I have looked at the mmCIF fields for sequence in the entity and entity_reference section and do not see a way to indicate any of the following: 1. The sequence is purely a guess based on the electron density maps. 2. Only part of the sequence is known and present in a sequence database. 3. Several parts of the sequence are known but not all of it; i.e. one has several stretches. 4. One only knows approximately how many residues comprise the macromolecule. 5. The authors could fit a stretch of backbone through some density but do not know which missing residues that stretch corresponds to. The PDB just released an entry like this. 6. For immunoglobulin FAB fragments, each chain will be linked to two separate PRI and/or SwissProt entries. Thus it will be necessary to be able to link one run of residues to one sequence entry and another run of residues to another sequence entry. 7. I am sure that other people will be able to think of other related representation problems. Frances C. Bernstein