Consensus Genetic Linkage Map Construction Based on One Common Parental Line for QTL Mapping in Wheat

: The consensus map is used for the veriﬁcation of marker order, quantitative trait locus (QTL) mapping and molecular marker-assisted selection (MAS) in wheat breeding. In this study, a wheat consensus genetic map named as Sp7A_G7A, was constructed using 5643 SNP markers in two double haploid (DH) populations of Spitﬁre × Bethlehem-7AS (Sp7A) and Gregory × Bethlehem-7AS (G7A), covering 4376.70 cM of 21 chromosomes (chr) with an average interval of 0.78 cM. The collinearity of the linkage maps with the consensus map of Con_map_Wang2014 and the physical map of wheat reference genome (IWGSC RefSeq v1.0) were analyzed based on the Spearman rank correlation coefﬁcients. As results, the three constructed genetic maps of Sp7A, G7A and Sp7A_G7A showed high collinearity with the Con_map_Wang2014 and the physical map, and importantly, the collinearity level between our constructed maps and the wheat physical map is higher than that between the Con_map_Wang2014 and the physical map. The seed coat color QTL detected in both populations under multiple environments were on the region (745.73–760.14 Mbp) of the seed color gene R-B1/Tamyb10-B1 ( TraesCS3B02G515900 , 3B: 757,918,264–757,920,082 bp). The validated consensus map will be beneﬁcial for QTL mapping, positional cloning, meta-QTL analysis and wheat breading.


Introduction
Common wheat (Triticum aestivum L.) (2n = 6x = 42, AABBDD) is an allohexaploid species derived from the hybridization of diploid Aegilops tauschii (2n = 2x =14, DD) and tetraploid wild emmer (2n = 4x =28, AABB) 10,000 years ago [1]; it is one of the four most important crops grown world-wide, supplying food for 35% of the world population [2]. Intense breeding activities for improving wheat varieties especially on yields over the past century have been carried out to meet the demand of the gradually increasing human population. Molecular markers have been widely used as an efficient tool for genetic analysis and positional cloning of plant species during the last three decades. Meanwhile,

DNA Extraction and Genotyping
DNA was extracted from a single plant of each DH line of Sp7A (191 lines) and G7A (218 lines) together with their parents according to a modified method of Zhou et al. [38]. The DNA concentration was measured by Nanodrop (Thermo Scientific, Waltham, MA, USA). The populations were genotyped using 12 K tGBS and 90 K wheat Infinium iSelect SNP arrays in Sp7A and G7A, respectively. SNP allele clustering and genotype calling were performed using the polyploid version of GenomeStudio software (Illumina, http://www.illumina.com). Genotype calling was performed using a default clustering algorithm as initially described in Wang et al. [24]. Genotyping data were corrected and filtered according to the following rules: SNP markers with low calling rate (<80%) were deleted together with distorted markers and double-cross markers. Afterwards, 2367 SNPs for the population of Sp7A and 3555 SNPs for the population of G7A were used to construct the single linkage maps, respectively. The physical positions of SNP markers were obtained by blasting the SNP-flanking sequences against the reference genome sequences of Chinese Spring released by the International Wheat Genome Sequencing Consortium (IWGSC RefSeq v1.0, http://www.wheatgenome.org/) with a filtering threshold of E-value < 1 × 10 −10 . For a specific marker with multiple physical positions, the position matching the linkage group of Wang et al. [24] and this study was chosen as the physical locations of the marker, which corrected the location information of the marker.

Construction the Single and Consensus Map
The QTL IciMapping V4.1 software [39] was used to group with LOD thresholds ≥ 3.0, then the SNP markers for the construction of individual linkage map for each population were ordered using "nnTwoOpt" algorithm. The consensus map of this study was constructed using MergeMap [40] to calculate the consensus marker orders of linkage groups according to the individual maps. Firstly, individual linkage maps were converted to acyclic graphs (DAGs) internally [41], and then, a consensus graph was merged on the basis of shared vertices, finally, each consensus DAG was simplified and linearized using a mean distance approximation to produce the final consensus map.

