Figure 1
Footprinting the SCL gene. Phylogenetic footprinting of the genomic region around the SCL gene. The alignment of Chapman et al.[5], with their experimentally determined regions of biological importance annotated, was footprinted in 100 bp windows with a 25 bp step using a dinucleotide model of evolution based on the HKY85 model[12]. This model contains terms for the frequency of each dinucleotide (taken from the complete 139 kb alignment) and for the transition/transversion ratio which is applied when the difference between the dinucleotides is a transition. The total branch length summed across the tree is plotted as e-length in red and the absolute value of the log likelihood (smaller is better) in blue. The yellow line indicates the level of conservation of the top 5% of windows for the entire 139 kb alignment. Local branch lengths are presented in 5 panels aligned with a stepped dendrogram representation of the phylogenetic tree. Annotations for each species are displayed below the graph, with the lower black lines representing sequence and white space gaps. Coloured annotations in the upper track are described below the mainplot. The fourth track is the derived ancestor of mouse and rat, and therefore has no sequence or annotation. The display of local branch lengths consists of a plot of the length at the lower bound of the 95% confidence estimate in salmon, and the upper bound of the 95% confidence estimate in green. The 95% confidence interval estimate for the branch length is represented by the white space between these graphs. Regions of high confidence conservation can be identified by looking for peaks in the lower salmon graph, and conversely regions of high confidence divergence can be identified by identifying hanging peaks in the upper green graph. Regions where no reliable branch length estimate can be given will appear white. Individual branch lengths can be compared to changes in annotation between branches. For example, the grey boxed region highlights a high confidence signal of divergence in dog between 75000 and 75300. This region correlates with part of an exon and a region of open chromatin in mouse, but is intronic in dog. This suggests that the open chromatin region will be altered or not present in dog, potentially altering regulation and function of this gene. Full analysis of the entire 139 kb region around the SCL gene is provided as a supplementary file.