Mapping of Two Major QTLs Controlling Flowering Time in Brassica napus Using a High-Density Genetic Map

Research on the flowering habit of rapeseed is important for the selection of varieties adapted to specific ecological environments. Here, quantitative trait loci (QTL) for the days-to-flowering trait were identified using a doubled haploid population of 178 lines derived from a cross between the winter type SGDH284 and the semi-winter type 158A. A linkage map encompassing 3268.01 cM was constructed using 2777 bin markers obtained from next-generation sequencing. The preliminary mapping results revealed 56 QTLs for the days to flowering in the six replicates in the three environments. Twelve consensus QTLs were identified by a QTL meta-analysis, two of which (cqDTF-C02 and cqDTF-C06) were designated as major QTLs. Based on the micro-collinearity of the target regions between B. napus and Arabidopsis, four genes possibly related to flowering time were identified in the cqDTF-C02 interval, and only one gene possibly related to flowering time was identified in the cqDTF-C06 interval. A tightly linked insertion–deletion marker for the cqFT-C02 locus was developed. These findings will aid the breeding of early maturing B. napus varieties.


Introduction
Flowering is an important growth stage of higher plants, as it coincides with the transition from vegetative growth to reproductive growth. This transformation process determines the timing of flowering, which is important for ensuring high yields and the quality of crops [1]. Several studies of the model plant Arabidopsis have been conducted to clarify the genetic architecture of flowering time regulation, and these studies have resulted in the identification of at least 300 genes involved in flowering initiation. Environmental plasticity in this trait was also described [2], and the analysis of several Arabidopsis flowering time mutants by Koornneef et al. (2004) greatly contributed to our understanding of flowering initiation [3]. Six flowering time pathways have been proposed based on plant genetics and molecular biology studies in the model species Arabidopsis thaliana: the vernalization pathway, the photoperiod pathway, the gibberellin pathway, the autonomous pathway, the ambient temperature pathway, and the endogenous pathway [4,5]. This research on Arabidopsis has provided a solid foundation for dissecting the genetic and molecular bases of flowering time control in crops, especially for Brassicaceae crops, which show a high degree of evolutionary conservation in their molecular pathways with Arabidopsis [6].
Rapeseed (Brassica napus L.) is an important member of the family Brassicaceae and a major source of high-quality vegetable oil and protein-rich animal feed worldwide [7]. The allotetraploid B. napus was formed on the Mediterranean coast through a few independent natural hybridization events between the ancestral diploids Brassica rapa and Brassica oleracea, as well as chromosome doubling approximately 7500 years ago [8]. Natural variation in flowering time has promoted the ecological adaptability of rapeseed and facilitated its spread to subtropical and temperate regions. Vernalization is one of the many factors affecting flowering time in B. napus. Rapeseed germplasm can be divided into three ecotypes according to the relative importance of a vernalization response for normal flowering: the spring type, semi-winter type, and winter type [9][10][11]. Spring-type accessions flower normally without vernalization treatment, semi-winter accessions require moderate vernalization to promote flowering, and winter-type accessions depend highly on prolonged low temperatures for flowering [10,11]. Even within ecotypes, flowering time can vary to mediate adaptation to different climates or farming systems. Therefore, achieving an optimal flowering time is an important objective for rapeseed breeders.