Map Validation
In order to confirm the marker order in the present consensus map, marker assignments to linkage group were compared with the corresponding positions in the consensus map constructed by Wang et al. [24] (Con_map_Wang2014) and the puta- tive physical positions in the wheat genome reference (IWGSC RefSeq v1.0, http:// www.wheatgenome.org/). The quality of the constructed individual linkage maps and consensus map were evaluated by the marker order consistency between our consensus map and Con_map_Wang2014, the collinearity between the linkage maps in this study and the wheat reference genome, the heat map and the uniform distribution of recombination fractions on the genome. The collinearity was evaluated by the Spearman rank correlation coefficient calculated by the R function cor.test. The heat maps of recombination fractions were constructed by the package "pheatmap" for R program (https://cran.r-project.org/web/packages/pheatmap/index.html).

Seeds Coat Color Parameters Measurement
Five hundred seeds of each DH line in the different replication for the three environments were used to measure the seeds coat color parameters L, a, and b with a SeedCount SC6000R analyser, a digital imaging systems specifically designed for the grain industry (Next Instruments, Ltd., Condell Park, NSW, Australia), using the Commission Internationale de l'Éclairage (CIE) L, a and b color system, respectively [42]. "L" designates the lightness of the sample (100 for white and 0 for black), "a" designates redness when positive or greenness when negative, and "b" designates yellowness when positive or blueness when negative. Each sample was scanned for three times, the mean values were used for statistical analysis and further analysis. Statistical analysis was carried out using SPSS 22.0 program (IBM SPSS Statistics, Chicago, IL, USA).

Inclusive Composite Interval Mapping of QTLs for Seeds Coat Color
QTL mapping was conducted to analyze the QTLs for seed coat color parameters using Inclusive Composite Interval Mapping (ICIM) [43] implemented by QTL IciMapping 4.1.0.0 (available at www.isbreeding.net). The walk speed for genome-wide scanning was set at 1 cM. The significant QTLs were calculated based on 1000 permutations at the 0.05 probability level at LOD threshold 3.

Common QTL across the Two DH Populations
Due to the differences in the two individual linkage maps, it was difficult to directly detect common QTLs across the two DH populations based on the QTL or marker position in each linkage map. Therefore, we assigned each QTL of the two population to the consensus map and the physical map. If the flanking markers of one QTL were 5 cM or 20 Mbp apart from the flanking markers of another QTL on both sides, the two QTLs were declared as common QTLs. The detected QTL was named as q + trait name + chromosome + the number of QTL on the chromosome, such as "qCIE-a-3B-1". "qCIE-a" indicated one QTL for seeds color parameters redness (CIE-a) in wheat, and "3B-1" indicates the first QTL on chromosome 3B [44]. Compared with the previously reported QTLs or gene for seeds coat color, the common QTLs in this study were used to validate the correctness and affection of the linkage maps and consensus map.

