Figure 1

Correspondence between frequency and percentage identity of interspecies alignments. Frequency is seen to be related to ANCOR parameters: for a given species and percentage identity, decreased frequency is observed for longer segments, as expected. Likewise, for a given length and percentage identity, ANCORs will tend to be rarer in a species that is more remote evolutionarily. Pairwise alignments of human versus mouse, chicken and Fugu were analyzed as described [10] to produce percentage identities for non-overlapping 100 base-pair blocks with maximum 10% gaps. A frequency value associated with a given percentage identity is the fraction of segments with equal or higher percentage identity out of all 100 bp segments of the human genome. This calculation was based on the fraction of such segments out of the total number of blocks analyzed for a given species, scaled by the fraction of the human genome that is aligned to this species: 40%, 2.5% and 1%, for mouse, chicken and Fugu respectively (details are in Additional data files 2 and 3). Dashed lines represent the same analysis for 50 bp blocks. The pairwise alignments were downloaded from the UCSC browser [36], and relate to the following assemblies: human, May 2004 (hg17); mouse, May 2004 (mm5); chicken, February 2004 (galGal2); and Fugu, August 2002 (fr1). In order to produce a comparable number of aligned blocks when analyzing different species, only human chromosome 17 alignments to mouse were considered, as compared to whole genome alignments in the case of chicken and Fugu.