SimTK: OpenMM: [#1450] option to fix the coordinates of the atoms

Share
Follow	Project Follow publicly Follow privately Unfollow See followers (63) List Map More info

Trackers

Date:
2011-04-18 16:50

Priority:
3

State:
Open

Submitted by:
David Zsolt Manrique (mdzs)

Assigned to:
Nobody (None)

Summary:
option to fix the coordinates of the atoms

Detailed description
Hi, i would suggest to include an option to restrict the coordinate freedom of some selected atoms. It would belong to the System class, (similarly to distance constrain) and could be set for example with a CoordConstraint(int atom_index,bool fix_x,bool fix_y,bool fix_z) function. It would influence the Force calculation by setting the constrained components to zero and in the integrator or in the optimizer the steps for constrained components are skipped. (So far it is possible to mimic the full coordinate constraint with very large mass, which is good as a dirty solution.)
Hi, i would suggest to include an option to restrict the coordinate freedom of some selected atoms. It would belong to the System class, (similarly to distance constrain) and could be set for example with a CoordConstraint(int atom_index,bool fix_x,bool fix_y,bool fix_z) function. It would influence the Force calculation by setting the constrained components to zero and in the integrator or in the optimizer the steps for constrained components are skipped. (So far it is possible to mimic the full coordinate constraint with very large mass, which is good as a dirty solution.)

Add A Comment:

Message
Date: 2012-10-30 21:08 Sender: Peter Eastman In version 4.1, you can fix an atom in place by setting its mass to 0. This is still less flexible that what you propose, since it doesn't let you constrain just one of the atom's coordinates, but it should be good enough for many purposes.
Date: 2011-04-18 16:56 Sender: Peter Eastman In addition to setting a large mass, you can use a CustomExternalForce to impose a restraint force on those atoms binding them to the desired locations. It's still not quite the same as a true constraint, but it's closer.

No Changes Have Been Made to This Item