Search

Figure S1. Predictive ability and bias for crown rust using various algorithms for genomic prediction. (PDF 101 kb)

Figure S2. Among-and-within-full-sib-family selection that incorporates an inexpensive genotyping assay to implement within-family selection using a high selection intensity. (PDF 227 kb)

Principal component analysis (PCA) based on the correlation matrix of 26 polar metabolites of parents and F1 samples of the Lolium mapping population. Red = P_m maternal line; blue = P_f paternal line; black = F1. (PDF 9 kb)

Principal component analysis (PCA) based on the correlation matrix of 26 polar metabolites of parents, F1 and F2 samples of the Lolium mapping population. Red = P_m maternal line; blue = P_f paternal line; black = F1; green = F2. (PDF 35 kb)

Principal component analysis (PCA) based on the correlation matrix of 26 polar metabolites of parents, F1 and F2 samples of the Lolium mapping population. Red = P_m maternal line; blue = P_f paternal line; black = F1; green = F2. (PDF 35 kb)

Selection of polar metabolites quantified using GC-MS grouped by retention index with listing of selection ions for the integration of peaks and identification based on comparison with the mass spectra of authentic standards derivatised as described in Materials and Methods. Alternatively, compounds were tentatively annotated based on comparison with the metabolomics gold standard ( http://chemdata.nist.gov/dokuwiki/doku.php?id=chemdata:amdis ). (DOCX 15Â kb)

Principal component analysis (PCA) based on the correlation matrix of 26 polar metabolites of parents and F1 samples of the Lolium mapping population. Red = P_m maternal line; blue = P_f paternal line; black = F1. (PDF 9 kb)

Figure S4. Heatmap of perennial ryegrass LG5 over six full-sib families. A Kruskal-Wallis test was performed on each marker to identify significant regions for heading. Using the perennial ryegrass genome zipper [26, 30] we identified a putative gene order for markers on LG5. These data were used to construct the heatmap for each family. A perennial ryegrass transcriptome-based genetic linkage map upon which GenomeZipper was based was used as reference to construct LG5 [26, 27] and placed above the heatmap. Each bar in the heatmap represents region between two genetic markers from the linkage map. The median Kruskal-Wallis test statistic was calculated for markers binned between markers on the genetic linkage map and used to construct the heatmap. Color of the heatmap illustrates the test-statistic of the Kruskal-wallis analysis. (PDF 32 kb)

Figure S9. Phylogenetic tree of candidate heading gene SOC1. The evolutionary history was inferred by using the Maximum Likelihood method based on the JTT matrix-based model [67]. The tree is mid-point rooted, drawn to scale, with branch lengths proportional to the number of substitutions per site. All positions containing gaps and missing data were eliminated. Evolutionary analyses were conducted in MEGA 6.06 [66]. All the associated Lolium and Arabidopsis proteins were highlighted. (PDF 42 kb)

Figure S11. Phylogenetic tree of candidate heading gene GI. The evolutionary history was inferred by using the Maximum Likelihood method based on the JTT matrix-based model [67]. The tree is mid-point rooted, drawn to scale, with branch lengths proportional to the number of substitutions per site. All positions containing gaps and missing data were eliminated. Evolutionary analyses were conducted in MEGA 6.06 [66]. Associated Lolium and Arabidopsis proteins were highlighted. (PDF 37 kb)

Figure S8. Phylogenetic tree of candidate heading gene PRR37. The evolutionary history was inferred by using the Maximum Likelihood method based on the JTT matrix-based model [67]. The tree is mid-point rooted, drawn to scale, with branch lengths proportional to the number of substitutions per site. All positions containing gaps and missing data were eliminated. Evolutionary analyses were conducted in MEGA 6.06 [66]. All the associated Lolium and rice proteins were highlighted. (PDF 38 kb)

Figure S10. Phylogenetic tree of candidate heading gene CO. The evolutionary history was inferred by using the Maximum Likelihood method based on the JTT matrix-based model [67]. The tree is mid-point rooted, drawn to scale, with branch lengths proportional to the number of substitutions per site. All positions containing gaps and missing data were eliminated. Evolutionary analyses were conducted in MEGA 6.06 [66]. Associated Lolium and rice proteins were highlighted. (PDF 37 kb)