Several flowering time genes and loci have been identified in B. napus over the past decades. There are two strategies that can be used to clone flowering time genes. The first strategy is homology cloning, which is based on the conservation of flowering time genes between Arabidopsis and Brassica. Several FLC (FLOWERING LOCUS C) paralogs in B. napus (BnFLC1-5) have been characterized by complementation tests in transgenic Arabidopsis using this approach [12]. Zou et al. (2012) also used this strategy to clone nine FLC homologs from the rapeseed genome and map them to six of 19 chromosomes [13]. Yin et al. (2021) cloned BnaFLC.A10 and BnaFLC.A2 in spring, winter, and semi-winter rapeseed and found that these two genes are required for the establishment of winter rapeseed crop phenotypes [14]. The second strategy involves using different populations to locate quantitative trait loci (QTLs) at the flowering stage in B. napus via forward genetic approaches. Parental linkage mapping analysis has been conducted in many of these studies. Long et al. (2007) showed that a major flowering time QTL, qFT10-4, which explained more than 50% of the phenotypic variation in flowering time during the spring environment, was located on chromosome A10 in the TN double haploid (DH) population [15]. Wang et al. (2009) identified six BnFT paralogs from the genome of B. napus, and three of them were associated with two major QTL clusters for flowering time [16]. Raman et al. (2013) constructed a DH population and found at least 20 QTLs in this population by QTL testing at the flowering stage [17]. Chen et al. (2018) successfully cloned the flowering time genes BnFLC.A2 and BnFLC.C2 using two pairs of near-isogenic lines [18]. Tudor et al. (2020) used the F 2 population and QTL-seq to locate two flowering genes, BnaFLC.A02 and BnaFT.A02, in the 10-Mb interval of the A02 chromosome [19]. Song et al. (2021) used the DH line to locate two major QTL sites for flowering time and identified four flowering-related candidate genes using transcriptome analysis, including EMF1, NF-YA1, HAP2B, and COL9 [20]. Xu et al. (2020) discovered a new QTL for flowering time on chromosome A09 using F 2 and its F 2:3 line [21]. Fang et al. (2022) used the DH line to analyze lines with different combinations of the three flowering genes BnFLC.A2, BnFLC.C2, and BnFLC.A3b and concluded that the potential heterosis of flowering genes can regulate the development of B. napus, early flowering (maturity), and high yield, which can facilitate the early flowering of crops [22]. Association mapping analyses of natural populations have been conducted in recent years. Many single-nucleotide polymorphism (SNP) loci have been found to be associated with flowering time genes, such as BnaFRI.A03, BnVIN3-C03, BnaFLC.A02, BnaFLC.A03, BnaFLC.C02, and BnaFLC.A10 et al. These SNP markers are significantly related to the flowering period and can be used for the wholegenome selection breeding of B. napus [23][24][25][26][27][28][29] [27].
In this study, we used a high-density genetic linkage map to identify two major flowering time QTLs on chromosomes C02 (cqDTF-C02) and C06 (cqDTF-C06), which are two environmentally stable QTLs. We also predicted candidate genes for these QTLs and developed a closely linked insertion-deletion (InDel) marker for the cqDTF-C02 locus. The results of this study enhance our understanding of the molecular basis for the regulation of flowering time and provide molecular markers that could be useful for the breeding of flowering time traits in rapeseed.

