The balance of nitrogen (N) availability directly affects plant health and, ultimately, productivity (De Bang et al., 2020). It is estimated that nearly half of the global increase in food production can be attributed to the application of N fertilizers (Tilman et al., 2002). Approximately half of the applied N fertilizers are absorbed by crops, while the remainder accumulates in the soil (Fan et al., 2012; He et al., 2021), leading to severe ecological issues such as soil acidification and water eutrophication (Zhang et al., 2022). Therefore, identifying crop germplasm resources capable of maintaining normal growth under reduced N fertilizer application is of great significance for energy conservation, emission reduction, and environmental protection.

Rapeseed (Brassica napus), as an important oilseed crop, is widely cultivated in regions such as China, Northern Europe, Canada, and Australia (Raboanatahiry et al., 2021; Tang et al., 2021). Regardless of the growth type—winter, spring, or semi-winter—rapeseed exhibits a high demand for N nutrition during the seedling stage (Shen et al., 2022). However, during the maturation stage, excessive N often disrupts the balance of nutrient elements, adversely affecting plant growth and development (Ren et al., 2017). Throughout the entire growth cycle of rapeseed, N plays a pivotal role in regulating growth, significantly influencing both yield and quality (Yahbi et al., 2022).

Genome-wide association studies (GWAS), a novel genetic analysis method based on linkage disequilibrium (Alseekh et al., 2021; Tibbs-Cortes et al., 2021), have become a powerful tool for understanding the genetic architecture of complex quantitative traits in crops. With the rapid advancement of sequencing technologies, GWAS has been widely applied to decipher the genetic basis of complex traits in major cereal crops such as rice (Jiang et al., 2021; Zhong et al., 2021), wheat (Saini et al., 2021), soybean (Niu et al., 2024), and maize (Sahito et al., 2024). In recent years, the declining cost of genome sequencing has significantly reduced the cost of SNP marker development, facilitating the successful application of GWAS in detecting quantitative trait loci (QTLs) associated with complex traits in rapeseed. These traits include seed quality (Zhao et al., 2022), harvest index (Qin et al., 2022), yield (Lu et al., 2017), erucic acid content (Xu et al., 2024), glucosinolate content (Tang et al., 2023), and oil content (Pal et al., 2021). Additionally, GWAS has played a crucial role in identifying genes associated with abiotic stress tolerance in rapeseed, including those related to nutrient (Knoch et al., 2024), salt (Wan et al., 2017), drought (Khanzada et al., 2020), low-temperature (Horvath et al., 2020), and heavy metal stresses (Wan et al., 2020).

In various crops, numerous QTLs mapping studies associated with low N tolerance have been reported (Kwon et al., 2021; Yang et al., 2021; Rosa-Martínez et al., 2023; Singh et al., 2023). However, reports on the cloning of low N tolerance genes remain limited. This is primarily due to the large genomic intervals of QTL regions, making it challenging to predict and isolate candidate genes. With its high resolution, GWAS enable the rapid identification and isolation of candidate genes at significant loci, providing a significant advantage in the selection of low N tolerance genes (Ahmed et al., 2024). In the study of low N tolerance candidate genes in major crops, the application of GWAS has already achieved notable progress. For example, Li et al. (2022) utilized 230 rice germplasm accessions and combined GWAS with transcriptomics to identify 411 candidate genes associated with low N stress tolerance from five QTLs, along with 2,722 differentially expressed genes responsive to low N signals. Notably, 24 genes were identified by both methods. Sanchez et al. evaluated multiple agronomic traits in 181 doubled haploid (DH) maize germplasm under both high and low N conditions (Sanchez et al., 2023). By combining 62,077 SNP markers, the study identified several SNPs associated with agronomic traits under these conditions, providing valuable insights for the development of molecular markers for high and low N stress tolerance in maize and the creation of maize germplasm materials with enhanced N tolerance. In recent years, several studies have advanced our understanding of the genetic basis of N use efficiency (NUE) in Brassica napus. Li et al. (2020) conducted comparative genome and transcriptome analyses on two genotypes with contrasting NUE, revealing that differences in N uptake and assimilation, as well as coordinated carbon and N metabolism, contribute to NUE variation. Wei et al. (Yang et al., 2020) integrated genome-wide association studies (GWAS) with gene expression profiling, identifying candidate genes associated with NUE and highlighting the complex regulatory networks involved. These studies underscore the importance of integrating genomic and transcriptomic approaches to elucidate the mechanisms underlying NUE. However, variations in experimental conditions, such as N levels, developmental stages, and phenotypic traits assessed, suggest that further research is needed to uncover additional loci and genes contributing to NUE in B. napus. Studies using GWAS to identify QTLs and genes associated with low N (LN) stress tolerance in rapeseed have also been reported. Under LN and high N (HN) conditions, GWAS combined with RNA-seq technology detected 14 QTLs associated with N efficiency traits in rapeseed. A total of 245, 540, and 399 differentially expressed genes (DEGs) were identified as specific to LN stress, HN conditions, and common to both LN and HN, respectively. Integrating GWAS, weighted gene co-expression networks, and DEG analysis revealed 16 genes involved in root development under LN conditions (Ahmad et al., 2022). Zeng et al. used 304 rapeseed inbred lines to evaluate three related traits under LN stress, identifying 11 significant SNPs associated with LN tolerance in rapeseed. These findings provide valuable information for understanding the genetic control of LN tolerance during the seedling stage in rapeseed (Zeng et al., 2021).

