The default R-Coffee is very similar to T-Coffee and relies on a three-steps strategy (i) generation of a standard library with proba_pair, (ii) prediction of the secondary structures of each sequence using RNAPlfold [37], (iii) RNA specific library extension using the method described in [23]. The rest of the procedure is identical to the standard T-Coffee. R-Coffee can be executed using the following command line:

t_coffee -seq RNA_set.fa -mode rcoffee -outfile RNA_rcoffee.aln

The rcoffee mode produces five different types of output files:

• RNA_rcoffee.dnd: guide tree guide tree used to assemble the progressive alignment.
• RNA_rcoffee.aln: final alignment given in ClustalW format.
• RNA_rcoffee.aln.html: final alignment colored in function of the CORE index from red (very consistent) to blue (poorly consistent).
• a series of .rfold files: files containing the secondary structure predicted by RNAPlfold for each sequence.
• RNA_rcoffee.template_list: file specifying the association of secondary template with each sequence.
  

By default the output of R-Coffee is in ClustalW format. It is however possible to reformat the MSA in Stockholm, a more versatile format that supports the inclusion of RNA secondary structures. For instance, the following command will generate a Stockholm formatted MSA containing the alifold [38] predicted secondary structure:

t_coffee -other_pg=seq_reformat -in RNA_rcoffee.aln -action +add_alifold -output stockholm_aln -out RNA_rcoffee_predicted.stk

If this structure is known, one can provide it rather than predict it using the following command (structure provided in alifold output format):

t_coffee -other_pg=seq_reformat -in RNA_rcoffee.aln -in2 1F6X_A_real.struc -input2 alifold -action +add_alifold -output stockholm_aln -out RNA_rcoffee_real.stk