Assessing Heat Stress Tolerance of Wheat Genotypes through Integrated Molecular and Physio-Biochemical Analyses

Abstract

Heat as an abiotic stress significantly impairs the sustainable productivity of wheat (Triticum aestivum L.). To determine the tolerance of genotypes to heat stress, a comprehensive approach should be used that integrates simultaneous phenotyping and genotyping analyses. The aim of this study is to identify local heat-tolerant genotypes using simple sequence repeat (SSR) markers and evaluate the selected genotypes under field conditions for their tolerance to heat stress. Of the 12 SSR markers that showed polymorphism, eight were associated with six important traits. The use of hierarchical cluster analysis (HC) based on SSR markers led to the identification of 13 genotypes that showed varying results and were grouped into three distinct heat tolerance classes: tolerant (T), moderately tolerant (MT), and sensitive (S). The results showed that heat stress had a significant effect on 19 traits under this study, with significant variation in tolerance to heat stress between genotypes. The tolerant genotypes exhibited a range of average thousand-kernel weight (TKW) values between 40.56 and 44.85, while the sensitive genotype (Yecora Rojo) had an average TKW of 35.45. Furthermore, the tolerant genotypes showed two to three times higher levels of antioxidants compared to the sensitive genotypes when exposed to heat stress. Among the traits analyzed, six showed a favorable combination of high heritability (>60%) and genetic gain (>20%). Through the integration of principal component analysis and stepwise multiple linear regression, it was determined that six traits (grain yield, 1000-kernel weight, plant height, intercellular carbon dioxide, flag leaf area, and grain filling duration) revealed differences between the 13 genotypes. HC analysis of the six traits resulted in the same division of genotypes into three main categories as observed in an HC analysis based on SSR markers. It is worth noting that Saudi wheat, including KSU106, KSU105, and KSU115 as local genotypes, in addition to the 16HTWYT-22 genotype, showed higher heat tolerance compared to the other genotypes tested, indicating its potential suitability for agriculture in Saudi Arabia. These results contribute to breeding programs focused on developing heat-tolerant wheat varieties and accelerate progress in wheat productivity improvement programs.

1. Introduction

Wheat (Triticum aestivum L.) is a cereal crop of great importance to world agriculture, but its agriculture often suffers from the harmful effects of heat stress. In Saudi Arabia, for instance, a rise in temperature by one degree Celsius can reduce crop productivity by 7–25% [1]. According to the 2014 Intergovernmental Panel on Climate Change (IPCC) report, there has been a consistent rise in atmospheric temperatures since the beginning of the 21st century, with projections of a further increase of approximately 1.0–1.7 °C [2]. Consequently, wheat growth and development have been increasingly challenged by more frequent and severe heat stress episodes as global climate change continues to unfold [3]. In connection with this, World Bank’s 2019 report indicated that the global population is growing at a rate of 1.1% per year [4]. As a result, the demand for food to feed this growing population is increasing. Thus, studies on wheat have attracted considerable attention in recent years. However, research regarding Saudi wheat genotypes, particularly their resistance to climate change, is very limited.

Climate factors, especially heat, play a pivotal role in shaping agricultural productivity and are recognized as one of the most critical abiotic stresses. When temperatures surpass a specific threshold, it leads to a phenomenon known as “heat stress” [5]. Heat stress occurring during the plant reproductive phase has crucial effects on both grain yield and quality by reducing the duration and activity of leaf photosynthesis, as well as altering the metabolism of sugars [6,7,8,9]. When exposed to severe heat stress, the metabolism of plants can incur detrimental effects due to the generation of reactive oxygen species (ROS) during metabolism, leading to the peroxidation of lipid membranes and the impairment of nucleic acids and proteins [10,11]. Heat stress shortens the period of canopy growth and reduces the plant’s leaf area, affecting the plant’s ability to intercept radiation and exchange carbon dioxide through the stomata. Stomatal closure and decreased activity of the Rubisco enzyme (which is crucial for photosynthesis) disrupts the photosynthetic process. As a result, the plant’s biomass accumulation and grain-filling period are reduced [12,13]. To alleviate these consequences, plants activate both enzymatic and non-enzymatic ROS-scavenging systems to neutralizing the damaging impact of ROS [14]. These systems encompass enzymes such as catalase (CAT), peroxidase (POD), and superoxide dismutase (SOD). Plants have been observed to exhibit enhanced tolerance to heat stress when their antioxidant capacity is increased [15,16]. Grasses that had acclimated to heat environment were found to have lower levels of ROS due to their higher antioxidant concentrations [17]. The increase in antioxidant capacity within plant cells may be triggered by certain signaling molecules [18,19,20]. This suggests that the synthesis, accumulation, and upregulation of antioxidant pathways are adaptive responses employed by plants to mitigate the damaging effects of heat stress [21,22,23]. This adaptive response helps protect plant cells from the damaging effects of ROS [24,25,26]. Studies on wheat genotypes have revealed a positive correlation between antioxidant enzyme levels and chlorophyll content, as well as a negative correlation between antioxidant enzymes and the membrane injury index across most growth stages [15,27]. This observation can be attributed to the ability of antioxidants to scavenge ROS, maintain chlorophyll and the cell membrane, and protect leaves from senescence, thereby ensuring the continuity of metabolism and maintenance of plant productivity [28,29]. Physiological traits associated with heat tolerance, such as antioxidant capacity, chlorophyll content, proline, glycine betaine, and leaf water content, can serve as valuable indirect selection tools for genetically improving wheat cultivars [30,31,32]. These physiological indicators tend to be less affected by environmental factors and exhibit greater genetic stability compared to traits directly associated with yield [33,34,35,36].

