Using guide networks in alignment construction

Most contemporary multiple alignment methods use a guide tree to progressively build an alignment from alignments in subtrees. Downstream analysis based on such an alignment, in particular phylogeny reconstruction, will be biased by the choice of guide tree, and may in many situations reflect this choice as much or even more than the signals contained in the data. Methods exist for extracting networks reflecting all the probable partitions in the data, and using such a network as guide in a progressive alignment method will alleviate many of the problems of guide tree based approaches. This project suggests investigating the feasibility of generalising guide tree based alignment methods to guide network alignment methods without significantly increasing the complexity.