Multi-locus genome-wide association studies of aluminum stress indices in shoot and root traits at seedling stage in Ethiopian durum wheat (Triticum turgidum ssp. durum) genotypes.

Multi-locus genome-wide association studies of aluminum stress indices in shoot and root traits at seedling stage in Ethiopian durum wheat (Triticum turgidum ssp. durum) genotypes - PMC Skip to main content An official website of the United States government Here's how you know Here's how you know Official websites use .gov A .gov website belongs to an official government organization in the United States. Secure .gov websites use HTTPS A lock ( Lock Locked padlock icon ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites. Search Log in Dashboard Publications Account settings Log out Search… Search NCBI Primary site navigation Search Logged in as: Dashboard Publications Account settings Log in Search PMC Full-Text Archive Search in PMC Journal List User Guide PERMALINK Copy As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health. Learn more: PMC Disclaimer | PMC Copyright Notice BMC Plant Biol . 2026 Mar 9;26:681. doi: 10.1186/s12870-026-08475-0 Search in PMC Search in PubMed View in NLM Catalog Add to search Multi-locus genome-wide association studies of aluminum stress indices in shoot and root traits at seedling stage in Ethiopian durum wheat ( Triticum turgidum ssp. durum) genotypes Tsegaye Abebe Tsegaye Abebe 1 College of Agriculture and Veterinary Medicine, Jimma University, P.O. Box: 307, Jimma, Ethiopia 2 College of Agriculture and Natural Resource, Gambella University, P.O. Box 126, Gambella, Ethiopia Find articles by Tsegaye Abebe 1, 2, ✉ , Wosene Gebreselassie Wosene Gebreselassie 1 College of Agriculture and Veterinary Medicine, Jimma University, P.O. Box: 307, Jimma, Ethiopia Find articles by Wosene Gebreselassie 1 , Temesgen Matiows Menamo Temesgen Matiows Menamo 1 College of Agriculture and Veterinary Medicine, Jimma University, P.O. Box: 307, Jimma, Ethiopia Find articles by Temesgen Matiows Menamo 1 , Tilahun Mekonnen Tilahun Mekonnen 3 Institute of Biotechnology, Addis Ababa University, P. O. Box: 1176, Addis Ababa, Ethiopia Find articles by Tilahun Mekonnen 3 , Behailu Mulugeta Behailu Mulugeta 4 Sinana Agricultural Research Center, Oromia Agricultural Research Institute, Bale-Robe, Ethiopia Find articles by Behailu Mulugeta 4 , Kassahun Tesfaye Kassahun Tesfaye 3 Institute of Biotechnology, Addis Ababa University, P. O. Box: 1176, Addis Ababa, Ethiopia 5 Bio and Emerging Technology Institute (BETin), P. O. Box: 5954, Addis Ababa, Ethiopia Find articles by Kassahun Tesfaye 3, 5 Author information Article notes Copyright and License information 1 College of Agriculture and Veterinary Medicine, Jimma University, P.O. Box: 307, Jimma, Ethiopia 2 College of Agriculture and Natural Resource, Gambella University, P.O. Box 126, Gambella, Ethiopia 3 Institute of Biotechnology, Addis Ababa University, P. O. Box: 1176, Addis Ababa, Ethiopia 4 Sinana Agricultural Research Center, Oromia Agricultural Research Institute, Bale-Robe, Ethiopia 5 Bio and Emerging Technology Institute (BETin), P. O. Box: 5954, Addis Ababa, Ethiopia ✉ Corresponding author. Received 2025 Nov 20; Accepted 2026 Feb 25; Collection date 2026. © The Author(s) 2026 Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/ . PMC Copyright notice PMCID: PMC13081369  PMID: 41796310 Abstract Background Soil acidity associated with aluminium toxicity is a major abiotic factors limiting durum wheat production in the highland areas of Ethiopia. Selection and use of Al tolerant genotypes is a sustainable strategy for boosting durum wheat production on acidic soils. Methods This study aimed to identify aluminum-tolerant durum wheat genotypes and associated genomic regions by evaluating 300 genotypes under both Al-stressed and non-stressed hydroponic conditions. Genome-wide association was estimated using 10,000 SNP markers for almunium stress indices and traits both in stress and non-stress conditions. Results The analysis of variance revealed highly significant differences among the genotypes for all studied traits. Alumimuim stress indices were estimated using nine methods in root and shoot tratis. Among eight stress indices, correlation analysis identified mean productivity, geometric mean productivity and aluminum adoption indices as effective measure for selecting aluminum tolerance genotypes. Genotypes grouped in to three major cluster based on their mean productivity, geometric mean productivity, aluminum adoption indices and traits in both almunium stress and non-stress conditions. Cluster one (20 genotypes) identified as aluminum tolerance group. Cluster two contain 30 moderatly tolerant genotypes and cluster three comprised 250 susceptible genotypes. Based on the functional annotations, 40 candidate genes associated with Al tolerance were identified using 10 QTNs detected by more than two model. Conclusions The functional characterization of these identified genes offers essential insights into the underlying genetic mechanisms that govern the response of durum wheat genotypes to aluminum stress. This knowledge offers a significant opportunity for the development of improved, aluminum-tolerant durum wheat varieties through further genetic research. Supplementary Information The online version contains supplementary material available at 10.1186/s12870-026-08475-0. Keywords: Aluminum toxicity, Durum wheat, Linkage disequilibrum, Stress indices Introduction Durum wheat (( Triticum turgidum ssp. durum (Desf.) is a crucial cereal crop and a rich source of dietary fiber, protein, vitamins, and minerals [ 1 ]. Despite its economic and nutritional importance the production and productivity of durum wheat are threatened by different biotic and abiotic stresses [ 2 ]. One of the limiting abiotic factors affecting durum wheat production in Ethiopia’s highlands is soil acidity linked to aluminum toxicity [ 3 ]. In acidic soils, aluminum (Al) is present in high concentrations [ 4 ]. Excess aluminum is harmful for durum wheat production [ 5 ]. Al toxicity has been reported 67% of the world’s acidic soils [ 6 ]. To overcome the problem of soil acidity, the Ethiopian government has embarked on a massive soil reclamation program using lime [ 7 ]. However, this method is unfeasible and uneconomical [ 8 ]. Thus, the selection and use of Al tolerant crops is a preferred method to address the problem of Al toxicity [ 9 ]. A suitable and high throughput phenotyping platform and efficient assessment methods appropriate for early-stage screening of varieties is a prerequisite to identify Al-toxicity tolerant crop varieties that will be eventually tested under field conditions [ 10 ]. Hydroponic assay is the most common method for screening plants for Al tolerance [ 11 ]. This approach simplifies control over nutrient availability, pH, and light conditions and allows easy access and non-destructive root measurements compared to phenotyping under pot and field conditions [ 12 ]. Genome-wide association studies have been recognized as a powerful approach to identifying genes that regulate complex crop traits [ 13 ]. The multi-locus (ML)-GWAS models are considered efficient and more reliable than single-locus (SL‐GWAS) models for mapping genomic regions because, in ML‐GWAS models, all‐marker effects are simultaneously estimated. Moreover, unlike SL‐GWAS models, these do not require testing of identified associations using stringent multiple testing corrections that generally result in the rejection of significant associations [ 14 ]. Various studies have been conducted on genome-wide association study of aluminum tolerance traits in bread wheat [ 15 ]. Despite several efforts made so far, in Ethiopian durum wheat genotypes, a genome-wide association study has not been conducted for the aluminum tolerance trait, and the candidate gene for the aluminum tolerance trait has not been identified. Therefore, this experiment was designed to screen a diverse set of Ethiopian durum wheat genotypes for Al 3+ tolerance using a rapid hydroponic assay and identifying genomic regions and candidate genes responsible for the aluminum tolerance related traits. Materials and methods Description of the study area The experiment was conducted at the plant tissue culture laboratory at Horticulture and Plant Science Department, College of Agriculture and Veterinary Medicine, Jimma university (JUCAVM) during 2023/2024. JUCAVM is situated in Oromia Regional State about 352 km away from Addis Ababa in the southwest part of Ethiopia with latitude and longitude of 7.667°N 36.833°E and an altitude is 1,780 m above sea level. Plant materials For this study, 300 durum wheat genotypes (283 landraces and 17 cultivars) collected from different geographical regions of Ethiopia were used (Supplementary Table 1; Fig. 1 ). The genotypes were obtained from the Ethiopian Biodiversity Institute (EBI). The