[modeller_usage] adding loops while retaining template structure

["Peter C. Lai" <>]

Hi everybody -

I'm modelling a membrane protein and I have externally refined my helices.
Now I want to have modeller build ab initio interhelical loops using modeller 
but my primary functional requirement is to keep the helices in the target 
exactly the same as the template. The default parameters for the model.build 
method aren't linking up the loops; the strands are just flying outwards from 
the helical bundle. I am guessing I need to define a whole set of restraints? 
Using anything other than INTERNAL_COORDINATES for build_method is making a 
complete mess of the model.

(I cannot use the automodel superclass because at the end, it is altering my 
helical structures in order to fit the loops.)

model.py:

log.verbose()
env = environ()
#all hydrogen model
env.libs.topology.read(file='$(LIB)/top_allh.lib')
env.libs.parameters.read(file='$(LIB)/par.lib')

#instantiate new alignment
aln = alignment(env)
aln.append(file='assem.ali', align_codes=('or17209h_helix', 'or17209h'))
#instantiate new model
mdl = model(env)
mdl.generate_topology(aln, sequence='or17209h')
#xfer the template coords
mdl.transfer_xyz(aln)
#build ab initio loops
mdl.build(initialize_xyz=False, build_method='INTERNAL_COORDINATES')
mdl.write(file='test.ini')


The way the .ali is arranged is that the template has the loops gapped out,
and the sequence is the complete protein.

Any ideas on how to get the loops to connect? I don't care about optimization
since I will be running MD in parameterized solvent anyway.

TIA
Peter

-- 
Peter C. Lai
Yale University School of Medicine
SenseLab