Q: Is it possible to built short-contact-free models (hydrogens
included) with Modeller?
>
The models do have hydrogen atoms, however, I noticed that most of the
hydrogens have
short-contacts within the same atom or residue OR with neighboring
residues and also have long-range clashes.
Your input files look OK (and similar to the canonical example, in
examples/all-hydrogens/ in the Modeller distribution).
> Assuming that there is everything fine with topology or parameter
> files, I wonder, the energy minimization/MD should have corrected
such > clashes.
Not necessarily. The main restraints used in optimization are
homology-derived (which don't apply to hydrogens anyway) and van der
Waals (which are small or nonexistent for hydrogens). Maybe you're
thinking of electrostatics? Modeller doesn't use electrostatics by
default, so you'd have to turn that on with