These materials consisting of 505 natural accessions of B. napus (Supplementary Table 1) (Tang et al., 2021) were used as research objects under control, low salt-alkali, and high salt-alkali stress conditions at seedling and mature stages.

Eight spots covering the entire salt-alkali land in Inner Mongolia of China were randomly selected, and 3 soil samples (10–20-cm depth) were taken in within the area (50–100 m²) of each spot. After the dried samples were filtered with a 1-mm sieve, 100 g of the filtered samples were transferred to a 1-L plastic bottle, and 500-ml carbon dioxide distilled water was added into the bottle (Water: Soil = 5:1). Suction and filtration were immediately performed after shaking for 8 min. We put 50–100 ml of the test solution into an evaporating dish with known weight, and then evaporated the evaporating dish to dryness. We wiped the outside of the evaporating dish with a filter paper and then put it in an oven at 100–105°C for 4 h. After it was transferred to a desiccator to cool for 30 min, it was weighed with an analytical balance. The drying step was repeated until the difference of the weights between the two drying steps was < .0003 g. Finally, we obtained weight of the dried residue (W). After adding 15% hydrogen peroxide solution, the residue continued to be heated in a water bath to remove organic matter. Then, the total amount of soluble salt (S) was obtained after drying. Salt content of the salt-alkali land was calculated as follows: soluble salt (%) = S (g)/W (g) ^*100.

For determination of pH value of salt-alkali land, 10 g of the dry soil sample after being filtered by a 1-mm mesh was weighed and transferred into a 25-ml beaker. We added 10-ml.01 mol/L CaCl₂ solution into the sample and mixed it well (Water: Soil = 1:1). And then the sample was placed for 10 min. pH value was measured with a pH meter (PHS-25, Shanghai Leici Instrument, Inc., Shanghai, China).

Wuyuan County (105°12′-109°53′ E, 40°13′-42°28′ N) of Bayan Nur City is located in the western part of the Inner Mongolia Autonomous Region, which belongs to the hinterland of the Hetao Plain. The ingredients of saline-alkali land are dominated by sulfate, chloride, chloride-sulfate or sulfate-chloride in the Hetao Plain. We designed 3 treatment groups, including control (Salt:0.1–0.25%, pH: 7.5–8.), low salt-alkali (Salt:0.35–0.53%, pH: 8.−8.5), and high salt-alkali (Salt:0.64–1.05%, pH: 8.3–9.) conditions (Supplementary Figure 1). Each germplasm of 505 B. napus accessions was sown in 4 rows, and the space between each row was 20 cm. Each row contained 8 individual plants, and the space between each individual plant was 20 cm. At the same time, the second technical repetition was set up in the adjacent plots in different salt-alkali lands (Figure 1A). The sowing time is May 7, 2017. Before sowing, 225 kg/hm² of complex fertilizer (N:P₂O₅:K₂O, 25:10:16; total nutrients ≥ 51%; Shandong Stanley Fertilizer Co., Ltd.) was applied as base fertilizer to maintain normal growth and development of plants on May 7, 2017. In addition, 150 kg/hm² of complex fertilizer as supplemental fertilizer was also applied on July 1, 2017. The irrigation was conducted on May 7, 2017, May 17, 2017, and July 1, 2017, respectively (Supplementary Table 30). We measured its salt content and pH value in control, low salt-alkali, and high salt-alkali stress conditions at seedling and mature stages (Supplementary Table 2). Then, some agronomic or growth indicators were determined (Supplementary Table 3) at the seedling (July 9, 2017) and harvest stages (September 9, 2017), respectively.

