Obviously, in there is no need to list (in PDB header or in .log file) groups for the first and second options, since it is obvious. In the third option a user provides atom selections for the groups, so they are listed in a parameter file and therefore listing them in .log or in PDB files will be redundant (although may not be a bad idea to keep track).
It would be redundant to list the group definitions for the third option in the PDB header; but isn't it equally redundant to list the R/Rfree and geometry statistics in the headers since they're in the log files as well?
Why don't you ask the authors of that published structure?
I will, but figured I'd get a quicker response here. Thanks, Pete
