Altitude‐Driven Stability and Sensitivity in Spike Morphology Traits of Worldwide Wheat Accessions

Summary statement

This study investigated the spike morphology traits of 528 wheat accessions across varying altitudinal environments and revealed that different genotypes employ distinct strategies to adapt to altitudinal changes. Genome‐wide association study was conducted to identify haplotypes associated with both the altitudinal sensitivity and stability of different spike morphology traits.

Wheat (Triticum aestivum L.), one of the most staple food crops cultivated worldwide, accounts for up to 20% of the total daily caloric intake in the human diet (http://faostat.fao.org). Improving grain yield has become an urgent priority in wheat breeding due to rapid population growth, which is increasing pressure on global food production (Liu et al. 2024). Covering an area 2.3‐fold larger than global arable land, mountainous regions represent a significant yet underperforming agricultural frontier (https://www.fao.org). Here, yields of staple crops such as wheat and maize are limited to merely 30%–50% of their climatic potential, highlighting a substantial yield gap (Agrometerology 1989), which underscores the urgent need to elucidate the phenotypic and genetic response mechanisms of crops across varying altitudinal gradients.

Altitude affects the developmental and physiological characteristics of plants through various abiotic stresses, including pressure, ultraviolet light, light intensity, temperature, humidity, precipitation and soil nutrients (Waseem et al. 2025). Wheat spike traits are highly susceptible to abiotic stress. As a critical determinant of grain yield components (Liu et al. 2024; Guo et al. 2024), wheat spike morphology remains poorly understood regarding its regulation by multiple abiotic factors across altitudinal gradients. We evaluated 13 spike morphology traits in 528 accessions across two locations with four altitude environments (Kangxian top altitude [KX‐T]: 32.53° N, 105.18° E, 1450 m, Kangxian base altitude [KX‐B]: 32.53° N, 105.18° E, 1000 m; Qingshui top altitude [QS‐T]: 34.32° N, 105.45° E, 1530 m, Qingshui base altitude [QS‐B]: 34.32° N, 105.45° E, 1430 m) (Figure 1A,B, Supporting Information S2: Table S1). This paired‐gradient design on a consistent mountain aspect minimizes confounding geographic variables, thereby isolating altitude as the primary environmental driver.

Figure 1.

Wheat spike morphology traits variation and genetic basis under different altitudinal environments. (A) The landscapes of seeding stage (left) and physiological maturity (right). (B) The wheat spike and their rachis are displayed. In the figure, 1–5, respectively, show the five sections from the apical to the basal, as well as the number of grains in each individual spikelet of each section. 6 represents the number of sterile spikelets at basal parts of spike. The white bar in the plots indicates 1 cm. (C) The number of genotypes with variations in wheat spike morphology traits. (D) Number of genomic regions for spikelet traits at high and low altitudes. Each circle represents an individual genomic region, and connected points of the same trait with different altitudes indicate co‐localized regions. (E) The phenotypic differences of the accessions harbouring Hap‐SW‐1 and Hap‐GWS‐1 in top and base altitude environments, respectively. Significance levels of phenotypic differences are indicated by ***(p < 0.001). B, base altitude; GNS, grain number per spike; GNSP‐A, grain number per spikelet at apical parts of spike; GNSP‐B, grain number per spikelet at basal parts of spike; GNSP‐C, grain number per spikelet at central part of spike; GNSP‐MA, grain number per spikelet at middle‐apical parts of spike; GNSP‐MB, grain number per spikelet at middle‐basal parts of spike; GWS, grain weight per spike; RW, rachis weight; SNS, fertile spikelet number per spike; SS‐A, sterile spikelet number at apical parts of spike; SS‐B, sterile spikelet number at basal parts of spike; SL, spike length; SW, spike weight; T, top altitude.