Heat tolerance is a multifaceted genetic trait governed by many genes, which makes the task of selecting and developing genotypes that exhibit both high productivity and heat tolerance complex. This complexity poses significant challenges for plant breeders in achieving sustainable food production [37,38]. Consequently, plant breeders face major challenges in achieving this goal, necessitating close collaboration with researchers and experts in related fields [39]. Molecular marker techniques have been widely used in genetic studies, such as molecular-assisted selection (MAS) and locus mapping quantitative traits (QTL) [40,41,42,43], with several genome-wide association studies (GWAS) focusing on agronomic traits and abiotic tolerance to establish links between phenotypic and genetic differences [44,45]. Laboratory tests provide more accurate data to evaluate heat tolerance in genotypes, due to the strong correlation between resistance to abiotic stresses observed in both field and laboratory conditions [46,47,48,49]. DNA markers offer the advantages of independence from environmental conditions, developmental stages, and organ types, thus enhancing the reliability of results [50,51,52,53]. Among various molecular markers, simple sequence repeat (SSR) markers are of great interest as they show high information, specificity, and neutrality, and they possess multiple allelic characteristics. They are widely distributed across the genome, show reproducibility in different environments, and possess common and high-throughput properties [54,55,56]. This would improve the breeding efficiency of yielding and heat-tolerant genotypes [43,57,58]. Therefore, it is essential to identify molecular markers associated with trait-specific genes [42,43,59,60,61,62,63,64,65,66]. Extensive research has been dedicated to identifying the genomic regions associated with heat tolerance in plants. These studies have utilized QTL analysis, which examines the relationship between genetic markers and observable traits indicative of heat tolerance. The key traits investigated include grain filling duration, thousand-grain weight, yield, and canopy temperature [67,68,69], as well as grain size, milling quality in wheat [70,71], and oil, starch, and protein concentrations in maize [72]. Additionally, researchers have studied traits related to greening and senescence [69,73], thylakoid membrane damage, plasma membrane damage, and chlorophyll content [33,74,75,76], as well as grain weight stability associated with the stay-green trait [66,77]. Acuna-Galindo et al. [65] identified 66 QTL associated with 81 different traits related to drought and heat stress adaptation. Numerous QTLs associated with heat tolerance have been identified across various developmental stages in plants [78,79].

Recent researches efforts have focused on the use of molecular markers and the examination of morphological, physiological, and biochemical parameters to select superior cultivars [80,81]. It is essential to have inexpensive, rapid, and easily measurable methods for detecting heat tolerance in a large number of genotypes [46,82]. Plant breeders and researchers often employ a range of multivariate analysis techniques and multidimensional methods, such as PCA, SMLR, PC analysis, and HC, to screen for and differentiate the sources of variation in their data [80,81,82,83]. These statistical tools are crucial for accurately verifying and selecting desirable traits in breeding programs. To achieve success in developing heat-tolerant wheat cultivars, it is essential to integrate agro-physiological traits into the breeding process [84,85,86]. By combining insights gained from these complementary approaches, including phenotypic evaluation and molecular marker analysis, this study aimed to identify the best-performing, heat-tolerant, high-yielding wheat genotypes suitable for cultivation in the Saudi environment with the help of molecular markers and physio-biochemical features, in addition to multiple analyses. This integrative approach can provide a more comprehensive understanding of the complex genetic and physiological mechanisms underlying heat adaptation in wheat. Ultimately, this knowledge could enable the development of improved, heat-tolerant, high-yielding wheat varieties that are promising for future breeding programs.

2. Materials and Methods

Sixty genotypes were used to investigate molecular markers associated with heat tolerance. Thirteen genotypes (four tolerance, four moderate tolerant, and five sensitive) were selected for field and physiological studies based on the molecular results (Table S1).

2.1. Genomic DNA Extraction and SSR Markers

Sixty wheat genotypes were germinated in growth chambers. Once the plants reached the fourth leaf stage, the leaves were collected and immediately frozen in liquid nitrogen, then ground using a pestle in a mortar. Genomic DNA was extracted using a Cetyltrimethylammonium ammonium bromide (CTAB) procedure, as described by Saghai Maroof et al. [87]. The quantity and quality of DNA were estimated using 0.8% agarose gels. The DNA was stored at −20 °C for further analysis. Additionally, the quality and concentration of DNA were estimated using a Nano Drop Spectrophotometer (ND-8000, Thermo-scientific Wilmington, DE, USA). The DNA samples were then diluted to a concentration of 20 ng/μL using ddH2O and stored at −20 °C for SSR fingerprinting.

Forty SSR markers associated with heat stress were utilized in this examination, as established by multiple researchers. PCR reactions were conducted using a 96 thermocycler and following the protocol described in reference [88]. The PCR reactions were performed in a 20 μL volume using the Promega Green master mix, which contains dNTPs, Taq polymerase, and a 10× PCR buffer. The reaction mixture included 1 mM MgCl₂, 15 pmol of each primer, and 100 ng of genomic DNA as the template. Following PCR amplification, the resulting products were subjected to electrophoresis on 3% agarose gels. Ethidium bromide (EtBr) was used to stain the gels and visualize the DNA fragments, which were then placed in a Gel doc system. A 100 base pair ladder served as a standard for size comparison. The DNA fragments that separated in the agarose gels were visually examined and scored. Each allele fragment was scored as either 1 (presence) or 0 (absence) for each marker. Out of the 40 evaluated SSR markers, 12 exhibited polymorphism. These markers were found to be associated with heat stress tolerance in 12 genotypes. The study evaluated 13 genotypes (DHL2, 16HTWYT-22, KSU115, Yecora Rojo, 16HTWYT-38, KSU105, Klassic, KSU106, 16HTWYT-30, 16HTWYT-20, 16HTWYT-9, 16HTWYT-12, and Line-47) in the field. The genotypes underwent various measurements and assessments as part of the study.

2.2. Phenotyping, and Data Recording

2.2.1. Experimental Design

Thirteen genotypes were planted at the College of Food and Agricultural Sciences farm in Riyadh, Saudi Arabia during the 2021/22 growing season. The experimental fields were planted on two different dates: November 25, representing normal (C) conditions, and December 25, representing heat stress (H) conditions. The genotypes were planted in rows measuring 2.0 m in length with a spacing of 0.20 m between them. The experimental design used a split-plot arrangement with three replicates. Sowing dates assigned the main plots, while different genotypes were represented in the sub-plots. Cultural practices were carried out following the recommended guidelines.

2.2.2. Agro-Physio-Biochemical Traits

This study examined 17 agro-physio-biochemical traits. The agronomic traits were measured using the methodology provided by Al-Ashkar et al. [30]. These traits included duration of grain filling (GFD, in days), grain filling rate (GFR in g/day), plant height (PH in cm), flag leaf area (FLA in cm), spikelets per spike (NSS per spike), spike length (SL in cm), number of grains per spike (NGS), number of spike per square meter (NS per m²), grain yield (GY ton ha⁻¹), and thousand-kernel weight (TKW g). Additionally, we evaluated the physiological trait of net photosynthetic rate (Pn) and intercellular carbon dioxide (Ci) measured using a portable photosynthesis system Li-640 (LI-COR, Inc., Lincoln, NE, USA), in addition to several biochemical traits, including polyphenol oxidase (PPO), superoxide dismutase (SOD), catalase (CAT), peroxidase (POD), and glycine betaine (GB), as shown below.

