This track displays the level conservation of the Zoo sequences at any given position. More precisely, it computes the p-value of observing a certain parsimony score for a given column of the multiple alignment, under the null model of neutral evolution.
1) Adam Siepel has inferrred the phylogenetic tree relating the Zoo species, as well as the length of each branches of the tree. I use this tree to measure the surprise of the conservation observed.
2) For each column of the multiple alignment produced by Webb Miller, we compute the parsimony score for the columm, using Adam's phylogenetic tree. We then compute a p-value for that parsimony score, under a null model of neutral evolution (see below for a description of this null model). Species with gaps in the column are ignored. A column will get a low p-value if the parsimony score for the set of non-gap species is surprisingly low, given the divergence of these species. For column i, let Xi be the negative log of the p-value for that column.
3) For position i of the alignment, the score plotted on the track is the average of the Xj's in a window of 10 bp around position i.
Many thanks to Adam Siepel for the phylogenetic inferrences and to Webb Miller for providing the multiple alignment of the Zoo sequences.