Construction of the Individual Maps and Consensus Map
After stringent filtration as described above, 2367 (Sp7A) and 3555 (G7A) SNP markers were used to construct the linkage maps, the details of the markers and the linkage maps were listed in the Table S1.
For Sp7A, 2367 SNP markers were grouped into 21 linkage groups corresponding to the 21 wheat chromosomes, covering 2838.18 cM across the whole genomes with an average interval of 1.20 cM; total map length for each chromosome ranged from 85.94 cM (Chr3D) to 186.92 cM (Chr5A), and chromosomes 5B and 6D showed the minimum (0.63 cM) and maximum (5.70 cM) average marker-intervals, respectively (Table 1 and Table S1, Figure S1. For G7A, 3555 SNP markers were grouped into 21 linkage groups corresponding to the 21 wheat chromosomes, covering 4381.92 cM across the whole genomes with an average interval of 1.23 cM. Total map length for each chromosome ranged from 120.40 cM (Chr4B) to 297.22 cM (Chr7D), and chromosomes 7A and 5D showed the minimum (0.66 cM) and maximum (6.58 cM) average marker-intervals, respectively (Table 1 and Table S1, Figure S2. MergeMap [40] was used to construct the consensus map by calculating the consensus marker orders of linkage groups according to the individual maps. For Sp7A_G7A, totally 5643 SNP markers were grouped into 21 linkage groups corresponding to the 21 wheat chromosomes, covering 4387.01 cM across the whole genomes with an average interval of 0.78 cM; total map length for each chromosome ranged from 121.97 cM (Chr4B) to 294.27 cM (Chr7D), and chromosomes 7A and 5D showed the minimum (0.43 cM) and maximum (3.49 cM) average marker-intervals, respectively (Table 1 and Table S1, Figure 1).
For Sp7A, 2367 SNP markers were grouped into 21 linkage groups corresponding to the 21 wheat chromosomes, covering 2838.18 cM across the whole genomes with an average interval of 1.20 cM; total map length for each chromosome ranged from 85.94 cM (Chr3D) to 186.92 cM (Chr5A), and chromosomes 5B and 6D showed the minimum (0.63 cM) and maximum (5.70 cM) average marker-intervals, respectively (Table 1 and Table S1, Figure S1).
For G7A, 3555 SNP markers were grouped into 21 linkage groups corresponding to the 21 wheat chromosomes, covering 4381.92 cM across the whole genomes with an average interval of 1.23 cM. Total map length for each chromosome ranged from 120.40 cM (Chr4B) to 297.22 cM (Chr7D), and chromosomes 7A and 5D showed the minimum (0.66 cM) and maximum (6.58 cM) average marker-intervals, respectively (Table 1 and Table S1, Figure S2).
MergeMap [40] was used to construct the consensus map by calculating the consensus marker orders of linkage groups according to the individual maps. For Sp7A_G7A, totally 5643 SNP markers were grouped into 21 linkage groups corresponding to the 21 wheat chromosomes, covering 4387.01 cM across the whole genomes with an average interval of 0.78 cM; total map length for each chromosome ranged from 121.97 cM (Chr4B) to 294.27 cM (Chr7D), and chromosomes 7A and 5D showed the minimum (0.43 cM) and maximum (3.49 cM) average marker-intervals, respectively (Table 1 and Table S1, Figure 1).

Heat Map
To estimate the quality of the linkage maps, the pair-wise recombination rates (r) of the SNP markers for Sp7A and G7A were calculated, and the heat maps were generated, respectively ( Figure S3). The value for the recombination rate was indicated by different colors ranging from yellow (lower) to purple (higher). As shown in Figure S3, the colors on and near diagonal lines for all the chromosomes are yellow, indicating that they have low recombination or high linkage disequilibrium, and the squares of different size along the diagonal lines indicate the existence of different sizes of LD blocks or linkage regions. From the marker density of different chromosomes, the Figure S3 shows the large variations of the recombination rate for different chromosomes in Sp7A and G7A.

Similar Recombination Patterns in Wheat Genome
The similar recombination trends showed in constructed genetic maps of Sp7A and G7A, in which the recombination rate of chromosome arms in the distal is higher than that in the proximal, while D genome showed the lowest recombination rate. In the genetic maps of Sp7A and G7A, each chromosome was divided into 10 intervals according to its physical positions. In each interval, the recombination rate between adjacent markers was calculated, and the sum of all the recombinant rates in each interval was presented in Figure 2. The results show that the recombination rates in the intervals near the two ends of each chromosome were high and in the middle region were low ( Figure 2). This result is consistent with those in previous studies that the recombination rates along chromosome arms show higher in high-recombination regions of distal than in low-recombination regions of proximal [45,46]. the diagonal lines indicate the existence of different sizes of LD blocks or linkage regions. From the marker density of different chromosomes, the Figure S3 shows the large variations of the recombination rate for different chromosomes in Sp7A and G7A.