Shihezi City (84°58′-86°24′ E, 43°26′-45°20′ N) is located in the northern part of Xinjiang Uygur Autonomous Region, which belongs to the Southern margin of Junggar Basin. The ingredients of saline-alkali land are dominated by soda, chloride, sulfate-chloride, chloride-sulfate, or sulfate in Shihezi City. In order to further verify and screen extreme materials under salt-alkali condition, we designed 2 treatment groups, including control soil (Salt:0.05–0.15%, pH: 6.8–7.3), and high salt-alkali soil (Salt:0.80–1.15%, pH: 8.5–10). Each germplasm of 505 B. napus accessions was sown in 4 rows, and the space between each row was 20 cm. Eight individual plants were grown in each row for one replicate, and the space between each individual plant was 20 cm. At the same time, another technical repetition was conducted in the adjacent plots. The sowing date is July 10, 2018. Before sowing, 225 kg/hm² of complex fertilizer (N:P₂O₅:K₂O, 25:10:16; total nutrients ≥ 51%; Shandong Stanley Fertilizer Co., Ltd.) was applied as base fertilizer to maintain normal growth and development of plants on July 10, 2017. In addition, 150 kg/hm² of complex fertilizer as supplemental fertilizer was also applied on August 4, 2017. The irrigation was conducted on July 10, 2017, August 4, 2017, and August 20, 2017, respectively (Supplementary Table 31). Then, we observed the response of the extreme materials at the seedling stage on September 19, 2018.

Some traits, such as plant height (PH), ground dry weight (GDW), and yield (Y), were measured (Supplementary Table 4). For GDW measurement, materials were harvested at seedling and mature stages, respectively, and dried at 85°C for at least 72 h before the samples were weighed. We collected the traits data from 5 biological replicates for each accession (Lv et al., 2016; Yu et al., 2017). The tolerance coefficient (TC) was calculated as the ratio between the agronomic and growth indexes under salt-alkali stress and control condition (Supplementary Table 5).

Combined with the response of these extreme materials to salt-alkali land in Xinjiang Uygur and Inner Mongolia Autonomous Region, the trait values of each variety were ranked, and then the top 10% most tolerant materials and the bottom 10% most sensitive materials were selected as extreme materials, respectively. Finally, a number of reliable extreme materials were screened out through Venn analysis between the two developmental stages.

The statistically significant differences between different data sets were evaluated by the Student's T-test (P < 0.05). We finally calculated the average of all traits with 5 replicates per accession. We used the scale package of R language (https://www.r-project.org/) to analyze data standardization and R language stat.desc packages to analyze phenotypic variation. The correlation coefficient was calculated using the corr.test of psych package (https://www.r-project.org/), and figures were plotted using the corrplot package of R language (https://www.r-project.org/).

In order to determine the variance components, we used the mixed-effects variance analysis of the avo package by R language (https://www.r-project.org/), including genetic effects (G_effect), treatment effects (E_effect), and interaction effects between genetic and treatment effects (G&E effect). Heritability is a commonly used reference index in crop genetic breeding. The broad sense heritability (H₂b) of all traits is calculated as follows: H₂b = σG2/(σG2 + σGE2/n + σe2/nr) where σG2 was the genotype variance, σe2was the error variance, and σGE2 was interaction between genetic and environmental effects. The estimates of H2b were analyzed by the AVOVA for ANOVA using the lmer function of the lme4 package by R environment (https://www.r-project.org/) (Chen et al., 2014; Li et al., 2020; Zhang et al., 2022).

Our laboratory collected 505 Brassica napus germplasm resources to construct a re-sequencing association population (Tang et al., 2021). The high-quality clean reads data were obtained by the BWA (v0.75) software (https://github.com/lh3/bwa) (Li and Durbin, 2009). The reference genome is derived from “Brassica v4.1” (“Darmor-bzh”) genome (http://www.Genoscope.cns.fr/brassicanapus/data/). We also performed GWAS with 7,671,951 SNPs (MAF >0.05) using GEMMA linear mixed models (Zhou and Stephens, 2012). It is worth noting that we divided the gene components into 300 kb bins. If there are multiple suggestive SNPs in the same bin, it is considered to be located at the same site, and the suggestive SNP with the smallest P-value is regarded as the lead SNP.

