By default, the loopmodel class selects all `loop' regions in your model for refinement, defined as any insertion in the alignment (i.e., a region of the target where template information is not available). You can override this and select any set of atoms of your choosing by redefining the loopmodel.select_loop_atoms() routine. (This routine should return a selection object; see Section 2.2.6 or Section 6.9 for further information.)
This example also demonstrates how to automatically assess each generated loop model.
from modeller import *
from modeller.automodel import *
log.verbose()
env = environ()
env.io.atom_files_directory = ['.', '../atom_files']
# Create a new class based on 'loopmodel' so that we can redefine
# select_loop_atoms
class MyLoop(loopmodel):
# This routine picks the residues to be refined by loop modeling
def select_loop_atoms(self):
# Two residue ranges (both will be refined simultaneously)
return selection(self.residue_range('19:', '28:'),
self.residue_range('45:', '50:'))
a = MyLoop(env,
alnfile = 'alignment.ali', # alignment filename
knowns = '5fd1', # codes of the templates
sequence = '1fdx', # code of the target
loop_assess_methods=assess.DOPE) # assess each loop with DOPE
a.starting_model= 1 # index of the first model
a.ending_model = 1 # index of the last model
a.loop.starting_model = 1 # First loop model
a.loop.ending_model = 2 # Last loop model
a.make() # do modeling and loop refinement