landraces represent different geographical regions and from four regional states of Ethiopia (Amhara, Oromia, SNNP, and Tigray). The released varieties were obtained from the DebreZeit Agricultural Research Center (DzARC) and Sinana Agricultural Research Center (SARC). Fig. 1. Open in a new tab Map of Ethiopia illustrating the collection sites of the 300 durum wheat genotypes used in the present study Treatments and experimental design The experiment was laid out using a randomized complete design (RCD). The experiment consisted of two treatment conditions (Aluminum treated or AlSO 4 .18H 2 O and zero aluminum or control). The control experiment was conducted simultaneously with the treated experiment, except for AlSO 4 .18H 2 O. Each genotype was replicated three times under both control and aluminum stress conditions, and each replicate consisted of an independent hydroponic culture unit (biological replication). To ensure precision and minimize measurement error, each trait was measured three times per seedling (technical replication). Sterilization and germination The seeds from each durum wheat genotype were surface sterilized with 1% sodium hypochlorite (NaOCl) solution and 70% Ethanol for 5 min, then the sterilized seeds were rinsed 5–10 times using sterile distilled water. The seeds were germinated in the culture room in a Petri dish lined with moistened soft for three days. Nutrient solution preparation and growth conditions A hydroponic nutrient solution medium was prepared according to [ 16 ]. It containing 500 µM KNO 3 , 500 µM CaCl 2 , 500 µM NH 4 NO 3 , 200 µM MgSO 4 7H 2 O, 100 µM KH 2 PO 4 , 46 µM H 3 BO 3 , 20 µM Fe: EDTA, 2 µM MnCl 2 4H 2 O, 1 µM ZnSO 4 7H 2 O, 0.5 µM NaMoO 4 2H 2 O and 0.3µMCuSO 4 5H 2 O. Al was added in the form AlSO4.18H2O. According to Wayima et al. (2019), 5µM Al concentration was used for the aluminum-treated experiment. The pH of both nutrient solutions was adjusted to 4.5 using 1 M HCl and/or 1 M NaOH solutions, under which aluminum becomes solubilized into its phytotoxic ionic form (Al³⁺). The nutrient solutions pH were monitored regularly using a calibrated digital pH meter to ensure stability of the target pH and to ensure reproducibility, reliability, and accurate assessment of plant responses to aluminum stress. Any variations were promptly fixed by carefully adding either 1 M HCl or 1 M NaOH. The nutrient solutions were replaced every day to minimize pH fluctuations and ensure continuous exposure of the seedlings to Al ions. After three days of continuous growth in the Petri dish, five representative seedlings from each genotype were selected, measured for the initial root and shoot lengths to see the effect of aluminum toxicity, and transferred into plastic cups containing the nutrient to screen for aluminum toxicity tolerance. Each seedlings was placed individually on cups with small holes at the bottom to allow roots to reach the nutrient solution inside a tray. After transfer, the seedlings were allowed to grow in a growth chamber for five days with a photoperiod of 14 h of light and 10 h of darkness. Data collection Shoot and root related traits In order to record root and shoot attributes, intact seedlings were removed from the nutrient solutions after growing continuously for five days after transfer to the nutrient media. The final root and shoot length was measured. The root and shoot length measurements were achieved by gently placing the seedling along a ruler on a bench. The root dry weight (RDW) and shoot dry weight (SDW) were measured after the samples were dried in an oven at 70 °C for 72 h. The root to shoot ratio was calculated by dividing root dry weight by shoot dry weight. 1 cm from the longest seminal root was excised to count the number of root hairs per plant. It was counted under a microscope (eyepiece resolution 2048 × 1536) with a USB digital camera. Genotypic data A single spike per genotype was gathered for genotyping. To represent each genotype five healthy seeds were extracted from each spike and placed in 3-liter pots in a greenhouse at the Swedish University of Agricultural Science (SLU), Alnarp, Sweden. by technical staff as part of a collaborative study. We gratefully acknowledge the technical staff at the Swedish University of Agricultural Sciences (SLU), Alnarp, Sweden, for their assistance with seed sowing and greenhouse management. Two week old seedlings of each genotype were used, which were then placed in each well of a 96-deep well plate and freeze-dried using the CoolSafe ScanVAC Freeze Dryer in accordance with Trait Genetics’ recommendations. For DNA extraction and subsequent genotyping, the freeze-dried samples in 96-well deep well plates were shipped to TraitGenetics (GmbH, Gatersleben, Germany). Trait Genetics’ lab extracted DNA from the leaf samples using a standared cetyltrimethylammonium bromide (CTAB) approach. Following the manufacturer’s instructions, the 300 genotypes were genotyped using an Illumina Infinium 25k wheat single nucleotide polymorphism (SNP) array. The initial pipeline used strict quality control procedures, removing low quality sequences according to the properties of the barcode region. The SNP markers were then found using the genome of Triticum turgidum, L. var. durum Desf. Statistical analysis The net average root and shoot length was calculated by subtracting the average final root and shoot length from the initial average root and shoot length. Before further analysis, data were evaluated using the Shapiro–Wilk test to assess if they fit into the normal distribution. The best linear unbiased prediction (BLUP) value was estimated using META-R package [ 17 ]. The estimated value of all traits was used for analysis. Analysis of variance (ANOVA) was performed using the generalized linear model (GLM) implemented in R with ‘aov’function. Pearson’s correlation analysis between the measured traits and alumnum tolerance indices was computed using the R package 4.1.1‘corrplot’ [ 18 ]. Heritability was estimated using the varability R package [ 19 ]. Numerous genetic parameters were estimated according to the formulas adopted from Johnson, Robinson [ 20 ]: Phenotypic variance was estimated using ( σ 2 p ) = σ 2 g + σ 2 e . Whereas, σ 2 p =phenotypic variance, σ 2 g =genotypic standard deviation, σ 2 e . = environmental variance. . Where, PCV = phenotypic coefficients of variation, and = grand mean for the traits. ; Where, GCV = genotypic coefficients of variation. Genetic advance (GA). Where σ Ph = phenotypic standard deviation, h 2 = broad sense heritability, K= the standardized selection differential at 5% selection intensity (k = 2.063) percent of the mean (GAM), The clustering was also estimated using cluster v2.1.6 and factoextra v1.0.7 R packages default algorism such as distance calculated using Euclidean data scaled, the dendrogram graph constructed using ward D 2 method and number of bootstrap was 1000 [ 21 ]. The number of cluster was estimated using silhouette method in factoextra R packages. Acidity indices were calculated using the following formula (Table 1 ). Table 1. Acid soil tolerance index Index Formula Outcome References Stress Susceptible Index (SSI) The accessions with SSI < 1 are more resistant to stress conditions Fischer and Maurer [ 22 ] Aluminum Tolerance Index (YSI) Genotype with high values of ATI specify stable under stress and non stress [ 23 ] Mean Productivity (MP) The accessions with high value MP are more desirable Rosielle and Hamblin [ 24 ] Geometric Mean Productivity (GMP) The accessions with high value of GMP index are more desirable Schneider, Rosales-Serna [ 25 ] Harmonic Mean Productivity (HMP) The accessions with high value of HMR index are more desirable Jafari, Paknejad [ 26 ] Aluminum Adaption Index (AAI) AAI=(Yns)(Yst)/ (µYns)(µYst) Highest values of AAI designate stress tolerant genotype [ 23 ] % Reduction (PCRD) PCRD = (Yns-YSt/ Yns)x 100 The accessions with low percent reduction are tolerant Open in a new tab Where; ns and st are a given genotype under non-stress and under stress, µst is the mean of all test genotypes under stress conditions and µns is the mean of all genotypes under non-stress soil conditions SSI Stress susceptibility index, ATI aluminum tolerance index, MP mean productivity, GMP geometric mean productivity, HMP harmonic mean productivity, AAI alumnium adoption index, PR percent reduction Genome-wide association analysis GWAS was conducted using the best linear unbiased estimate value of six traits and 10,000 SNP markers. Association analyses were performed using multi-locus GWAS methods, it includes multi-locus random-SNP-effect MLM (mrMLM) [ 27 ], fast multi-locus random-SNP-effect EMMA (FASTmrEMMA) [ 28 ], Iterative Sure Independence Screening EM-Bayesian LASSO (ISIS EM-BLASSO) [ 29 ], polygenic-background-control-based Kruskal–Wallis test plus empirical Bayes (pKWmEB) [ 30 ], fast mrMLM (FASTmrMLM) [ 29 ], and polygenic