Protein alignments instead of nucleotide

[bmichanderson]

Given the input to the pipeline will often be exons, it may be preferable to run at least the initial alignments using the protein residues, then back-translating to nucleotide for the remainder of the pipeline and tree building.

One tool I've used in the past is pal2nal.pl, a Perl script from 10.1093/nar/gkl315 and available at http://www.bork.embl.de/pal2nal/

To use it, you need the protein alignments and the original nucleotide fasta files (keeping the same naming convention or order). I noticed that translating the nucleotide alignments with Biopython leads to internal stop codons sometimes (a separate issue for HybPiper) being represented as a "*", but MAFFT removes them, so I had to manually replace them with "X" prior to alignment (e.g. with sed) or pal2nal complains.