Understanding the genetic architecture of drought stress tolerance in bread wheat at seedling and reproductive stages is crucial for developing drought-tolerant varieties. In the present study, 192 diverse wheat genotypes, a subset from the Wheat Associated Mapping Initiative (WAMI) panel, were evaluated at the seedling stage in a hydroponics system for chlorophyll content (CL), shoot length (SLT), shoot weight (SWT), root length (RLT), and root weight (RWT) under both drought and optimum conditions. Following that, a genome-wide association study (GWAS) was carried out using the phenotypic data recorded during the hydroponics experiment as well as data available from previously conducted multi-location field trials under optimal and drought stress conditions. The panel had previously been genotyped using the Infinium iSelect 90K SNP array with 26,814 polymorphic markers. Using single as well as multi-locus models, GWAS identified 94 significant marker-trait associations (MTAs) or SNPs associated with traits recorded at the seedling stage and 451 for traits recorded at the reproductive stage. The significant SNPs included several novel, significant, and promising MTAs for different traits. The average LD decay distance for the whole genome was approximately 0.48 Mbp, ranging from 0.07 Mbp (chromosome 6D) to 4.14 Mbp (chromosome 2A). Furthermore, several promising SNPs revealed significant differences among haplotypes for traits such as RLT, RWT, SLT, SWT, and GY under drought stress. Functional annotation and in silico expression analysis revealed important putative candidate genes underlying the identified stable genomic regions such as protein kinases, O-methyltransferases, GroES-like superfamily proteins, NAD-dependent dehydratases, etc. The findings of the present study may be useful for improving yield potential, and stability under drought stress conditions.

Context: Water stress is fast becoming a major limiting factor for wheat production. Hence, identifying drought tolerant genotypes is critical for sustaining the food supply chain. However, there are no phenotypic markers or statistical models available that may be employed for the efficient selection of field grown drought tolerant wheat genotypes. Objectives: We evaluated wheat genotypes to (1) identify novel sources of drought tolerance (2) understand underlying adaptation mechanisms of drought tolerance (3) identify phenotypic markers and a stable model for the selection of drought tolerant genotypes. Methods: One hundred ninety-six diverse wheat genotypes were evaluated at three different locations in India: Banaras Hindu University (BHU), Varanasi (E1 and E3: control; E2 and E4: drought); Agharkar Research Institute (ARI), Pune (E5 and E7: control; E6 and E8: drought) and Borlaug Institute for South Asia (BISA), Jabalpur (E9 and E11: control; E10 and E12: drought) for various agronomic, physiological and yield traits for two consecutive years. Drought was imposed at the heading stage (Z59) by withholding irrigation for four weeks until the moisture reading reached <45% than the control (100%). Results: The performance of all genotypes significantly declined under drought at all the locations. Normalized difference vegetation index (NDVI) significantly correlated (r = 0.41** and 0.36**) with the grain yield under drought during maturity. At the same time, there was no association under control conditions (r = 0.07 and 0.10) at the BHU center during 2020-21 and 2021-22, respectively. Stress indices, such as geometric mean productivity (GMP) and stress tolerance index (STI), showed a high correlation (r= 0.89** and r = 0.88**, respectively) with the grain yield under drought and were effective in differentiating drought tolerant genotypes. GGE bi-plots discriminated the environments (observed obtuse angle between E3 with E6 and E9, E4 with E6 and E9) having negative relation and cross-over interaction for grain yield. Further, the multi-trait stability index (MTSI) identified 29 stable genotypes across all environments and was predicted as the most accurate model due to its fewer Root Mean Square Prediction Difference (RMSPD) values. Conclusion: NDVI is a useful high throughput screening tool under drought and MTSI is an effective method for selecting stable wheat genotypes across different water stressed locations. Implications: The identified tools (NDVI), method (MTSI), and tolerant genotypes appear to be valuable resources that together will be useful in the ongoing breeding programs to enhance the drought tolerance of wheat.