The P-value should be 1/marker number, 0.05/marker number or 0.01/marker number, which were named suggestive P-Value, significant P-value, and high-significant P-Value. High-density SNPs markers can improve the accuracy of candidate gene positioning, but there is often a certain linkage effect between adjacent SNP markers, so we often combined inter-linked SNP markers for one molecular marker. Therefore, we turned the final merged SNP markers, which were called effective SNP markers. The number of SNP markers used to calculate the P-value was far less than the total SNPs markers. Previous studies have shown that B. napus can be divided into two groups: spring and semi winter. In addition, semi winter was divided into Semi Winter 1 and Semi Winter 2 (Tang et al., 2021). The average number of SNP markers for different sub-species is about 7.38 million (Oberved_number) and 1.08 million (Effective_number), respectively. Effective SNPs account for about 14% of total SNPs. The average threshold is about 6 [~1. e-06, -log₁₀ (P-value)] between different subspecies. This value was previously used by our laboratory in GWAS analysis of the same 505 B. napus accessions (Tang et al., 2021). Therefore, this value is usually considered as a recognized threshold in the 505 B. napus accessions. The suggestive and significant P-value threshold of the entire population was set to be 1. e-06. All SNPs exceeded the significance criterion were assessed for location of candidate genes. If coding regions were present, the potential impact on protein of each SNP was subsequently determined using the “Allele finder” facility Gene Ontology analysis. The candidate genes with differential gene expression ratios between the control treatment and stress treatment (R ≥ 2 or ≤ 0.5) (https://bigd.big.ac.cn/) (Zhang et al., 2019) were identified within 200 kb upstream or downstream of the lead SNP. We finally determined the protein function corresponding to the candidate gene in Brassica napus based on the homologous gene information on Arabidopsis (Chalhoub et al., 2014).

In order to evaluate the response of the 505 B. napus accessions to salt-alkali stress, we determined the related traits at seedling and mature stages under control (Salt:0.1–0.25%, pH: 7.5–8., CK), low salt-alkali (Salt:0.35–0.53%, pH: 8.−8.5, L), and high salt-alkali (Salt:0.64–1.05%, pH: 8.3–9., H) stress conditions, respectively (Figure 1A; Supplementary Figure 1). Plant height (S_PH_CK, S_PH_L, and S_PH_H) and aboveground dry weight (S_GDW_CK, S_GDW_L, and S_GDW_H) were measured at the seedling stage (Supplementary Table 4). In addition, the traits, such as plant height (M_PH_CK, M_PH_L, and M_PH_H), aboveground dry weight (M_GDW_CK, M_GDW _L, and M_GDW_H) and yield (M_Y_CK, M_Y_L, and M_Y_H) were obtained at the mature stage (Supplementary Table 4). We found that salt-alkali stress significantly inhibited the growth, development, and even yield of crops (Supplementary Figure 2).

Moreover, in order to better reflect the salt-alkali stress response, we also focused on the tolerance coefficients (TCs) of traits, which were calculated from the ratio of all traits between salt-alkali stress and control condition, denoted by the suffix type “trait_R1” under low salt-alkali stress conditions and “trait_R2” under high-salt-alkali stress conditions (Supplementary Table 5) (Guo et al., 2018). Descriptive statistics provided a good description of the phenotypic variation on 505 B. napus accessions under control condition and salt-alkali stress conditions. “Mean,” ‘Max,” “Min,” “SD,” “SE,” and “CV” were used to calculate the descriptive statistics of the average value of all traits (Supplementary Table 6). We found that the coefficient of variation (CV, %) of most traits reached more than 20%. Most traits of the 505 B. napus accessions show significant variations between control condition and salt-alkali stress conditions. Interestingly, large variations in plant height (M_PH), aboveground dry weight (M_GDW), and yield (M_Y) were observed among the accessions under salt-alkali stress conditions at the mature stage, whereas plant height (S_PH) and ground dry weight (S_GDW) showed small variations under the control condition at the seedling stage (Supplementary Table 6). These results suggest that the developmental stage has a big effect on the phenotypic variation under salt-alkali stress conditions.