The levels of antioxidant enzymes were measured, including catalase (CAT), peroxidase (POD), superoxide dismutase (SOD), and polyphenol oxidase (PPO). Fresh leaf samples weighing 0.5 g were utilized. The leaves were crushed in liquid nitrogen and suspended in a buffer solution consisting of 50 mM potassium phosphate buffer (pH 7.8) and 1% (w/v) polyvinyl polypyrrolidone for extraction. The samples were then centrifuged at 14,000 rpm for 10 min at a temperature of 4 °C. The enzyme extract used to assess the activity of CAT, POD, PPO, and SOD in subsequent assays was obtained from the supernatant, which was prepared according to the instructions provided in [89,90,91].

The CAT activity was measured following the protocol described by Aebi [89]. The assay was performed in a 3 mL reaction mixture containing the following components: 1 mL of 0.1 M PB (pH 7.2), 1 mL of 75 mM H₂O₂ solution, and 1 mL of the enzyme extraction. The decrease in absorbance at 240 nm was recorded every 20 s for a total duration of 3 min. The SOD activity was measured according to Kono [92]. A reaction mixture was prepared that contained the following components: 1.3 mL of 50 mM sodium carbonate buffer, 500 μL of 96 μM NBT, 100 μL of 0.6% Triton X-100, and 100 μL of 20 mM hydroxylamine hydrochloride. This reaction mixture was incubated at room temperature for 2 min; then, 70 μL of enzyme extract was added. The absorbance of the mixture was immediately read at 560 nm, reading the absorbance every 20 s for 1–2 min. Reaction components for POD 3.0 mL PB (100 mM, pH 7.0), 50 μL guaiacol solution (20 mM), 30 μL H₂O₂ solution (12.3 mM) and 100 μL of enzyme extract were mixed well, and the absorbance was read at 436 nm. The absorbance was read up to 3 min [93]. The PPO activity was determined using the method outlined by Duckworth and Coleman [91]. The reaction was carried out in a mixture containing 0.03 mL of the enzyme solution and 1.74 mL of a 20 mM catechol solution (was prepared using a 50 mM PB, pH 6.8 at 25 °C). The absorbance of the reaction mixture was then measured at a wavelength of 420 nm.

The antioxidant enzyme activities were all measured based on protein concentration of the samples. The protein quantification was carried out following the protocol described by Bradford [94], utilizing bovine serum albumin (BSA) as the calibration standard.

The levels of Glycine betaine (GB) were determined by analyzing ground leaf samples with liquid nitrogen, following the methodology described by Sallam et al. [81]. Subsequently, 1 mg of the ground sample was transferred to a glass tube, and 1.5 mL of 2 N H₂SO₄ was added. The tube was then placed in a water bath set at 60 °C for 10 min to extract Glycine betaine. After centrifugation at 14,000 rpm for 10 min, the resulting supernatants were collected for further analysis. To determine the concentration of GB, 125 μL of the supernatant sample was mixed with 50 μL of a cold solution of potassium tri-iodide (KI-I₂). The KI-I₂ solution was prepared by dissolving 15.7 g of iodine and 20 g of potassium iodide in 100 mL of distilled water. The mixture was left at a temperature of 0–4 °C for 16 h and then centrifuged at 14,000 rpm for 15 min. The upper liquid phase was discarded. To dissolve any remaining small crystals in the tube chamber, 1.4 mL of 1,2-dichloroethane was added, and the solution was incubated for approximately 2–2.5 h. The resulting samples were analyzed using a spectrophotometer at a wavelength of 365 nm (U-2000, Hitachi Instruments, Tokyo, Japan) to determine the concentration of GB. A standard curve was prepared using betaine stock solution with concentrations of 1, 2, 4, 6, and 8 μL, which served as references to calculate the GB concentration in the samples.

2.2.3. Quality Traits

Two traits (protein content (PC) and gluten index (GI)) were estimated as explained in detail by the Committee [95].

2.3. Statistical Analysis

For agro-phenotypic analysis, the data of various parameters were analyzed by analysis of variance (ANOVA) and genetic parameters 19 traits were implemented using SAS v9.2 software (SAS Institute, Inc., Cary, NC, USA). The variance (mean squares) of data for 19 traits was used to compute variance components that are used to compute genetic parameters (genetic variance (σ²G), residual variance (σ²e), phenotypic variance (σ²Ph), heritability (h² %), genotypic coefficient of variability (G.C.V %), phenotypic coefficient of variability (Ph.C.V. %), genetic advance (GA), and genetic gain (GG)), as described by Al-Ashkar et al. [96]. Principal component analysis (PCA) was carried out based on data provided by the correlation matrix to find out the variable contributing the most to the variance and the components loading the most on the variables. PCA identified important traits that were located in the first two components. In SMLR (stepwise multiple linear regression), PC (path coefficient), (hierarchical cluster), and LD (liner discriminant) analysis, 19 traits were used in SMLR to determine the key traits that contribute to enhancing and developing the variable of interest (GY), after which PC analysis was used to divide variation into direct and indirect effects. The effective indices (six out of nineteen traits) were used in the HC analysis to evaluate the phenotypic dissimilarity matrix between 13 genotypes, characterized into three groups For the Euclidean distance and ward’s method of agglomeration, LD analysis was employed to validate the genotype tolerance categories (the six indices used as quantitative variables) with the three categories (as quantitative variables). In the case of genotyping analysis, SSR bands were scored (present (1) or absent (0) to create a binary matrix). The genetic dissimilarity (matrix of pairwise) between genotypes was calculated used the coefficient of Jaccard dissimilarity. Agglomerative HC analysis was implemented using the unweighted pair group average method (UPGAM). Statistical analyses (PCA, SMLR, PC, HC, and DFA) were implemented through XLSTAT statistical package software (vers. 2019.1, Excel Add-ins soft SARL, New York, NY, USA).

3. Results

3.1. Genotypic Analysis Based on SSR Markers

Based on the outcomes of the stepwise marker analysis, a hierarchical clustering (HC) approach was employed to categorize the 13 genotypes into three main groups. The clustering was performed using the marker that exhibited correlations with the parameters under control, heat stress, and relative change, as depicted in

Table 1. Selection of influential marker (independent variables) with six studies traits (dependent variables) for control, heat, and relative change based on SMLR analysis.