Similar Recombination Patterns in Wheat Genome
The similar recombination trends showed in constructed genetic maps of Sp7A and G7A, in which the recombination rate of chromosome arms in the distal is higher than that in the proximal, while D genome showed the lowest recombination rate. In the genetic maps of Sp7A and G7A, each chromosome was divided into 10 intervals according to its physical positions. In each interval, the recombination rate between adjacent markers was calculated, and the sum of all the recombinant rates in each interval was presented in Figure 2. The results show that the recombination rates in the intervals near the two ends of each chromosome were high and in the middle region were low (Figure 2). This result is consistent with those in previous studies that the recombination rates along chromosome arms show higher in high-recombination regions of distal than in low-recombination regions of proximal [45,46].

The Collinearity of Linkage Maps with the Previous Reported Map and the Wheat Reference Genome
To evaluate the collinearity of the linkage maps in current study with the previously reported consensus map of Con_map_Wang2014 and the physical map of wheat reference genome (IWGSC RefSeq v1.0), the genetic Spearman rank correlation coefficient for each chromosome was calculated according to genetic orders of the shared markers between the three linkage maps in the present study and Con_map_Wang2014, respectively (Table 1), and the Spearman rank correlation coefficient of each chromosome was calculated between the genetic and the physical orders of the markers for the three linkage maps, respectively ( Table 1). The consecutive curves and the circos graph of collinearity among genetic map, Con_map_Wang2014 and the physical map were showed in Figures S4 and S5 and Figure 3, respectively. The linkage maps of Sp7A, G7A, and the consensus map of Sp7A_G7A constructed in this study showed high collinearity with Con_map_Wang2014 and the physical map of wheat genome ( Figures S4 and S5 and Figure 3). tween the genetic and the physical orders of the markers for the three linkage maps, respectively ( Table 1). The consecutive curves and the circos graph of collinearity among genetic map, Con_map_Wang2014 and the physical map were showed in Figures S4 and S5 and Figure 3, respectively. The linkage maps of Sp7A, G7A, and the consensus map of Sp7A_G7A constructed in this study showed high collinearity with Con_map_Wang2014 and the physical map of wheat genome ( Figures S4 and S5 and Figure 3). Among all the twenty-one Spearman rank correlation coefficients between the linkage map of Sp7A and Con_map_Wang2014, 12 were larger than 0.90, ranging from 0.369 (Chr6D) to 0.989 (Chr6A) showing high collinearity (Table1, Figure S4A,C). A larger correlation coefficient revealed between the linkage map of Sp7A and the wheat reference genome with 19 larger than 0.90, 13 larger than 0.99, and ranging from 0.174 (Chr6D) to 1.000 (Chr4A, Chr5A, Chr6A, and Chr7A) (Table 1, Figure S4B,C) Similarly, the high collinearity presented between the linkage map of G7A and Con_map_Wang2014. There were 12 Spearman rank correlation coefficients larger than 0.90, ranging from 0.107 (Chr6D) to 0.996 (Chr1A and Chr4B) (Table 1, Figure S5A,C). It is Among all the twenty-one Spearman rank correlation coefficients between the linkage map of Sp7A and Con_map_Wang2014, 12 were larger than 0.90, ranging from 0.369 (Chr6D) to 0.989 (Chr6A) showing high collinearity (Table 1, Figure S4A,C). A larger correlation coefficient revealed between the linkage map of Sp7A and the wheat reference genome with 19 larger than 0.90, 13 larger than 0.99, and ranging from 0.174 (Chr6D) to 1.000 (Chr4A, Chr5A, Chr6A, and Chr7A) (Table 1, Figure S4B,C) Similarly, the high collinearity presented between the linkage map of G7A and Con_map_Wang2014. There were 12 Spearman rank correlation coefficients larger than 0.90, ranging from 0.107 (Chr6D) to 0.996 (Chr1A and Chr4B) (Table 1, Figure S5A,C). It is also noticed that a higher collinearity between the linkage map of G7A and physical map with 18 correlation coefficients larger than 0.90, ranging from 0.151 (Chr6D) to 0.999 (Chr7D) (Table 1, Figure S5B,C).
The high collinearity showed the high degree of marker order consistency among three constructed genetic maps of current study, Con_map_Wang2014 and the physical map of wheat reference genome IWGSC RefSeq v1.0. It is noteworthy that the collinearity between the linkage maps of this study and physical map was higher than that between the consensus map of Wang et al. [24] and the physical map (Table 1). One interesting phenomenon on Chr6D was found that the collinearity among the three linkage maps and Con_map_Wang2014 as well as physical map was much lower in both two populations ( Table 1, Figures S4 and S5 and Figure 3).

