Problem with creating fragment library

Dear Mark,

Thanks for using the code! We are working on your problem, we'll get back to you as soon as possible.

Eliseo.

Mark,
Thanks for bringing this up - we will get back with you after we look into this a bit more.

-Ed

Hi Mark,

The attachment functionality on this forum does not appear to be working. Can you paste your pdb and ff files for cyclopentane directly onto the forum?

A rigid molecule will only have 1 configuration in the fragment library. The original library_setup.py doesn't handle rigid molecules. You can create the data file by hand, or we have a new version of library_setup.py that, for rigid molecules, will reformat your pdb file into a data file. I can send the new script to you if you email me directly.

Ryan
rmullen2 at nd dot edu

Hi Mark,

Thanks for posting your files. The format of your pdb and ff file look fine, but the posted mcf is for cyclohexane. You can post your cyclopentane.mcf or you can use the one at the bottom of my post. However, I will warn you, even using the TraPPE parameters on the website, I get really compressed cyclopentane structures. This appears to be an issue with the TraPPE dihedral energies. For example, I generated a fragment library for UA cyclopentane, UA cyclohexane and all atom cyclopentane (amber 03 forcefield). All three give similar angle energies, but the UA cyclopentane dihedral energies are much higher:

......................................angle energy....................dihedral energy
UA cyclopentane........................237...................................395389 kJ/mol
UA cyclohexane..........................343.....................................38387
AA cyclopentane.........................311......................................5879

I would suggest contacting the Siepmann group to see if the reported dihedral parms are correct.

Converting TraPPE dihedral parms to Cassandra (cyclic alkanes)

The TraPPE website reports dihedral parms in K, whereas Cassandra requires kJ/mol. In addition, the TraPPE functional form for cyclic alkane dihedrals is slightly different than the one they used for linear alkanes (which is what Cassandra uses). Cassandra uses the dihedral functional:

E = a0 + a1 ( 1 + cos f) + a2 (1 - cos 2f) + a3 (1 + cos 3f)

whereas the TraPPE functional for cyclic alkanes is

E = c0 + c1 * cos f + c2 * cos 2f + c3 * cos 3f

To convert, use the following formulas:

a0 = (c0 - c1 + c2 - c3) * 8.314 / 1000
a1 = c1 * 8.314 / 1000
a2 = - c2 * 8.314 / 1000
a3 = c3 * 8.314 / 1000

So the parameters for cyclopentane should be a0 = 4.174, a1 = 381.729, a2 = -137.337, a3 = 12.438. The same parameters for cyclohexane should by a0 = 13.959, a2 = -29.174 (your a1 and a3 are fine).

Cyclopentane.mcf

!*******************************************************************************
!Molecular connectivity file for c5cyclo.pdb
!*******************************************************************************
!Atom Format
!index type element mass charge type parameters
!type="LJ", parms=epsilon sigma

# Atom_Info
5
1 C_s1 C 14.027 0.0 LJ 56.300 3.880 ring
2 C_s1 C 14.027 0.0 LJ 56.300 3.880 ring
3 C_s1 C 14.027 0.0 LJ 56.300 3.880 ring
4 C_s1 C 14.027 0.0 LJ 56.300 3.880 ring
5 C_s1 C 14.027 0.0 LJ 56.300 3.880 ring

!Bond Format
!index i j type parameters
!type="fixed", parms=bondLength

# Bond_Info
5
1 1 2 fixed 1.540
2 1 5 fixed 1.540
3 2 3 fixed 1.540
4 3 4 fixed 1.540
5 4 5 fixed 1.540

!Angle Format
!index i j k type parameters
!type="fixed", parms=equilibrium_angle
!type="harmonic", parms=force_constant equilibrium_angle

# Angle_Info
5
1 2 1 5 harmonic 31250.0 105.50
2 1 2 3 harmonic 31250.0 105.50
3 2 3 4 harmonic 31250.0 105.50
4 3 4 5 harmonic 31250.0 105.50
5 1 5 4 harmonic 31250.0 105.50

!Dihedral Format
!index i j k l type parameters
!type="none"
!type="CHARMM", parms=a0 a1 delta
!type="OPLS", parms=c0 c1 c2 c3
!type="harmonic", parms=force_constant equilibrium_dihedral

# Dihedral_Info
5
1 1 2 3 4 OPLS 4.174 381.729 -137.337 12.438
2 1 5 4 3 OPLS 4.174 381.729 -137.337 12.438
3 2 1 5 4 OPLS 4.174 381.729 -137.337 12.438
4 2 3 4 5 OPLS 4.174 381.729 -137.337 12.438
5 3 2 1 5 OPLS 4.174 381.729 -137.337 12.438

# Improper_Info
0

!Fragment Format
!index number_of_atoms_in_fragment branch_point other_atoms

# Fragment_Info
1
1 5 1 2 3 4 5


# Fragment_Connectivity
0


END