Traits	Treatments	Markers	R2 Par.	R2 Com.	p-Value *
GY	Control	wmc527	0.790	0.790	<0.0001
	Heat	wmc326	0.606	0.606	0.002
		gwm337	0.180	0.786	0.016
	Relative change	wmc326	0.671	0.671	0.001
		gwm337	0.133	0.805	0.026
TKW	Control	wmc154	0.472	0.472	0.010
	Heat	wmc326	0.764	0.764	<0.0001
	Relative change	wmc326	0.672	0.672	0.001
FLA	Heat	wmc326	0.699	0.699	0.0001
	Relative change	wmc326	0.866	0.866	<0.0001
		wmc65	0.049	0.915	0.037
PH	Control	wmc65	0.328	0.328	0.041
	Heat	wmc326	0.359	0.359	<0.0001
		wmc54	0.318	0.677	<0.0001
		wmc65	0.205	0.882	<0.0001
		gwm369	0.083	0.965	0.002
	Relative change	gwm369	0.603	0.603	0.002
GFD	Control	wmc65	0.736	0.736	0.000
	Heat	wmc326	0.873	0.873	<0.0001
		wmc65	0.048	0.921	<0.0001
		wmc54	0.052	0.972	0.000
		wmc74	0.017	0.989	0.005
		gwm337	0.005	0.994	0.047
	Relative change	wmc326	0.746	0.746	<0.0001
		gwm337	0.125	0.871	0.001
		wmc527	0.062	0.933	0.006
		wmc54	0.032	0.965	0.028
Ci	Control	wmc326	0.462	0.462	0.011
	Heat	wmc326	0.764	0.764	<0.0001
	Relative change	wmc326	0.656	0.656	0.001

Coefficient partial determination (R² Par.), cumulative coefficient determination (R² Com.), * means p-value of coefficient partial determination. Abbreviation as described in Section 2.

. The first category, denoted as “T” for tolerance, encompassed four genotypes (KSU105, KSU106, 16HTWYW-22, and KSU115); the second category, labeled as “MT” for moderate tolerant, included four genotypes (16HTWYW-38, 16HTWYT-20, 16HTWYT-9, and DHL2). Lastly, the third category, named “S” for sensitive, covered five genotypes (16HTWYW-30, 16HTWYW-12, Line47, Yecora Rojo, and Klassic) (Figure 1). The clustering of the 13 genotypes based on phenotypic distance was further evaluated for its association with genetic distance using the Mantel test. The test revealed a significant positive correlation (r = 0.440, p < 0.012, and alpha = 0.05) between the phenotypic distance and genetic distance. These positive correlations highlight the importance of utilizing SSR markers as an effective tool for identifying tolerant genotypes at the early stages of a breeding program.

3.2. Exploring the Relationship between SSR Markers and Agro-Physio-Biochemical Traits

The study employed stepwise multiple linear regression (SMLR) analysis to identify the SSR markers that had the strongest association with the agrophysical and biochemical traits measured under control, temperature stress conditions and relative rate of change (RC). The analysis revealed that specific SSR markers were significantly associated with 6 out of the 19 agrophysical and biochemical traits that were examined (Table 1). The cumulative R² value under control conditions ranged from 0.328 for PH to 0.790 for GY; under heat conditions, it ranged from 0.357 for PH to 0.994 for GFD, and the RC value ranged from 0.656 for Ci to 0.965 for GFD. Interestingly, wmc326 was significantly correlated with all six traits under heat stress and under relative change except for plant height. It also showed significant correlation with Ci under all conditions and with FLA under heat stress and under relative change. The results also showed that six molecular markers (wmc326, wmc65, wmc54, wmc74, gwm337, and wmc527) were correlated with GFD under heat stress, control, and relative change and had the highest cumulative R² 0.994. The plant height trait was also associated with five markers (Table 1).

3.3. Phenotypic Analysis for Assessing Terminal Heat Tolerance in Wheat Genotyp

Significant variations (p < 0.01) were observed between the optimal and heat-stressed treatments (I) for 16 measured traits, as indicated by ANOVA. Similarly, highly significant variations (p < 0.01) were observed for the genotypes (G) and the interaction between treatment and genotype (I × G) across the same nineteen traits (

Table 2. Analysis of 19 traits estimated in 13 wheat genotypes under heat stress and control.

Source	DF	PPO	CAT	POD	SOD	GB	Pn	Ci	FLA	GFD	GFR
H	1	0.35 **	103.99 **	215.662 **	4.54 **	33.28 **	1067.31 **	171,829.81 **	3924.31 **	284.62 **	308.07 **
Error A	2	0.002	0.005	0.668	0.004	0.003	0.537	24.453	6.888	0.205	0.059
G	12	0.023 **	0.470 **	5.706 **	0.031 **	1.568 **	6.901 **	1821.177 **	277.864 **	69.78 **	6.026 **
H*G	12	0.009 **	0.172 **	4.637 **	0.018 **	0.611 **	2.964 **	501.600 **	61.524 **	26.10 **	2.500 **
Error B	48	0.001	0.012	0.13	0.004	0.021	0.595	77.837	9.595	2.896	0.327
Genetic Parameters
σ2G		0.002	0.05	0.178	0.002	0.16	0.656	219.929	36.057	7.281	0.587
σ2e		0.001	0.002	0.022	0.001	0.003	0.099	12.973	1.599	0.483	0.054
σ2Ph		0.004	0.078	0.951	0.005	0.261	1.15	303.529	46.311	11.631	1.004
h2 %		59.396	63.425	18.72	42.134	61.043	57.052	72.457	77.858	62.601	72.63
G.C.V. %		14.666	12.67	12.815	11.203	15.933	8.656	7.161	14.884	6.6	6.07
Ph.C.V. %		19.03	15.909	29.619	17.259	20.392	11.46	8.413	16.868	8.342	7.123
GA		0.076	0.366	0.376	0.063	0.643	1.26	26.005	10.915	4.398	9.342
GG %		23.284	20.785	11.422	14.981	25.643	13.469	12.557	27.055	10.757	10.65
Source	DF	PH	SL	NG	NS	NSS	TKW	GY	PC	GI
Rep	2	1.423	2.013	19.828	217.782	1.59	1.34	0.027	0.745	28.057
H	1	2820.013 **	29.538 **	4078.154 **	233,317.385 **	108.513 **	997.064 **	92.966 **	430.755 **	1829.522 **
Error A	2	17.321	0.346	41.734	160.731	0.359	0.237	0.013	0.779	52.474
G	12	233.902 **	11.959 **	207.238 **	5597.440 **	7.504 **	48.428 **	1.629 **	7.021 **	731.008 **
H*G	12	64.013 **	0.733 ns	76.916 *	2051.051 **	2.957 **	20.011 **	0.728 **	3.685 **	179.671 **
Error B	48	8.191	0.471	33.133	87.701	0.808	1.2	0.015	0.723	17.454
Genetic Parameters
σ2G		28.315	1.871	21.72	591.065	0.758	4.736	0.15	0.556	91.889
σ2e		1.365	0.079	5.522	14.617	0.135	0.2	0.002	0.121	2.909
σ2Ph		38.984	1.993	34.54	932.907	1.251	8.071	0.271	1.17	121.835
h2 %		72.633	93.872	62.885	63.357	60.592	58.679	55.29	47.517	75.421
G.C.V. %		6.071	14.302	12.52	5.04	5.485	5.175	5.918	5.471	12.42
Ph.C.V. %		7.123	14.762	15.788	6.332	7.046	6.756	7.958	7.936	14.301
GA		9.342	2.73	7.613	39.864	1.396	3.434	0.593	1.059	17.149
GG %		10.658	28.545	20.452	8.264	8.795	8.167	9.064	7.768	22.22