Phenotypic Variation
The seeds coat color parameters (including: Brightness (L), redness (a), and yellowness (b)) were measured by SeedCount SC6000R analyser (Next Instruments, Ltd., Australia). The large color variations were observed in both populations in all environments (Table 2). For Sp7A, across all three environments, the average brightness was 50.72, ranging from 43.20 to 59.90, while the redness was 6.33, ranging from 3.60 to 9.50 and yellowness was 18.27 varied from 13.90 to 24.70. For G7A, across two environments, the average brightness was 51.00, ranging from 44.20 to 57.60 while the redness averaged 6.38, ranging from 3.6 to 8.6 and yellowness was 19.47 varied from 15.90 to 23.05. The broad-sense heritability of three seed coat color parameters was calculated for two populations in different environments ( Table 2). The result showed high heritability (>90%) in both populations and environments, indicating the high genetic variations within the populations. The correlation analysis of the seeds coat color parameters in different populations was showed in Figure 4. The three parameters (redness, yellowness, brightness) of seeds coat color showed high correlation with each other in the two populations at different environments. The parameter redness (a) was negatively correlated with yellowness (b) and brightness (L), while b was positively correlated to L. In addition, the correlations coefficients among the three parameters of seed coat color for the two populations were similar in different environments, indicating the stability of seed coat color and the accuracy of the measurement.

QTL Analysis for Seeds Coat Color in Two Populations
For the DH population of Sp7A, in the three trials (Sp7A_K, Sp7A_W and Sp7A_GH), a total of 11 significant QTLs were detected (Table 3, Figure 5), which included five QTLs of brightness (L), three QTLs for each redness (a) and yellowness (b), respectively. One stable QTL qCIE-L-3B-2 located at 97 cM of Chr3B of Sp7A, were co-detected for brightness in three environments with LOD value ranging from 20.72 to 36.12, explaining 41.0-72.3% of phenotypic variation. Stable QTL qCIE-a-3B-1 located at 59-61 cM, were detected for redness in Sp7A_K and Sp7A_W with LOD value of 4.22-47.59, explaining 4.2-50.2% of phenotypic variation, moreover, stable QTL qCIE-a-3B-2 located at 97 cM same as qCIE-L-3B-2, were detected for redness in three environments with LOD value of 34.39-53.50, explaining 30.0-80.5% of phenotypic variation. Stable QTL qCIE-b-3B-1 were detected at 97 cM for yellowness in three environments with LOD value of 7.05-24.02, explaining 14.5-63.6% of phenotypic variation (Table 3, Figure 5). Sp7A G7A Figure 4. The correlation of the three parameters brightness (L), redness (a), and yellowness (b) of seeds coat color in the two DH populations Sp7A and G7A.