background-control-based least angle regression plus empirical Bayes (pLARmEB) [ 31 ]. A LOD score of ≥ 3.00 was used to declare the genomic regions strongly related to QTNs. Furthermore, SNP markers that were repeatedly detected in at least two models were designated as reliable [ 32 ]. The resulting -log10 (P) values obtained from the ML-GWAS approaches were utilized to create Manhattan and Q–Q plots using the mrMLM.GUI V4.0.12 package [ 28 ]. Linkage Disequilibrium (LD) and linkage disequilibrium decay analysis The correlation coefficient (r 2 ) between a pair of SNPs was computed using the TASSEL software (version 5.0). The fitted curve to the r² values and the average r² for each chromosome were determined by finding the intersection of the fitted curve with a specific threshold. Along each chromosome the physical distance was plotted against the average r 2 for the entire genome and each chromosome. The LD decay rate was determined as the rate at which the r 2 value dropped to half its greatest value to determine the point at which the markers are weakly linked, we utilized a cut-off value of r 2 = 0.2; this cut-off indicates a limit on the size of the QTL for pairs of markers. Identification of candidate genes The SNP markers significantly associated with the traits of interest were searched in the Phytzome database using durum wheat (Triticum turgidum(Svevo.v1) reference genome. To identify genes related to significant QTNs, genome-wide LD decay of 4.032 Mb up and downstream from the SNP position was used to search the candidate genes for each trait. Results Variations in shoot and root traits The analysis of variance revealed the presence of significant variations ( p <0.001) among the genotypes for all measured traits (Table 2). The shoot length ranged from 2.15 to 8.93 cm, with a mean of 5.29 cm, root length ranged from 1.02 to 6.47 cm, with a mean of 2.7 cm, shoot dry weight varied from 89 to 353.33 mg, with a mean of 178.50 mg, root dry weight range from 25.67 to 78.67 mg, with a mean of 65.68 mg, number of root hairs in one centimeter ranged from 20 − 79, with a mean of 45.8, and root to shoot ratio varied from 0.24 to 0.67, with a mean 0.37. The phenotypic coefficient of variation (PCV) ranged from 9.05% for shoot dry weight to 41.66% for root dry weight. The genotypic coefficient of variation (GCV) varied from (8.09%) for shoot dry weight to 41.66% for root dry weight. Broad sense heritability ranged from 80% for shoot dry weight to 98% for shoot length. Genetic advance (GA) ranged from (0.04%) for shoot dry weight to (11.41%) for root to shoot ratio. The genetic advances as a percent of the mean (GAM) ranged from (14.93%) for shoot dry weight to (85.86%) for root dry weight (Table 2 ). Table 2. Estimates of range, mean, genetic parameters of variance, and heritability of durum wheat genotypes grown under hydroponic conditions Traits Unit Mean Max Min PCV GCV ECV H 2 (%) GA GAM (%) SL Cm 5.29*** 8.93 2.15 13.27 13.17 1.61 0.98 1.35 39.70 RL Cm 2.70*** 6.47 1.02 39.97 39.93 1.70 0.99 1.40 82.19 SDW Mg 178.50*** 353.33 89.00 9.05 8.09 2.87 0.80 0.04 14.93 RDW Mg 65.68*** 178.67 25.67 41.66 41.66 3.50 0.89 0.036 85.86 RS No 0.37*** 0.67 0.24 20.43 19.83 4.92 0.94 11.41 39.65 RH No 45.80*** 79.00 20.00 21.51 20.82 5.12 0.93 0.15 41.53 Open in a new tab SL Shoot length, RL Root length, SDW Shoot dry weight, RDW Root dry weight, RH Root hair number, Min minimum, Max maximum, GCV genotypic coefficient of variation, PCV phenotypic coefficient of variation, ECV environmental coefficient of variation, H2% broad-sense heritability, GA genetic advance, GAM genetic advance percentage of mean, cm centimeter, mg millgram ***Means a significance leve0l at p < 0.001 Correlation of measured traits Pearson’s correlation analysis showed a significant strong, moderate, and weakly positive correlation among all measured traits. The correlation coefficients between the six traits ranged from 0.46 to 0.92. The highest correlation (0.94) was observed between RL and SDW, followed by RDW and SDW (0.91). The lowest correlation was observed between RS and SL (0.46). The second lowest correlation was recorded between RS and SDW (0.51) (Fig. 2 ). Fig. 2. Open in a new tab Heatmap of Pearson correlation coefficient matrix between the root and shoot traits of durum wheat genotypes, RL, root length; SL, shoot length; RDW, root dry weight; SDW, shoot dry weight, RH, root hairs; RS, root to shoot ratio Correlation between traits and aluminum indices To identify the most informative aluminum‑tolerance index for further analyses, we computed Pearson correlation coefficients between the measured root and shoot traits and eight commonly used tolerance indices. These indices included Mean Productivity (MP), Geometric Mean Productivity (GMP), Stress Susceptibility Index (SSI), Tolerance (TOL), Aluminum Tolerance Index (ATI), Harmonic Mean Productivity (HMP), Percent reduction (PR), and Aluminum Addoption indices (AAI) (Fig. 3 ). Correlations were evaluated separately for control and aluminum‑stress conditions.The results showed that MP, GMP and AAI exhibited consistently strong, positive, and statistically significant correlations with both root and shoot traits in the control and stress condition. Therfore, those indices were retained for further analyses, including the cluster analysis. In contrast, the remaining five indices (SSI, TOL, ATI, HMP, and PR) displayed weak correlations with the measured traits and these correlations were not statistically significant ( p > 0.05), these indices were excluded from subsequent analyses. Fig. 3. Open in a new tab Pearson correlation coffeicient b/n traits and aluminum tolerance indices, soot length stress (SLS), shoot length non stress (SNLS), root length stress (RLS), root length non stress (RLNS), shoot dry weight stress (SDWS), shoot dry weight non stress (SDWNS), root dry weight stress (RDWS), root dry weight non stress (RDWNS), root hairs stress (RHS), root hairs non stress (RHNS), root to shoot ratio stress (RSS), root to shoot ratio non stress (RSNS), stress susceptibility index (SSI), aluminum tolerance index (ATI), mean productivity (MP), geometric mean productivity (GMP), harmonic mean productivity (HMP), tolerance index (TOL), alumnium adoption index (AAI), percent reduction (PR) Clustering based on measured traits and aluminum tolerance indices Cluster analysis was carried out using measured traits and three selected indices. The genotypes clustered into three groups (Fig. 4 & Supplementary Table 2). Cluster I comprised 18 accessions and 2 cultivars with high values for the measured tratis and stress indices (MP, GMP, and AAI). Especially, accessions #3540, 5563 and 5249 in clusters C-I were identified as the most desirable due to their high tolerance levels. (Fig. 5 , Table 3 & Supplementary Table 2). Fig. 4. Open in a new tab Clutering analysis of 300 durum wheat genotypes using Hierarchical clustering method Cluster II consisted of 30 genotypes ( 27 accesstions and 3 modern cultivars) characterized by medium values of measured traits and medium aluminum tolerance indices value (Supplementary Table 2). This cluster was categorized as moderately tolerance to aluminum toxicity. On the other hand, cluster III consisted 238 accessions and 12 modern cultivars, this cluster characterized by low value of the measured traits and low aluminum tolerance indices value, the accessions and cultivars in this cluster was classified as susceptible to aluminum toxicity (Table 3 ). Table 3. Cluster summary based on stress indices and shoot and root traits grown in aluminum stress and control or without aluminum conditions Traits/Indices C-I C-II C-III Max Min Mean Max Min Mean Max Min Mean RLS 4.48 3.12 3.74 2.86 2.12 2.4 1.98 1.02 1.45 RLNS 6.47 4.85 5.38 5.04 3.97 4.46 5.32 2.34 3.46 MP 5.29 3.99 4.56 3.93 3.05 3.43 3.49 1.71 2.46 GMP 5.16 3.89 4.48 3.77 2.9 3.27 2.97 1.58 2.24 AAI 4.22 2.4 3.19 2.26 1.34 1.7 1.4 0.4 0.82 SLS 5.77 4.6 5.05 4.56 3.94 4.17 3.94 2.15 3.17 SL NS 8.93 8.55 8.74 8.55 8.28 8.43 8.27 4.74 6.93 MP 6.78 6.3 6.5 6.29 6.05 6.15 6.04 3.54 5.1 GMP 6.6 5.98 6.2 5.98 5.63 5.79 5.58 3.33 4.72 AAI 1.78 1.46 1.58 1.46 1.3 1.37 1.28 0.45 0.92 SDWS 0.23 0.14 0.18 0.14 0.11 0.12 0.1 0.09 0.08 SDWNS 0.35 0.27 0.3 0.27 0.26 0.26 0.26 0.2 0.24 MP 0.41 0.28 0.32 0.27 0.24 0.25 0.24 0.2 0.22 GMP 0.29 0.19 0.23 0.19 0.17 0.18 0.17 0.14 0.16 AAI 3.04 1.4 1.91 1.39 1.07 1.2 1.07 0.7 0.92 RDWS 0.14 0.07 0.1 0.07 0.05 0.05 0.05 0.03 0.04 RDWNS 0.18 0.12 0.15 0.12 0.1 0.11 0.1 0.06 0.08 MP 0.23 0.14 0.17 0.12 0.1 0.11 0.1 0.06 0.08 GMP 0.16 0.1 0.12 0.09 0.07 0.08 0.07 0.04 0.05 AAI 6.57 2.62 3.86 2.03 1.31 1.57 1.3 0.43 0.8 RSS 0.61 0.53 0.55 0.52 0.43 0.46 0.41 0.25 0.35 RSNS 0.55 0.48 0.52 0.48 0.41 0.44 0.4 0.24 0.34 MP 0.31 0.24 0.28 0.24 0.18 0.2 0.16 0.07 0.12 GMP 0.56 0.49 0.53 0.49 0.42 0.44 0.4 0.26 0.34 AAI 2.33 1.81 2.09 1.77 1.34 1.46 1.32 0.5 0.89 RHS 46 37.67 41.78 37.67 