Multiple sequence alignment for phylogenetic purposes

Abstract

I have addressed the biological rather than bioinformatics aspects of molecular sequence alignment by covering a series of topics that have been under-valued, particularly within the context of phylogenetic analysis. First, phylogenetic analysis is only one of the many objectives of sequence alignment, and the most appropriate multiple alignment may not be the same for all of these purposes. Phylogenetic alignment thus occupies a specific place within a broader context. Second, homology assessment plays an intricate role in phylogenetic analysis, with sequence alignment consisting of primary homology assessment and tree building being secondary homology assessment. The objective of phylogenetic alignment thus distinguishes it from other sorts of alignment. Third, I summarise what is known about the serious limitations of using phenetic similarity as a criterion for automated multiple alignment, and provide an overview of what is currently being done to improve these computerised procedures. This synthesises information that is apparently not widely known among phylogeneticists. Fourth, I then consider the recent development of automated procedures for combining alignment and tree building, thus integrating primary and secondary homology assessment. Finally, I outline various strategies for increasing the biological content of sequence alignment procedures, which consists of taking into account known evolutionary processes when making alignment decisions. These procedures can be objective and repeatable, and can involve computerised algorithms to automate much of the work. Perhaps the most important suggestion is that alignment should be seen as a process where new sequences are added to a pre-existing alignment that has been manually curated by the biologist.