QTL Analysis for Seeds Coat Color in Two Populations
For the DH population of Sp7A, in the three trials (Sp7A_K, Sp7A_W and Sp7A_GH), a total of 11 significant QTLs were detected (Table 3, Figure 5), which included five QTLs of brightness (L), three QTLs for each redness (a) and yellowness (b), respectively. One stable QTL qCIE-L-3B-2 located at 97 cM of Chr3B of Sp7A, were co-detected for brightness in three environments with LOD value ranging from 20.72 to 36.12, explaining 41.0-72.3% of phenotypic variation. Stable QTL qCIE-a-3B-1 located at 59-61 cM, were detected for redness in Sp7A_K and Sp7A_W with LOD value of 4.22-47.59, explaining 4.2-50.2% of phenotypic variation, moreover, stable QTL qCIE-a-3B-2 located at 97 cM same as qCIE-L-3B-2, were detected for redness in three environments with LOD value of 34.39-53.50, explaining 30.0-80.5% of phenotypic variation. Stable QTL qCIE-b-3B-1 were detected at 97 cM for yellowness in three environments with LOD value of 7.05-24.02, explaining 14.5-63.6% of phenotypic variation (Table 3, Figure 5).  For the DH population of G7A, in the two field trials (G7A_W and G7A_M), a total of 9 significant QTLs were detected (Table 3, Figure 5). Four QTLs were for each brightness and redness, while one QTL for yellowness. One stable QTL qCIE-L-3B-2 located at 145-156 cM of Chr3B in G7A, were co-detected for brightness in two environments with LOD value ranging from 7.71 to 41.60, explaining 35.9-70.7% of phenotypic variation. In three environments, QTL qCIE-a-3B-2 and qCIE-b-3B-1 located at 143-156 cM were detected with LOD value over 15, explaining 57.6-84.1% and 58.7-68.0% of phenotypic variations, respectively (Table 3, Figure 5). The red and cyan region is the QTL location region of seeds coat color parameters for Sp7A and G7A, respectively.  For the DH population of G7A, in the two field trials (G7A_W and G7A_M), a total of 9 significant QTLs were detected (Table 3, Figure 5). Four QTLs were for each brightness and redness, while one QTL for yellowness. One stable QTL qCIE-L-3B-2 located at 145-156 cM of Chr3B in G7A, were co-detected for brightness in two environments with LOD value ranging from 7.71 to 41.60, explaining 35.9-70.7% of phenotypic variation. In three environments, QTL qCIE-a-3B-2 and qCIE-b-3B-1 located at 143-156 cM were detected with LOD value over 15, explaining 57.6-84.1% and 58.7-68.0% of phenotypic variations, respectively (Table 3, Figure 5).
Overall, QTL qCIE-L-3B-2, qCIE-a-3B-2 and qCIE-b-3B-1 were co-detected with a large LOD value and PVE% in single environment, which were located at 97 cM and 143-156 cM on Chr3B in both Sp7A and G7A populations, collinear to the same region 149-160 cM of consensus map and 745.73-760.14 Mbp of wheat genome (Table 3, Figure 5). L, a, and b, 3 parameters of seeds coat color, showed high correlation with each other (Figure 4), therefore, the region of the three QTLs was significantly associated with seed coat color. Moreover, one gene R-B1/Tamyb10-B1 (TraesCS3B02G515900, 3B: 757,918,264-757,920,082 bp) located in the region of the QTLs for seeds color of this study ( Figure 5), was reported to control seed coat color in wheat [47,48].

Consensus Map Increased the Mapping Resolution
MergeMap [40] was used to construct the consensus map by calculating the consensus marker orders of linkage groups according to the individual map. The consensus map (Sp7A_G7A) from the two populations had an average interval of 0.78 cM between two adjacent markers, lower than that observed in the two individual maps, which was 1.20 cM for Sp7A and 1.23 cM for G7A). Overlapping regions between individual maps were enriched by additional markers, the gaps between adjacent markers were observed smaller in Sp7A_G7A, and the density of Sp7A_G7A was also increased. High consistency of marker order among Sp7A_G7A of this study and the Con_map_Wang2014 as well as the physical map was confirmed by pairwise Spearman rank correlation coefficients, respectively, which evaluated the degree of marker order correspondence (Table 1). Therefore, either the marker orders of individual maps or consensus map were proved more reliable. The percentages of the three sub-genome lengths (A = 37.2%, B = 30.2%, D = 32.6%) in the present consensus map were closer to those (A = 34.1%, B = 31%, D = 34.9%) in Wang et al. [24].

