Similarity distribution within three different genomic regions. Percentage identity was calculated as described in Figure 1 for non-overlapping 50 bp blocks of human-mouse alignments. The frequency of blocks with a given identity level was calculated out of all blocks analyzed in a specified genomic region. The genomic regions are: the complete human chromosome 21; the human orthologous regions of two mouse gene deserts, MMU3 and MMU19 ; full upstream and downstream intergenic regions flanking DACH1 gene . Blocks of 100% identity consist of around 1.5% of the DACH1-flanking DNA, whereas their fraction is only 0.1% within the two other regions. For comparison, the distribution of identity values for coding segments in human chromosome 17 (selected to obtain adequate statistics) is shown, with an intermediate level fraction of 0.8% of the blocks of 100% identity. The latter distribution is seen to be similar to the computed distribution for functional regions depicted in Figure 5. The coordinates of coding exons were extracted using UCSC Table Browser > from the primary table 'refGene'. Genomic coordinates of the selected regions and statistical properties of the distributions are given in Additional files 4 and 5.