Discovered QTLs evince new potential gene roles”

Comment: Line 627-628: Concluding a gene/QTL/marker effect, should require concrete experimental results. Just annotation based on literature does not make proper sense. Yes, you can relate but concluding is different and should be based on sound demonstrated results. Same for the functional role for the protein (where are results ? – line 630-631).

Response: We completely agree with the concerns of the reviewer. We have only referred to our results as ‘potential’ genes of interest based on the original association signal in the QTL and some downstream analysis. We have only included these sections as some exposition about the genes that our QTLs of interest house are of relevance. We had acknowledged this even in the first version of the manuscript in the conclusions in lines 658-659:

“We affirm that we must be careful about the immediate interpretations of our findings as they have not yet been subjected to systematic case-control studies after gene cloning.”

We have made changes throughout the manuscript explicitly mentioning that our genes might ‘potentially’ have such roles. E.g. Line 33, 70, 451, 461, 465, 473, 480 and 486. It was never our intention to claim more than our findings, the functional roles were brought up simply to put forward the descriptive information from the annotations, expression studies and SNP consequences.

The exploration of population structure, correlation between different phenotypes of different groups, the discovered QTLs and the genes of interest from all the mapped genes are the primary findings of this study. On top of this, the git repository is an important asset for researches doing population structure and GWAS studies as all our methods have been elaborately documented there.

References:

1. Formentin, E., et al., Transcriptome and Cell Physiological Analyses in Different Rice Cultivars Provide New Insights Into Adaptive and Salinity Stress Responses. Frontiers in Plant Science, 2018. 9(204).

2. Li, F., et al., Genetic Basis Underlying Correlations Among Growth Duration and Yield Traits Revealed by GWAS in Rice (Oryza sativa L.). Frontiers in Plant Science, 2018. 9(650).

3. Yuan, J., et al., Genetic basis and identification of candidate genes for salt tolerance in rice by GWAS. Scientific Reports, 2020. 10(1): p. 9958.

4. Schläppi, M.R., et al., Assessment of Five Chilling Tolerance Traits and GWAS Mapping in Rice Using the USDA Mini-Core Collection. Frontiers in Plant Science, 2017. 8(957).

5. Thapa, R., et al., Genome-Wide Association Mapping to Identify Genetic Loci for Cold Tolerance and Cold Recovery During Germination in Rice. Frontiers in Genetics, 2020. 11(22).

6. Ma, X., et al., Genome-Wide Association Study for Plant Height and Grain Yield in Rice under Contrasting Moisture Regimes. Frontiers in Plant Science, 2016. 7(1801).

7. Wang, X., et al., New Candidate Genes Affecting Rice Grain Appearance and Milling Quality Detected by Genome-Wide and Gene-Based Association Analyses. Frontiers in Plant Science, 2017. 7(1998).

8. Volante, A., et al., Genome-Wide Analysis of japonica Rice Performance under Limited Water and Permanent Flooding Conditions. Frontiers in Plant Science, 2017. 8(1862).

9. Zhao, M., et al., Mining Beneficial Genes for Aluminum Tolerance Within a Core Collection of Rice Landraces Through Genome-Wide Association Mapping With High Density SNPs From Specific-Locus Amplified Fragment Sequencing. Frontiers in Plant Science, 2018. 9(1838).

10. Pariasca-Tanaka, J., C. Baertschi, and M. Wissuwa, Identification of Loci Through Genome-Wide Association Studies to Improve Tolerance to Sulfur Deficiency in Rice. Frontiers in Plant Science, 2020. 10(1668).

11. Crowell, S., et al., Genome-wide association and high-resolution phenotyping link Oryza sativa panicle traits to numerous trait-specific QTL clusters. Nature Communications, 2016. 7(1): p. 10527.

12. Amin, M., et al., Over-expression of a DEAD-box helicase, PDH45, confers both seedling and reproductive stage salinity tolerance to rice (Oryza sativa L.). Molecular Breeding, 2012. 30(1): p. 345-354.

13. Gregorio, G., D. Senadhira, and R. Mendoza, Screening rice for salinity tolerance, vol 22, IRRI discussion paper series. International Rice Research Institute, 1997.

14. Huang, M., et al., BLINK: a package for the next level of genome-wide association studies with both individuals and markers in the millions. GigaScience, 2019. 8(2).

Attachment

Submitted filename: Response to reviewers.docx

Click here for additional data file.^{(31.3KB, docx)}

PLoS One. doi: 10.1371/journal.pone.0259456.r003

Decision Letter 1

Prasanta K Subudhi

18 Aug 2021

PONE-D-21-14535R1

Novel QTLs for salinity tolerance revealed by genome-wide association studies of biomass, chlorophyll and tissue ion content in 176 rice landraces from Bangladesh

PLOS ONE

Dear Dr. Seraj,

Although manuscript has been improved significantly, there are still issues pointed by the Reviewer 3 which need to be addressed. Specifically, candidate genes selection based on RNA-Seq data need careful attention.

Please submit your revised manuscript by Oct 02 2021 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.
A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.
An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

We look forward to receiving your revised manuscript.

Kind regards,

Prasanta K. Subudhi, Ph.D.

Academic Editor

PLOS ONE

Additional Editor Comments (if provided):

Major revision

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: (No Response)

Reviewer #3: (No Response)

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

Reviewer #1: Yes

Reviewer #3: Partly

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #3: No

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

Reviewer #1: Yes

Reviewer #3: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

Reviewer #1: Yes

Reviewer #3: Yes

**********

6. Review Comments to the Author

Reviewer #1: The manuscript by Alam et al. now titled “Novel QTLs for salinity tolerance revealed genome-wide association studies of biomass, chlorophyll and tissue ion content in 176 rice landraces from Bangladesh” is a revised version of a previously submitted paper. I describes a GWAS mapping analysis of 11 phenotypes yielding 13 QTL and list of associated candidate genes.

The manuscript has been substantially edited and most of my concerns were addressed. However, there are still issues that need to be addressed to improve the quality of this manuscript.

Line 28: take out “sophisticated”. This is a qualified personal opinion and not a quantified fact. There are still a number of such embellishing qualifiers that need to be removed (see below).

Line 63: take out “elaborately”. This is an opinion.

Line 68: edit to “…publicly available gene expression profiles….”, because you did not do this for the current paper.

