Fig. 2

Aggregate normal vs. cancer codon and codon pair usage comparison for select tissues. A, B Scatter plots comparing codon usage between prostate adenocarcinoma and normal prostate tissue (A) and between cholangiocarcinoma and normal bile duct tissue (B). Each red point represents a codon. Codons above the black diagonal line are more frequent in cancer tissue than normal tissue. The mean square error (MSE) value is noted in the top left of the graph. A higher MSE value indicates more difference between codon usage in the primary tumor tissue and codon usage in normal tissue. C, D Principal component analysis for codon (C) and codon pair (D) usage in normal lung tissue, non-small cell lung cancer tissues, and genomics. Genomic codon and codon pair usage values are not transcriptome weighted. E, F Euclidean distance dendrograms based on tissue-specific codon usage (E) or codon pair usage (F)