The Collinearity of the Consensus Map
The collinearity of the linkage maps in current study with the Con_map_Wang2014 and the physical map of wheat reference genome (IWGSC RefSeq v1.0) was evaluated through Spearman rank correlation coefficient (Table 1, Figures S4 and S5 and Figure 3). Most chromosomes showed high collinearity among three constructed genetic maps of current study, the Con_map_Wang2014 and the physical map. Moreover, the collinearity between our constructed maps and the physical map of wheat is higher than that between the Con_map_Wang2014 and the physical map (Table 1), which illustrated that the linkage maps and the consensus map of our study were reliable for further study including QTL mapping and MSA. In this study, the collinearity of Chr6D between the linkage maps for the two populations and the physical maps was low, suggesting a chromosomal inversion existing on Chr6D. It may also be caused by fewer markers on Chr6D. More polymorphic markers are needed to confirm the order on 6D. Similar result had been reported in Ma et al. [49]. The collinear analysis is a useful tool to validate the correctness of the constructed maps, to help with the correction of the exact marker location and to find chromosomal rearrangement including inversion.

Application of the Integrated Consensus Map
The current linkage maps have been validated in two aspects. Firstly, the linkage map was verified based on its characteristic analysis, including the heat map of recombination, the distribution of recombination fractions on the genome, and the marker order consistency between our consensus map and Con_map_Wang2014 as well as the collinearity between the linkage maps in this study and the physical map of wheat reference genome (Table 1, Figures S4 and S5 and Figure 3). Secondly, the linkage map was validated by the QTL co-location for seeds coat color in the two populations based on the constructed individual genetic maps and consensus map.
Wheat seeds coat color, associated with pre-harvest sprouting (PHS), is a very important trait for wheat breeding. Red-grained wheats are usually more tolerant to PHS than the white-grained wheats [50,51]. R-B1/Tamyb10-B1 (TraesCS3B02G515900, 3B: 757,918,264-757,920,082 bp), on chromosome 3B, controls seeds coat color and shows multiple effects on wheat PHS resistance by accumulating red pigment catechins that inhibit seed germination [47,52]. Many papers of QTL mapping and cloning have co-located the QTL or gene for seeds coat color on chromosome 3B around or at the physical region of Tamyb10-B1 [48,53,54]. In this study, we detected the QTLs for seeds coat color parameters (L, a, and b) in the similar region (745.73-760.14 Mbp) on Chr3B both in the two populations and multiple environments based on the individual DH population map and consensus map with high LOD values of 7.71-53.99, explaining 14.5-84.1% of phenotypic variation (Table 3, Figure 5), moreover, the consensus map narrowed the co-located region to (750.14-760.14 Mbp) ( Figure 5). This is further proof that the accuracy of the individual Sp7A and G7A map, and consensus map, and those maps are reliable and functional for QTL mapping. The SNP markers on the consensus map are derived from genes, and will be beneficial to association mapping, meta-QTL analysis and positional cloning, and utilized in wheat breeding.

Conclusions
A total of 5643 SNP markers in two DH populations were grouped into 21 linkage groups corresponding to the 21 wheat chromosomes for constructing a consensus genetic map, covering 4376.70 cM across the whole genomes with an average interval of 0.78 cM. The relatively high collinearity of linkage maps with the Con_map_Wang2014 and the physical map of wheat reference genome, and the QTLs for seeds coat color parameters detected in this study validated the quality of the linkage maps. The two single genetic maps of Sp7A and G7A, and the consensus map of Sp7A_G7A will be very useful for QTL mapping, positional cloning, meta-QTL analysis and wheat breading.
Supplementary Materials: The following are available online at https://www.mdpi.com/2073-4 395/11/2/227/s1. Figure S1. The genetic linkage map constructed by 2367 SNP markers in a DH population derived from a cross between Spitfire and Bethlehem-7AS. Figure S2. The genetic linkage map constructed by 3555 SNP markers in a DH population derived from a cross between Gregory and Bethlehem-7AS. Figure S3. The heat map of the matrices of pair-wise recombination fractions indicated by SNP markers for each chromosome. (A) The DH population of Sp7A; (B) the DH population of G7A; the axes of X (horizontal) and Y (vertical) represent the markers on each chromosome, the diagonal indicates that the recombination rate of the same marker is 0.0, and the cell color indicates the recombination rate of the two markers. Figure S4