Differences with ClustalW2

The current implementation provides different results from the original ClutalW mainly for identity score and some features not described in the original papers:

• Identity score: it was not possible to replicate the distance score of the original ClustalW. The scoring function is for sure "1 minus the number of identities in the best alignment divided by the length of the shorter sequence" (checked in the original ClustalW code). What is different is the result of the pairwise alignment, some domains seem to be cut, probably is a local alignment with gotoh feature. This is an essential point for the MSA, therefore it is impossible to obtain the same result.

• Many important variables are scaled in the original code (like gap, weights, etc.), also the normalization of the weights of the guides tree is different from that one described in the reference paper.

• Weight matrices: In the original papers the distances are measured directly from the guide tree to select the right substitution matrix, in the current implementation the matrix is selected by the average percent indentity of the sequences. This feature is imported.

• Initial gap penalties: depending on the substitution matrix choosed a scaling factor is applied (e.g. 0.5 for Blosum30). This feature is imported.