The figure shows the ratio between observed and expected number of mononucleotide repeats (y-axis) as a function of their length (x-axis) in all open reading frames of E. coli, S. cerevisiae, and C. elegans. Observed and expected numbers were summed over all the genes in a genome. Each line represents repeats of one nucleotide. While very short repeats occur at about the expected frequencies, longer repeats are consistently rarer than expected. Dots mark cases where the difference between observed and expected number of repeats is significant (at p < 0.05, two-sided test, based on 1,000 randomizations). Open dots indicate that the observed number is higher than the expected number; filled dots indicate that the observed number is lower than the expected number.

We analyzed codons choice for the four amino that can be encoded by a homogenous codon (NNN, N ∈ {A, C, G, T}) as well as by one or more heterogeneous codons (NNX, X ≠ N). The ratio of homogeneous and heterogeneous codons encoding these amino acids (y-axis) is plotted in dependence of the number of nucleotides constituting the homogeneous codon (N) immediately following (x-axis). This ratio decreases at positions followed by runs of the critical nucleotide (N) of increasing length. The figure is based on all occurrences of these four amino acids from open reading frames of E. coli, S. cerevisiae, and C. elegans. Each line represents one nucleotide. The y-axis is standardized by the genome-wide average ratio between homogeneous and heterogeneous codons for that particular nucleotide. Solid lines between x = i and x = i+1 indicate that the ratio between homogeneous codon and heterogeneous codons differs significantly between positions followed by i and i+1 of the critical nucleotide (p < 0.05, chi-square test). Dotted lines indicate that the difference is not significant.