The lines depict the ratio between the observed and expected number of mononucleotide repeats (summed over all genes in the genome) as a function of their length

The expected numbers were calculated with a null-model that conserved the amino acid sequence and the gene-specific codon frequencies. The areas comprise 95% of the data from the randomized genomes. For repeats of three (for A, C and G) or six (for T) nucleotides and longer, the lines lie below this area, indicating that such repeats are significantly under-represented.<p><b>Copyright information:</b></p><p>Taken from "Stabilization of the genome of the mismatch repair deficient by context-dependent codon choice"</p><p>http://www.biomedcentral.com/1471-2164/9/249</p><p>BMC Genomics 2008;9():249-249.</p><p>Published online 28 May 2008</p><p>PMCID:PMC2430213.</p><p>