Most of the unigenes that could be mapped

Most of the unigenes that could be mapped to the pseudomolecules with high sequence similarity hits (threshold value 1E-30) could be mapped to both A and C genomes, enabling a detailed analysis of the collinearity of the genomes. The positions in the respective GSK503 of the 57,246 unigenes that could be mapped to both are illustrated in Fig. 1. This shows the expected high degree of conservation of gene order, but there are numerous segmental rearrangements differentiating them. The extent of rearrangement is perhaps surprising given that they shared a common ancestor only 3.7 MYA [9], but this genome plasticity may be accounted for by their polyploid origins [10]. There is also a background of unigenes mapping to non-homoeologous positions in the A and C genomes, including two prominent segments shadowing each of the collinear blocks. To investigate the possible basis of these, 60 unigenes anchored to linkage group A1 were selected randomly and sequence similarity in the A and C genomes assessed by visual inspection of BLAST alignments. The results of this analysis (Additional file 8) showed that the apparent non-collinear mapping is most often an artefact caused by repetitive sequences, but sequences missing from one genome (i.e. copy number variation) are common, too. The latter explains the “shadow” collinear blocks as the best sequence match remaining after loss of a homoeologous sequence will be that of a paralogue; the hexaploid basis of each of the A and C genomes then results in two such paralogous segments being detected.
Acknowledgements This work was supported by UK Department for Environment, Food and Rural Affairs (Defra IF0144).
Specifications table
Value of the data
Supplementary methods
MRM data
Data Supplementary Fig. 1 is a Venn diagram showing differentially-expressed peptides between treatment responders and non-responders in the discovery stage for subjects treated with PR (peg-interferon/ribavirin) and T/PR (telaprevir/peg-interferon/ribavirin). Differentially-expressed peptides are defined as those with false discovery rate less than 0.05 and fold differential expression greater than 2. The overlap between the two treatments is highly statistically significant (p<2×10−16). Supplementary Fig. 2 shows ROC curves for predicting SVR for PR (A) and T/PR (B) subjects using only the five identified peptides as predictors. The model was fitted using only data from the PR subjects and is independently validated on the T/PR subjects. AUCs for A and B are 0.97 and 0.81, respectively. Supplementary Fig. 3 shows the ROC curve for predicting SVR for PR subjects in using MRM data. The AUC for this curve is 0.8. Supplementary Fig. 4 is a comparison of PR week 4 viral declines for the (A) discovery and (B) MRM analysis stages. Supplementary Table 1 lists subject characteristics according to treatment group and response (SVR). Supplementary Table 2 lists oefficients in the model predicting treatment outcome from discovery phase. Supplementary Table 3 lists the clinical trials from which the samples used in the proteomic analyses were obtained. Supplementary Table 4 lists the numbers of differentially-expressed components identified from statistical comparisons in the discovery stage. Supplementary Table 5 lists the gene symbols for the proteins identified in the discovery stage. Supplementary Table 6 lists the coefficients in the model predicting treatment outcome from discovery stage data using only components that are identified. Supplementary Table 7 lists references for liver expression of the proteins that are significantly-differentially expressed between responders to PR. Supplementary Table 8 lists pairwise pearson correlations among differentially-expressed proteins and other clinical covariates. Supplementary Table 9[2–7,10–12] are the references for Fig. 4 in [1].
Value of the data