Phenotyping of Two Parents and the 158A-SGDH Population
The DH lines 158A and SGDH284 were used in this study (Figure 1a,b) in the semiwinter rape environment (Fengyang, Anhui). The difference in flowering time between these two materials was approximately 7-13 days, and this difference was significant ( Table 1). We crossed the early flowering 158A (female parent) and the late flowering SGDH284 (male parent) to obtain the F 1 generation. The anthers of the F 1 generation were collected for microspore culture and doubled with colchicine to obtain the DH population, which was named 158A-SGDH in this study. Days-to-flowering (DTF) phenotypic data were collected from the 178 DH lines in three experiments conducted over three years (Tables 1 and S1).
In this study, we used a high-density genetic linkage map to identify two major flowering time QTLs on chromosomes C02 (cqDTF-C02) and C06 (cqDTF-C06), which are two environmentally stable QTLs. We also predicted candidate genes for these QTLs and developed a closely linked insertion-deletion (InDel) marker for the cqDTF-C02 locus. The results of this study enhance our understanding of the molecular basis for the regulation of flowering time and provide molecular markers that could be useful for the breeding of flowering time traits in rapeseed.

Phenotyping of Two Parents and the 158A-SGDH Population
The DH lines 158A and SGDH284 were used in this study (Figure 1a,b) in the semiwinter rape environment (Fengyang, Anhui). The difference in flowering time between these two materials was approximately 7-13 days, and this difference was significant (Table 1). We crossed the early flowering 158A (female parent) and the late flowering SGDH284 (male parent) to obtain the F1 generation. The anthers of the F1 generation were collected for microspore culture and doubled with colchicine to obtain the DH population, which was named 158A-SGDH in this study. Days-to-flowering (DTF) phenotypic data were collected from the 178 DH lines in three experiments conducted over three years ( Table 1 and Table S1).  The flowering phenotype of the 158A-SGDH population showed a normal distribution and ranged from 145 to 171 days for the six replicates over three years, indicating that the DTF phenotype is a quantitative trait controlled by multiple genes (Figures 2 and S1). Moreover, the DTF of the 158A-SGDH population for the period of 2019-2020 was significantly earlier than that for the period of 2018-2019, which might have stemmed from the warmer ambient temperature in the winter of 2019 ( Figure S1). There was a significant positive correlation in the DTF phenotype among the six replicates in the three environments (r 2 = 0.78-0.95, p < 0.001) ( Figure 2). The broad-sense heritability of DTF was 91.52%. The significance level by t test. *** p < 0.001. a Mean ± SEM, SEM means standard error of mean.
The flowering phenotype of the 158A-SGDH population showed a normal distribution and ranged from 145 to 171 days for the six replicates over three years, indicating that the DTF phenotype is a quantitative trait controlled by multiple genes (Figure 2 and Figure S1). Moreover, the DTF of the 158A-SGDH population for the period of 2019-2020 was significantly earlier than that for the period of 2018-2019, which might have stemmed from the warmer ambient temperature in the winter of 2019 ( Figure S1). There was a significant positive correlation in the DTF phenotype among the six replicates in the three environments (r 2 = 0.78-0.95, p < 0.001) ( Figure 2). The broad-sense heritability of DTF was 91.52%.

High-Density Genetic Linkage Map Construction
To construct a high-density linkage map, next-generation sequencing (NGS) was performed on a total of 178 individuals from the DH population, and a total of 360.87 Gb clean data were generated (Table S2). The average quality score of Q20 was over 98.69%, the GC content was approximately 39.09% (Table S2), and the average sequencing depth was 1.70× for every individual (Table S2). For the parents 158A and SGDH284, 5,064,588,459 bp and 5,362,611,753 bp of clean data were detected, and the sequencing depths were 3.67× and 4.19×, respectively (Table S2).
According to the NGS data of the two parental lines and 178 individuals, a total of 8,780,700 SNPs were identified, and 946,690 SNPs were used for map construction. A total of 13,395 bin markers were obtained based on an analysis of every 15 SNP, each of which was considered a bin [30], and these were subsequently used for genotyping in 178 individuals. Finally, a total of 2777 bin markers were anchored to 19 linkage groups (Table S3). The total length of the linkage groups was 3268.01 cM, and the average distance between the markers was 1.18 cM (Figure 3, Tables S3 and S4). Of the 19 linkage groups, the longest was A01 (273.90 cM), followed by A09 (259.02 cM), and the shortest was A06 (86.89 cM). The linkage group with the largest number of bin markers was A01 (254 bins), followed by A04 (243 bins), and linkage group C07 had the lowest number of bins (56 bins) (Tables S3 and S4).
8,780,700 SNPs were identified, and 946,690 SNPs were used for map construction. A total of 13,395 bin markers were obtained based on an analysis of every 15 SNP, each of which was considered a bin [30], and these were subsequently used for genotyping in 178 individuals. Finally, a total of 2777 bin markers were anchored to 19 linkage groups (Table  S3). The total length of the linkage groups was 3,268.01 cM, and the average distance between the markers was 1.18 cM (Figure 3, Tables S3 and S4). Of the 19 linkage groups, the longest was A01 (273.90 cM), followed by A09 (259.02 cM), and the shortest was A06 (86.89 cM). The linkage group with the largest number of bin markers was A01 (254 bins), followed by A04 (243 bins), and linkage group C07 had the lowest number of bins (56 bins) (Tables S3 and S4).
To assess the quality of the genetic map, we analyzed the correlation between the bin markers in the genetic map and their positions on the physical map ( Figure S2, Tables S5, and S6). The mean value of the Pearson correlation coefficients between the genetic map and the physical map was 93.8%, indicating high collinearity between them ( Figure S2 and Table S7). The above results indicate the robustness and accuracy of the constructed genetic map.  To assess the quality of the genetic map, we analyzed the correlation between the bin markers in the genetic map and their positions on the physical map ( Figure S2, Tables S5 and S6). The mean value of the Pearson correlation coefficients between the genetic map and the physical map was 93.8%, indicating high collinearity between them ( Figure S2 and Table S7). The above results indicate the robustness and accuracy of the constructed genetic map.

