Algorithms for forest pattern matching

Abstract

Ordered labelled trees are trees where the left-to-right order among siblings is significant. An ordered labelled forest is a sequence of ordered labelled trees. Given an ordered labelled forest F ("the target forest") and an ordered labelled forest G ("the pattern forest"), the forest pattern matching problem is to find a sub-forest F' of F such that F' and G are the most similar over all possible F'. In this paper, we present efficient algorithms for the forest pattern matching problem for two types of sub-forests: closed subforests and closed substructures. As RNA molecules' secondary structures could be represented as ordered labelled forests, our algorithms can be used to locate the structural or functional regions in RNA secondary structures. © Springer-Verlag Berlin Heidelberg 2010.