Line 211: take out “elaborate” (see comments above).

Lines 235-238: use past tense, not present tense, to describe your results.

Lines 252-253: change sentence to “We observed better trait correlations under salt stress conditions than control conditions”. The “fortification” term is too confusing.

Lines 264-28): I still think this paragraph is too long and has unnecessary information. Simply state that non-Gaussian distributions were normalized before GWAS analyses.

Line 281: remove “extensive”. Again, this is an unnecessary qualification/personal opinion.

Line 292: spell out CDP (Comprehensive Data for Plants). Actually, my recommendation is to change the naming to something reflecting salts stress since CDP is a very idiosyncratic.

Lines 295-296: remove the sentence starting with “The suggestive association…”. There is no follow-up to the statement.

Line 304: change “connotations” to “evaluations”.

Lines 330-331: change “imparts sufficient prospect in the role…” to “suggest a role..”.

Line 241: change “To select a set of informative candidate genes with authentic functional roles..” to simply “To select a set of candidate genes…”. The rest are unnecessary qualifications.

Line 425: change “traditionalist” to “traditional”.

Line 428: remove “deliberately”.

Lines 431-433: modify the sentence starting with “The lack of…” to “The lack of this correlation in shoots under stress conditions suggests that sodium exclusion from rice leaves might be critical for survival in salt stress”. At the point, a correlation is not an “elegant demonstration” without further analyses to “demonstrate” causation.

Line 437: remove “highly”.

Line 457: change “causing a nonsense consequence” to “generating a stop codon”.

Line 477: remove the first part of the sentence (“With rigorous….significant thresholds”) and start sentence simply with “We mapped…”.

Line 494: remove “greatly”.

Reviewer #3: The paper has merit and acceptable for publication. However, it needs major revision and should be shortened to focus on QTL detection for salinity tolerance by GWAS. The candidate gene extraction by outsourced RNASeq data is highly speculative in relation to QTLs detected in this study.

RNAseq Data were outsourced from MSU, while Vialone Nano and Baldo RNAseq data were from Formentin 2018. Gene expression and downstream analysis is not appropriate to be included in this paper because RNA gene expression is 1.) genotype-specific, 2.) tissue-specific, 3.) developmental stage-specific, and most importantly, 4.) treatment-specific. While you may have the same ID or variety, the available RNA expression data is only specific to the materials they tested during that time. Moreover, RNA expression studies needs further validation because RNAs are unstable and transient naturally. For this study, to outsource RNASeq data is not acceptable because even MSU RNA Seq data were not collected during salinity stress of the test samples, but mostly during developmental stage of rice and tissue-specific.

Scientific papers are inherently difficult to read and authors should strive to make it more comprehensible by using more simple words to encourage readers. For example line 253-261 states “Rectangle C1 and S1 compare correlations between biomass traits in control and stress conditions, C2 and S2 compare correlations between chlorophyll, SES and biomass in control and stress conditions. C3 and S3 compare correlations between chlorophyll and SES in control and stress. The improved linear correlations between biomass, chlorophyll and SES traits imply that the encumbrance of salt stress on a plant prevents the disproportionate gain of biomass. Since plant biomass and chlorophyll content is markedly reduced under salt stress (supplementary table 3), we can conclude that salt stress does not affect all phenotypes for all genotypes at the same rate and the genetic advantages that any genotype has in terms of tissue growth (root/shoot) or chlorophyll accumulation will be compromised by the effect of abiotic stress at a greater magnitude than traits which have near-baseline values.”

The correlations of traits do not mean causation or effects, it is simply telling a trend between the two traits. Why is such a long confusing conclusion here? What exactly the authors would like to convey based on correlations? It’s a fact that traits like biomass and chlorophyll content are affected by salt stress! The paper would be more meaningful if they measured the reduction of traits in control and stressed genotypes. It would be interesting to find out which varieties had been least affected by the stress and by how much, compare to other genotypes, or an analyses like some sort of index of growth reduction comparison among genotypes with regards to stress and control treatments.

The QTL naming is not reflecting of the association of a trait to a loci. The CDP is too broad and not meaningful to associate to salinity tolerance or any specific trait.

In Discussions, “Biomass traits and chlorophyll content could be valuable indices for the screening of salinity tolerance”—this is already known fact.

Please state the criticism/problem to SES that needs to be addressed. Line 407-408 is intriguing or dramatic and maybe unnecessary to be written.

In discussion, again, high and low correlations are not causation or effects but simply statistical trends. It should also be noted if the correlations are negative or positive. Authors should refrain from drawing too much generalization or speculations unless supported with hard proof validation especially that this study is dealing with diverse germplasms.

Discussion 4.1 is too general and mostly known fact by previous studies. The paper would be more meaningful if authors discussed the traits of varieties used in response to salt stress. This study is nothing new except for the varieties used, so it would be nice to have those information available to the readers in comparison to other known salinity tolerance study.

The paper is claiming novel QTLs, therefore, the focus of discussion should be those novel QTLs, their significance, effects- positive or negative effect, enhancing tolerance or sensitivity, variance explained by the QTLs, occurrence, and usefulness in future salt tolerance introgression. Authors may further discuss the similarity and differences of detected QTLs in relation to previous other QTLs.

In Discussion 4.2, Lines 452-455, candidate genes discussed in the paper were selected on the basis of (a) functionally impairment of gene in the population and (b) individuals carrying the genotype of the functional allele are quantifiably different in terms of phenotype. The study would be meaningful and therefore has merit to discuss candidate genes in details if candidate genes were re-sequenced to confirm the functional SNPs, and validated by gene expression like qRT-PCR. At this point, candidate genes as underlying genes controlling salinity tolerance is premature assumptions despite employing extensive statistical analysis. Gene expression is study-specific and should be tested in contrasting genotypes.