Candidate Gene Prediction of the Two Major Flowering time QTL Regions
According to the B. napus ZS11 reference genome (http://yanglab.hzau.edu.cn/BnIR/ accessed on 10 March 2022), 195 genes were predicted in the 1.03-Mb cqDTF-C02 region. Based on the micro-collinearity of the target region between B. napus and Arabidopsis, we detected four genes that might be associated with flowering time: BnaC02G0032100ZS (AT5G08370/AGAL2), BnaC02G0038900ZS (AT5G10130/DFC), BnaC02G0039100ZS (AT5G10140/FLC), and BnaC02G0046300ZS (AT5G11530/EMF1) ( Figure 5, Table S9). In addition to the above four genes that might be related to flowering time, 29 genes with unknown functions were identified in this segment (Table S9) The FPKM values of BnaC02G0038900ZS, BnaC02G0039100ZS, and BnaC02G0046300ZS were all less than 5 in the 158A parent; however, the FPKM values of these genes were higher in the SGDH284 parent (20.64, 277.51, and 31.31, respectively) (Table S11).
According to the ZS11 reference genome, 65 genes were identified in the 530-kb interval in the cqDTF-C06 region. We found that one of these genes might be related to flowering time, namely BnaC06G0229100ZS (AT3G61250/AtMYB17) ( Figure 5, Table S10). The FPKM value of BnaC06G0229100ZS was less than 5 in both the 158A and SGDH284 parents (Table S11). In addition, 15 genes with unknown functions were identified in this QTL region (Table S10).

InDel Markers Linked to Flowering Time in the Target Intervals
InDel markers for cqDTF-C02 and cqDTF-C06 were developed based on the Illumina sequencing data of 158A and SGDH284 using the ZS11 reference genome. We developed nine InDel markers for the cqDTF-C02 locus, and one of the primers with parental polymorphism and a strong band pattern was C2-5. We analyzed a small population of 10 extremely early and 10 extremely late flowering plants with the C2-5 marker and found that the marker was associated with flowering time ( Figure S3). The flowering times of six replicates of the 158A-SGDH population were studied using the InDel marker C2-5 at the cqDTF-C02 locus. The flowering times of the lines carrying this locus and those without this locus significantly differed ( Figure 6). Primers that were polymorphic between the parents could not be developed for the cqDTF-C06 locus.

Importance of Flowering Time in the Crops
An important goal of crop breeding is to improve crops to better adapt to the local environment and climate, so as to breed new varieties [32]. Previous studies have shown that early flowering and early maturity of B. napus had a significant positive correlation, and early maturity of B. napus could be bred through the selection of early flowering materials [33,34]. Flowering time genes have been reported to regulate crop yield in rice, tomato, and rapeseed [22,[35][36][37][38][39]. Therefore, the identification of new flowering regulatory loci can help to elucidate the genetic basis of flowering and then select varieties adapted to different geographic regions. In this study, we identified two main QTLs for flowering time, and found that the trait had a high heritability (91.52%) based on three years of six replicates of flowering data, which is consistent with the results of the study by Li et al.