* = significant at p ≤ 0.05, ** = significant at p ≤ 0.01, ns = insignificant. Abbreviation as described in Section 2.

). The (h²) for the traits exhibited high levels for twelve traits (60.592% ≤ h² ≤ 93.872%) and moderate levels for six traits (59.396% ≤ h² ≤ 42.134%). The GG values were high for seven traits (28.545% ≤ GG ≤ 20.452%) and moderate for seven traits (14.984% ≤ GG ≤ 10.567%). The PCV and GCV The values exhibited both convergence and divergence across different traits. moreover, for all traits, the σ²G was smaller than the σ²Ph (Table 2).

3.4. Principal Component Analysis of Agro-Physio-Biochemical Traits under Control and Heat Stress

The first three principal component analyses (PCAs) showed that the eigenvalues were greater than 1 and together explained 87.018% of the total variance for the 19 traits studied. PCA1 and PCA2 showed 63.752% and 16.861% of the total variance, respectively. PCA1 (values of ≥0.337) was associated with 16 traits (PPO, CAT, SOD, Pn, Ci, FLA, GFD, GFR, PH, SL, NGS, NSS, NS, TKW, GY, and PC), while PCA2 was primarily related to two traits (GB and POD), and PCA3 was specifically linked to the GI trait (

Table 3. PCA of 13 wheat genotypes: eigenvalues proportion, and cumulative variance for the first three components of measured traits under heat.

	PCA1	PCA2	PCA3
Eigenvalue	12.113	3.204	1.217
Variability (%)	63.752	16.861	6.404
Cumulative %	63.752	80.613	87.018
Eigenvectors:
POD	0.412	0.495	0.000
SOD	0.691	0.244	0.000
PPO	0.337	0.307	0.212
CAT	0.745	0.208	0.007
GB	0.331	0.561	0.012
Pn	0.974	0.000	0.001
Ci	0.951	0.003	0.001
FLA	0.739	0.184	0.013
GFD	0.416	0.415	0.016
GFR	0.713	0.054	0.003
PH	0.507	0.019	0.289
SL	0.374	0.286	0.006
NS	0.875	0.031	0.004
NSS	0.700	0.180	0.001
NGS	0.623	0.002	0.061
TKW	0.793	0.149	0.003
GY	0.918	0.021	0.000
Pc	0.785	0.018	0.015
GI	0.231	0.026	0.572

Values in bold indicate related traits, Abbreviation as described in Section 2.

). The correlations between traits on PCA1 and PCA2, as illustrated in Figure 2, are indicative of the relationships between these traits. The PCA analysis accounted for a total of 19 traits. PCA1 displayed a positive correlation with thirteen traits and a negative correlation with six traits. On the other hand, PCA2 exhibited a positive correlation seventeen traits and a negative correlation with two traits. It is noteworthy that all genotypes subjected to heat stress demonstrated negative correlations with PCA. Furthermore, seven of these genotypes exhibited negative correlations with PCA2 as well.

3.5. SMLR and PC Analysis for the Preformance of Yield Traits

To prioritize yield as the primary objective, both stepwise multiple linear regression (SMLR) and path coefficient (PC) analysis were employed to identify genetically influenced traits. The SMLR analysis yielded findings that were noteworthy with regard to the contributions of various traits. Among the traits under consideration, TKW, PH, and Ci demonstrated significant contributions, with values of 0.868, 0.066, and 0.025 respectively (

Table 4. Stepwise regression analysis for grain yield, thousand-kernel weight, and flag leaf area (dependent index) with three yield-related traits (independent index).

Stepwise Regression					Path Coefficient
Dependent	Source				Partitioning the Correlation				R2
		Regression Coefficient	p-Value	R2 Par.	R2 Com.	Direct Effect	Indirect Effect	Correlation Value	Direct Effect
GY	Intercept	0.011
	TKW	0.585	0.004	0.868	0.868	0.494	0.437	0.931	0.244
	PH	0.530	0.005	0.066	0.934	0.343	0.474	0.817	0.118
	Ci	0.316	0.047	0.025	0.958	0.259	0.581	0.840	0.067
	Total direct effect								0.429
	Total indirect effect								0.529
	Total R2				0.958				0.958
	Residual				0.205				0.205
TKW	Intercept	0.010
	FLA	0.481	<0.001	0.780	0.780
	Total direct effect
	Total indirect effect
	Total R2				0.780				0.780
	Residua
FLA	Intercept	−0.117
	GFD	0.794	0.011	0.699	0.699	0.511	0.325	0.836	0.261
	Ci	0.951	0.012	0.146	0.845	0.502	0.331	0.833	0.252
	Total direct effect								0.513
	Total indirect effect								0.332
	Total R2				0.845				0.845
	Residual				0.394				0.394

). Collectively, these three traits accounted for 0.958 of the total contribution (with a residual value of 0.205), as shown in Table 4. Given the strong correlation observed between TKW and yield, further investigation was conducted into traits associated with TKW. Specifically, FLA exhibited a substantial contribution value of 0.780. Given the strong correlation between FLA and TKW, an additional analysis was conducted to explore the traits linked to FLA and determine their direct and indirect effects on the same trait. In this analysis, two traits were identified as strongly associated with leaf area: GFD and Ci, contributing 0.699 and 0.146, respectively. Collectively, these two traits accounted for 0.845 of the total contribution, with a residual value of 0.394.