After trait values of each strain were ranked, the top 10% most tolerant and bottom 10% most sensitive materials were selected as extreme materials, respectively. A number of reliable extreme materials have been screened out through Venn analysis between the two developmental stages (Figure 1B; Supplementary Figure 3). Combining the response of these extreme materials on salt-alkali land in Xinjiang Uygur and Inner Mongolia Autonomous Region of China, we finally identified that “X182,” “X252,” “X148,” “X200,” “X278,” and “X398” were as resistant varieties, and “X188,” “X568,” “X648,” “X552,” and “X756” were as sensitive varieties (Figure 1C; Supplementary Figure 4) under the two different salt-alkali stress environments. These extreme materials are valuable germplasm resources for the genetic improvement of salt-alkali tolerance in B. napus.

We also performed normal distribution detection on the original absolute value (Supplementary Table 7) and TCs of all traits for GWAS (Supplementary Table 8), respectively. The absolute value of all traits conformed to normal or lognormal distributions and the TCs of all traits for GWAS also belonged to normal distributions. The lognormal distribution of S_PH and S_GDW may be attributed to the significant phenotypic variations among 505 B. napus accessions. Correlation analysis for all traits at seedling and mature stages was calculated and presented in Supplementary Table 9. The two traits at the seedling stage and the three traits at the mature stage showed significant correlations with “r” under control, low, and high salt-alkali stress conditions, which ranged from 0.021 to 0.982 between each other (Figure 2A; Supplementary Figures 5, 6; Supplementary Table 9). It is worth noting that the S_PH as strongly correlated with the S_GDW (r ≥ 0.5) and the M_GDW was strongly correlated with the M_Y (r ≥ 0.6), whereas the M_PH was weakly correlated with the M_GDW and M_Y (r ≥ 0.16) under control, low, and high salt-alkali stress conditions. Moreover, the strong positive correlations among control, low, and high salt-alkali stress conditions were observed for almost all the traits, and their correlations were from 0.497 to 0.869 while the correlation of all traits between seedling and mature stages was found to be relatively low. These results indicated that the genetic mechanism of salt-alkali stress response at the two different developmental stages was likely to be different.

Heritability is a key factor of decision-making in crop breeding, and it is essential for the effective design of breeding programs (Chen et al., 2014). Therefore, we also calculated the broad heritability (H₂b) of all traits at seedling and mature stages. As shown in Supplementary Table 10, the variations in 505 B. napus accessions were largely determined by genotype and heritability estimates, which ranged from 0.759 to 0.889 (Figure 2B; Supplementary Table 10). Moreover, the genetic effect and treatment effect (G_effect and E_effect, P ≤ 0.05) could significantly represent the genotype effect and environment effect using the AVOVA by R environment (Figure 2B; Supplementary Table 10). Except for the S_GDW, the G&E_effects of other traits were found to be extremely significant. These results indicated that the response to salt-alkali stress is dynamic, which is largely determined by complex multiple genetic effects, and the mechanism of adaption to salt-alkali stress in B. napus is complicated.

The 15 absolute value and 10 TCs (Supplementary Tables 4, 5) were performed using the GEMMA linear mixed models for GWAS of all traits (Figures 3A,B; Supplementary Figures 7A–C) (Zhou and Stephens, 2012). About 3,536 (~0.046% of in total) strongly associated SNPs (–log₁₀ (P-value) ≥ 6.) were chosen to further map the QTLs with the significant effect SNPs (Supplementary Table 11). A total of 1,170 SNPs (CK_SNPs), 1,207 SNPs (L_SNPs), and 762 SNPs (H_SNPs) were identified for absolute value of all traits under control, low salt-alkali, and high salt-alkali conditions, respectively. Moreover, 367 SNPs (R1_SNPs) and 30 SNPs (R2_SNPs) were identified for TCs under low salt-alkali and high salt-alkali conditions, respectively (Figure 3C; Supplementary Table 12). Based on these results, we obtained different numbers of specific and co-localized SNPs for different traits under different salt-alkali conditions. Among these SNPs, 740 specialized SNPs were detected, including 169 CK_SNPs, 136 L_SNPs, 49 H_SNPs, 364 R1_SNPs, and 22 R2_SNPs, respectively. About 366 co-localized SNPs (CK_SNPs and L_SNPs) and 9 co-localized SNPs (H_SNPs and R2_SNPs) were located on Chromosome A03, respectively. We also identified 70 (L_SNPs and H_SNPs), 632 SNPs (CK_SNPs, L_SNPs, and H_SNPs) and 3 co-localized SNPs (CK_SNPs, L_SNPs, H_SNPs, and R1_SNPs) located on Chromosome A10, respectively. Additionally, 20 co-localized SNPs on Chromosome A10 were simultaneously identified by M_GDW_L, M_PH_CK, M_PH_H, M_PH_L, and M_Y_L; 8 co-localized SNPs on Chromosome A03 were simultaneously identified by M_GDW_H, M_GDW_R2, M_Y_H, and M_Y_R2 (Figure 3C; Suppleentary Table 12). We finally obtained 9 main effect and co-localized QTLs related to salt-alkali stress response (Supplementary Table 13).