34.67 35.59 34.33 20 26.91 RHNS 79 74.67 76.13 74.67 70.33 72.14 70 46.67 60.97 MP 60.67 54 55.32 53.83 51.83 52.71 51.83 33.83 44.37 GMP 58.78 50.91 52.43 50.88 48.61 49.75 48.55 30.88 40.8 AAI 1.9 1.43 1.52 1.43 1.3 1.36 1.3 0.53 0.93 No of geno 20 30 250 Decision Tolerante Intermidate Suseptible Open in a new tab Rls Root length under stress, Rlns root length under non stress, Sls shoot length under stress, Slns shoot length under non stress, Sdws shoot dry weight under stress, shoot dry weight under non stress (Sdwns), root dry weight under stress (Rdws), root dry weight under non stress (Rdwns), root to shoot ratio under stress (Rss), root to shoot ratio under non stress (Rsns), number of root hairs under stress (Rhs), number of root hairs under non stress (Rhns), mean productivity (MP), geometric mean productivity (GMP), aluminum adoption index (AAI) Fig. 5. Open in a new tab Examples of seedlings grown for 8 days under 5 μM Al3+-containing nutrient solution. Accession 3540 (Fig. A ) was scored as tolerant, and accession 5576( Fig. B ) was scored as Susceptible Genome wide association analysis Prior to further analysis, the genotype data were filterd using TASSEL 5.2.80 software [ 33 ].The initial dataset for the Genome-Wide Association Study (GWAS) comprised 25,000 Single Nucleotide Polymorphism (SNP) markers genotyped across a panel of 300 durum wheat genotypes. To ensure the reliability and informativeness of the markers used for association mapping, the dataset underwent a stringent, multi-stage quality control filtering process. First, SNPs with a low minor allele frequency (MAF) are generally considered as rare alleles with less power in detecting marker-trait associations (MTAs) and are prone to genotyping error. Thus, SNP loci with missing data above 5% or with minor allele frequency (MAF) below 5% were removed from further analysis. Further filtering of the remaining SNP loci was conducted based on the level of observed heterozygosity (Ho), and loci with Ho greater than 0.01 were excluded. This step was crucial to eliminate potential genotyping errors, residual heterozygosity. Following this comprehensive quality control procedure, a final, high-quality dataset consisting of 10,000 SNPs was retained. This refined marker set was subsequently utilized for the GWAS analysis. GWAS analysis was performed in R Statistical to map associations using MP, GMP and AAI indices, shoot and root traits. The SNPs were distributed across the A and B genomes of durum wheat. The B genome exhibited a greater quantity of SNPs (Fig. 6 ). Generally, in the current study, 52.01% of the SNPs were located on the B genome and 47.99% on the A genome. Earlier studies also found more SNPs on the B genome than A genome of bread wheat and durum wheat [ 34 ]. Fig. 6. Open in a new tab Distribution of 10,000 filtered single nucleotide polymorphisms (SNPs) across the durum wheat genome Genome wide associations analysis identified multiple genomic regions associated with shoot and root traits (SL, RL, RDW, RH, and RS) and three stress indices (MP, GMP and AAI). A total of 55 QTNs were detected using six mr-MLM model (Table 4 ; Figs. 7 and 8 ). Out of the total QTNs, 30 were detected only in Al stress condition, 10 QTNs were detected only in non- stress condition, 14 QTNs were identified by stress tolerance index of their corresponding traits (Table 4 ; Figs. 7 and 8 ). From the total QTNs, 4 QTNs were detected by 2 models, two QTNs was detected by 3 models, one QTN was detected by 4 models, one QTN was detected by 5 models and 1 QTN was detected by all six models, the remaining 45 QTNs were detected by single model (Figs. 7 and 8 ). Chromosome 5B exhibit the highest number of identified QTNs. The LOD value ranged from 3.01 to 5.59 and the proportion of phenotypic variance explained (r 2 ) by each QTN ranged from 0.000000283 to 12.83%. The result focused on QTNS detected by more than two models, especially on 10 QTNs. Table 4. List of significant SNP markers identified for root and shoot traits detected by more than two model evaluated for durum wheat genotype showing physical position, association level ( P - value), phenotypic variation (R2) explained by the locus, additives SNP effect and minor allele frequencies (MAF) SNP(MTAs) CH POS(bp) QTN effect LOD score P -value R 2 (%) MAF MD SNPs Traits AX-158,592,916 7B 695,259,812 -0.629 3.17 0.00013 12.81 0.14 FASTmrMLM RLS AX-158,592,916 7B 695,259,812 -0.0002 3.33 0.00009 8.73 0.14 FASTmrEMMA AX-158,592,916 7B 695,259,812 -0.2167 3.33 0.00009 2.77 0.14 pLARmEB AX-158,592,916 7B 695,259,812 -0.631 3.17 0.00013 12.83 0.14 ISIS EM-BLASSO BS00077791_51 7 A 612,331,060 -7.36E-05 3.29 0.00001 8.06 0.20 pLARmEB RLNS BS00077791_51 7 A 612,331,060 -0.5162 3.20 0.00012 6.04 0.20 pKWmEB BS00022805_51 2B 731,961,061 -0.4548 3.35 8.63E-05 4.75 0.45 mrMLM BS00022805_51 2B 731,961,061 -0.4933 5.45 5.44E-07 5.56 0.45 ISIS EM-BLASSO AX-94,463,626 5B 577,635,486 -0.2347 3.07 0.00017 2.58 0.47 FASTmrEMMA RLGMP AX-94,463,626 5B 577,635,486 -0.1177 3.07 0.00017 4.91 0.47 pKWmEB AX-94,463,626 5B 577,635,486 -0.1196 3.07 0.00017 2.68 0.47 ISIS EM-BLASSO BS00064085_51 5B 654,276,618 -0.1794 3.54 4.4554 5.91E-06 6.59 mrMLM SLS BS00064085_51 5B 654,276,618 -0.1497 3.14 4.4554 5.91E-06 4.59 FASTmrMLM AX-111,562,104 4B 538,585,114 -0.1031 3.72 3.53E-05 3.98 0.36 pKWmEB SLNS AX-111,562,104 4B 538,585,114 -0.1283 3.44 6.85E-05 1.72 0.37 ISIS EM-BLASSO BS00065128_51 5B 579,360,050 -0.1364 3.12 0.00015 6.42 0.47 pKWmEB SLMP BS00065128_51 5B 579,360,050 -0.1273 3.17 0.00013 3.46 0.47 ISIS EM-BLASSO BS00065128_51 5B 579,360,050 -0.1038 3.47 6.36E-05 2.34 0.47 pLARmEB SLGMP BS00065128_51 5B 579,360,050 -3.60E-05 3.57 4.99E-05 2.83E-07 0.47 ISIS EM-BLASSO BS00065128_51 5B 579,360,050 -0.1276 3.57 4.99E-05 3.54 0.47 FASTmrMLM BS00064085_51 5B 654,276,618 -0.1862 3.01 0.00011 6.73 0.34 mrMLM BS00064085_51 5B 654,276,618 -0.3111 3.01 0.00011 4.70 0.33 FASTmrEMMA AX-158,524,110 3 A 23,919,840 -4.5252 3.19 0.00013 4.84 0.45 mrMLM RDWNS AX-158,524,110 3 A 23,919,840 -3.4131 3.19 0.00013 2.75 0.45 FASTmrMLM AX-158,524,110 3 A 23,919,840 -6.6908 3.19 0.00013 2.60 0.45 FASTmrEMMA AX-158,524,110 3 A 23,919,840 -3.4129 3.35 8.57E-05 2.75 0.45 pLARmEB AX-158,524,110 3 A 23,919,840 -3.4131 3.19 0.00013 4.84 0.45 pKWmEB AX-158,524,110 3 A 23,919,840 -3.4131 3.19 0.00013 2.75 0.45 ISIS EM-BLASSO AX-158,524,110 3 A 23,919,840 -0.0143 3.65 4.12E-05 4.85 0.45 mrMLM RSNS AX-158,524,110 3 A 23,919,840 -0.0107 3.65 4.12E-05 2.75 0.45 FASTmrMLM AX-158,524,110 3 A 23,919,840 -0.0209 3.20 0.00012 2.55 0.45 FASTmrEMMA AX-158,524,110 3 A 23,919,840 -0.0143 3.65 4.12E-05 4.85 0.45 pLARmEB AX-158,524,110 3 A 23,919,840 -0.011 3.85 2.52E-05 4.85 0.45 pKWmEB AX-158,524,110 3 A 23,919,840 -0.0107 3.46 6.50E-05 2.74 0.45 ISIS EM-BLASSO AX-158,540,954 2 A 745,265,681 1.7386 5.59 3.86E-07 8.51 0.36 mrMLM RHMP AX-158,540,954 2 A 745,265,681 3.0351 3.92 2.13E-05 6.25 0.36 FASTmrEMMA AX-158,540,954 2 A 745,265,681 1.1663 3.22 0.00012 3.84 0.36 ISIS EM-BLASSO AX-110,508,154 5B 592,460,023 1.2932 4.42 6.40E-06 4.60 0.39 pLARmEB RHGMP AX-110,508,154 5B 592,460,023 5.86E-05 3.86 2.48E-05 6.57 0.39 pKWmEB AX-158,540,954 2 A 745,265,681 0.0782 4.92 1.94E-06 7.52 0.356 mrMLM RHAAI AX-158,540,954 2 A 745,265,681 0.133 3.33 9.10E-05 5.25 0.355 FASTmrEMMA Open in a new tab SNP, Single nucleotide polymorphism, MTAs, Marker traits associations, Chr Chromosome, POS Physical position of SNP, bp Base pair, MAF Minor allele frequency, R2 Phenotypic variance explained, SL Shoot length stress condition, SLNS Shoot length non stress condition, RL Root length stress condition, RLNS Root length non stress condition, RDWNS Root dry weight non stress condition, RSNS Root to shoot ratio non stress condition, MP Mean productivity, GMP Geometric mean productivity, AAI Aluminum adoption index Fig. 7. Open in a new tab Significant QTNs co-detected simultaneously by using two or more multi-locus GWAS methods using shoot and root traits in durum wheat genptypes. RLS, Root length in stress condtion; RLNS, Root length in non stress condition; SLS, Shoot length in stress condition; SLNS, Shoot length in non stress condition; RDWNS, Root dry weight in non stress condition; RSRNS, Root to shoot ratio in non stress condition. The horizontal line indicates the significant threshold level Fig. 8. Open in a new tab Significant QTNs co-detected simultaneously by using two or more multi-locus GWAS methods using shoot and root traits and Stress indices in durum wheat genptypes. RLAAI, Root length traits of aluminum adoption indices; SLMP, Shoot length traits of mean productivity; SLGMP, Shoot length trait of geometric mean productivity; RHMP, Number of root hairs trait of mean productivity; RHGMP, Number of root hairs trait of geometric mean productivity; RHAAI, Number of root hairs trait of aluminum adoption indice. The horizontal line indicates the significant threshold level Genome-wide