Table 1.
Polymorphism of 7,720 SNP markers based on 410 inbred tomato accessions in the SolCAP germplasm collection.
Figure 1.
Principal component analysis (PCA) based on 4,393 SNP markers.
The PCA was conducted separately using data for all sub-populations of the SolCAP germplasm (A) and data for only the three large-fruited cultivated sub-populations consisting of the processing, fresh market, and vintage accessions (B). The processing accessions are indicated Δ (red); fresh market, ○ (blue); vintage, + (green); cultivated cherry, □ (violet); landrace, × (gold); wild cherry, ▪ (gray); and S. pimpinellifolium, ◊ (black).
Table 2.
Pairwise estimates of Fst (θ) between sub-populations.
Table 3.
Descriptive statistics for genetic diversity within sub-populations.
Table 4.
Distribution of polymorphic SNP markers in seven sub-populations of the SolCAP germplasm.
Figure 2.
Rarefaction analysis to estimate the number of polymorphic markers in each sub-population.
Accessions are coded as in Figure 1 with processing indicated by red; fresh market by blue; vintage by green; cultivated cherry, violet; landrace, gold; wild cherry, gray; and S. pimpinellifolium, black. Curves are plotted with standard deviations indicated by vertical bars.
Figure 3.
Minor allele frequency (MAF) patterns, PCA loadings, and Fst outliers on chromosomes 1 to 6.
The minor allele was determined relative to allele calls for all 410 inbred accessions based on 7,310 SNPs, and then MAF was estimated and graphed for processing (Proc), fresh market (FM), and vintage (Vint) and S. pimpinellifolium (Pimp) sub-populations (A). The 28 tomato genes listed in Table S9 are positioned based on their coding sequences (arrow) and flanking markers (dotted line). The Y-axis represents allele frequency and the X-axis represents physical positions of the SNPs oriented with respect to the tomato genome sequence [102]. PCA loadings for PC 1 and PC 2 are graphed with candidates for loci under positive selection based on Fst outlier analysis [27], [28] (B). The candidate loci are indicated by dots (•) with a color scheme indicating pairwise comparisons that were significant: red for Proc vs. FM; green for Vint vs. Proc; blue for Vint vs. FM; violet for Proc vs. FM and Vint; sky blue for FM vs. Proc and Vint; and gold for Vint vs. Proc and FM. All other loci are indicated by X (black).
Figure 4.
Minor allele frequency (MAF) patterns, PCA loadings, and Fst outliers on chromosomes 7 to 12.
The minor allele was determined relative to allele calls for all 410 inbred accessions based on 7,310 SNPs, and then MAF was estimated and graphed for processing (Proc), fresh market (FM), and vintage (Vint) and S. pimpinellifolium (Pimp) sub-populations (A). The 28 tomato genes listed in Table S9 are positioned based on their coding sequences (arrow) and flanking markers (dotted line). The Y-axis represents allele frequency and the X-axis represents physical positions of the SNPs oriented with respect to the tomato genome sequence [102]. PCA loadings for PC 1 and PC 2 are graphed with candidates for loci under positive selection based on Fst outlier analysis [27], [28] (B). The candidate loci are indicated by dots (•) with a color scheme indicating pairwise comparisons that were significant: red for Proc vs. FM; green for Vint vs. Proc; blue for Vint vs. FM; violet for Proc vs. FM and Vint; sky blue for FM vs. Proc and Vint; and gold for Vint vs. Proc and FM. All other loci are indicated by X (black).
Figure 5.
Linkage disequilibrium (LD) decay on chromosomes 1 to 4.
LD measures r2 against genetic map distance between pairs of SNP markers within each chromosome for processing, fresh market, and vintage sub-populations. Decay curves of locally weighted scatterplot smoothing (LOESS) [66] are represented by red, and decay curves of non-linear regression (NLR) [67] are represented by blue. Horizontal dashed and solid lines indicate the baseline r2values estimated using the 95th percentile method (0.23 for processing; 0.12 for fresh market; and 0.11 for vintage) and a fixed r2value of 0.2, respectively.
Figure 6.
Linkage disequilibrium (LD) decay on chromosomes 5 to 8.
LD measures r2 against genetic map distance between pairs of SNP markers within each chromosome for processing, fresh market, and vintage sub-populations. Decay curves of locally weighted scatterplot smoothing (LOESS) [66] are represented by red, and decay curves of non-linear regression (NLR) [67] are represented by blue. Horizontal dashed and solid lines indicate the baseline r2values estimated using the 95th percentile method (0.23 for processing; 0.12 for fresh market; and 0.11 for vintage) and a fixed r2value of 0.2, respectively.
Figure 7.
Linkage disequilibrium (LD) decay on chromosomes 9 to 12.
LD measures r2 against genetic map distance between pairs of SNP markers within each chromosome for processing, fresh market, and vintage sub-populations. Decay curves of locally weighted scatterplot smoothing (LOESS) [66] are represented by red, and decay curves of non-linear regression (NLR) [67] are represented by blue. Horizontal dashed and solid lines indicate the baseline r2values estimated using the 95th percentile method (0.23 for processing; 0.12 for fresh market; and 0.11 for vintage) and a fixed r2value of 0.2, respectively.
Table 5.
Chromosome by chromosome linkage disequilibrium (LD) analysis within three representative sub-populations of cultivated tomato.