In order to identify the candidate genes associated with different traits under different salt-alkali stress conditions, we finally focused on the SNPs identified by GWAS for TCs of all traits, such as R1_SNPs (Figure 3C; Supplementary Tables 12, 14), R2_SNPs (Figure 3C; Supplementary Tables 12, 15), H_SNPs & R2_SNPs (Figure 3C; Supplementary Tables 12, 16), and CK_SNPs & L_SNPs & H_SNPs & R1_SNPs (Figure 3C; Supplementary Tables 12, 17). By analyzing the differential gene expression between control and stress conditions (R ≥ 2 or ≤ 0.5) (https://bigd.big.ac.cn/) (Zhang et al., 2019), 222 genes were identified within 200 kb upstream or downstream of the lead SNPs (Supplementary Table 18), such as 36 genes located on ChrA03 from 10,625,336 to 10,741,817 bp for M_Y_R1 and from 10,625,336 to 10,741,817 bp for M_GDW_R1 (Supplementary Figures 8A–C; Supplementary Table 14) and 15 candidate genes located on ChrA03 from 8,751,415 to 9,293,723 bp for M_Y_R2 and from 9,292,526 to 9,293,723 bp for M_GDW_R2 (Supplementary Figures 9A–C). GO analysis revealed that most genes were associated with the cellular process, the metabolic process, biological regulation, and response to stimulus (Supplementary Figure 10). Moreover, we also found that the most of these candidate genes were significantly differentially expressed under salt stress according to the RNA-seq data (PRJCA008229, https://ngdc.cncb.ac.cn/omix/) (Zhang et al., 2022). The relative expression of some candidate genes under normal and salt conditions was taken as an example (Figure 3D; Supplementary Table 19), which was consistent with the previous study (Zhang et al., 2019), showing the accuracy of candidate gene prediction. These results suggest that the accuracy of candidate gene prediction and the tolerance of B. napus to salt-alkali stress are regulated by a complex genetic mechanism.

A total of 36 genes within the QTL region on ChrA02 from 6,927,771 to 7,018,686 bp, which was named as to be qSAT.A02.1 within 200 kb upstream and downstream of the candidate SNPs of M_PH_R1 (Figures 4A,B; Supplementary Tables 14, 18), and the expression of some genes were found to be significantly induced by many abiotic stresses (https://bigd.big.ac.cn/) (Zhang et al., 2019). Among these genes, we finally focused on the BnaA02g12820D, whose functions related to salt-alkali stress response have not been reported. BnaA02g12820D named as BnABA4 is a homologous gene of Arabidopsis ABA4, which is involved in the photoprotection of PSII and plays a vital role in neoxanthin production in abscisic acid biosynthesis. Based on our RNA-seq data, we found that BnABA4 was significantly induced by salt stress (Figure 3D). Moreover, the results from the expression pattern showed that BnABA4 was highly expressed in siliques and stems (Figure 4C).

