Pfam uses #=GC seq_cons, Easel expects #=GC RF

[cryptogenomicon]

Consensus column annotation in Pfam is on a #=GC seq_cons line. Easel expects reference annotation on a #=GC RF line. The Pfam seq_cons lines are handled by Easel as "unparsed annotation", which is usually fine. An example where it's not fine is in esl_msa_SequenceSubset(), which only propagates recognized GC annotations, not unparsed ones, for fear of propagating annotation that would not be valid for a subsetted alignment. Elena and Grey ran across this.