Importance of Flowering Time in the Crops
An important goal of crop breeding is to improve crops to better adapt to the local environment and climate, so as to breed new varieties [32]. Previous studies have shown that early flowering and early maturity of B. napus had a significant positive correlation, and early maturity of B. napus could be bred through the selection of early flowering materials [33,34]. Flowering time genes have been reported to regulate crop yield in rice, tomato, and rapeseed [22,[35][36][37][38][39]. Therefore, the identification of new flowering regulatory loci can help to elucidate the genetic basis of flowering and then select varieties adapted to different geographic regions. In this study, we identified two main QTLs for flowering time, and found that the trait had a high heritability (91.52%) based on three years of six replicates of flowering data, which is consistent with the results of the study by Li et al. (2018) and Helal et al. (2021) [11,27]. The high heritability at the flowering time will help us to carry out marker-assisted selection breeding for this trait.

Construction of the High-Density Genetic Map for the 158A-SGDH Population
QTL mapping is an effective method for analyzing complex quantitative traits. The construction of high-density genetic linkage maps can increase the accuracy of QTL mapping. In recent years, many rapeseed studies have successfully mapped flowering time QTLs using high-density maps [20,21,[23][24][25][26][27][28][29]40]. In this study, after constructing a DH population with 178 lines, we constructed a high-density genetic linkage map of this DH population using NGS technology. The map was 3268.01 cM in length and contained 2777 bin markers with an average marker interval of 1.18 cM (Figure 3 and Table S3). However, some linkage groups did not completely cover the physical map, such as A03, A07, A09, C02, C03, C04, C06, and C08 ( Figure S2). This might be explained by the low individual sequencing depth of the DH population; the average sequencing depth was 1.70× for each individual (Table S3). To increase the accuracy of the analysis, the sequencing depths for the 158A and SGDH284 parents were 3.67× and 4.19×, respectively (Table S2).

Analysis of Two Major Flowering Time QTLs in the 158A-SGDH Population
A total of 56 QTLs for flowering time were detected by QTL analysis. Two major QTLs, cqDTF-C02 and cqDTF-C06, were identified by meta-analysis (Table 2); these QTLs could explain 13.54-32.04% and 6.07-16.03% of the phenotypic variation, respectively ( Figure 5, Table 2). These two major QTLs were detected in six and five replicates, respectively, and they were also detected stably and in the same direction, indicating that the two QTLs changed little among years; this facilitated the fine mapping of QTLs and gene cloning ( Figure 5, Table 2). The cqDTF-C02 locus was detected in different populations, indicating that this QTL shows extensive allelic variation among different varieties [17,18,23]. Raman et al. (2016) conducted a GWAS analysis and found that FLC2 (Bna.FLC.A02/C02) was responsible for 22% of the variation in flowering time under non-vernalized conditions in a diverse population of spring-type rapeseed [23]. Chen et al. (2018) used a pair of near-isogenic materials to map a major QTL in the C02 linkage group, which could explain 47.9% of the phenotypic variation [18]. They experimentally showed that BnFLC.C2 was the target gene in this locus. Another major QTL, cqDTF-C06, detected in five replicates in the three years of this study but not in previous studies, was identified as a novel QTL for flowering time. cqDTF-C06 is likely derived from a novel mutation, and more natural germplasm should be examined to determine whether it represents a rare mutation and can be exploited for the breeding of early maturing rapeseed.
These findings indicate that our experimental results are reliable. However, there were no overlapping loci with cqDTF-C06, suggesting that this might be a new QTL for flowering time. Thus, our findings have increased our knowledge of the QTLs for flowering time in rapeseed.

