In AQUA all atoms are assigned an external name (the original name) and an internal name. Matching is done through AtomLIB libraries in $AQUADATADIR.
While the explicit listing of names in these files could be seen as an overkill, it does give a rigorous control of what is being called what. It definitely brings out non-conforming atom names; some cases are captured by a special routine and flagged in aqpc.log as
At present this need not cause any further problems, although it may be that some pseudo-atoms can not be calculated later.
It is possible to run the program with user-defined libraries without having to recompile. I expect that this feature may become more important, once information on atom types will be utilized. Also the library mechanism provides a way to invert stereospecific mappings.
At present the code should be able to handle a coordinate file in conjunction with its own restraint file (meaning a file produced by the same software package), i.e. X-PLOR coordinate files with X-PLOR restraint files, etc. Also PDB coordinate entries can be processed against the corresponding MR restraint entries (in fact, this is part of the Biotech Validation Project, see Background to AQUA).
Cross-hybridization, using different formats for restraints and coordinates, should now also work, since names of restraint atoms are searched for in the internal name list of the structure, cf. 'qhelp names'. (This change was introduced in version 0.50).
Currently for BIOSYM and X-PLOR some wildcard options are mapped onto specific names of atoms appearing in the AtomLIB, if the restraint atom can not be found. These "mapped" cases are listed in 'aqpc.log'. Possible mappings are described elsewhere, see 'qhelp guess'.
These extra mappings can be switched off by using the set guess off option, see 'qhelp options'.