16.marker.83-86

The 3242 map has had redundant markers removed. The 3242r map retains all redundant markers.

The mapping population was derived from an initial cross between the orange-rooted carrot line ‘Flavor Select’ and a purple-rooted carrot derived from the cross (2566B × Turkish) × 9304. The ultimate purple root source in this cross referred to as ‘Turkish’ is a sibling of P9547 which, like P9547, had purple phloem, xylem, and petioles. A single F₁ plant was self-pollinated to produce the F₂ population ‘3242,’ which segregated for purple color in the root phloem and xylem and in leaf petioles.

A total of 507 F₂ plants from the 3242 population grown in Mendoza in 2014 (N = 253) and 2015 (N = 254) were phenotyped for anthocyanin pigmentation in the root phloem and xylem, and in leaf petioles. Genotyping by sequencing was performed at the Biotechnology Center of the University of Wisconsin. After DNA digestion with ApeKI methylation-sensitive restriction enzyme, samples were barcoded and pooled. Sequencing was carried out on an Illumina HiSeq 2000 apparatus, using single-end 100 nt reads and v3 ‘sequencing by synthesis’ reagents. The data were analyzed using the TASSEL-GBS pipeline version 4.3.7. A linkage map was constructed using Joinmap 4.0 software. Markers with more than 20% missing data were excluded from the analysis, leaving 24,507 SNP markers. Markers that significantly deviated from the expected 1:2:1 codominant segregation ratio, as evidenced by Chi-square tests (p < 0.05), were also excluded. For chromosome 3 of the 3242 map, which included a cluster of markers with significant segregation distortion, the linkage group was constructed using first the non-distorted markers, followed by the addition of the rest of the SNP markers.

Linkage groups were obtained at a LOD > 10. Regression mapping algorithm and Haldane’s mapping function were used to calculate genetic distances between marker loci. The goodness-of-fit jump was set to 6, and the rest of the parameters used the default settings in Joinmap. The carrot genome assembly v2.0 was used as a reference to identify marker locations, and linkage groups were anchored to the carrot chromosomes through these physical locations of GBS markers.

For each quantitative trait, normality of the residuals was tested by Shapiro–Wilk test. A normal model was assumed a priori to test for normality, and if the test was significant (p ≤ 0.05), a nonparametric model was used to detect and map QTL. The R package R/qtl (Broman et al. 2003) was used for QTL analysis with the interval mapping method.

The 5171 map has had redundant markers removed. The 5171r map retains all redundant markers.

The mapping population was derived from an initial cross between the orange-rooted carrot line ‘Flavor Select’ and a purple-rooted carrot derived from the cross (2566B × Turkish) × 9304. The ultimate purple root source in this cross referred to as ‘Turkish’ is a sibling of P9547 which, like P9547, had purple phloem, xylem, and petioles. A single F₁ plant was self-pollinated to produce the F₂ population ‘3242,’ which segregated for purple color in the root phloem and xylem and in leaf petioles. From this F₂, a derivative F₄ population called ‘5171’ was produced.

A total of 110 F₄ plants from population 5171 were phenotyped for anthocyanin pigmentation in the root phloem and xylem, and in leaf petioles. Genotyping by sequencing was performed at the Biotechnology Center of the University of Wisconsin. After DNA digestion with ApeKI methylation-sensitive restriction enzyme, samples were barcoded and pooled. Sequencing was carried out on an Illumina HiSeq 2000 apparatus, using single-end 100 nt reads and v3 ‘sequencing by synthesis’ reagents. The data were analyzed using the TASSEL-GBS pipeline version 4.3.7. A linkage map was constructed using Joinmap 4.0 software. Markers with more than 20% missing data were excluded from the analysis, leaving 24,507 SNP markers. Markers that significantly deviated from the expected 1:2:1 codominant segregation ratio, as evidenced by Chi-square tests (p < 0.05), were also excluded. For chromosome 3 of the 3242 map, which included a cluster of markers with significant segregation distortion, the linkage group was constructed using first the non-distorted markers, followed by the addition of the rest of the SNP markers.

Linkage groups were obtained at a LOD > 10. Regression mapping algorithm and Haldane’s mapping function were used to calculate genetic distances between marker loci. The goodness-of-fit jump was set to 6, and the rest of the parameters used the default settings in Joinmap. The carrot genome assembly v2.0 was used as a reference to identify marker locations, and linkage groups were anchored to the carrot chromosomes through these physical locations of GBS markers.

For each quantitative trait, normality of the residuals was tested by Shapiro–Wilk test. A normal model was assumed a priori to test for normality, and if the test was significant (p ≤ 0.05), a nonparametric model was used to detect and map QTL. The R package R/qtl (Broman et al. 2003) was used for QTL analysis with the interval mapping method.