Based on the results of the SMLR, the effects of the three traits (TKW, PH, and Ci) that showed an association with GY were divided into direct and indirect effects through the path coefficient analysis. The TKW trait affected the GY for 0.244 (as a direct effect) out of 0.958, while the total direct and indirect impacts for the three traits were 0.429 and 0.529, respectively. When FLA was considered as the dependent variable, GFD trait affected leaf area for 0.261 (as a direct effect) out of a total of 0.845, while the total direct and indirect effects of the two traits (GFD and Ci) were 0.513 and 0.332, respectively (Table 4).

3.6. Hierarchical Clustering and Linear Discriminant Analysis

Based on the results of the SMLR and PC (which showed that five traits were associated with the GY and contributed to significant effects on the GY, whether directly or indirectly), we conducted a hierarchical clustering analysis of the 13 genotypes based on their performance using RC for these five traits (TKW, PH, Ci, FLA, and GFD) in addition to the GY. The results showed that the hierarchical distribution of the genotypes was distributed into three distinct groups. The first category, labeled as “T” for tolerant, encompassed four genotypes (KSU106, KSU105, 16HTWYT-22, and KSU115). The second category, denoted as “S” for sensitive, included seven genotypes (16HTWYT-30, Yecora Rojo, 16HTWYT-20, Klassic, 16HTWYT-12, Line47, and 16HTWYT-9). Finally, the third category, named “MT” for moderate tolerant, consisted of two genotypes (DHL2 and 16HTWYT-38) (Figure 3). To increase the credibility of the categorizations, an LD analysis was performed. This analysis aimed to validate the categories of the three groups (T, MT, and S) both before and after the analysis. The results of the LD analysis showed that the prior and posterior categories were completely consistent across all 13 genotypes, achieving a 100% accuracy rate. The membership probability values, which indicate the degree of compatibility between the prior and posterior categories, were (>0.5), with a value of 1 for all genotypes included in the study (

Table 5. Posterior probability of membership in heat groupings through linear discriminant.

Genotypes		Classification				Cross-Validation
	Prior	Posterior	Membership Probabilities			Posterior	Membership Probabilities
			Pr(MT)	Pr(S)	Pr(T)		MT	S	T
16HTWYT-30	S	S	0.000	1.000	0.000	S	0.000	1.000	0.000
DHL2	MT	MT	1.000	0.000	0.000	MT	1.000	0.000	0.000
16HTWYT-20	S	S	0.000	1.000	0.000	S	0.000	1.000	0.000
16HTWYT-38	MT	MT	1.000	0.000	0.000	MT	1.000	0.000	0.000
16HTWYT-9	S	S	0.000	1.000	0.000	S	0.000	1.000	0.000
KSU105	T	T	0.000	0.000	1.000	T	0.000	0.000	1.000
16HTWYT-12	S	S	0.000	1.000	0.000	S	0.000	1.000	0.000
Yecora Rojo	S	S	0.000	1.000	0.000	S	0.000	1.000	0.000
16HTWYT-22	T	T	0.000	0.000	1.000	T	0.000	0.000	1.000
KSU115	T	T	0.000	0.000	1.000	T	0.000	0.000	1.000
Klassic	S	S	0.000	1.000	0.000	S	0.000	1.000	0.000
Line 47	MT	MT	1.000	0.000	0.000	MT	1.000	0.000	0.000
KSU106	T	T	0.000	0.000	1.000	T	0.000	0.000	1.000

).

4. Discussion

Abiotic stresses significantly impact plant growth and productivity, with heat and drought being the primary factors affecting global wheat production [97,98]. Climate projections indicate that global average temperatures are expected to rise by approximately 6 °C by the end of the 21st century [99,100]. Research has shown that even a 1 °C increase in temperature could result in a 6% decline in worldwide wheat production, particularly if the warming occurs during the plant’s reproductive phase, resulting in substantial losses in grain yield [101,102]. The integration of genotyping and molecular marker development holds great potential to enhance our understanding of these issues and provide valuable resources for selecting and breeding new wheat varieties that are tolerant to environmental stresses, including heat stress, which is particularly relevant for Saudi Arabia [40,103,104,105]. These advancements could play a critical role in ensuring sustainable wheat production in the face of a changing climate. The surrounding environment can influence the genotypes, and variations in soil conditions can negatively impact the field evaluation of these genotypes. However, the DNA within the cells remains unaffected by these external factors. This distinction is what sets selection based on molecular markers apart from other approaches [25,106,107,108].

The marker trait association (MTA) approach represents a promising avenue for plant breeding, offering a potential solution to the limitations inherent in linkage mapping [109,110]. In the present study, a set of 60 genotypes was screened using an SSR marker. From this screening, 13 selected genotypes displayed varying degrees of heat tolerance. These molecular markers, which are linked to quantitative trait loci (QTLs) in Marker-Assisted Selection (MAS), play a crucial role in determining key agro-physio-biochemical traits under heat stress [61,62,111,112,113,114]. In this particular study, we conducted HC analysis based on SSR markers, resulting in the categorization of 13 genotypes into three distinct categories (Figure 1). By employing a complementary approach that combined phenotyping analysis (based on phenotypic distance) and genotyping analysis (based on genetic distance), we revealed a significant positive correlation (r = 0.440 and p < 0.012) through the Mantel test. The strong correlation observed between the phenotypic and genotypic traits of the genotypes studied underscores the close association between these characteristics. This is a significant finding, as it emphasizes the effectiveness of SSR markers as a valuable tool for identifying tolerant genotypes early on in a breeding program. The results of this study align with numerous previous reports that have documented similar findings [61,62,81,114]. Interestingly, in the genotyping analysis, two genotypes (16HTWYT-20 and 16HTWYT-9) exhibited greater genetic distance compared to the categories determined by phenotyping analysis (Figure 1 and Figure 2). However, out of the 13 genotypes assessed, 11 (16HTWYT-30, 16HTWYT-12, KSU106, KSU105, Line-47, DHL2, 16HTWYT-22, KSU115, 16HTWYT-38, Yecora Rojo, and Klassic) demonstrated complete concordance between genotyping and phenotyping analyses. The congruence lends credibility and precision to both techniques employed in the evaluation and sorting of genotypes, facilitating the selection of the optimal genotypes capable of tolerating varied environmental conditions.