Figure S2. Heatmap of perennial ryegrass LG1 over six full-sib families. A Kruskal-Wallis test was performed on each marker to identify significant regions for heading. Using the perennial ryegrass genome zipper [26, 30] we identified a putative gene order for markers on LG1. These data were used to construct the heatmap for each family. A perennial ryegrass transcriptome-based genetic linkage map upon which GenomeZipper was based was used as reference to construct LG1 [26, 27] and placed above the heatmap. Each bar in the heatmap represents region between two genetic markers from the linkage map. The median Kruskal-Wallis test statistic was calculated for markers binned between markers on the genetic linkage map and used to construct the heatmap. Color of the heatmap illustrates the test-statistic of the Kruskal-wallis analysis. (PDF 31 kb)

Figure S6. Phylogenetic tree of candidate heading genes PHYA, PHYB and PHYC. The evolutionary history was inferred by using the Maximum Likelihood method based on the JTT matrix-based model [67]. The tree is mid-point rooted, drawn to scale, with branch lengths proportional to the number of substitutions per site. All positions containing gaps and missing data were eliminated. Evolutionary analyses were conducted in MEGA 6.06 [66]. All the associated Lolium and Arabidopsis proteins were highlighted. (PDF 59 kb)

Figure S3. Heatmap of perennial ryegrass LG3 over six full-sib families. A Kruskal-Wallis test was performed on each marker to identify significant regions for heading. Using the perennial ryegrass genome zipper [26, 30] we identified a putative gene order for markers on LG3. These data were used to construct the heatmap for each family. A perennial ryegrass transcriptome-based genetic linkage map upon which GenomeZipper was based was used as reference to construct LG3 [26, 27] and placed above the heatmap. Each bar in the heatmap represents region between two genetic markers from the linkage map. The median Kruskal-Wallis test statistic was calculated for markers binned between markers on the genetic linkage map and used to construct the heatmap. Putative ortholog of LpGI, was identified in the phylogenetic analysis and placed onto LG3 using genetic positions from genome zipper. The genetic positions of these orthologs were extrapolated over the heatmap. Color of the heatmap illustrates the test-statistic of the Kruskal-wallis analysis. (PDF 26 kb)

Figure S5. Heatmap of perennial ryegrass LG6 over six full-sib families. A Kruskal-Wallis test was performed on each marker to identify significant regions for heading. Using the perennial ryegrass genome zipper [26, 30] we identified a putative gene order for markers on LG6. These data were used to construct the heatmap for each family. A perennial ryegrass transcriptome-based genetic linkage map upon which GenomeZipper was based was used as reference to construct LG6 [26, 27] and placed above the heatmap. Each bar in the heatmap represents region between two genetic markers from the linkage map. The median Kruskal-Wallis test statistic was calculated for markers binned between markers on the genetic linkage map and used to construct the heatmap. Putative ortholog of LpCRY2 was identified in the phylogenetic analysis and placed onto LG6 using genetic positions from genome zipper. The genetic positions of these orthologs were extrapolated over the heatmap. Color of the heatmap illustrates the test-statistic of the Kruskal-wallis analysis. (PDF 35 kb)

Figure S12. Phylogenetic tree of candidate heading gene VRN1. The evolutionary history was inferred by using the Maximum Likelihood method based on the JTT matrix-based model [67]. The tree is mid-point rooted, drawn to scale, with branch lengths proportional to the number of substitutions per site. All positions containing gaps and missing data were eliminated. Evolutionary analyses were conducted in MEGA 6.06 [66]. Associated Lolium and barley proteins were highlighted. (PDF 40 kb)

Figure S1. Scatter plot for heading vs aftermath heading. Linear regression was done using aftermath heading as response variable and heading as explantory variable for family G18. Early heading genotypes showed tendency of higher aftermath heading. (PDF 10 kb)