BnABA4 consisting of five introns and six exons shows rich SNP variations, leading to synonymous and nonsense variation (Supplementary Table 20). The haplotypes of BnABA4 were grouped into 6 haplotypes, including hap.A, hap.B, hap.C, hap.D, hap.E, and hap.F types (Supplementary Table 21). Both the absolute and relative values of plant height (M_PH) of hap.B type at the mature stage were significantly higher than other haplotypes under the control, low or high salt-alkali stress condition. We thus considered the accessions with hap.B types of BnABA4 as extremely resistant haplotypes under salt-alkali stress conditions (Figures 4D–H). On the contrary, both the absolute and relative values of M_PH of hap.D type at the mature stage were significantly less than other haplotypes under control, low or high salt-alkali stress condition (Figures 4D–H). We thus considered the accessions with the hap.D types of BnABA4 as extremely sensitive haplotypes under salt-alkali stress conditions. Therefore, BnABA4 was considered as a candidate gene on the QTL qSAT.A02.1 for salt-alkali stress response, showing a rich SNP variation in 505 B. napus accessions. The identified extreme haplotypes related to salt-alkali tolerance would be useful for breeding salt-alkali-tolerant B. napus in the future.

There were 22 genes within the QTL region on ChrA02 from 7,918,159 to 8,015,641 bp, which was named as to be qSAT.A02.2 within 200 kb upstream and downstream of the candidate SNPs of M_PH_R1 (Figures 5A,B; Supplementary Tables 14, 18), and the expressions of some genes were found to be significantly induced by many stresses (https://bigd.big.ac.cn/) (Zhang et al., 2019). Among these genes, we finally focused on the BnaA02g13920D and BnaA02g14320D, whose functions related to salt-alkali stress have not been reported. BnaA02g13920D and BnaA02g14320D were named as BnBBX14 and BnVTI12, which is the area homologous gene of Arabidopsis BBX14 and VTI12, respectively. BnBBX14 encodes a transcription factor of B-box zinc finger type with a CCT domain, which normally localizes to the transgolgi network. However, the BnVTI12 belongs to the VTI gene family, which normally localizes to plasma membrane. Based on our RNA-seq data, we also found that BnBBX14 and BnVTI12 were significantly induced by salt stress (Figure 3D; Supplementary Table 19). Additionally, BnBBX14 and BnVTI12 were found to be expressed in various tissues. The expression level of BnBBX14 was relatively high (Figure 5C) in siliques and stems. However, the expression level of BnVTI12 in flowers was higher than other tissues (Supplementary Figure 11A).

BnBBX14 consisting of one intron and two exons shows rich SNP sequence variations, leading to synonymous, missense, and disruptive inframe deletion variation (Supplementary Table 22). The haplotypes of BnBBX14 were grouped into 6 haplotypes, including hap.A, hap.B, hap.C, hap.D, hap.E, and hap.F types (Supplementary Table 23). Except for hap.A types accessions of plant height (M_PH) under the control condition, both the absolute and relative values of M_PH of the hap.A type at the mature stage were significantly higher than other haplotypes under the control, low or high salt-alkali stress condition (Figures 5D–H). We thus considered the accessions with hap.A type of BnBBX14 as extremely resistant materials under salt-alkali stress conditions. BnVTI12 consisting of four introns and five exons exhibits rich SNP variations, leading to synonymous and splice region variants (Supplementary Table 24). The haplotypes of BnVTI12 were grouped into 5 haplotypes, including hap.A, hap.B, hap.C, hap.D, and hap.E types (Supplementary Table 25). Both the absolute and relative values of plant height (M_PH) of the hap.E type at the mature stage were significantly higher than other haplotypes. We thus considered the accessions with hap.E type of BnVTI12 as extremely resistant haplotypes under salt-alkali stress conditions. Except for M_PH trait absolute value of hap.A types accessions under control, low, and high salt-alkali stress condition, the relative values of hap.A type under low and high salt-alkali stress conditions were significantly higher than other haplotypes. We thus considered the accessions with hap.A type of BnVTI12 as extremely resistant haplotypes under salt-alkali stress conditions. Therefore, BnBBX14 and BnVTI12 were considered as two candidate genes on qSAT.A02.2 for salt-alkali stress response, showing a rich SNP variation among 505 B. napus accessions. The identified extreme haplotypes related to salt-alkali tolerance would be useful for breeding salt-alkali-tolerant B. napus in the future.

