While I understand the desire for simplicity in moving between PDB and mmCIF, what I am suggesting is hardly a major parse. The remark records are already being parsed to the extent of extracting the remark number (see the example). Once you have done that, throwing away the information in the columns up to that point, and in the sequence fields at the end of the record, is no additional work, and produces a record that is cleaner and much more consistent with the rest of an mmCIf entry. To see examples done this way, take a look at the *.tcif files pointed to from the MANIFEST subpage of the pdb2cif subpage on http://ndbserver.rutgers.edu/software I am not talking about changing the definitions in the dictionary, just making the example consistent with the phrase "text of remark" which has meant the information in columns 12-70 of a PDB remark for many years (see page 8 of the 1992 PDB format description). None of the recent changes in the internal format of remark records have changed the fact that the text of the remark in columns 12-70. -- H. J. Bernstein