Environmental monitoring revealed distinct microclimates across the altitudinal gradients (Supporting Information S1: Figure S1). While all spike morphology traits exhibited broad genotypic variation and high heritability (Supporting Information S1: Figure S2), their distributions differed: most were normally distributed, whereas SS‐A and SS‐B followed a binary distribution (Supporting Information S1: Figure S3). High‐resolution correlation analysis identified GNSP‐C and GNSP‐MB as key determinants of SW, and these trait associations were highly consistent across all four geographic locations (Supporting Information S1: Figure S4). Approximately 21% (112) of the 528 accessions showed no changes in the measured spike traits at different altitudes, 203 genotypes (39%) exhibited significant differences in one or two traits between altitude environments, whereas 14% (74) of accessions showed significant differences in six or more traits (Figure 1C). These differential response patterns among spike morphology traits across accessions suggest that different accessions employ variable strategies to cope with the challenges imposed by varying altitude environments. With increasing altitude, a significant reduction in SNS, SS and SL was observed in most of the genotypes (Supporting Information S1: Figure S5), suggesting that the accessions may modulate multiple reproductive‐trait developmental processes to adapt to diverse altitude environments and the differences in the response characteristics of different traits to altitude. Unlike the universal negative yield response to altitude documented by De Flaviis et al. (2023) in a small cohort (n = 6), our results likely reflect the broader genetic variation and expanded scale of our study. There were 307 accessions showing an increase in one or more traits and 268 accessions showing a decrease in one or more traits due to the rise in altitude environment (Supporting Information S1: Figure S5). Many accessions were repeatedly counted, indicating that substantial genotype‐dependent plasticity in response to elevational variation, highlighting distinct adaptive strategies among accessions. We identified 112 accessions with stable spikelet phenotypes across altitudes, indicating the presence of adaptive mechanisms that buffer against environmental variation (Figure 1C, Supporting Information S2: Table S2). These accessions serve as a critical resource for elucidating the genomic architecture of adaptation and for developing climate‐resilient wheat varieties.

GWAS across the top and base altitude environments identified 6253 and 12 533 significant marker–trait associations, respectively (Supporting Information S1: Figure S6). We delineated 163 and 212 distinct genomic regions for the 13 spike morphology traits under top‐ and base‐altitude conditions using LD, respectively (Supporting Information S2: Table S3). These genomic regions were distributed across the A (61 and 100), B (60 and 53) and D (42 and 59) genomes (Supporting Information S1: Figure S7). A greater number of spike morphology‐associated genomic regions were detected in the A subgenome compared to the B and D subgenomes, suggesting that the detection of these markers is independent of the intrinsic marker density of the subgenomes (Supporting Information S1: Figure S8). We speculate that the A subgenome harbours a greater abundance of genes governing spike morphology, a finding corroborated by Liu et al. (2024). Among the significant signals identified in top and base altitude environments, the significant SNPs detected on Chromosome 7 accounted for 90% of those within the A subgenome. This suggests that chromosome 7A may harbour critical regulatory regions governing environmental adaptation in wheat. Consistent with this, Z. Wang et al. (2021) identified environment‐associated regions on 7A (211.7–272.9 Mb). Given that wheat adapts to environmental conditions by regulating its growth period, and cloned vernalization‐related genes such as TaVRT‐2 and VRN‐A3 are localized to Chromosome 7A, these findings further corroborate the validity of our hypothesis. Furthermore, we conducted GWAS on trait differences and ratios between altitudes to detect minor‐effect and altitude‐dependent loci (Supporting Information S1: Figure S9). The difference‐based analysis proved more powerful, identifying 103 genomic regions compared to 88 from the ratio‐based analysis (Supporting Information S2: Table S4). The most significant signals were observed for GNSP‐MA, GNSP‐C and SL in the difference analysis, and for GNSP‐MA, GNSP‐C, GWS and SL in the ratio analysis.

Genomic regions were classified into two categories: altitude‐stable regions, which were detected across two environments, and altitude‐sensitive regions, which were identified only under specific altitude conditions or through comparative analyses between altitudes. Co‐localized genomic regions were identified for all traits except GNSP‐A (Figure 1D). This conservation suggests environment‐independent genetic regulation for these traits, in contrast to the altitude‐sensitive control of GNSP‐A. The most extensive co‐location of genomic regions for GNSP‐B and GNSP‐MB across altitudes confirms their high stability, consistent with our phenotypic observations of spike morphology. Haplotype analysis of the nine stable loci confirmed their stability, as phenotypic values showed no significant differences between top‐ and base‐altitude environments, except for GWS (Supporting Information S1: Figure S10). Some specific haplotypes were identified that showed increased grain number, greater grain weight and reduced floral sterility. Accessions harbouring a combination of these favourable haplotypes exhibited remarkable altitude stability and represent valuable genetic resources for breeding varieties with tailored altitude adaptability (Supporting Information S2: Table S5). Altitude‐sensitive genomic regions analysis identified haplotypes adapted to distinct environments. Within two key regions on chromosome 4B (124 226 565–141 585 467 and 131 585 467–141 585 467), two haplotypes were identified for SW and GWS. Haplotypes Hap‐SW‐1 and Hap‐GWS‐1 were associated with top‐altitude adaptation (Figure 1E). Notably, 81% of accessions carrying Hap‐SW‐1 also harboured Hap‐GWS‐1 (Supporting Information S2: Table S6).