The paper is too lengthy for the methods, and discussion was too short to emphasize the actual findings of the study. While extensive data mining was conducted, the claims regarding candidate genes are not fully supported by current findings. Further validation is needed. However, if the paper is shorted into QTL detection alone, the study is meritorious of publication. I suggest to remove the candidate gene mining by RNASEq, unless fully supported by re-sequencing, functional SNP study and gene expression analysis. We uphold to scientific and rigorous standard regardless of negative or positive results. It is no excuse to say that this study started out as a student project and that it is time constraints and with limited budget. However, once published, this paper would stand as scholastic achievement and pride of the student and all authors included.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #3: No

PLoS One. 2021 Nov 5;16(11):e0259456. doi: 10.1371/journal.pone.0259456.r004

Author response to Decision Letter 1

24 Sep 2021

Reviewer #1:

Comment: The manuscript by Alam et al. now titled “Novel QTLs for salinity tolerance revealed genome-wide association studies of biomass, chlorophyll and tissue ion content in 176 rice landraces from Bangladesh” is a revised version of a previously submitted paper. I describes a GWAS mapping analysis of 11 phenotypes yielding 13 QTL and list of associated candidate genes.

The manuscript has been substantially edited and most of my concerns were addressed. However, there are still issues that need to be addressed to improve the quality of this manuscript.

Line 28: take out “sophisticated”. This is a qualified personal opinion and not a quantified fact. There are still a number of such embellishing qualifiers that need to be removed (see below).

Response: Removed in line 28.

Comment: Line 63: take out “elaborately”. This is an opinion.

Response: Removed in line 62.

Comment: Line 68: edit to “…publicly available gene expression profiles….”, because you did not do this for the current paper.

Response: Added in line 67.

Comment: Line 211: take out “elaborate” (see comments above).

Response: Section has been removed.

Comment: Lines 235-238: use past tense, not present tense, to describe your results.

Response: New lines 233-234 have been changed to past tense.

Comment: Lines 252-253: change sentence to “We observed better trait correlations under salt stress conditions than control conditions”. The “fortification” term is too confusing.

Response: Changed in line 250.

Comment: Lines 264-28): I still think this paragraph is too long and has unnecessary information. Simply state that non-Gaussian distributions were normalized before GWAS analyses.

Response: We have kept the information about data distributions, the test for normalcy and rationale for transformation. We have removed everything else. The new section is in lines 260-273.

Comment: Line 281: remove “extensive”. Again, this is an unnecessary qualification/personal opinion.

Response: Removed in line 292.

Comment: Line 292: spell out CDP (Comprehensive Data for Plants). Actually, my recommendation is to change the naming to something reflecting salts stress since CDP is a very idiosyncratic.

Response: The original dataset generated in our laboratory, from which data was used to prepare the first version of this manuscript, has been titled as such. The authors plan to carry out more GWAS studies with their data using the computation pipeline that has been established in this manuscript. The authors wish to maintain this naming convention for the sake of consistency and convenience.

Comment: Lines 295-296: remove the sentence starting with “The suggestive association…”. There is no follow-up to the statement.

Response: Sentence was removed in line 307.

Comment: Line 304: change “connotations” to “evaluations”.

Response: The authors feel that the word ‘connotations’ is more appropriate here.

Comment: Lines 330-331: change “imparts sufficient prospect in the role…” to “suggest a role..”.

Response: Changed in line 344.

Comment: Line 241: change “To select a set of informative candidate genes with authentic functional roles..” to simply “To select a set of candidate genes…”. The rest are unnecessary qualifications.

Response: Section has been moved to supplementary notes. So, left unchanged.

Comment: Line 425: change “traditionalist” to “traditional”.

Response: Changed in line 377.

Comment: Line 428: remove “deliberately”.

Response: Removed in line 380.

Comment: Lines 431-433: modify the sentence starting with “The lack of…” to “The lack of this correlation in shoots under stress conditions suggests that sodium exclusion from rice leaves might be critical for survival in salt stress”. At the point, a correlation is not an “elegant demonstration” without further analyses to “demonstrate” causation.

Response: The authors concede. The change has been made in lines 384-285.

Comment: Line 437: remove “highly”.

Response: Changed in line 389.

Comment: Line 457: change “causing a nonsense consequence” to “generating a stop codon”.

Response: Section has been moved to supplementary.

Comment: Line 477: remove the first part of the sentence (“With rigorous….significant thresholds”) and start sentence simply with “We mapped…”.

Response: Changed in line 426.

Comment: Line 494: remove “greatly”.

Response: Changed in line 443.

Reviewer #3:

Comment: The paper has merit and acceptable for publication. However, it needs major revision and should be shortened to focus on QTL detection for salinity tolerance by GWAS. The candidate gene extraction by outsourced RNASeq data is highly speculative in relation to QTLs detected in this study.

Response: We appreciate the new perspectives and valuable insights from the reviewer. In this revision, we omit outsourced expression studies from the main manuscript and add more detail about inter-variety variations in Bangladeshi landraces in response to salt stress.

Comment: RNAseq Data were outsourced from MSU, while Vialone Nano and Baldo RNAseq data were from Formentin 2018. Gene expression and downstream analysis is not appropriate to be included in this paper because RNA gene expression is 1.) genotype-specific, 2.) tissue-specific, 3.) developmental stage-specific, and most importantly, 4.) treatment-specific. While you may have the same ID or variety, the available RNA expression data is only specific to the materials they tested during that time. Moreover, RNA expression studies needs further validation because RNAs are unstable and transient naturally. For this study, to outsource RNASeq data is not acceptable because even MSU RNA Seq data were not collected during salinity stress of the test samples, but mostly during developmental stage of rice and tissue-specific.

Response: The reviewer raises some valid concerns. We agree that the impact of the manuscript up to QTL mapping is far greater than its remainder which yields speculative conclusions at best. We have streamlined the manuscript further to accommodate the views of the reviewer. Instead of completely deleting the expression analysis and functional consequence studies, we have moved the complete section to supplementary notes should any keen reader be interested in going over them.

Comment: Scientific papers are inherently difficult to read and authors should strive to make it more comprehensible by using more simple words to encourage readers. For example line 253-261 states “Rectangle C1 and S1 compare correlations between biomass traits in control and stress conditions, C2 and S2 compare correlations between chlorophyll, SES and biomass in control and stress conditions. C3 and S3 compare correlations between chlorophyll and SES in control and stress. The improved linear correlations between biomass, chlorophyll and SES traits imply that the encumbrance of salt stress on a plant prevents the disproportionate gain of biomass. Since plant biomass and chlorophyll content is markedly reduced under salt stress (supplementary table 3), we can conclude that salt stress does not affect all phenotypes for all genotypes at the same rate and the genetic advantages that any genotype has in terms of tissue growth (root/shoot) or chlorophyll accumulation will be compromised by the effect of abiotic stress at a greater magnitude than traits which have near-baseline values.”

