Which multiple alignment algorithm should I use?

Four different multiple alignment algorithms are available in Geneious under Align/Assemble→Multiple Align. It is important to consider the size of your dataset when choosing which one to use.  Below is a brief overview of each algorithm. 

Geneious aligner

The Geneious aligner is a progressive pairwise aligner, similar to ClustalW (below).  It is the slowest algorithm in Geneious and recommended for small alignments (e.g. fewer than 50 sequences, less than 1 kb in length).

ClustalW

Clustalw is also a progressive aligner.  For details of the algorithm, see http://www-bimas.cit.nih.gov/clustalw/clustalw.html.  Clustalw is a little faster than Geneious, but should still be restricted to small alignments.

MUSCLE 

MUSCLE is a good choice for medium-large alignments of a up to a few thousand taxa.   The algorithm is described at http://nar.oxfordjournals.org/content/32/5/1792.full.pdf+html and a full manual is available at http://www.drive5.com/muscle/manual/index.html.  The defaults are optimized for best accuracy.  However, you can reduce the run-time on large alignments without too much reduction in accuracy by reducing the maximum number of iterations.  

MAFFT (plugin)

The MAFFT aligner is the best choice for large alignments (e.g. up to 30,000 sequences, or fewer long sequences).  Full details are given at http://mafft.cbrc.jp/alignment/software/.

For long sequences, the algorithm performs best if sequences are closely related.  For highly divergent sequences, a whole genome aligner like Mauve or LASTZ may be more efficient.  The MAFFT plugin can be installed by going to Tools→Plugins.

 

Have more questions? Submit a request

Comments