Given the importance of molecular markers and their independence from the surrounding environment, their association with crop, physiological, and biochemical traits is of great importance and reliability for plant breeders. Our results showed a strong association between some SSR markers (wmc326, wmc65, wmc54, wmc74, gwm337, and wmc527) and six important traits (GY, TKW, FLA, PH, GFD, and Ci) (Table 1). These traits can be effective indicators in screening and selecting genotypes for their tolerance to heat stress and for future studies in breeding and improvement programs [33,61,66,67,115,116].

Because productivity under field conditions is the main goal of breeding programs, field evaluation of the 13 selected genotypes was conducted under heat stress and control conditions by evaluating 19 agro-physio-biochemical traits (Table 2). The results revealed significant differences in performance between the genotypes under the control and heat stress conditions, indicating that these genotypes were highly susceptible to the effects of heat stress (Table 2) [23,25]. Moreover, the genotypes exhibited a high degree of genetic diversity, with varying performance across all the measured traits. The heat-tolerant genotypes (KSU106, KSU105, 16HTWYW-22, and KSU115) displayed the highest mean values for the assessed traits. For instance, the heat-tolerant genotype KSU115 recorded the highest value of TKW, GY, and FLA of 42.20 g, 6.77 ton ha⁻¹, and 54.9 cm², respectively, compared to the sensitive genotype Yecora Rojo, which had 35.4 g, 4.7 ton ha⁻¹, and 25.21 cm², respectively, under the same heat stress conditions [57,81,84]. Furthermore, the heat-tolerant genotypes exhibited the highest mean values for antioxidants and glycine betaine. The mean values for antioxidants in the heat-tolerant genotypes were twice as high as those recorded in the sensitive genotypes [23,24].The findings indicate that heat-tolerant plant genotypes possess superior physiological and biochemical mechanisms to mitigate the detrimental effects of heat stress [117,118,119]. This enables them to outperform heat-sensitive genotypes. The differential response of heat-tolerant genotypes to heat stress is attributed to their enhanced capacity to produce antioxidants (CAT, POD, and SOD) when exposed to heat stress [27,109]. These antioxidants protect cellular membranes and proteins by scavenging ROS [24,26,120]. As a result, they preserve chlorophyll levels and leaf health, allowing photosynthesis to continue unimpeded. This helps maintain high wheat yields under stressful heat conditions, in contrast to heat-sensitive genotypes that are negatively impacted by the stress and lack sufficient antioxidant levels to provide adequate protection. This genetic diversity can be leveraged to develop effective phenotypic selection strategies for breeding heat-tolerant cultivars, particularly for quantitative traits as reported in previous studies [121,122,123]. The strong interaction observed between the genotypes and the environmental conditions was a key factor for the traits evaluated in this study. This highlights the necessity for conducting extensive evaluations of genotypes across a diverse range of environments in order to thoroughly assess and characterize their heat-related characteristics [124]. Phenotypic variation among the genotypes played a notable role in the dominant genotype–environment interaction for all traits, as evidenced by the findings presented in Table 2. The level of heat tolerance improvement depends on the extent of genetic variation and heritability of the traits. (h²) provides valuable information about the proportion of genetic contribution relative to phenotypic variation, enabling predictions of validity and the reliability of phenotypic values [80,125,126,127]. Previous studies by Al-Ashkar et al. [80] and Burton [128] have demonstrated that combining h², (GCV), and (GA) allows for credible assessments of the expected GG through phenotypic selection.

The h² values exceeding 60.0% and the GG values exceeding 20.0% together indicate that the variance is due to genetic factors, making traits with high values of this parameter effective and reliable indicators for the selection of genotypes tolerant to different environmental conditions [127]. The GY trait is highly environmentally sensitive and has low h² and is not reliable for direct selection of genotypes. Therefore, combining other traits can serve as powerful indicators for indirect selection of improved grain yield under heat stress conditions. This integrated, multi-faceted approach enables plant breeders to make more well-informed and balanced decisions when selecting desirable genotypes for their breeding programs [80].

In our study, we found that six traits (CAT, GB, FLA, SL, NGS, and GI) had high h² (>60.0%) and high GG (>20.0%), and seven traits (PPO, SOD, Pn, Ci, GFR, GFD and PH) had intermediate h² (>30.0%) and GG (>10.0%). And the remaining traits displayed one of the following scenarios: low h² coupled with low GG, intermediate heritability but low genetic gain, or vice versa (Table 2). These results indicate that the first and second set of traits are reliable screening indicators for selecting genotypes for tolerance to different environmental stresses because they have high to moderate h² and high to moderate GG [127,129]. These traits demonstrate effectiveness and promise as screening criteria in breeding programs. Utilizing them can facilitate the selection of genotypes with stress tolerance, which can then be used to transfer these advantageous traits to other genotypes through breeding. Notably, all the measured traits demonstrated a genotypic coefficient of variation (GCV) lower than the phenotypic coefficient of variation (PCV), despite the preference of breeders for higher GCV values over PCV values [130].

The PCA is a valuable tool for identifying significant traits positioned within the first two components, as demonstrated by loadings ≥ 0.337 in Table 3. Genotypes that exhibited heat tolerance (KSU106, 16 HWY-22, KSU105, and KSU110) demonstrated notable correlations with the majority of physiological traits, as depicted in Figure 2. By integrating genetic parameters and PCA outcomes, we discovered five traits (CAT, GB, FLA, SL, and NGS) that exhibited high heritability and genetic gain values while being located within the first two components. Consequently, these traits serve as efficient screening criteria [80,82].