A total of 27 genes within the QTL region were identified on ChrA03 from 5,751,090 to 6,079,821 bp, which was named as to be qSAT.A03.1 within 200 kb upstream and downstream of the candidate SNPs of M_PH_R1 (Figures 6A,B; Supplementary Table 14), and the expressions of some genes was found to be significantly induced by many stresses (https://bigd.big.ac.cn/) (Zhang et al., 2019). Among these genes, we finally focused on the BnaA03g12450D and BnaA03g12800D, whose functions related to salt-alkali stress response have not been reported. BnaA03g12450D and BnaA03g12800D were named as BnPYL8 and BnCRR1, which are the homologous gene of Arabidopsis PYL8/RCAR3 and CRR1, respectively. BnPYL8 encoding an ABA receptor interacts with protein phosphatase 2Cs ABI1 and ABI2 in ABA signaling, and BnCRR1 is involved in electron transfer between PSI and PSII and plays a vital role in the biogenesis or stabilization of the NAD(P)H dehydrogenase. Based on our RNA-seq data, we also found that BnPYL8 and BnCRR1 were significantly induced by salt stress (Figure 3D; Supplementary Table 19). In addition, BnPYL8 and BnCRR1 were found to be expressed in various tissues. The expression level of BnPYL8 was relatively high in roots (Figure 6C), whereas the expression level of BnCRR1 in the silique was higher than that in other tissues (Supplementary Figure 12A).

BnPYL8 consisting of three introns and four exons exhibits rich SNP variations, leading to synonymous, missense, and splice region variation (Supplementary Table 26). The haplotypes of BnPYL8 were grouped into 6 haplotypes, including hap.A, hap.B, hap.C, hap.D, hap.E, and hap.F types (Supplementary Table 27). Both the absolute and relative values of plant height (M_PH) at the mature stage in the hap.A, hap.B, and hap.C type accessions were significantly higher than other haplotypes, while those in the hap.E type accessions were significantly less than other haplotypes under the control, low or high salt-alkali stress condition. Both the absolute and relative values of plant height (M_PH) at the mature stage in the hap.A, hap.B, and hap.C type accessions were significantly higher than other haplotypes, while those in the hap.E type accessions were significantly less than other haplotypes under the control, low or high salt-alkali stress condition (Figures 6D–H). We could consider the accessions with hap.A, hap.B, and hap.C types of BnPYL8 as extremely resistant haplotypes and the accessions with hap.E type of BnPYL8 as extremely sensitive haplotypes under salt-alkali stress conditions. In addition, BnCRR1 consisting of nine introns and ten exons shows rich SNP variations, leading to synonymous, missense, and splice region variation (Supplementary Table 28). The haplotypes of BnCRR1 were grouped into 6 haplotypes, including hap.A, hap.B, hap.C, hap.D, hap.E, and hap.F types (Supplementary Table 29). Except for M_PH trait absolute value of hap.F types accessions under high salt-alkali stress conditions, both the absolute and relative values in the hap.F type accessions were significantly less than other haplotypes under control, low, and high salt-alkali stress conditions (Supplementary Figures 12B–F). We thus considered the accessions with hap.F type of BnCRR1 as extremely sensitive haplotypes under salt-alkali stress conditions. On the contrary, both the absolute and relative values of plant height in the hap.A, hap.D, and hap.E type accessions were significantly higher than other haplotypes under control, low, and high salt-alkali stress conditions. We thus considered the accessions with hap.A, hap.D, and hap.E types of BnCRR1 as extremely resistant haplotypes under salt-alkali stress conditions. According to the above results, BnPYL8 and BnCRR1 were considered as two candidate genes on qSAT.A03.1 for salt-alkali stress response, which showed a rich SNP variation among 505 B. napus accessions. The identified extreme haplotypes associated with salt-alkali tolerance would be helpful for breeding salt-alkali-tolerant B. napus in the future.