Hello AMBER developers,
I had the opportunity to meet a few of you at the recent CTBP summer
school at UCSD. I had mentioned that I wanted to contribute to the
AMBER project a script to simplify the setup of systems for softcore
TI, and so Ross Walker has kindly subscribed me here.
The attached script takes the prmtop/rst files from two systems, A and
B, where B differs from A only in some small perturbed region, and:
- Determines perturbed regions by finding all atoms that do not
share atom types or coordinates between A and B. Minor differences in
precision, say from PDB conversions, are tolerated.
- Ensures non-perturbed coordinates and box dimensions are exactly
equal by copying from A to B.
- Generates updated B.prmtop/rst and accompanying scmask parameters.
- Generates PDB files with the perturbed regions "highlighted" (user
can use VMD etc. to color by temp factor or occupancy to visualize
what the script guessed)
To generate a softcore-TI-friendly system, the user would:
1. Set up system A, generate A.pdb with ambpdb -aatm and modify it to
yield B.pdb
2. In LEaP:
B = loadpdb B.pdb
saveamberparm B B.prmtop B.crd
3. softcore_setup.py A.prmtop A.rst B.prmtop B.rst
This generates B.SC.prmtop/B.SC.rst and A.SC.pdb/B.SC.pdb.
4. Use A.prmtop/A.rst/B.SC.prmtop/B.SC.rst in simulations
The only caveat is that (AMBER 10's) LEaP does not write
SOLVENT_POINTERS or ATOMS_PER_MOLECULE blocks in step 2. I hacked my
copy of LEaP to address this but wonder if there's a better way.
I think this simplifies things significantly and hopefully it would
have a place with AMBER. Comments, suggestions, etc. are welcome!
Thanks,
--Tom
Thomas T. Joseph
MD/PhD student
Mount Sinai School of Medicine, New York, NY
Received on Wed Aug 19 2009 - 22:25:15 PDT