Heat tolerance is a complex genetic trait that is significantly influenced by environmental factors. Consequently, the sole reliance on GY as an indicator of genotypes’ heat tolerance index is inadequate [80]. To ensure the accuracy of our findings, we employed additional statistical analyses, specifically SMLR and PC analyses. These analyses provide valuable insights into the dependent correlations with independent variables [129,131]. The results of the SMLR and PC analyses demonstrated a strong association positive between TKW, PH, and Ci with GY, with an R2 value of 0.958 and a significance level of p < 0.047. Among these traits, TKW contributed the most (0.868), followed by PH (0.066) and Ci (0.025), as indicated in Table 4. Similar results were found in [132,133,134]. Given the substantial correlation between TKW and yield, we proceeded to investigate the traits associated with TKW (R² = 0.780, p < 0.0001). Notably, FLA emerged as a trait strongly linked to TKW, with a contribution value of 0.780. Given the pivotal role of leaf area in determining the weight of a thousand grains, we included it as a dependent attribute in order to investigate the associated traits; this is consistent with [81,84,135,136]. The SMLR analysis demonstrated that GFD and Ci were significantly correlated with FLA (R² = 0.845, p < 0.012). The contribution rates of GFD and Ci were 0.699 and 0.146, respectively. To distinguish between direct and indirect impacts, we proceeded with a PC analysis. The analysis yielded noteworthy results when considering grain yield. The TKW exhibited a direct effect of 0.244, indicating a significant correlation between TKW and its direct influence on GY. Moreover, when the dependent variable was the FLA trait, the GFD trait displayed a direct effect of 0.261. These direct effects provide valuable insights into the association between the respective traits [30,137].

Research has shown that identifying key traits influencing crop yield and selecting for those strongly correlated with grain yield can be an effective breeding strategy. However, simply estimating the association between yield and its components may not be enough to fully understand the importance of each component in determining final grain yield. More advanced statistical techniques, such as SMLR and PC analysis, can provide deeper insights into the relationships between independent variables and GY. These methods can decompose the associations into direct and indirect effects on GY [138,139]. Plant breeders have also employed path analysis to help identify the most useful traits to target as selection criteria for improving overall crop yield [140,141]. By utilizing these more sophisticated analytical approaches, breeders can gain a more nuanced understanding of the complex factors driving grain yield.

Our result, as shown in Table 4, reveals that SMLR and PC analysis identified five traits (TKW, FLA, GFD, Ci, and PH) that were positively associated with GY. These traits hold significant value, as they provide breeders with important insights into the overall productivity of a given genotype. The strong positive correlation we observed between thousand-kernel weight and yield aligns with findings from previous studies [142,143]. Multiple analyses (SMLR and PC) in our study also identified a strong positive association between FLA, TKW, and GFD, which identified the importance of FLA and its direct and indirect effect on GY (Table 4). The FL is a critically important tissue for plant growth and development. As the uppermost leaf containing high concentrations of chlorophyll, the flag leaf plays a vital role in light interception, photosynthesis, and the storage of assimilates [144,145]. These key physiological functions make the flag leaf essential for crop productivity and the accumulation of carbohydrates in the grains [146,147]. Research has shown that the flag leaf’s photosynthetic rate is positively correlated with grain yield determination in rice [148,149]. The leaf area of a plant significantly impacts its overall growth, development, productivity, and quality by influencing light capture [150,151]. In wheat, the length and width of the flag leaf have been found to be positively correlated with important yield components like the number of spikelets per plant and grain weight [152,153]. Similarly, in rice, positive correlations have been reported between leaf thickness, grain weight per inflorescence, and the number of spikelets per inflorescence [154]. Interestingly, we found that the traits identified by SMLR and PC analysis as positively associated with yield (Table 4) were the same traits for which molecular marker showed strong associations (Table 1), making them reliable indicators for selection for heat tolerance and a topic that opens up prospects for future breeding programs to tolerate different environmental conditions and improve quality. By comprehending and quantifying the contributions of these traits to grain yield, we can pave the way towards enhancing the yield potential of wheat under heat stress conditions [155,156].

To further explore the relationships among the 13 genotypes under study, we conducted a HC analysis incorporating the five traits (TKW, PH, Ci, FLA, and GFD) identified through the previous SMLR and PC analyses, in addition to GY. We used the relative rate of change for these traits, and the results showed that the genotypes were distributed into three groups based on their performance under heat stress conditions. The S category comprised seven genotypes, while the T and MT categories included four and two genotypes, respectively. It is noteworthy that the HC analysis has been frequently utilized by researchers for the purpose of ranking heat-tolerant wheat genotypes [122,128,129,157,158]. We performed LD analysis to validate the accuracy of the classes. LD analysis showed complete consistency between the preceding and subsequent classes among the 13 genotypes used (Table 5). Furthermore, cross-validation showed that the preceding and subsequent classes were identical. Thus, LD analysis represents a unique statistical tool, providing selection criteria for the accuracy and reliability of genotype resources with respect to heat tolerance [111,112,113,159,160,161,162].

5. Conclusions

This study used a multivariate analytical approach to investigate genotypic and phenotypic characteristics associated with heat tolerance in wheat genotypes. SSR markers, linked to quantitative trait loci (QTLs) in marker-assisted selection (MAS), play a crucial role in determining key agronomic physical and biochemical traits under thermal stress. The results identified six traits that exhibited a combination of high heritability (>60%) and genetic gain (>20%), suggesting significant potential for selection and improvement. The SMLR and PC analyses collectively identified six major traits (GY, TKW, FLA, GFD, PH, and Ci) that were strongly associated with molecular markers and effectively showed phenotypic differences between genotypes, which were then used as the basis for HC analysis. The HC analysis divided the 13 tested genotypes into three distinct classes, a classification fully supported by LD and SSR marker analysis. The Mantel test revealed significant associations between a various of morphological, physiological, and biochemical traits and the SSR genetic markers evaluated in this study. These results highlight the potential utility of these trait–marker associations as effective screening criteria for identifying heat-tolerant wheat genotypes. Based on the study’s findings, the researchers identify four wheat genotypes that exhibited exceptional tolerance to heat stress. These included three local genotypes (KSU105, KSU106, KSU115) and one global genotype (16HTWYT-22). Notably, these heat-tolerant genotypes displayed the highest values across all the measured traits, even under the heat stress conditions. For example, the KSU115 genotype recorded the greatest TKW of 42.20 g, the highest GY of 6.77 ton ha⁻¹ and the largest FLA of 54.9 cm². Additionally, these genotypes possessed a remarkable capacity to scavenge oxygen free radicals, attributed to their high antioxidant content, which was particularly evident under heat stress conditions. These genotypes are promising candidates for incorporated into future breeding programs aimed at further evaluation their tolerance to other abiotic stresses, such as drought and salinity, by examining relevant physiological parameters and studying the gene expression of genes associated with tolerance to environmental stresses.