association analysis identified four significant marker–trait associations (MTAs) related to shoot length (detected under stress and/or non‑stress conditions) and two MTAs associated with the aluminum‑tolerance indices (MP and GMP). All detected loci map to chromosomes 4B and 5B (Table 4 ). The phenotypic variance explained by the associated SNPs ranged from 0.000000283 to 6.73. From the identified MTAs, QTN ( BS00064085_51) was identified in shoot length of stress condition and one indices (GMP) on chromosome 5B. It explain 0.00000591–6.73% of the total phenotypic variance. Another siginificant QTN (AX-111562104) was identified in shoot length traits of non stress condition in chromosome 4B. The phenotypic variance explained by associated SNP ranged from 1.72 to 3.98%. Additionaly, one QTN ( AX-94463626) was identied in two stress indices (MP and GMP) on chromosome 5B (Table 4 ; Figs. 7 and 8 ). The GWAS scan detected four significant MTAs for root length under both stress and non-stress conditions, as well as one stress indices (AAI) on chromosome 2B, 5B, 7A and 7B. It accounting for 2.58–12.83 of the total phenotypic variance for these trait. Specifically, QTN (AX-158592916) was identified in root length traits of stress condition in chromosome 7B. QTN (BS00077791_51 and BS00022805_51) were detected in root length traits of non stress condition in chromosome 2B and 7B. QTN ( AX-94463626) was identified in root length traits of aluminum adoption indice in chromosome 5B (Table 4 ; Figs. 7 and 8 ). The GWAS revealed one QTN (AX-158524110) for root dry weight of non stress condition and root to shoot ratio of non stress condition in chromosome 3A (Table 4 ; Figs. 7 and 8 ). It explain 2.55–4.85% of the total phenotypic variance. This QTN was simultaneously detected in two traits. one significant QTN (AX-158540954) was detected in number of root hairs trait of mean productivity and aluminum adoption indice in chromosome 2A. Lastely, the Genome wide association analysis idietified one QTN (AX-110508154) for number of root hairs trait of geometric mean prodctivity indice in chromosome 5B (Table 4 ; Figs. 7 and 8 ). Linkage disequilibrium decay LD was estimated by calculating the squared allele frequency correlation (r 2 ) among all possible pairs of markers for each of the 14 chromosomes. Among all possible pairs of SNPs on each chromosome 499,815 pairs were found in LD (Supplementary Table 3). Correlation coefficient ( r 2 ) values were then plotted against genetic distance (Mbp) across the whole genome, LD decay across the whole genome level was estimated using background LD (r 2 = 0.2) as a threshold. The LD decay of the whole genome was 4.032Mbp (Fig. 9 ). Fig. 9. Open in a new tab Scatter plot showing linkage disequilibrium (LD) decay in the whole genome by plot ting (r2) against genetic distance (bp) in 300 durum wheat accessions Candidate genes and annotation A total of 40 candidate genes were found upstream and downstream of the LD decay regions of these nine QTNs (Table 5 ). Among these genes ten genes were linked to AX-158,592,916, six to BS00077791_51, Five to AX-94,463,626, four to BS00064085_51, four to AX-111,562,104, three to 7B_45377044, three to BS00065128_51, three to AX-158,524,110, one to AX-158,540,954, and one to AX-110,508,154.These genes were annotated with various protein functions. For instance, The TRITD_3Av1G012300 and TRITD_2Bv1G246550 genes were annotated as Wall-associated receptor kinase1. TRITD_5Bv1G232300 and TRITD_5Bv1G207280 genes were annotated as Citrate synthase. TRITD_5BvG231180, TRITD_2BvG233580, TRITD_3Av1G010820 genes were annotated as heavy metal detoxification protein (Table 5 ). Table 5. List of 40 candidate genes for 10 important QTNs associated with Al toxicity tolerance No Genes mRNA 1 TRITD_5Bv1G231430 Zinc finger protein 2 TRITD_5Av1G021790 3 TRITD_1Bv1G177000 4 TRITD_1Bv1G176000 5 TRITD_3Av1G010550 6 TRITD_2Bv1G240790 7 TRITD_2Bv1G232170 8 TRITD_2Bv1G231800 Cytochrome P450 9 TRITD_5Av1G025370 10 TRITD_5Bv1G206390 11 TRITD_2Bv1G242520 ABC transporter protein 12 TRITD_2Bv1G243610 MADS-box Transcription factor 13 TRITD_7Bv1G015610 14 TRITD_7Bv1G015570 15 TRITD_5Bv1G206870 NAC transcription factors 16 TRITD_5Bv1G206810 17 TRITD_5Av1G021830 18 TRITD_5Av1G022710 F-box protein 19 TRITD_3Av1G011460 20 TRITD_2Bv1G241620 21 TRITD_7Bv1G015170 F-box protein 22 TRITD_5Bv1G201710 WRKY transcription factor 23 TRITD_5Av1G238820 24 TRITD_2Bv1G233400 25 TRITD_1Bv1G003380 Glutathione S-transferase 26 TRITD_2Bv1G233050 27 TRITD_5Bv1G20602 28 TRITD_5Av1G022810 29 TRITD_5Bv1G231180 Heavy metal detoxification protein 30 TRITD_2Bv1G233580 31 TRITD_3Av1G010820 32 TRITD_3Av1G012300 Wall-associated receptor kinase 1 33 TRITD_2Bv1G246550 34 TRITD_5Bv1G232300 Citrate synthase 35 TRITD_5Bv1G207280 36 TRITD_3Av1G010630 Receptor protein kinase 37 TRITD_1Bv1G002880, 38 TRITD_1Bv1G003280 Leucine Rich protein 39 TRITD_5Bv1G233900 Transmembrane protein 40 TRITD_1Bv1G176940 Open in a new tab TRITD_5Bv1G231430,TRITD_5Av1G021790,TRITD_1Bv1G177000,TRITD_1Bv1G176000,TRITD_3Av1G2010550,TRITD_2Bv1G240790 and TRITD_2Bv1G232170 genes were annotated as zinc finger protein. TRITD_2Bv1G242520 gene was annotated as ABC transporter protein. The TRITD_2Bv1G243610, TRITD_7Bv1G015610 and TRITD_7Bv1G015570 genes were annotated as MAD-box transcription factor (Table 5 ). The TRITD_5Bv1G206870 and TRITD_5Bv1G206810 genes were annotated as NAC transcription factor.TRITD_5Av1G022710 TRITD_3Av1G011460, TRITD_2Bv1G241620 and TRITD_7Bv1G015170 genes were annotated as F-box protien. The TRITD_5Bv1G201710, TRITD_5Av1G238820 and TRITD_2Bv1G233400 genes were annotated as WRKY transcription factor. The TRITD_5Bv1G206870 and TRITD_5Bv1G206810 genes were annotated as NAC transcription factor. TRITD_5Av1G022710 TRITD_3Av1G011460, TRITD_2Bv1G241620 and TRITD_7Bv1G015170 genes were annotated as F-box protein (Table 5 ). The TRITD_5Bv1G201710, TRITD_5Av1G238820 and TRITD_2Bv1G233400 genes were annotated as WRKY transcription factor. The TRITD_1Bv1G003380, TRITD_2Bv1G233050 and TRITD_5Bv1G20602 genes were annotated as Glutathione S- transferase. TRITD_3Av1G010630 and TRITD_1Bv1G003280 genes were annotated as receptor protein kinase. The TRITD_5Bv1G233900 and TRITD_1Bv1G0176940 genes were annotated as transmembrane protein (Table 5 ). Discussion The production and productivity of durum wheat is threatened by different biotic and abiotic factors [ 2 ]. Among the factors, soil acidity associated with aluminum toxicity is one of the limiting abiotic factors that affect durum wheat production in the highland areas of Ethiopia [ 3 ]. In this study, a hydroponic system was used to identify Al tolerance durum wheat genotypes and cultivars under 0µM AlSO 4 .18H 2 O and 5µM AlSO 4 .18H 2 O. Hydroponic systems are suitable for early growth and seedling screening under submerged conditions. Various findings have confirmed that hydroponic conditions are suitable for screening of Al tolerance genotypes because there are no soil-related challenges such as disease and salinity [ 35 ]. The result revealed significant variation in Root and shoot traits under Al stress and with out Al condition The present study evaluated 300 Ethiopian durum wheat genotypes under hydroponic conditions using an optimal concentration of 5 µM AlSO 4 .18H 2 O and 0 µM AlSO 4 .18H 2 O. The 5 µM Al concentration was selected based on a previous Al tolerance screening study conducted on durum wheat [ 3 , 36 ]. All studied phenotypic traits revealed significant ( p < 0.001) variations among the genotypes. This shows that the Ethiopian durum wheat gene pool contains a significant amount of genetic varations, indicating the possibility of increasing durum wheat production under Al stress condition. The existence of phenotypic variability in Al tolerance has also been reported in Ethiopian durum wheat genotypes [ 3 ]. Estimation of genetic parameter GCV assesses genetic diversity among various phenotypes, whereas PCV measures the total relative variance. The values of PCV and GCV can be categorized as low (0–10%), moderate (10–20%), or high (> 20%). The majority of the shoot and root traits revealed high GCV and PCV except shoot length and shoot dry weight. This high GCV value suggested better improvement in the selection of traits. Closer values of the PCV to the GCV estimates for all traits, thereby showing little environmental effect on the expression of the traits. According to [ 37 ], Heritability estimates can be classified as low (less than 40%), medium (40–59%), moderately high (60–79%), and extremely high (more than 80%). Based on this, all traits showed high heritability; this indicates that the variabilities is in a high proportion from genetic background. This result suggests that selection could be easy. Heritability knowledge alone is not sufficient for an efficient selection process. Thus, estimations of heritability combined with genetic advancement are thought