Occasionally the PDB has 'negative' sequence positions (for example: -1. write the author for additional examples of negative numbers) for diverse reasons. If one of these PDBs is handed to the "fasta creation" utilities in rna_tools, it pukes -- it thinks that it's looking at a malformed tag (because it expects A:1-10, which has one hyphen, not A:-1-10).
Simple fix that I don't want to do right now.