Fig. 7

Codon employment shift as a commonality of pathological changes. a. The behavior of GP1 genes in different stages of cancer. The graphs indicate that GP1 genes are associated to early cancer stages, and that they become even more strongly favored as the cancer progresses, although this tails off in the last metastasis stages, when the cancer tissue attains its maximal development. The symbols show the average differences in CEC in cancer versus controls for the GP1 genes, in several stages of non-small cell lung cancer [27] or colorectal cancer [28], normalized to the maximal change observed in the respective study. b-c. Gene ontology semantic signatures of differential expression b and CorrCEC analysis c. Red spots indicate significantly enriched “biological process” GO categories, while blue spots indicate de-enriched categories. The symbol size is proportional to the magnitude of the change. The significant enrichment here refers to GO categories containing significant amounts of genes whose sequence codon compositions correlate better to the codon employment observed in disease; de-enrichment refers to GO categories containing genes whose codon compositions anti-correlate to the codon employment observed in disease. We plotted the significantly enriched GO categories using their X- and Y- semantic coordinates to represent their semantic signature (Additional file 6; [24]). For more detailed information about these graphs see Additional file 5 and the examples detailed in Additional files 8, 9, 10, 11, 12, 13, 14, 15, 16 and 17. The inset on the right represents a scatter plot of the enrichment scores for the two analyses, and confirms the broad overlap between the codon employment-predicted expression and the actual mRNA expression in cancer tissue