MSA programs were chosen based on different algorithmic approaches beyond download
availability and popularity. All programs with their versions, URL for download and main
algorithms are presented in Table 1. The programs were run using their default parameters for
protein alignment with three exceptions: 1) MAFFT run in “auto” mode where, given the size
of the dataset analyzed, the L-INS-i (iterative refinement with consistency from local
pairwise alignment) method was mostly selected among the others; 2) POA run with the
BLOSUM62 substitution matrix since there is no default matrix for this program and the
input sequence order was preserved; 3) T-Coffee initially run in single-core mode and, when
necessary, the multi-core parameter was adjusted to use all sixteen hardware processors as
specified in “Computational cost assessment” section.