Quality on simulated data set. On a simulated data set MSA programs are evaluated by means of three different pair-wise and column-wise alignment error statistics as well as relative error in alignment length. In terms of alignment accuracy on simulated data, ProGraphMSA adopts a position in the midfield after MSA tools with iterative refinement. Surprisingly, Prank has the strongest bias in alignment length, probably due to using p-distances and thus not detecting distant homologies, while ProGraphMSA's results are closest to the reference alignment lengths.
Szalkowski BMC Bioinformatics 2012 13:129 doi:10.1186/1471-2105-13-129