Analysis of Candidate Genes Associated with the Two Major Flowering Time QTLs
In Arabidopsis, the molecular network of flowering time was summarized by Srikanth and Schmid (2011) [4]. Cultivated Brassica species, such as B. rapa, B. oleracea, and B. napus, all belong to the family Brassicaceae [41]. Because Brassica crops are closely related to the model plant Arabidopsis, many studies have analyzed the flowering genes of B. napus and compared them with those in Arabidopsis. In our study, we mapped two major flowering time QTLs, cqDTF-C02 and cqDTF-C06, in rapeseed using a QTL mapping strategy.
In the cqDTF-C02 region, we identified four genes involved in flowering: AT5G08370/AGAL2, AT5G10130/DFC, AT5G10140/FLC, and AT5G11530/EMF1. The corresponding genes in rapeseed are BnaC02G0032100ZS, BnaC02G0038900ZS, BnaC02G0039100ZS, and BnaC02G00-46300ZS ( Figure 5, Table S9). The above four candidate genes in rapeseed are homologous to the genes that inhibit flowering in Arabidopsis, including AT5G10140/FLC and AT5G11530/EMF1. The genes that promote flowering are AT5G08370/AGAL2 and AT5G10130/DFC. The AT5G08370/AGAL2 gene is a member of the glycoside hydrolase family 27 (GH27) and functions as an α-galactosidase. Mutations of this gene delay flowering by up to 3 weeks, and the total number of leaves in the rosette was 17-21 in mutant plants compared to 10-13 in wild-type plants [42]. AT5G10130/DFC is located downstream of the FLC gene, which Swiezewski et al. (2007) found could lead to reductions in FLC mRNA and early flowering in Arabidopsis [43]. AT5G10140/FLC (FLOWERING LOCUS C) is a key regulator of flowering time in the vernalization pathway; FLC encodes a MADS-box transcription factor that delays flowering by blocking the transcription of FT, SOC1, and FD [44][45][46]. AT5G11530/EMF1 is a subunit of a polycomb repressive complex, which is involved in the repression of FT expression, and thus, prevents photoperiod-independent flowering. EMF1 also regulates AG, AP3, PI, SPL, and MIR172 expression [47].
In previous studies, FLC has been identified as the key gene involved in the vernalization pathway of the Arabidopsis flowering regulatory network; it has thus been well studied. The FLC copy located on C02 has been mapped to several QTLs [17,18,23], including the main QTL cqDTF-C02 in this study. Candidate genes with high homology to the flowering time gene AT5G10140 (FLC) of Arabidopsis were also detected, including BnaC02G0039100ZS. Transcriptome data analysis of the 158A and SGDH284 parents at the seedling stage showed that the expression of BnaC02G0039100ZS differed in the parents. The FPKM value was less than 5 in the 158A parent, and the FPKM value in the SGDH284 parent was 277.51. FLC homologous genes on A2 and C2 have been shown to alter the flowering time of rape [18,23]. In Arabidopsis, FLC directly binds to the promoter region of SOC1 and to the first intron of FT, which prevents the expression of these genes [38,48]. The repression of SOC1 and FT expression delays the expression of the floral meristem identity genes LFY and AP1, which increases the length of the vegetative phase and thus delays flowering [49][50][51]. Therefore, BnaC02G0039100ZS, which is homologous to AT5G10140 (FLC), might be the target gene of the cqDTF-C02 region.
In the cqDTF-C06 region, only one gene related to flowering, AT3G61250/AtMYB17, was observed, and the corresponding gene in rapeseed is BnaC06G0229100ZS ( Figure 5, Table S10). AT3G61250/AtMYB17 is a target of the meristem identity regulator LEAFY (LFY); it plays a role in the meristem identity transition from vegetative growth to flowering [52,53].
Flowering-related genes might also be found in the cqDTF-C02 and cqDTF-C06 QTL segments, as they contained 29 and 15 unknown genes, respectively (Tables S9 and S10). We plan to construct near-isogenic lines to further narrow the candidate interval through fine mapping. Additional work is needed to functionally verify possible candidate genes; genetic complementation experiments also need to be conducted.

Plant Materials
A DH population consisting of 178 lines was derived from a cross between a male SGDH284 parent and a female 158A parent; these two DH lines were selected from a microspore culture of winter rapeseed Sollux and semi-winter rapeseed zhongyou9988, respectively. The Sollux was provided by Dr. Xiyuan Ni (Zhejiang Academy of Agricultural Sciences), and the zhongyou9988 was a commercial variety bred by the Oil Crops Research Institute of the Chinese Academy of Agricultural Sciences [31].

