As input Unres server can use protein sequence given using one letter code
(with XX to mark a new chain) or PDB files.
Only standard codes of aminoacids are recognized in PDB files.
Unres server requires input PDB files with continuous (without breaks)
protein chains. PDB files with gaps in the structure have to be first prepared
by filling up all missing residues. There is a plan to add such service to
the UNRES server but currently, a user has to model missing fragment using
external software or online servers (for example Modeller software, Modloop
server).
Disulfide bonds are read from PDB based on SSBOND records and for multichain
protein COMPND record with propers CHAIN: tokens listing all chains in the
PDB file.
A protein structure with disulfide bonds and no corresponding
SSBOND records will result in clashes and a very high energy
which can crash calculations.
TER records in PDB file are read to recognize chain's ends.