to be more beneficial. According to [ 38 ], genetic advancement was characterised as low (< 10%), moderate (10%–20%), and high (> 20%) as a percentage of the mean. The result demonstrated substantial genetic progress and heritability were identified for all traits. It shows less influence of environment on the manifestation of certain features and the prevalence of additive gene action in their inheritance. Correlation analysis revealed strong positive association between shoot and root traits Perarson correlation analysis showed that all the traits were positive srongly and weakly correlated with each other. The RL and SDW were strongly correlated. The strong correlation between RL & SDW might be due to the role of roots in providing water and nutrients to the shoots [ 39 ]. The lowest (0.46) correlation was observed between SL and RSR. When exposed to aluminum toxicity, the plant’s immediate adaptive response is to spend energy in the root system to detoxify aluminum by organic acid exudation or to explore deeper. This greater investment in root maintenance and growth occurs even when stress inhibits shoot elongation, resulting in a much higher R/S ratio. This adaptive shift results in a modest correlation between shoot length and root-to-shoot ratio. Generally, in the current study, majority of the traits showed positive strong correlation, It suggests that enhancing one trait can frequently lead to improvements in others. It makes breeding programs more easier and cost-effective. Correlation analysis revealed strong positive and weak association between shoot and root traits and stress indices The findings indicated that there were both weak negative and strong positive associations between the stress indices and root and shoot traits. Strong positive relationships were found between the three stress indices (Aluminum Adoption Index (AAI), Geometric Mean Productivity (GMP) and Mean Productivity (MP)) and shoot and root traits. These strong positive correlations observed in the three indices (MP, GMP, and AAI) accurately reflect plant performance in both aluminum-stressed and non-stressed conditions. Genotypes that score high for these indices also have high root and shoot growth under both stress and non-stress. Thus, the indices work well for selecting plants that are tolerant and thrive in stressed condition. In contrast, SSI, TOL, ATI, HMP, and PR showed weak associations with shoot and root traits, implying that these indices were less reliable in evaluating plant performance under aluminum stress and non-stress conditions. Cluster analysis identified tolerant genotypes In the present study genotypes were clustered into three groups. Cluster I comprised 18 accessions and 2 cultivars with high values for MP, GMP, and AAI, indicating a strong tolerance to aluminum toxicity. Remarkably, accessions 3540, 5563, and 5249 from clusters C-I were considered to be the most desirable due to their high tolerance levels. Even when exposed to aluminum stress, these genotypes maintain vigorous root and shoot growth. Cluster II comprised 27 accessions and 3 cultivars. In this cluster the genotypes exhibited medium for the measured traits and medium aluminum tolerance indices value.The genotypes found in this cluster were chatagorized as intermediate tolerance of aluminum toxicity. Conversely, cluster III consisted 12 cultivars and 238 accesstions, this cluster was characterized by low value of the measured traits and low aluminum tolerance indices value, the cultivars and accessions in this cluster were categorized as susceptible. As a result, the accesstions and cultvar categorized in this cluster are not used in acidic soil prone area. Multi-locus genome-wide association study Finding genomic regions linked to complex traits has become possible with the use of multi-locus genome-wide association studies (ML-GWAS) [ 40 ]. ML-GWAS models provide a more dependable and efficient framework by calculating the effects of all markers at the same time. The genetic architecture of complex traits may become more fully understood as a result of this simultaneous estimation. Therefore, ML-GWAS for aluminum toxicity tolerance indices, including geometric mean productivity (GMP), mean productivity (MP), and aluminum adoption index (AAI) can offer important information about the genetic basis of aluminum tolerance in durum wheat. Our investigation of the genetic variants linked to these aluminum-tolerance indices allowed us to pinpoint potential genes and genomic region that support durum wheat tolerant to aluminum toxicity. A total of 54 QTNs were identified for all shoot, root traits and aluminum tolerance indices in both stress-stressed and non-stressed conditions. Four significant QTNs (BS00064085_51, AX-111562104, BS00065128_51, and BS00064085_51) associated with shoot length and stress indices were identified on chromosomes 4B and 5B. This finding suggests that multiple genetic regions throughout the durum wheat genome play a role in influencing shoot length under both normal and stress conditions, as well as aluminum tolerance indices. Notably, both markers (BS00064085_51and BS00065128_51) were located on the same chromosome (5B), although they are different markers.The result showed the possibility of a linkage or interaction between these two genomic regions responsible for shoot length .Four significant QTNs were identified that influence root length under both stress and non-stress conditions, as well as the stress indices (AAI) on chromosome 2B, 5B, 7A, and 7B. Four different SNPs (AX-158592916, BS00077791_51, BS00022805_51, and AX-94463626) were detected for root length in both stress and non stress as well as aluminum adoption indice in separate chromosome, These suggest that multiple genetic loci contribute to root growth under Al stress condition. It provide potential targets for future breeding programs aimed at enhancing aluminum tolerance in Ethiopian durum wheat genotypes. The number of root hairs per plant trait on chromosomes 2A and 5B was linked to two major QTNs (AX-158540954 and AX-110508154). The outcome showed that several genetic loci throughout the durum wheat genome influenced this trait. This demonstrates the intricate genetic regulation of root hair growth. The quantity of root hairs is crucial for the intake of nutrients and water; in stressful conditions, these structures become vital for plant life. The correlation between certain genomic regions and root hairs under stress indicates that these regions may affect the root’s reaction to Al, and it is a useful predictor of the plant’s overall performance under stressful conditions. One significant QTN( AX-158524110) was associated with the root-to-shoot ratio (RSR) traits of non stress condition on chromosome 3A. This indicates that the root-to-shoot ratio controlled by single genome region. This QTN ( AX-158524110) was also associated with root dry weight of non stress condittion, it indicate the existance of high correlation between the traits and also this region are responsible for multiple traits. LD and LD decay of each and whole genome was computed LD was estimated by calculating the squared allele frequency correlation (r 2 ) among all possible pairs of markers for each of the 14 chromosomes and entire genome. Among all possible pairs of SNPs on each chromosome, 499,815 pairs were found in LD. Correlation coefficient ( r 2 ) values were then plotted against genetic distance (Mbp) across the whole genome, LD decay across the whole genome level was estimated using background LD (r 2 = 0.2) as a threshold. The LD decay of the whole genome was 4.032Mbp . Candidate genes contributing to aluminum toxicity tolerance in durum wheat In this study, several candidate genes associated with aluminum (Al) toxicity tolerance were identified in durum wheat. These genes are functionally linked to well-established Al tolerance mechanisms, including cell wall modification, organic acid biosynthesis and transport, intracellular detoxification, membrane stabilization, and stress-responsive transcriptional regulation. TRITD_3Av1G012300, TRITD_2Bv1G246550 genes encode wall-associated kinase (WAKs). Through cell wall alterations, WAK1 contributes to aluminum tolerance in Arabidopsis. These changes impact the cell wall’s ability to bind and trap aluminum ions, reducing their detrimental effects on root growth [ 41 ]. Several studies have shown that alteration in cell wall composition is crucial for enhancing Al tolerance [ 42 – 45 ]. Specifically, STOP1 regulates the transcription of atALMT1 and atMATE, which encode transporters responsible for the exudation of organic acids such as malate and citrate from root apices. These organic acids chelate aluminum ions in the rhizosphere, thereby protecting root tissues from aluminum toxicity [ 46 ]. The genes TRITD_5Bv1G232300 and TRITD_5Bv1G207280, encoding citrate synthase, It is directly associated with citrate