Response: The reviewer is correct that correlations of traits do not explicitly mean causation or effects but we disagree with their opinion that correlation is ‘simply a trend’. The word ‘trend’ in statistics is more appropriate for fields like economics where only directional patterns are observed.

We want to take this opportunity to break down the section quoted here by the reviewer. To make things simple, let us look at the correlation between Shoot weight and Chlorophyll A in control and stress conditions. Please download the notebook at: https://github.com/DeadlineWasYesterday/Cat-Does-Plant/blob/master/Figures%20and%20tables/Revision%202%20argument.pdf to follow this demonstration. There should be a download button on the top right of the preview in this webpage.

Please look at figure 1 in this notebook. It shows that shoot weight is reduced under salt stress. Figure 2 shows a less dramatic, but identifiable reduction in chlorophyll A content. From this, we can conclude that shoot weight and chlorophyll A content is reduced under salt stress. Figures 3 and 4 linearly model the nature of this reduction. We can observe a clear linear relationship under stress conditions which was less prominent in the absence of salt. This is what the color depth of the rectangles in figure 3 of our manuscript signifies.

In figure 5 we have plotted the distribution of the ratio of shoot weight and chlorophyll A content from supplementary table 3. This shows that both the magnitude and the spread of the shoot weight to chlorophyll A ratio is reduced in stress conditions. The bidirectional bar chart found in the last figure (out[10]), shows a dramatic difference between the dispersal of shoot weight to chlorophyll A ratios in our experimental varieties. From this bar chart, we can see that plants that had a high ratio in control conditions, exhibit smaller ratios in stress conditions.

This is the exact conclusion we have drawn in the section quoted from the manuscript by the reviewer. The correlation heatmap on figure 3, the p values in supplementary table 5 and the density plots in supplementary figure 2 directly demonstrate this phenomenon. To paraphrase: Salt does not affect all genotypes equally. In salt stress, phenotypes correlate better linearly and the ratios become more proportional, which is to say that any advantage that a genotype has in terms of disproportionate tissue growth or chlorophyll accumulation will be compromised first under salt stress in general.

We beg the reviewer to not undermine our analyses by calling them ‘merely a trend’. A sufficient amount of thought and attention was put into this. The authors feel that there is ample data in the manuscript to support their claim.

We have added lines 263-275 and a new figure 4 to bring forward the effects of salt stress on phenotypes of different varieties.

Comment: The QTL naming is not reflecting of the association of a trait to a loci. The CDP is too broad and not meaningful to associate to salinity tolerance or any specific trait.

Response: The rationale behind the original naming convention was that this was a diverse set of phenotype data for GWAS analysis. We had filtered out quite a few before writing the original manuscript and as per the recommendations of the reviewers, traits other than those concerned with salinity tolerance were left out from the first revision. The naming is consistent throughout the manuscript, images and supplementary files and we hope to keep them consistent in our following work using data that have been excluded from this manuscript. Additionally, the QTL names with qCDP are hardcoded in all the sourcecode, making our experiments easily reproducible using data that was made available online.

Comment: In Discussions, “Biomass traits and chlorophyll content could be valuable indices for the screening of salinity tolerance”—this is already known fact.

Response: There are a number of recent studies on salinity tolerance where the experimental design exclusively focuses on ion content and ion ratios and/or overlooks simpler observations such as biomass and chlorophyll content [1-3]. The authors agree that the effects of salt stress on biomass and chlorophyll is quite obvious. In spite of this, our study as a whole brings forward the numerical relationships between these phenotypes in response to salinity tolerance. The quoted section in particular reflects the basic impression of our journey throughout the study. A lot of it builds on the referenced manuscript by Pires et al. 2015 [4]. The authors wish to keep the section as is.

Comment: Please state the criticism/problem to SES that needs to be addressed. Line 407-408 is intriguing or dramatic and maybe unnecessary to be written.

Response: SES is a composite score which is prone to human error. The principal component we derived from stress biomass and chlorophyll content in this revision would be more numerically robust and a better index for salt injury than an evaluation made by the naked eye. Of course, SES scoring is less laborious and still applicable in many scenarios. The authors feel that a comprehensive study of salt tolerance will still be incomplete without recording biomass and chlorophyll content. In light of this, we have appended the following section at the end of our conclusions in line 435: “Although it is well established that biomass and chlorophyll related traits alongside tissue ion content aids the assessment of salinity tolerance, our studies show that the numerical values of plant biomass and chlorophyll content could manifest themselves into a more objective index of salt injury unabated by human error.”

Comment: In discussion, again, high and low correlations are not causation or effects but simply statistical trends. It should also be noted if the correlations are negative or positive. Authors should refrain from drawing too much generalization or speculations unless supported with hard proof validation especially that this study is dealing with diverse germplasms.

Response: The blue color in figure 3 and supplementary figure 1 indicates a positive correlation and the red color indicates negative. Out of the significant correlations shown in figure 3, only SES shares negative correlation with the remaining phenotypes. The authors have addressed part of this comment in detail in the response to the second comment.

Comment: Discussion 4.1 is too general and mostly known fact by previous studies. The paper would be more meaningful if authors discussed the traits of varieties used in response to salt stress. This study is nothing new except for the varieties used, so it would be nice to have those information available to the readers in comparison to other known salinity tolerance study.

Response: The authors feel differently about section 4.1. We feel that the contents of the section have more depth than the reviewer has surmised here. The distributions of our phenotypes have been observed. We have discussed why we did not derive phenotypes from sodium and potassium ratios, which are very commonly used in the screening of salinity tolerance. We feel that our study puts forward ample evidence to allow us to emphasize the value of biomass and chlorophyll data in the study of salinity tolerance.

Comment: The paper is claiming novel QTLs, therefore, the focus of discussion should be those novel QTLs, their significance, effects- positive or negative effect, enhancing tolerance or sensitivity, variance explained by the QTLs, occurrence, and usefulness in future salt tolerance introgression. Authors may further discuss the similarity and differences of detected QTLs in relation to previous other QTLs.

