Fig 1. Substitution matrix and vector of ancestral probabilities.

We assume a simplified phylogenetic tree in which the branches leading to the extant organisms (human, mouse and dog) trifurcated from their last common ancestor. For each position within a SS, a separate substitution matrix P was constructed for each of the three phylogenetic branches (only the human substitution matrix is shown for simplicity). Rows in P correspond to nucleotides in the ancestral genome, and columns, to nucleotides in the descendant genome (here, human). An element of P is the conditional probability p(X|Z) that a nucleotide Z at this position in the ancestral genome was replaced with nucleotide X in the descendant genome. Elements of the vector of ancestral probabilities p(Z) are probabilities of each nucleotide in the ancestral genome.