Field Experiment and Phenotypic Measurements
The 178 DH lines, together with their parental lines SGDH284 and 158A, were grown under natural field conditions at the Experimental Station of Anhui Science and Technology University, Fengyang, China. All the lines were arranged in a completely randomized block design with one replicate in 2018-2019, three replicates in 2019-2020, and two replicates in  1/20.2 (Fengyang, 20201/20.2 (Fengyang, -2021, were used to study the DTF trait. The DTF of the parents and the DH population were recorded as the number of days from the day of sowing to the day when 50% of the rapeseed plants exhibited at least one open flower in a plot [10]. The broad-sense heritability was estimated using linear mixed models in the lme4 package [54].

Genome Sequencing and Genotyping
Genomic DNA from the parents of the 178 DH lines was extracted using a modified cetyltrimethyl ammonium bromide method [55] and stored at −20 • C. The sequencing was conducted by Personal Biotechnology Co., Ltd. Shanghai, China. Using NGS, a library of 400 inserts was constructed for the parents and the 178 DH lines. Paired-end sequencing of these libraries was then conducted using the Illumina NovaSeq sequencing platform. The BWA-MEM (0.7.12-r1039) [56] program was then used to compare the filtered high-quality data against the reference genome (the B. napus ZS11 reference genomes [57,58]) using default parameters. Picard 1.107 software was used to sort the SAM files and convert them into BAM files. Reads near InDels are most prone to mapping errors. To minimize the identification of SNPs caused by mapping errors, the reads near InDels were re-compared to improve the accuracy of the SNP calling. The Indel realigner command in the GATK program was used to re-compare all reads near InDels to improve the accuracy of SNP prediction. After genome sequencing, a total of 8,780,700 SNPs were obtained through analysis and screening. To construct a genetic map, the 8,780,700 SNPs obtained were filtered by retaining ones in which (1) the sites are homozygous in the two parents and inconsistent between the parents; (2) sites where the offspring's sequencing depth is greater than 2; and (3) sites where the offspring's deletion rate is less than 0.5. After filtering, a total of 946,690 SNPs were obtained. All 15 SNPs were considered a bin, and 15 SNPs were used as a sliding window from the beginning to the end of the linkage group [30]. Finally, 13,395 bin markers were obtained. Fifteen SNPs corresponded to 30 alleles, and these 30 loci were defined as 0 if the genotype of parent 1 was greater than or equal to 24, defined as 2 if the genotype of parent 2 was greater than or equal to 24, and 1 if it was between the two [30]. In the QTL mapping, genotype 1 was treated as a deletion type. Genotypes were converted into the input file format of MSTmap software.

Linkage Map Construction and QTL Mapping
The software MSTmap was used to classify and map linkage groups [59]. The parameters were set as follows: the cut-off p-value was 1E−6, the no-map-dist was 30.0, and the no-map-size was 2. The values of the recombination fractions were converted into genetic map distances (cM) via the Kosambi mapping function [60].
QTL analyses for the DTF trait were conducted by composite interval mapping [61] using WinQTL cartographer 2.5 software (http://statgen.ncsu.edu/qtlcart/WQTLCart. htm/ accessed on 18 May 2021). The walking speed was set to 1 cM. The LOD threshold for the DTF trait was determined by permutation analysis with 1000 repetitions. Significant QTLs repeatedly identified in different environments were integrated into consensus QTLs by a QTL meta-analysis using BioMercator 4.2 [62,63]. These consensus QTLs were divided into two groups: major QTLs consisting of R 2 > 20% or R 2 > 10% in two or more environments, and the rest of the QTLs, which were designated as minor [64].