Response: This is primarily a QTL detection and mapping study. Molecular dissection of these QTLs requires thorough and extensive laboratory work, which the authors have been planning for the near future. The topics brought up by the reviewer are very relevant and will be the subject of our subsequent studies.

Comment: In Discussion 4.2, Lines 452-455, candidate genes discussed in the paper were selected on the basis of (a) functionally impairment of gene in the population and (b) individuals carrying the genotype of the functional allele are quantifiably different in terms of phenotype. The study would be meaningful and therefore has merit to discuss candidate genes in details if candidate genes were re-sequenced to confirm the functional SNPs, and validated by gene expression like qRT-PCR. At this point, candidate genes as underlying genes controlling salinity tolerance is premature assumptions despite employing extensive statistical analysis. Gene expression is study-specific and should be tested in contrasting genotypes.

Response: These sections have been removed in accordance with the comments from the reviewer.

Comment: The paper is too lengthy for the methods, and discussion was too short to emphasize the actual findings of the study. While extensive data mining was conducted, the claims regarding candidate genes are not fully supported by current findings. Further validation is needed. However, if the paper is shorted into QTL detection alone, the study is meritorious of publication. I suggest to remove the candidate gene mining by RNASEq, unless fully supported by re-sequencing, functional SNP study and gene expression analysis. We uphold to scientific and rigorous standard regardless of negative or positive results. It is no excuse to say that this study started out as a student project and that it is time constraints and with limited budget. However, once published, this paper would stand as scholastic achievement and pride of the student and all authors included.

Response: We have revised the manuscript as per the suggestions of the reviewer.

References:

1. Warraich, A.S., et al., Rice GWAS reveals key genomic regions essential for salinity tolerance at reproductive stage. Acta Physiologiae Plantarum, 2020. 42(8): p. 134.

2. Kumar, V., et al., Genome-wide association mapping of salinity tolerance in rice (Oryza sativa). DNA Research, 2015. 22(2): p. 133-145.

3. Batayeva, D., et al., Genome-wide association study of seedling stage salinity tolerance in temperate japonica rice germplasm. BMC Genetics, 2018. 19(1): p. 2.

4. Pires, I.S., et al., Comprehensive phenotypic analysis of rice (Oryza sativa) response to salinity stress. Physiologia Plantarum, 2015. 155(1): p. 43-54.

Attachment

Submitted filename: Response to reviewers 2.docx

Click here for additional data file.^{(24.7KB, docx)}

PLoS One. doi: 10.1371/journal.pone.0259456.r005

Decision Letter 2

Prasanta K Subudhi

11 Oct 2021

PONE-D-21-14535R2Novel QTLs for salinity tolerance revealed by genome-wide association studies of biomass, chlorophyll and tissue ion content in 176 rice landraces from BangladeshPLOS ONE

Dear Dr. Seraj,

Specifically, following things need to be addressed.1. Since there are some associations between genes and QTLs mentioned in section 3.4 and conclusion section include a sentence 'Supplementary analysis of gene expression and functional annotations assumes

432 potential roles for a number of genes within the identified QTLs', authors are advised to include few lines in the discussion section regarding this. 2. It seems there are some supplementary tables and figures not cited in the text. Please ensure that all supplementary information is cited.

Please submit your revised manuscript by Nov 25 2021 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.
A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.
An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Prasanta K. Subudhi, Ph.D.

Academic Editor

PLOS ONE

Journal Requirements:

Please review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the rebuttal letter that accompanies your revised manuscript. If you need to cite a retracted article, indicate the article’s retracted status in the References list and also include a citation and full reference for the retraction notice.

Additional Editor Comments (if provided):

Minor revision

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

PLoS One. 2021 Nov 5;16(11):e0259456. doi: 10.1371/journal.pone.0259456.r006

Author response to Decision Letter 2

13 Oct 2021

Comment from the editor:

1. Since there are some associations between genes and QTLs mentioned in section 3.4 and conclusion section include a sentence 'Supplementary analysis of gene expression and functional annotations assumes potential roles for a number of genes within the identified QTLs', authors are advised to include few lines in the discussion section regarding this.

2. It seems there are some supplementary tables and figures not cited in the text. Please ensure that all supplementary information is cited.

Response:

1. We have added lines 424 to 452 regarding the expression and functional studies.

2. We have cited all the supplementary information in the text:

Supplementary table 1 cited in line 81.

Supplementary table 2 cited in line 236.

Supplementary table 3 cited in line 237.

Supplementary table 4 cited in line 246.

Supplementary table 5 cited in line 249.

Supplementary table 6 cited in line 283.

Supplementary table 7 cited in line 318.

Supplementary table 8 cited in line 327.

Supplementary table 9 cited in line 329.

Supplementary table 10 cited in line 427.

Supplementary table 11 cited in line 430.

Supplementary table 12 cited in line 432.

Supplementary table 13 cited in line 433.

Supplementary table 14 cited in line 434.

Supplementary figure 1 cited in line 248.

Supplementary figure 2 cited in line 277.

Supplementary figure 3 cited in line 287.

Supplementary figure 4 cited in line 289.

Supplementary figure 5 cited in line 307.

Supplementary figure 6 cited in line 324.

Supplementary figure 7 cited in line 430.

Supplementary text 1 cited in line 426.

Attachment

Submitted filename: Response to editor.docx

Click here for additional data file.^{(12.9KB, docx)}

PLoS One. doi: 10.1371/journal.pone.0259456.r007

Decision Letter 3

Prasanta K Subudhi

20 Oct 2021

Novel QTLs for salinity tolerance revealed by genome-wide association studies of biomass, chlorophyll and tissue ion content in 176 rice landraces from Bangladesh

PONE-D-21-14535R3

Dear Dr. Seraj,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Prasanta K. Subudhi, Ph.D.

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

S1 Text. Gene expression and functional studies within the identified QTLs.

(DOCX)

Click here for additional data file.^{(40.4KB, docx)}

S1 Fig. Extended correlation matrix of observed traits.

(JPG)

Click here for additional data file.^{(5.3MB, jpg)}

S2 Fig. Density plots for basic and derived phenotypes.

(JPG)

Click here for additional data file.^{(809.5KB, jpg)}

S3 Fig. Changes in phenotype distributions after applying an ideal warping function.

