[modeller_usage] problems with filling missing residue without moving exing ones

[Francesco Pietra <>]

I could nicely fill missing residues 519-522 and 683-684 by using the
normal protocol at /wiki/Missing residue. However, as I would like the
existing residues not being modified, I used a script:



from modeller import *
from modeller.automodel import * # Load the automodel class

log.verbose()
env = environ()

# directories for input atom files
env.io.atom_files_directory = ['.', '../atom_files']

class MyModel(automodel):
     def select_atoms(self):
         return selection(self.residue_range('518', '521'),
                          self.residue_range('681', '683'))

a = MyModel(env, alnfile = 'alignment.ali',
              knowns = myPDB, sequence = 'myPDB_fill')
a.starting_model= 1
a.ending_model  = 1

a.make()


based on the last part of the wiki. Consistently, the error is:

.................................
 File "fill_resid_no-move.py", line 20, in ?
    a.make()
  File "/usr/lib/modeller9.10/modlib/modeller/automodel/automodel.py",
line 106, in make
    atmsel = self._check_select_atoms()
  File "/usr/lib/modeller9.10/modlib/modeller/automodel/automodel.py",
line 490, in _check_select_atoms
    atmsel = self.select_atoms()
  File "fill_resid_no-move.py", line 12, in select_atoms
    return selection(self.residue_range('518', '521'),
  File "/usr/lib/modeller9.10/modlib/modeller/coordinates.py", line
304, in residue_range
    start = self.residues[start]._num
  File "/usr/lib/modeller9.10/modlib/modeller/coordinates.py", line
234, in __getitem__
    (self.offset, self.length, self.suffix))
  File "/usr/lib/modeller9.10/modlib/modeller/util/modutil.py", line
23, in handle_seq_indx
    int_indx = lookup_func(*args)
  File "/usr/lib/modeller9.10/modlib/modeller/coordinates.py", line
298, in _indxres
    raise KeyError("No such residue: %s" % indx)
KeyError: 'No such residue: 518'


I tried different selections as '518', '521'  always getting error in
finding the first residue in this '...' list.


alignment.ali is constructed, as in the wiki, of first the structure,
then the sequence. At any event, the alignment.ali is the same that
works well with the script that allows all atoms to move. I serched
the web; hope this is no duplication.

Thanks for advice

francesco pietra