Candidate Gene Analysis within Target Intervals
All the sequences of the annotated genes from the candidate regions were downloaded from the B. napus reference genome, ZS11 (http://yanglab.hzau.edu.cn/BnIR and http: //cbi.hzau.edu.cn/bnapus/index.php/ accessed on 10 March 2022) [65,66], and aligned against the genome sequence of Arabidopsis by local BLAST analysis.
The genomic sequences of the predicted genes were submitted to NCBI (http://www. ncbi.nlm.nih.gov/ accessed on 12 March 2022) and TAIR (www.arabidopsis.org/ accessed on 12 March 2022) for homolog search and basic functional analysis. Genes from B. napus with the highest sequence similarity to genes in Arabidopsis were designated as orthologous genes. Arabidopsis genes associated with flowering time traits were retrieved. Finally, their B. napus homologs within the CIs of target QTLs were designated as putative candidate genes for the QTLs.

Development of InDel Markers in the Candidate Regions
InDel polymorphic sites between two parents in the vicinity of the major QTL region were selected and converted into PCR-based markers according to the re-sequence data. Polymorphism analysis between parents was performed using these markers. Markers with clear polymorphic bands between the parents were then used to analyze a small population of 10 extremely early flowering individuals, 10 extremely late flowering individuals, and both parents. For the cqDTF-C02 site, one marker with better bands was developed and named C2-5. The sequence of the left primer was 5 -CGTGTCAAGTCTGCATTGTTGT-3 , and that of the right primer was 5 -TTCCTGCCTTATCCATCCCA-3 .

Conclusions
We obtained 12 reproducible consensus QTLs during the flowering stage of rapeseed through preliminary mapping and meta-analysis of the QTLs of the 158A-SGDH population, and there were two major consensus QTLs, cqDTF-C02 and cqDTF-C06. The analysis of the two major QTL regions revealed five possible candidate genes in the two QTL intervals during the flowering stage. We also developed a molecular marker closely linked to cqDTF-C02, and this marker could be useful for marker-assisted selection. The results of this study will aid the cloning of flowering time-related genes of rapeseed, the development of functional markers of flowering time-related genes, and the breeding of early maturing rapeseed varieties. Molecular marker-assisted selection has the advantages of high accuracy, high efficiency, and low cost. This study provided excellent molecular markers for the improvement of early flowering and early maturity breeding of rapeseed, which could be directly applied to targeted marker-assisted breeding and the production of rapeseed.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/plants11192635/s1, Table S1: Days to flowering of 178 individuals in the DH population in Fengyang in 2018-2019/2019-2020/2020-2021; Table S2: The sequence data quality of 178 individuals in the DH population and two parents; Table S3: Distribution of SNPs on 19 genetic linkage groups; Table S4: The bin marker information and the genotype of the 178 individuals in DH population and two parents; Table S5: The genetic position of all the mapped bin markers in each linkage group; Table S6: The physical position of all the bin markers in each chromosome; Table S7: The R 2 value between the genetic map and physical map in each linkage group; Table S8: QTL detection of days to flowering in 158A-SGDH population; Table S9: The candidate gene information in the major cqDTF-C02 QTL region; Table S10: The candidate gene information in the major cqDTF-C06 QTL region; Table S11: The FPKM values of flowering timerelated genes in two major QTL segments in parental 158A and SGDH284 seedling stage; Figure S1. Frequency distribution of phenotypes at flowering stage in Brassica napus 158A-SGDH population; Figure S2. Collinearity analysis of the genetic map and physical map in B. napus; Figure S3. A small population of 10 extremely early flowering and 10 extremely late flowering plants were analyzed by the cqDTF-C02 locus-specific InDel marker C2-5. Remarks: Tables S1-S7 and Tables S9-S11 are in  the Excel file named Supplementary Tables; Table S8 and Figures S1-S3 are in the Word file named Supplementary Information.
Author Contributions: All authors planned and supervised the research. L.C. and Z.F. conceptualized and designed the study. L.C. conducted the experiment and wrote the manuscript, W.L. and Z.F. provided the experimental materials, W.H., Y.W., J.T., J.G. and B.H. performed the flowering time investigations, L.C. analyzed the data, X.C. and Y.S. supervised the experiment, and L.C. and Z.F.