Physical map of the R2 region combining the location of genetic markers anchored to the region, to the location of genomic sequences and R-genes. The physical location of the sequence of genetic markers used to map quantitative and qualitative resistance at the R2 region is represented on the top of the figure (STM3016 from Milbourne et al. [42]; T1430, C2_At5g04810, TG123, TG370_F, T437_R, cLPT5-B19 from Bombarely et al.[43]; Th21 from Park et al. [44], C237 from Moloney et al.[39]; 11_4_f, 107O01_52, 40SSR2_f from Destefanis et al. unpublished results). Three potato genotypes (HB, DM and RH from the top to the bottom) are separated from the orthologous tomato region by the measuring scale. BAC and scaffold sequences are shown as solid lines, gaps are displayed as dotted lines. R2GHs are represented as filled boxes of assigned orientation. Red boxes represent full length open reading frames; black boxes designate analyzed pseudogenes, while grey boxes are pseudogenes not included in the analysis. The dotplot was obtained from the alignment of the R2 sequence (accession FJ536325) and the potato or tomato BAC/scaffold sequences. Putative orthologous pairs of R2GHs are inter-connected by dotted lines. Transposable elements are represented on the sequences as triangles of assigned orientation. Light gray-shaded areas cover three R-gene clusters, the proximal cluster C, followed by cluster B and the distal cluster A. The small cluster specific to DM, for which there is no apparent orthologous cluster in tomato (and for which we have no coverage in RH and HB) is illustrated on the right. The figure is drawn to scale. (PDF 197 kb)

Because Cd is a non-essential element with toxic effects on humans, while the other Group 12 transition element, Zn, is an essential micronutrient for human health, simultaneous selection of reduced Cd and increased Zn, if possible, would be an important objective of potato breeding. To develop cultivars with greater Zn and less Cd concentrations in the tubers, it is important to understand the physiological and genetics mechanisms that gives rise to differences in tuber Cd and Zn concentrations. To achieve these objectives, two potato cultivars, which accumulate larger tuber Cd and Zn concentrations (Lady Rosetta) or smaller tuber Cd and Zn concentrations (Cara), were used in this study. The patterns of Cd concentration, Cd content and dry weight accumulation of the two cultivars were examined at different stages of plant growth. The data suggest that differences in total Cd uptake and in Cd partitioning among organs are the mechanisms governing differences in tuber -Cd accumulation in the two cultivars. The cultivar exhibiting less concentrations of tuber Cd exhibited less root-to-shoot and shoot-to-tuber translocation, driven by larger root and shoot biomass which retained more Cd in roots and shoots, respectively, reducing its movement to the tubers. Greater remobilization and more efficient tuber loading was observed in the cultivar which accumulated more tuber Cd, indicating that remobilization of Cd from leaves to tubers was a major factor, not only in tuber-Cd loading, but also in the establishment of differences in tuber-Cd. Regardless of cultivar differences, the concentration of Cd in the tuber was very low compared to that in other organs, suggesting that, despite its large phloem mobility, Cd tends to be sequestered in the shoots. Like Cd, the two cultivars differed in Zn concentration, Zn content and dry weight partitioning between organs. The differences in tuber Zn concentration between the two cultivars were associated with differences in internal Zn distribution between stem and tuber. Lady Rosetta translocated more Zn to the tubers while Cara accumulated relatively more Zn in the stems. The greater stem Zn concentration of Cara was associated with greater stem biomass rather than preferential allocation of Zn to the stem. Further examination of reciprocal grafted plants and selected F1 genotypes derived from a cross between the two cultivars showed that partitioning between tuber and shoot was an important variable for explaining differences in tuber Cd and Zn concentration. A segregating population, comprising 188 F1 progeny derived from a cross between the two cultivars, was genotyped using the SolCap 8303 SNP array, and evaluated for Cd, Zn and maturityrelated traits. Linkage and QTL mapping were performed using TetraploidSNPMap software, which incorporates all allele dosage information. The final genetic map comprised 3755 SNP markers with an average marker density of 2.94 per cM. Tuber Cd and Zn concentrations were measured in the segregating population over two years. QTL mapping identified four loci for tuber Cd concentration on chromosomes 3, 5, 6 and 7, which explained genetic variance ranging from 5% to 33%, and five loci for tuber Zn concentration on chromosome 1, 3, 5 and, 6, explaining from 5% to 38% of genetic variance. Among the QTLs identified for tuber Cd concentration, three loci coincided with tuber Zn concentrations. The major effect QTL for both tuber Cd and Zn concentration coincided with the maturity locus on chromosome 5, where earliness was associated with increased tuber concentration of both elements. Coincident minor-effect QTLs for Cd and Zn, sharing the same direction of effect, were also found on chromosomes 3 and 6, and these were unrelated to maturity. The results indicate partially overlapping genetic control of tuber Cd and Zn concentration in the cross, involving both maturity-related and non-maturity-related mechanisms.