To better understand the genetic control of LN tolerance traits in rapeseed, we performed a GWAS on 275 rapeseed accessions using 1,135,744 quality-controlled SNP markers derived from a high-density physical map. Using the linear mixed effects model (LMM), we identified 71 SNPs associated with five agronomic traits related to LN tolerance during the seedling stage. 20 candidate genes related to N metabolism within a 50 kb upstream and downstream region of the significant SNPs. This study provides a deeper understanding of the impact of LN stress on rapeseed seedling growth and offers valuable insights for developing molecular markers related to LN tolerance and screening LN-tolerant rapeseed materials.

Crop tolerance to LN stress is a quantitative trait regulated by multiple genes, with a clear interaction between the related genes and the external environment (Liu et al., 2021). This results in the instability of LN stress tolerance traits, which can be influenced by various factors. Therefore, precise phenotyping methods are key to obtaining accurate phenotypic data. The semi-automated hydroponic system used in this study can accommodate more than 3,000 rapeseed seedlings, including nutrient solution replenishment, nutrient solution circulation, and ventilation systems. The system maintains plants under relatively constant and uniform light, temperature, and nutrient conditions, ensuring that the obtained phenotypic data is relatively accurate and has already been applied in research (Zeng et al., 2021). Additionally, three parallel trials were conducted across three water tanks, and the multiple comparison analysis indicated no significant differences between the three trials ( Supplementary Table S5 ), suggesting that the phenotypic data obtained using this system is reliable.

Under LN stress, the results of this study revealed extensive genetic variation in five phenotypic traits (NL, SL, RL, SFW, and RFW), all of which are quantitative traits controlled by multiple genes. Correlation analysis indicated that, under LN conditions, NL showed a significant positive correlation with SFW, and SFW was also significantly positively correlated with RFW. These findings suggest that RFW and SFW may be key phenotypic characteristics associated with LN tolerance. These results are consistent with previous studies, highlighting the crucial role of root growth in LN tolerance (Freschet et al., 2021; Lv et al., 2021).

