Summary: Normalisation of Affine Gap Costs used in Optimal Sequence Alignment.
Department of Computer Science,
Partially supported by an Australian Research Council grant.
revised Sep 1992
Abstract. It is shown how to normalise the costs of an alignment algorithm that employs affine or linear
gap costs. The normalised costs are interpreted as the log probabilities of the instructions of a finitestate
editmachine. This gives an explicit model relating sequences that can be linked to processes of mutation
Keywords: affine, alignment, edit distance, gap, indel, linear, model, normalisation.
Alignment algorithms are in common use in Molecular Biology and in other disciplines[$Need,
$Sank, $Myer] to calculate an editdistance between a pair of sequences, such as two DNA or protein
sequences. A cost is assigned to an alignment by a cost function, often having one or more parameters or
weights. An alignment algorithm finds an optimal alignment having minimum cost under its cost function.
(Equivalently a similarity score can be assigned and the algorithm finds an alignment having a maximum