Multiple sequence alignment: Difference between revisions

Revision as of 19:28, 11 June 2006

Revision as of 19:31, 11 June 2006

Line 7:

==Multiple sequence alignment programs and techniques==

A common approach for multiple sequence alignment is to progressively align sequences using a guide tree. Initially, each sequence at the leaves is represented as a trivial alignment of a single sequence. Then, at each internal node, the alignments at its children are merged into an alignment of the alignments. At the end, the root contains an alignment on all the sequences. This is called ''progressive alignment''. Usually, these alignemnts which occur at the interior nodes are done as [[pairwise alignment]]s where each of the two children alignments is treated as a single sequence. Two sequences can be aligned using dynamic programming techniques such as [[Needleman-Wunsch]] and a scoring[[substitution matrix]] such as [[BLOSUM]] or [[Margaret Dayhoff]]'s [[Point accepted mutation|PAM]] (Point Accepted Mutation).

There are many computer programs to produce multiple sequence alignments starting with a collection of unaligned sequences ([[Clustal|ClustalW]], [[DIALIGN]], [[MAVID]], [[Threaded Blockset Aligner|TBA]], [[T-Coffee]]) and to represent graphically those alignments ([[Clustal|ClustalW]], [http://www.sanger.ac.uk/Software/Pfam/help/belvu_setup.shtml Belvu]).