The Partition Function Variant of Sankoff's Algorithm

Many classes of functional RNA molcules are characterized by highly conserved secondary structures but little detectable sequence similarity. Reliable multiple alignments can therefore be constructed only when the shared structural features are taken into account. Sankoff’s algorithm can be used to construct such structure-based alignments of RNA sequences in polynomial time. Here we extend the approach to a probabilistic one by explicitly computing the partition function of all pairwisely aligned sequences with a common set of base pairs. Stochastic backtracking can then be used to compute e.g. the probability that a prescribed sequence-structure pattern is conserved between two RNA sequences. The reliability of the alignment itself can be assessed in terms of the probabilities of each possible match.