The polygenic nature of spike architecture is evidenced by pleiotropic genomic region effects. We identified 98 genomic regions influencing more than two traits (Supporting Information S1: Figure S11), consistent with previous reports of co‐localized genomic regions for different spike characteristics (S. Wang et al. 2023). There were two and nine regions associated with three traits co‐localized, which were identified in the top altitude and the base altitude, respectively. The traits (GNSP‐C, GNSP‐A, GNSP‐MB) with a strong correlation to the phenotype exhibit a phenomenon of co‐localization. The top altitude did not identify any regions that were associated with four or more traits. A pleiotropic region governing four spike morphology traits, GWS, SW, SL and GNSP‐MB, was identified on Chromosome 2D (31 443 586–39 851 097) under base altitude conditions (Supporting Information S2: Table S3). This observed genetic overlap may arise from either: tightly linked loci governing distinct traits or genuine pleiotropic effects of individual genes. Remarkably, Zhao et al. (2023) also identified PPD‐D1 within this region, a gene known to regulate photoperiod sensitivity and facilitate adaptation to high‐altitude environments. This suggests a tight linkage between genes controlling flowering time and spike architecture, which may reflect environmental adaptive selection in wheat. Specifically, at high altitudes, the flowering period is adjusted to avoid potential freezing injury, whereas at low altitudes, it may be modulated to achieve larger spikes. This study identified key genomic regions and superior haplotypes that regulate wheat spike morphology in response to altitude, providing a molecular basis for breeding varieties with enhanced stability and adaptability across diverse elevations.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1: The temperatures and humidities of four locations. Figure S2: Proportion of variance explained by genotype, environment, genotype × environment interaction (G×E), and residual effects for wheat morphology traits. Figure S3: Distribution of 13 spike morphology traits under different environments. Figure S4: Correlation between different spike morphology traits under different altitude environments. Figure S5: The number of accessions exhibiting significant differences among 13 spike morphology traits between top altitude (BLUEs of KX‐T and QS‐T) and base altitude (BLUEs of KX‐B and QS‐B) environments. Figure S6: GWAS based on top and base altitude to dissect spike morphology traits using wheat accessions. Figure S7: The distribution of the genomic regions associated with spike morphology traits mapping strategies. Figure S8: Genome‐wide distribution of SNP density in the wheat population. Figure S9: GWAS based on compared value to dissect spike morphology traits using wheat accessions. Figure S10: Haplotype effect and superior germplasm analysis. Figure S11: The number of regions of different morphology traits.

Table S1: Summary of the 528 worldwide wheat accessions. Table S2: Accessions that did not show any change in 13 spike morphology traits under different altitude conditions. Table S3: 375 genomic regions significantly associated with 13 wheat morphology traits based on top and base altitudes. Table S4: 191 genomic regions significantly associated with 13 wheat morphology traits basied on difference and ratio value of different altitude. Table S5: Accessions corresponding to each haplotype for different phenotypes. Table S6: Accessions harboring different haplotypes.

Wang, H. , Zhang Y., He K., et al. 2026. “Altitude‐Driven Stability and Sensitivity in Spike Morphology Traits of Worldwide Wheat Accessions.” Plant, Cell & Environment 49: 4819–4822. 10.1111/pce.70473.

Data Availability Statement

The genotypes of 528 wheat accessions used in this study have been deposited in the Genome Variation Map (https://bigd.big.ac.cn/gvm) under accession numbers GVM000463 and GVM000315 (Niu et al. 2023; Zhao et al. 2023).