The results of GWAS are influenced by factors such as species, population size, target traits, marker density, and genetic diversity within the population (Zhu et al., 2008), with population size and genetic diversity playing major roles. It is also important to note that population structure and relatedness can significantly contribute to false positives in GWAS, thus greatly affecting the accuracy of the experiment (Kichaev et al., 2019). In this study, the linear mixed model (LMM) was employed, which is particularly effective in accounting for population structure and kinship, ensuring the robustness of the results and providing a reliable foundation for further analysis. In total, 71 significant SNP markers associated with LN tolerance were detected, with several SNPs located near genes related to plant N metabolism. Based on the results of GWAS, 20 candidate genes were identified within a 50kb upstream and downstream region of the significant SNPs, all of which are implicated in N compound transport. GO enrichment analysis indicated that these genes are primarily associated with the GO term GO:0071705 (N compound transport), suggesting their potential role in N metabolism. N metabolism is crucial for plant growth and development, particularly under LN stress, where the regulation of N absorption, transport, and assimilation is essential. These identified genes are involved in the absorption of N (e.g., NPF family genes), its transport (e.g., PUP family genes), and possibly its reutilization (e.g., certain ATPases and transporters). The expression of these genes may contribute to improving N use efficiency in canola under LN conditions, thereby enhancing its tolerance to N deficiency. It is important to note that the NRT1/PTR (NPF) family comprises transporters with diverse substrate specificities, including nitrate, peptides, and phytohormones (Corratgé-Faillie and Lacombe, 2017). For instance, BnaC01G0289500ZS, encoding Protein NRT1/PTR FAMILY 2.10 (NPF2.10), is part of the NPF family, which includes members known to participate in nitrate uptake and transport in plants (O’Brien et al., 2016) Another prominent group includes genes encoding proteins involved in autophagy, RNA metabolism, and DNA replication. For example, BnaA05G0042900ZS encodes Cysteine protease ATG4a (ATG4A), which plays a role in autophagy, a crucial process for nutrient recycling under N-limited conditions, as reported in members of the same family in barley (Chen, 2019) and rice (Fan et al., 2020). Similarly, BnaC04G0427700ZS encodes DEAD-box ATP-dependent RNA helicase 56 (RH15), potentially involved in stress-responsive RNA metabolism (Nawaz and Kang, 2017), while BnaC08G0205400ZS encodes Replication protein A 70 kDa DNA-binding subunit B, which might regulate DNA repair and replication processes under stress (Capel et al., 2020). Additionally, BnaC05G0316100ZS and BnaC05G0316200ZS encode Lysine histidine transporter-like proteins (AATL1), suggesting an enhanced role for amino acid transporters in N remobilization during stress (Parmar et al., 2019). BnaC09G0524600ZS encodes a WAT1-related protein (At5g13670), known to regulate cell wall modification and secondary metabolism, processes critical for maintaining root architecture and function under LN conditions (Chen et al., 2021). Further research on these genes will provide valuable insights into their mechanisms of action and offer a theoretical foundation for improving N use efficiency in B. napus.

There have been studies utilizing GWAS methods to detect candidate genes associated with traits under LN stress in rapeseed. In this study, no significant SNP was found to be physically close to a previously reported locus Zeng et al. (2021) used 304 rapeseed inbred lines to assess plant height (PH), shoot fresh mass (SFM), leaf length (LL), and leaf width (LW) under LN stress (N: 0.3 mM). Ahmad et al. (2022) measured five root morphology traits and eight biomass-related traits in 327 B. napus lines under LN stress (0.5 mM). Haelterman et al. (2024) screened 300 B. napus inbred lines under low (N: 0.1 mM) and high (N: 5 mM) nitrate conditions, identifying 319 significant SNPs associated with 15 biomass and root morphology traits. These SNPs were subsequently integrated into 80 candidate QTLs. To further assess the genetic contribution of the candidate genes, haplotype analysis was conducted using SNPs. For some genes, no polymorphic sites were detected, and thus haplotypes could not be defined. For the remaining genes, although haplotypes were identified, most showed one haplotype or other haplotypes included only a limited number of accessions (n ≤ 6), making statistical comparisons unreliable. The haplotype classifications for all candidate genes are provided in Supplementary Table S6 . Compared with previous studies, our GWAS identified a different set of SNPs associated with seedling-stage traits under LN stress. Unlike Ahmad et al (Ahmad et al., 2022), who integrated RNA-Seq to identify gene expression changes, our study focused on trait-based GWAS combined with haplotype analysis, revealing new candidate genes such as NPF2.10, ATG4a, and AATL1. Zeng et al. (2021) primarily investigated shoot traits and detected SNPs not overlapping with ours, possibly due to differences in germplasm, N levels, or sampling time points. Moreover, while Haelterman et al. (2024) concentrated on nitrate-specific root traits, our hydroponic system allowed for standardized seedling phenotyping, leading to the detection of 71 significant SNPs and 20 N-related candidate genes. To support the functional relevance of these genes, we examined their expression profiles using a previously obtained transcriptome dataset from B. napus seedlings subjected to LN treatment. Notably, BnaA09G0386000ZS was significantly upregulated under LN conditions, suggesting its involvement in nitrogen-responsive regulatory pathways. This result provides transcriptomic evidence that complements our association analysis and supports the potential role of BnaA09G0386000ZS in contributing to nitrogen use efficiency (NUE). However, it is important to note that the transcriptome dataset was derived from an independent experiment with a different nitrogen treatment regime and developmental stage, and thus the data were used only for reference. Further expression validation, such as qRT-PCR under comparable treatment conditions, and functional studies will be necessary to confirm the roles of this and other candidate genes in LN stress tolerance. These findings expand the current catalog of genetic variants contributing to N use efficiency in B. napus and offer new molecular targets for breeding.