Fig. 6

Codon usage shift as a hallmark of human pathology. a-b. The Pearson correlation coefficients between the codon percentages in the transcript composition and the respective transcript amounts (CECs) in healthy subjects and in cancer patients a. The graph indicates the means ± SEM for the codon coefficients for the mRNA amounts in control subjects. The SEMs are here often smaller than the symbol sizes. Data points show the averages of 24 (X-axis) or 25 (Y-axis) controls subjects analyzed. The black line is the identity line, not a fit to the data, to illustrate the fact that most codons are aligned on the identity line. The inset represents the codon employment shift, expressed as the average percentage change of the CEC. For a detailed description of the CEC, please refer to Materials and Methods. The data analyzed for this panel has been previously published [26]. b. Similar graph comparing healthy subjects with cancer patients. In cancer there is a clear shift in codon employment in the A−/U-direction. All AU3 codons are above the identity line, while the GC3 codons are, correspondingly, below the identity line. The inset summarizes this effect as the average codon employment shift (expressed in % of the control CECs, over the different AU3 or GC3 synonymous codons). The data used for this panel is derived from several published cancer studies [27,28,29,30,31,32,33,34]. The error bars indicate the variation between independent studies