biosynthesis In transgenic canola and yeast, overexpression of citrate synthase genes raises citrate levels and improves aluminum tolerance by chelating Al 3 [ 47 ]. Similar result were reported by Ray et al. 2009 identified the efflux of Al-dependent release of malate and citrate anions from wheat root apices which chelate the toxic Al 3 Cations thus protecting them from damage. TRITD_1Bv1G003380, TRITD_5Bv1G20602, TRITD_5Av1G022810, TRITD_2Bv1G233050 genes enode Glutathione S-transferase. TRITD_3Av1G010820, TRITD_5Bv1G231180, TRITD_2Bv1G233580 genes encode heavy metal transport/detoxification protein and TRITD_2Bv1G242520 gene encode ABC transporter ATP-binding protein. ABC transporters. These proteins facilitate cellular defense mechanisms and aluminum detoxification, allowing crops to withstand aluminum toxicity [ 48 – 50 ].TRITD_3Av1G010630,TRITD_1Bv1G002880genes encode receptor protein kinase and TRITD_1Bv1G003280 gene encode leucine rich repeat protein. The genes identified in the present study act as sensors of extracellular Al stress and transmit information to transcriptional and metabolic pathways, facilitating the coordinated control of organic acid production, detoxification, and cell wall remodeling [ 51 ]. TRITD_5Bv1G206390,TRITD_5Av1G025370,TRITD_2Bv1G231800 genes encode cytochrome P450, enzyme, Cytochrome P450 enzymes play a crucial role in plant stress response, including biosynthesis and stress regulation, which could relate to enhancing aluminum tolerance by regulating sterol content in plant root tips [ 52 ]. TRITD_7Bv1G015610, TRITD_7Bv1G015570, TRITD_2Bv1G243610 genes encode MADS box transcription factor. GsMAS1, a MADS-box transcription factor from wild soybean, enhances aluminum stress tolerance in Arabidopsis by upregulating specific genes like atALMT1 and STOP2, indicating a potential role for aluminum toxicity tolerance [ 53 ]. TRITD_5Bv1G206870, TRITD_5Bv1G206810, TRITD_5Av1G021830 genes encode NAC domain-containing protein. In Arabidopsis, ANAC017 transcription facto regulates aluminum tolerance by regulating cell wall-modifying genes, specifically XTH31 [ 54 ]. Conclusion In the present study, a highly controlled hydroponic system that allowed efficient discrimination between Al-sensitive and Al-tolerant durum wheat genotypes was used. The results from this study are very useful for planning future durum wheat breeding programs, particularly in Soil acidity affected areas. Significant genetic variation in shoot and root traits was observed among the genotypes. The genotypes were classified into three clusters based on their tolerance levels. Cluster I containing the most tolerant accessions. Remarkably, accesstions 3540 and 5563 were the most tolerant, while Cluster III comprised the most susceptible ones. The Genome wide association analysis identified forty candidate genes and 54 significant QTNs associated with aluminum tolerance related traits in durum wheat genotypes. These genes are essential for detoxifying heavy metals and preserving the integrity of cell walls, especially under conditions of aluminum stress. altering the release of organic acids, triggering defense mechanisms, increasing the effectiveness of phosphorus utilization, scavenging reactive oxygen species, storing heavy metals in vacuoles, chelating aluminum, sensing external cues, starting intracellular signaling cascades, and mediating signal transduction pathways. The identification of these genes and their roles offers important insights into the molecular mechanisms underlying aluminum stress responses in durum wheat. Further research on these genes may lead to the development of improved durum wheat genotypes that are tolerant to aluminum toxicity. Supplementary Information Supplementary Material 1. (47.6KB, docx) Acknowledgements The author would like to thank the Ethiopian Biodiversity Institute (EBI), DebreZeit Agricultural Research Center (DzARC) and Sinana Agricultural Research Center (SARC) for providing us with the Durum wheat genotypes and passport data. The authors would like to thank the Institute of Biotechnology, Addis Ababa University (AAU) for providing of SNP data. This study was supported by the ministry of education. Authors’ contributions All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by Tsegaye Abebe, Wosene Gebreselassie, Temesgen Menamo, Tilahun Mekonnen, Behailu Mulugeta, and Kassahun Tesfaye. The first draft of the manuscript was written by Tsegaye Abebe and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript. Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Data availability All materials used in this finding attached as supplementary materials. Declarations Ethics approval and consent to participate Not applicable. This study did not involve clinical trials, human participants, or animals. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References 1. Zingale S, Spina A, Ingrao C, Fallico B, Timpanaro G, Anastasi U, et al. Factors affecting the nutritional, health, and technological quality of durum wheat for pasta-making: A systematic literature review. Plants. 2023;12(3):530. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 2. Ballesta P, Mora F, Del Pozo A. Association mapping of drought tolerance indices in wheat: QTL-rich regions on chromosome 4A. Scientia Agricola. 2019;77(2):e20180153. 3. Wayima EF, Ligaba-Osena A, Dagne K, Tesfaye K, Machuka EM, Mutiga SK, et al. Screening of diverse Ethiopian durum wheat accessions for aluminum tolerance. Agronomy. 2019;9(8):440. [ Google Scholar ] 4. Mello JCd, Tonial IB, Lucchetta L. Aluminum accumulation in the wheat production chain: a review. Food Sci Technol. 2023;43:e116022. [ Google Scholar ] 5. Ofoe R, Thomas RH, Asiedu SK, Wang-Pruski G, Fofana B, Abbey L. Aluminum in plant: Benefits, toxicity and tolerance mechanisms. Front Plant Sci. 2023;13:1085998. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 6. Jaiswal SK, Naamala J, Dakora FD. Nature and mechanisms of aluminium toxicity, tolerance and amelioration in symbiotic legumes and rhizobia. Biol Fertil Soils. 2018;54:309–18. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 7. Brhane H, Haileselassie T, Tesfaye K, Hammenhag C, Ortiz R, Geleta M. Hydroponic Optimization and Screening of Aluminum Tolerance on Finger Millet (Eleusine coracana (L.) Gaertn.) Accessions and Cultivars. Agronomy. 2023;13(6):1596. [ Google Scholar ] 8. Matonyei TK, Barros BA, Guimaraes RG, Ouma EO, Cheprot RK, Apolinário LC, et al. Aluminum tolerance mechanisms in Kenyan maize germplasm are independent from the citrate transporter ZmMATE1. Sci Rep. 2020;10(1):7320. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 9. Negusse H, Cook DR, Haileselassie T, Tesfaye K. Identification of Aluminum Tolerance in Ethiopian Chickpea (Cicer arietinum L.) Germplasm. Agronomy. 2022;12(4):948. [ Google Scholar ] 10. Abate E, Hussein S, Amelework A, Shaff JE, Laing M, Tadele Z, et al. Investigation of Al-toxicity tolerance in tef (Eragrostis tef) under hydroponic system using root growth measurement and haematoxylin staining methods. Aust J Crop Sci. 2022;16(8):1047–59. [ Google Scholar ] 11. Choudhary AK, Singh D. Screening of pigeonpea genotypes for nutrient uptake efficiency under aluminium toxicity. Physiol Mol Biology Plants. 2011;17:145–52. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 12. Hede A, Skovmand B, Ribaut JM, González-de‐León D, Stølen O. Evaluation of aluminium tolerance in a spring rye collection by hydroponic screening. Plant Breeding. 2002;121(3):241–8. [ Google Scholar ] 13. Naeem M, Zia MAB, Yousaf MF, Asim A. An overview of genome-wide association mapping studies in Poaceae species (Model crops. Wheat and Rice; 2022. [ DOI ] [ PubMed ] 14. Zhong H, Liu S, Sun T, Kong W, Deng X, Peng Z, et al. Multi-locus genome-wide association studies for five yield-related traits in rice. BMC Plant Biol. 2021;21(1):364. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 15. Farokhzadeh S, Fakheri BA, Mahdi Nezhad N, Tahmasebi S, Mirsoleimani A. Genetic dissection of spike-related traits in wheat (Triticum aestivum L.) under aluminum stress. Genet Resour Crop Evol. 2020;67(5):1221–43. [ Google Scholar ] 16. Delhaize E, Ryan PR, Hebb DM, Yamamoto Y, Sasaki T, Matsumoto H. Engineering high-level aluminum tolerance in barley with the ALMT1 gene. Proceedings of the National Academy of Sciences. 2004;101(42):15249-54. [ DOI ] [ PMC free article ] [ PubMed ] 17. Alvarado G, Rodríguez FM, Pacheco A, Burgueño J, Crossa J, Vargas M, et al. META-R: A software to analyze data from multi-environment plant breeding trials. Crop J. 2020;8(5):745–56. [ Google Scholar ] 18. Nakagawa S, Lagisz M, O’Dea RE, Pottier P, Rutkowska J, Senior AM et al. orchaRd 2.0: An R package for visualising meta-analyses with orchard plots. 2023. 19. Elias M, Chere D, Lule D, Serba D, Tirfessa A, Gelmesa D, et al. Multi-locus genome‐wide association study reveal genomic regions underlying root system architecture traits in Ethiopian sorghum germplasm. Plant Genome. 