(JPG)

Click here for additional data file.^{(324.5KB, jpg)}

S4 Fig. Residual plots for basic and derived phenotypes.

(JPG)

Click here for additional data file.^{(1.1MB, jpg)}

S5 Fig. Collage of Manhattan plots for traits with biomass associations.

(JPG)

Click here for additional data file.^{(7.4MB, jpg)}

S6 Fig. Observed extent of linkage disequilibrium.

(JPG)

Click here for additional data file.^{(300.1KB, jpg)}

S7 Fig. Gene expression heatmaps from S1 Text.

(JPG)

Click here for additional data file.^{(4.3MB, jpg)}

S1 Table. Data on vegetation and population structure of 176 plants from the study.

(XLSX)

Click here for additional data file.^{(24KB, xlsx)}

S2 Table. Heritability of traits calculated from genomic kinship and broad-sense heritability calculated from variance components.

(XLSX)

Click here for additional data file.^{(11.1KB, xlsx)}

S3 Table. Observed phenotype means for each genotype.

(XLSX)

Click here for additional data file.^{(69.6KB, xlsx)}

S4 Table. Phenotype values for genotypes estimated by genomic prediction.

(XLSX)

Click here for additional data file.^{(91.5KB, xlsx)}

S5 Table. P values for entire correlation matrix in S1 Fig.

(XLSX)

Click here for additional data file.^{(27.6KB, xlsx)}

S6 Table. Details on phenotype distributions before and after filtering.

(XLSX)

Click here for additional data file.^{(12.4KB, xlsx)}

S7 Table. Details on QTLs from previous studies that were found to have similar connotations as our findings.

(XLSX)

Click here for additional data file.^{(11.9KB, xlsx)}

S8 Table. All significant and suggestive marker associations that fall within the sequence range of known genes.

(XLSX)

Click here for additional data file.^{(59.3KB, xlsx)}

S9 Table. All significant and suggestive marker association within known gene regions from validation tests by the CMLM model in GAPIT.

(XLSX)

Click here for additional data file.^{(58.6KB, xlsx)}

S10 Table. Gene expression data for candidate genes.

(XLSX)

Click here for additional data file.^{(85KB, xlsx)}

S11 Table. Gene expression data from leaves and roots of a sensitive and a tolerant variety.

(XLSX)

Click here for additional data file.^{(172.6KB, xlsx)}

S12 Table. Effects of all 28188 possible substitution events within candidate gene regions.

(XLSX)

Click here for additional data file.^{(2MB, xlsx)}

S13 Table. Results of One-way ANOVA carried out for all markers within candidate gene regions.

(XLSX)

Click here for additional data file.^{(223.6KB, xlsx)}

S14 Table. Results of student’s t test carried out for all markers within candidate gene regions.

(XLSX)

Click here for additional data file.^{(972KB, xlsx)}

Attachment

Submitted filename: Response to reviewers.docx

Click here for additional data file.^{(31.3KB, docx)}

Attachment

Submitted filename: Response to reviewers 2.docx

Click here for additional data file.^{(24.7KB, docx)}

Attachment

Submitted filename: Response to editor.docx

Click here for additional data file.^{(12.9KB, docx)}

Data Availability Statement

All data and source code can be found at https://github.com/DeadlineWasYesterday/Cat-Does-Plant/.

[pone.0259456.ref001] 1.German Environment Agency, F.M.f.E.C.a.D., Ten million hectares of arable land worldwide are ’lost’ every year: less and less fertile and healthy soil. 2015, Umweltbundesamt: Germany https://www.umweltbundesamt.de/en/press/pressinformation/ten-million-hectares-of-arable-land-worldwide-are.

[pone.0259456.ref002] 2.Olesen J.E., et al., Impacts and adaptation of European crop production systems to climate change. European Journal of Agronomy, 2011. 34(2): p. 96–112. [Google Scholar]

[pone.0259456.ref003] 3.Gupta G.S.J.R.E.S., Land degradation and challenges of food security. 2019. 11: p. 63. [Google Scholar]

[pone.0259456.ref004] 4.Hossain M.S.J.I.R.J.B.S., Present scenario of global salt affected soils, its management and importance of salinity research. 2019. 1: p. 1–3. [Google Scholar]

[pone.0259456.ref005] 5.Eynard A., Lal R., and Wiebe K., Crop Response in Salt-Affected Soils. Journal of Sustainable Agriculture, 2005. 27(1): p. 5–50. [Google Scholar]

[pone.0259456.ref006] 6.Tsioumani E., The State of the World’s Biodiversity for Food and Agriculture: a Call to Action? Environmental Policy and Law, 2019. 49: p. 110–112. [Google Scholar]

[pone.0259456.ref007] 7.Solis C.A., et al., Back to the Wild: On a Quest for Donors Toward Salinity Tolerant Rice. 2020. 11(323). [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0259456.ref008] 8.McCouch S., et al., Feeding the future. 2013. 499(7456): p. 23–24. [DOI] [PubMed] [Google Scholar]

[pone.0259456.ref009] 9.Kromdijk J. and Long S., One crop breeding cycle from starvation? How engineering crop photosynthesis for rising CO 2 and temperature could be one important route to alleviation. Proceedings of the Royal Society B: Biological Sciences, 2016. 283: p. 20152578. doi: 10.1098/rspb.2015.2578 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0259456.ref010] 10.De Leon T.B., Linscombe S., and Subudhi P.K., Molecular Dissection of Seedling Salinity Tolerance in Rice (Oryza sativa L.) Using a High-Density GBS-Based SNP Linkage Map. Rice, 2016. 9(1): p. 52. doi: 10.1186/s12284-016-0125-2 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0259456.ref011] 11.Yu J., et al., Genome-wide association study and gene set analysis for understanding candidate genes involved in salt tolerance at the rice seedling stage. Molecular Genetics and Genomics, 2017. 292(6): p. 1391–1403. doi: 10.1007/s00438-017-1354-9 [DOI] [PubMed] [Google Scholar]

[pone.0259456.ref012] 12.Zhao Y., et al., New alleles for chlorophyll content and stay-green traits revealed by a genome wide association study in rice (Oryza sativa). Scientific Reports, 2019. 9(1): p. 2541. doi: 10.1038/s41598-019-39280-5 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0259456.ref013] 13.Kawahara Y., et al., Improvement of the Oryza sativa Nipponbare reference genome using next generation sequence and optical map data. Rice, 2013. 6(1): p. 4. doi: 10.1186/1939-8433-6-4 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0259456.ref014] 14.Pires I.S., et al., Comprehensive phenotypic analysis of rice (Oryza sativa) response to salinity stress. 2015. 155(1): p. 43–54. [DOI] [PubMed] [Google Scholar]

[pone.0259456.ref015] 15.Mansueto L., et al., Rice SNP-seek database update: new SNPs, indels, and queries. Nucleic acids research, 2017. 45(D1): p. D1075–D1081. doi: 10.1093/nar/gkw1135 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0259456.ref016] 16.annieart0, Free Bangladesh Map Illustration. https://www.vecteezy.com/vector-art/122269-free-bangladesh-map-illustration.

