mmCIF CA trace not quite correct

[Eric Pettersen]

For structures that change from full-atomic to chain-trace at some point, in mmCIF the switch to using "missing structure" pseudobonds occurs one residue too late. For example in each chain of 1grl, it switches to CA-only starting at residue 303, which means that the 302->303 cross residue bond needs to be a missing structure pseudobond, but it isn't.

In a possibly related problem, the mmCIF reader for 4bml, makes some CA-CA bonds covalent and the rest as missing structure (e.g. /A:153-153 is covalent).