2024;17(2):e20436. [ DOI ] [ PMC free article ] [ PubMed ] 20. Johnson HW, Robinson H, Comstock R. Estimates of genetic and environmental variability in soybeans 1. Agron J. 1955;47(7):314–8. [ Google Scholar ] 21. Kassambara A, Mundt F. Package ‘factoextra’. Extract and Visualize the Results of Multivariate Data Analyses. 2016. 22. Fischer R, Maurer R. Drought resistance in spring wheat cultivars. I. Grain yield responses. Aust J Agric Res. 1978;29(5):897–912. [ Google Scholar ] 23. Howeler R, editor. Identifying plants adaptable to low pH conditions. Plant-Soil Interactions at Low pH: Proceedings of the Second International Symposium on Plant-Soil Interactions at Low pH, 24–29 June 1990, Beckley West Virginia, USA; 1991: Springer. 24. Rosielle A, Hamblin J. Theoretical aspects of selection for yield in stress and non-stress environment 1. Crop Sci. 1981;21(6):943–6. [ Google Scholar ] 25. Schneider KA, Rosales-Serna R, Ibarra‐Perez F, Cazares‐Enriquez B, Acosta‐Gallegos JA, Ramirez‐Vallejo P, et al. Improving common bean performance under drought stress. Crop Sci. 1997;37(1):43–50. [ Google Scholar ] 26. Jafari A, Paknejad F, JAMI AM. Evaluation of selection indices for drought tolerance of corn (Zea mays L.) hybrids. 2009. 27. Wang S-B, Feng J-Y, Ren W-L, Huang B, Zhou L, Wen Y-J, et al. Improving power and accuracy of genome-wide association studies via a multi-locus mixed linear model methodology. Sci Rep. 2016;6(1):19444. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 28. Wen Y-J, Zhang H, Ni Y-L, Huang B, Zhang J, Feng J-Y, et al. Methodological implementation of mixed linear models in multi-locus genome-wide association studies. Brief Bioinform. 2018;19(4):700–12. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 29. TambaCL, Zhang Y-M. A fast mrMLM algorithm for multi-locus genome-wide association studies. biorxiv. 2018:341784. 10.1101/341784. 30. Ren W-L, Wen Y-J, Dunwell JM, Zhang Y-M. pKWmEB: integration of Kruskal–Wallis test with empirical Bayes under polygenic background control for multi-locus genome-wide association study. Heredity. 2018;120(3):208–18. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 31. Zhang J, Feng J-Y, Ni Y, Wen Y, Niu Y, Tamba C, et al. pLARmEB: integration of least angle regression with empirical Bayes for multilocus genome-wide association studies. Heredity. 2017;118(6):517–24. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 32. Chaurasia S, Singh AK, Songachan L, Sharma AD, Bhardwaj R, Singh K. Multi-locus genome-wide association studies reveal novel genomic regions associated with vegetative stage salt tolerance in bread wheat (Triticum aestivum L). Genomics. 2020;112(6):4608–21. [ DOI ] [ PubMed ] [ Google Scholar ] 33. Bradbury PJ, Zhang Z, Kroon DE, Casstevens TM, Ramdoss Y, Buckler ES. TASSEL: software for association mapping of complex traits in diverse samples. Bioinformatics. 2007;23(19):2633–5. [ DOI ] [ PubMed ] [ Google Scholar ] 34. Baloch FS, Alsaleh A, Shahid MQ, Çiftçi V, Sáenz de Miera E, Aasim L. A whole genome DArTseq and SNP analysis for genetic diversity assessment in durum wheat from central fertile crescent. PLoS ONE. 2017;12(1):e0167821. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 35. Delhaize E, Ryan PR, Randall PJ. Aluminum tolerance in wheat (Triticum aestivum L.)(II. Aluminum-stimulated excretion of malic acid from root apices). Plant Physiol. 1993;103(3):695–702. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 36. Wayima EF. Classification of Ethiopian soils with pH. J Soil Sci Environ Manag. 2019. In press. 37. Khosa J, Dhatt A. Studies on genetic variability and heritability in bulb onion (Allium cepa L.) in North-Western plains of India. J Hortic Sci. 2013;8(2):255–8. [ Google Scholar ] 38. Johnson HW, Robinson H, Comstock R. Estimates of genetic and environmental variability in soybeans. 1955. 39. Zhao Q, Guo J, Shu M, Wang P, Hu S. Impacts of drought and nitrogen enrichment on leaf nutrient resorption and root nutrient allocation in four Tibetan plant species. Sci Total Environ. 2020;723:138106. [ DOI ] [ PubMed ] [ Google Scholar ] 40. Rakitsch B, Lippert C, Stegle O, Borgwardt K. A Lasso multi-marker mixed model for association mapping with population structure correction. Bioinformatics. 2013;29(2):206–14. [ DOI ] [ PubMed ] [ Google Scholar ] 41. Sivaguru M, Ezaki B, He Z-H, Tong H, Osawa H, Baluška F, et al. Aluminum-induced gene expression and protein localization of a cell wall-associated receptor kinase in Arabidopsis. Plant Physiol. 2003;132(4):2256–66. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 42. Riaz M, Yan L, Wu X, Hussain S, Aziz O, El-Desouki Z, et al. Excess boron inhibited the trifoliate orange growth by inducing oxidative stress, alterations in cell wall structure, and accumulation of free boron. Plant Physiol Biochem. 2019;141:105–13. [ DOI ] [ PubMed ] [ Google Scholar ] 43. Xu JM, Wang ZQ, Jin JF, Chen WW, Fan W, Zheng SJ, et al. FeSTAR2 interacted by FeSTAR1 alters its subcellular location and regulates Al tolerance in buckwheat. Plant Soil. 2019;436(1):489–501. [ Google Scholar ] 44. Zhu CQ, Cao XC, Zhu LF, Hu WJ, Hu AY, Abliz B, et al. Boron reduces cell wall aluminum content in rice (Oryza sativa) roots by decreasing H2O2 accumulation. Plant Physiol Biochem. 2019;138:80–90. [ DOI ] [ PubMed ] [ Google Scholar ] 45. Sawaki Y, Iuchi S, Kobayashi Y, Kobayashi Y, Ikka T, Sakurai N, et al. STOP1 regulates multiple genes that protect Arabidopsis from proton and aluminum toxicities. Plant Physiol. 2009;150(1):281–94. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 46. Kochian LV, Hoekenga OA, Pineros MA. How do crop plants tolerate acid soils? Mechanisms of aluminum tolerance and phosphorous efficiency. Annu Rev Plant Biol. 2004;55(1):459–93. [ DOI ] [ PubMed ] [ Google Scholar ] 47. Anoop VM, Basu U, McCammon MT, McAlister-Henn L, Taylor GJ. Modulation of citrate metabolism alters aluminum tolerance in yeast and transgenic canola overexpressing a mitochondrial citrate synthase. Plant Physiol. 2003;132(4):2205–17. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 48. Houde M, Diallo AO. Identification of genes and pathways associated with aluminum stress and tolerance using transcriptome profiling of wheat near-isogenic lines. BMC Genomics. 2008;9(1):400. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 49. Huang CF, Yamaji N, Mitani N, Yano M, Nagamura Y, Ma JF. A bacterial-type ABC transporter is involved in aluminum tolerance in rice. Plant Cell. 2009;21(2):655–67. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 50. Zhao F-J, Tang Z, Song J-J, Huang X-Y, Wang P. Toxic metals and metalloids: Uptake, transport, detoxification, phytoremediation, and crop improvement for safer food. Mol Plant. 2022;15(1):27–44. [ DOI ] [ PubMed ] [ Google Scholar ] 51. Ding ZJ, Xu C, Yan JY, Wang YX, Cui MQ, Yuan JJ, et al. The LRR receptor-like kinase ALR1 is a plant aluminum ion sensor. Cell Res. 2024;34(4):281–94. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 52. Pandian BA, Sathishraj R, Djanaguiraman M, Prasad PV, Jugulam M. Role of cytochrome P450 enzymes in plant stress response. Antioxidants. 2020;9(5):454. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 53. Zhang X, Li L, Yang C, Cheng Y, Han Z, Cai Z, et al. GsMAS1 encoding a MADS-box transcription factor enhances the tolerance to aluminum stress in Arabidopsis thaliana. Int J Mol Sci. 2020;21(6):2004. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 54. Tao Y, Wan JX, Liu YS, Yang XZ, Shen RF, Zhu XF. The NAC transcription factor ANAC017 regulates aluminum tolerance by regulating the cell wall-modifying genes. Plant Physiol. 2022;189(4):2517–34. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Associated Data This section collects any data citations, data availability statements, or supplementary materials included in this article. Supplementary Materials Supplementary Material 1. (47.6KB, docx) Data Availability Statement All materials used in this finding attached as supplementary materials. Articles from BMC Plant Biology are provided here courtesy of BMC ACTIONS View on publisher site PDF (3.1 MB) Cite Collections Permalink PERMALINK Copy RESOURCES Similar articles Cited by other articles Links to NCBI Databases Cite Copy Download .nbib .nbib Format: AMA APA MLA NLM Add to Collections Create a new collection Add to an existing collection Name your collection * Choose a collection Unable to load your collection due to an error Please try again Add Cancel Follow NCBI NCBI on X (formerly known as Twitter) NCBI on Facebook NCBI on LinkedIn NCBI on GitHub NCBI RSS feed Connect with NLM NLM on X (formerly known as Twitter) NLM on Facebook NLM on YouTube National Library of Medicine 8600 Rockville Pike Bethesda, MD 20894 Web Policies FOIA HHS Vulnerability Disclosure Help Accessibility Careers NLM NIH HHS USA.gov Back to Top