[pone.0259456.ref017] 17.Amin M., et al., Over-expression of a DEAD-box helicase, PDH45, confers both seedling and reproductive stage salinity tolerance to rice (Oryza sativa L.). Molecular Breeding, 2012. 30(1): p. 345–354. [Google Scholar]

[pone.0259456.ref018] 18.Cock J., Yoshida S., and Forno D.A., Laboratory manual for physiological studies of rice. 1976: Int. Rice Res. Inst. [Google Scholar]

[pone.0259456.ref019] 19.Gregorio G., Senadhira D., and Mendoza R., Screening rice for salinity tolerance, vol 22, IRRI discussion paper series. International Rice Research Institute, 1997. [Google Scholar]

[pone.0259456.ref020] 20.IRRI I.J.I.R.R.I., Philippine, Standard evaluation system for rice. 2002: p. 1–45. [Google Scholar]

[pone.0259456.ref021] 21.Endelman J.B., Ridge Regression and Other Kernels for Genomic Selection with R Package rrBLUP. 2011. 4(3): p. 250–255. [Google Scholar]

[pone.0259456.ref022] 22.Fusi N., et al., Warped linear mixed models for the genetic analysis of transformed phenotypes. Nature Communications, 2014. 5(1): p. 4890. doi: 10.1038/ncomms5890 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0259456.ref023] 23.Browning B.L., Zhou Y., and Browning S.R., A One-Penny Imputed Genome from Next-Generation Reference Panels. The American Journal of Human Genetics, 2018. 103(3): p. 338–348. doi: 10.1016/j.ajhg.2018.07.015 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0259456.ref024] 24.Raj A., Stephens M., and Pritchard J.K., fastSTRUCTURE: Variational Inference of Population Structure in Large SNP Data Sets. 2014. 197(2): p. 573–589. [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0259456.ref025] 25.Bradbury P.J., et al., TASSEL: software for association mapping of complex traits in diverse samples. Bioinformatics, 2007. 23(19): p. 2633–2635. doi: 10.1093/bioinformatics/btm308 [DOI] [PubMed] [Google Scholar]

[pone.0259456.ref026] 26.Letunic I. and Bork P., Interactive Tree Of Life (iTOL): an online tool for phylogenetic tree display and annotation. Bioinformatics, 2007. 23(1): p. 127–128. doi: 10.1093/bioinformatics/btl529 [DOI] [PubMed] [Google Scholar]

[pone.0259456.ref027] 27.Huang M., et al., BLINK: a package for the next level of genome-wide association studies with both individuals and markers in the millions. GigaScience, 2019. 8(2). doi: 10.1093/gigascience/giy154 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0259456.ref028] 28.Tang Y., et al., GAPIT version 2: an enhanced integrated tool for genomic association and prediction. 2016. 9(2): p. 1–9. [DOI] [PubMed] [Google Scholar]

[pone.0259456.ref029] 29.Turner S.D.J.B., qqman: an R package for visualizing GWAS results using QQ and manhattan plots. 2014: p. 005165. [Google Scholar]

[pone.0259456.ref030] 30.Chang C.C., et al., Second-generation PLINK: rising to the challenge of larger and richer datasets. GigaScience, 2015. 4(1). doi: 10.1186/s13742-014-0042-5 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0259456.ref031] 31.Mai N.T.P., et al., Discovery of new genetic determinants controlling the morphological plasticity in rice root and shoot under phosphate starvation using GWAS. 2020: p. 2020.10.31.363556. [DOI] [PubMed] [Google Scholar]

[pone.0259456.ref032] 32.To H.T.M., et al., Unraveling the Genetic Elements Involved in Shoot and Root Growth Regulation by Jasmonate in Rice Using a Genome-Wide Association Study. Rice, 2019. 12(1): p. 69. doi: 10.1186/s12284-019-0327-5 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0259456.ref033] 33.Zhang J., et al., QTL mapping and candidate gene analysis of ferrous iron and zinc toxicity tolerance at seedling stage in rice by genome-wide association study. BMC Genomics, 2017. 18(1): p. 828. doi: 10.1186/s12864-017-4221-5 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0259456.ref034] 34.Tao Y., et al., Genome-wide association mapping of aluminum toxicity tolerance and fine mapping of a candidate gene for Nrat1 in rice. PLOS ONE, 2018. 13(6): p. e0198589. doi: 10.1371/journal.pone.0198589 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0259456.ref035] 35.Zhang Y., et al., QTL identification for salt tolerance related traits at the seedling stage in indica rice using a multi-parent advanced generation intercross (MAGIC) population. Plant Growth Regulation, 2020. 92(2): p. 365–373. [Google Scholar]

[pone.0259456.ref036] 36.Cui Y., Zhang F., and Zhou Y., The Application of Multi-Locus GWAS for the Detection of Salt-Tolerance Loci in Rice. 2018. 9(1464). [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0259456.ref037] 37.Frouin J., et al., Tolerance to mild salinity stress in japonica rice: A genome-wide association mapping study highlights calcium signaling and metabolism genes. PLOS ONE, 2018. 13(1): p. e0190964. doi: 10.1371/journal.pone.0190964 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0259456.ref038] 38.Diop B., et al., Bridging old and new: diversity and evaluation of high iron-associated stress response of rice cultivated in West Africa. Journal of Experimental Botany, 2020. 71(14): p. 4188–4200. doi: 10.1093/jxb/eraa182 [DOI] [PubMed] [Google Scholar]

[pone.0259456.ref039] 39.Dingkuhn M., et al., Crop-model assisted phenomics and genome-wide association study for climate adaptation of indica rice. 2. Thermal stress and spikelet sterility. Journal of Experimental Botany, 2017. 68(15): p. 4389–4406. doi: 10.1093/jxb/erx250 [DOI] [PubMed] [Google Scholar]

[pone.0259456.ref040] 40.Delorean, E.E.J.-.-C.T. and Dissertations, Detecting durable resistance to rice bacterial blight. 2016.

