Custom setup of MORP.pert causes incorrect behavior of OpenMMFrEnergyST

If a solute perturbation is defined in MORPH.pert, but the initial and final 
non-bonded parameters are identical, and there are no perturbations in bond, 
angle or dihedrals, then the gradients should be zero. 

This is indeed observed, but only if the atom types do not involve dummy atoms 
("du"). 

julien@node004:methanol-discharge-solvated$ head MORPH.pert
version 1
molecule LIG
    atom
        name C
        initial_type    c3
        final_type      c3
        initial_charge  0.11670
        final_charge    0.11670
        initial_LJ      3.39967  0.10940
        final_LJ        3.39967  0.10940

julien@node004:methanol-discharge-solvated$python debug.py 0.500 1
julien@node004:methanol-discharge-solvated$ cat gradients-0.5000.dat 
    1        -0.0000000000
    2        -0.0000000000
    3        -0.0000000000
    4        -0.0000000000
    5        -0.0000000000
    6        -0.0000000000
    7        -0.0000000000
    8        -0.0000000000
    9        -0.0000000000
   10        -0.0000000000

* Then edit of MORPH.pert for the first atom
julien@node004:methanol-discharge-solvated$ head MORPH.pert
version 1
molecule LIG
    atom
        name C
        initial_type    c3
        final_type      du
        initial_charge  0.11670
        final_charge    0.11670
        initial_LJ      3.39967  0.10940
        final_LJ        3.39967  0.10940
julien@node004:methanol-discharge-solvated$python debug.py 0.500 1
julien@node004:methanol-discharge-solvated$ cat gradients-0.5000.dat 
    1         0.7942053587
    2         1.1477001496
    3         1.0496873523
    4         0.9552297854
    5         1.1364659597
    6         0.9471439288
    7         1.2354505913
    8         0.8188560249
    9         1.0472695641
   10         0.9962898923

Original issue reported on code.google.com by julienm...@googlemail.com on 12 Jul 2013 at 10:33

chrys-athome / sire

Custom setup of MORP.pert causes incorrect behavior of OpenMMFrEnergyST #7