Physical clustering of AGA and GAA codons in genes.

<p>(<b>A</b>) To illustrate clustering or nonrandom spacing of AGA and GAA codons in several genes, AGA and GAA usage in each gene was calculated in a sliding window size of 15 codons (see <a href="http://www.plosgenetics.org/article/info:doi/10.1371/journal.pgen.1005706#sec011" target="_blank">Methods</a>) and plotted in the histogram along the length of the gene. Pink peaks correspond to physical clusters of AGA and GAA codons while blue peaks represent regions deficient in these codons. (<b>B</b>) The close spacing of AGA and GAA codons is illustrated for all AGA- and GAA-enriched genes in a heat map showing AGA and GAA usage calculated in a sliding window size of 15 codons. For clarity, the map shows regions of genes within 250 codons downstream of the start codon or upstream of the stop codon. (<b>C</b>) Boxplot of the number of 3-mer codon runs identified in each gene (actual) and that of shuffled sequences maintaining the codon composition of each gene (randomized). (<b>D</b>) Boxplot of the number of 3-mer codon runs identified in significantly down-regulated proteins and in other proteins. (<b>E</b>) Similar to (<b>D</b>), but only proteins enriched with AGA/GAA codons in both groups were considered. Asterisks in panels C, D and E represent outliers.