[pone.0259456.ref041] 41.Schläppi M.R., et al., Assessment of Five Chilling Tolerance Traits and GWAS Mapping in Rice Using the USDA Mini-Core Collection. 2017. 8(957). [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0259456.ref042] 42.Rohila J.S., et al., Identification of Superior Alleles for Seedling Stage Salt Tolerance in the USDA Rice Mini-Core Collection. 2019. 8(11): p. 472. [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0259456.ref043] 43.Misra G., et al., Whole genome sequencing-based association study to unravel genetic architecture of cooked grain width and length traits in rice. Scientific Reports, 2017. 7(1): p. 12478. doi: 10.1038/s41598-017-12778-6 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0259456.ref044] 44.Huang X., et al., Genome-wide association studies of 14 agronomic traits in rice landraces. Nature Genetics, 2010. 42(11): p. 961–967. doi: 10.1038/ng.695 [DOI] [PubMed] [Google Scholar]

[pone.0259456.ref045] 45.Mather K.A., et al., The extent of linkage disequilibrium in rice (Oryza sativa L.). 2007. 177(4): p. 2223–2232. [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0259456.ref046] 46.Bhowmik S., et al., Identification of salt tolerant rice cultivars via phenotypic and marker-assisted procedures. 2007. 10(24): p. 4449–4454. [DOI] [PubMed] [Google Scholar]

[pone.0259456.ref047] 47.GARCIA A., et al., Sodium and potassium transport to the xylem are inherited independently in rice, and the mechanism of sodium: potassium selectivity differs between rice and wheat. 1997. 20(9): p. 1167–1174. [Google Scholar]

[pone.0259456.ref048] 48.Negrão S., et al., Recent updates on salinity stress in rice: from physiological to molecular responses. 2011. 30(4): p. 329–377. [Google Scholar]

[pone.0259456.ref049] 49.Golldack D., et al., Salinity stress-tolerant and -sensitive rice (Oryza sativa L.) regulate AKT1-type potassium channel transcripts differently. Plant Mol Biol, 2003. 51(1): p. 71–81. doi: 10.1023/a:1020763218045 [DOI] [PubMed] [Google Scholar]

[pone.0259456.ref050] 50.Kawasaki S., et al., Gene expression profiles during the initial phase of salt stress in rice. Plant Cell, 2001. 13(4): p. 889–905. doi: 10.1105/tpc.13.4.889 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0259456.ref051] 51.Formentin E., et al., Transcriptome and Cell Physiological Analyses in Different Rice Cultivars Provide New Insights Into Adaptive and Salinity Stress Responses. Frontiers in Plant Science, 2018. 9(204). doi: 10.3389/fpls.2018.00204 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0259456.ref052] 52.Schläppi M.R., et al., Assessment of Five Chilling Tolerance Traits and GWAS Mapping in Rice Using the USDA Mini-Core Collection. Frontiers in Plant Science, 2017. 8(957). doi: 10.3389/fpls.2017.00957 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0259456.ref053] 53.Jeong J. and Choi G., Phytochrome-interacting factors have both shared and distinct biological roles. Molecules and cells, 2013. 35(5): p. 371–380. doi: 10.1007/s10059-013-0135-5 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0259456.ref054] 54.Krusell L., et al., Shoot control of root development and nodulation is mediated by a receptor-like kinase. Nature, 2002. 420(6914): p. 422–6. doi: 10.1038/nature01207 [DOI] [PubMed] [Google Scholar]

PERMALINK

Novel QTLs for salinity tolerance revealed by genome-wide association studies of biomass, chlorophyll and tissue ion content in 176 rice landraces from Bangladesh

Md Nafis Ul Alam

G M Nurnabi Azad Jewel

Tomalika Azim

Zeba I Seraj

Roles

Abstract

1. Introduction

2. Methods

2.1 Data and source code

2.2 Plant material and growth conditions

2.3 Phenotyping

2.4 Phenotype statistics

2.5 Preprocessing genotype data

2.6 Population structure estimation

2.7 Genome-wide association studies

3. Results

3.1 Observation of population structure

Fig 1. Map of Bangladesh showing the locations where 114 out of 176 rice varieties from this study have been traditionally propagated and cultivated.

Fig 2. Population structure estimation results for 176 varieties.

3.2 Comprehensive analysis of phenotypes

Table 1. Basic and derived phenotypes.

Fig 3. Linear correlations (r) between observed and predicted phenotypes.

Fig 4. Complementarity between SES scores and principal component 1 derived from biomass and chlorophyll content in stress conditions.

3.3 Genome-wide association studies

Table 2. Identified major QTLs.

Fig 5. Manhattan and Q-Q plots for qCDP1.1, qCDP12.1 and qCDP12.3.

Fig 6. Manhattan and Q-Q plots for qCDP3.1, qCDP5.1, qCDP6.1, qCDP8.1 and qCDP9.1.

Fig 7. Manhattan and Q-Q plots for qCDP7.1, qCDP9.1 and qCDP9.3.

Fig 8. Manhattan and Q-Q plots for qCDP10.1 and qCDP12.2.

3.4 QTL mapping

4. Discussion

4.1 Biomass traits and chlorophyll content could be valuable indices for the screening of salinity tolerance

4.2 Some Bangladeshi landraces exhibit high degrees of salt tolerance

4.3 Potential genes of interest were found in the identified QTLs

5. Conclusion

Supporting information

Acknowledgments

Data Availability

Funding Statement

References

Decision Letter 0

Prasanta K Subudhi

Roles

Author response to Decision Letter 0

Decision Letter 1

Prasanta K Subudhi

Roles

Author response to Decision Letter 1

Decision Letter 2

Prasanta K Subudhi

Roles

Author response to Decision Letter 2

Decision Letter 3

Prasanta K Subudhi

Roles

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases