[gmx-users] Comparing ff in Topology file by PRODRG and gromos G45a3
Justin A. Lemkul
jalemkul at vt.edu
Wed Feb 9 20:23:06 CET 2011
XUEMING TANG wrote:
> Hi there
>
> This is maybe an old question. I want to ask a general procedure of
> generate topology file with gromo96 force field by Dundee Prodrg.
> Take SDS for an example, I want to use gromos96 G45a3 forcefield. After
> the file is generated by dundee, I compare the forcefield parameter of
> G45a3 with what Dundee output for(bond length, angle, dihedral, imp,
> potential). There are some differences (see below). While from papers
> author usually just mentioned that the file is generated by Prodrg
> without detailed information. According to Dundee website, the
An unfortunate practice. PRODRG topologies are questionable, at best, and often
the time is not taken to properly refine them, or at least describe what changes
were made when writing methodology.
> forcefield is gromos which is outdated. Should I change all the
> forcefield parameter by hand for any molecule I want to use with gromos
> ff. Or there is any shortcut I can try. I understand that charge
All of this depends on what you mean by "GROMOS" force field. There are many
such parameter sets. The original PRODRG was based on GROMOS87, which is indeed
extremely outdated. The PRODRG 2.5 (beta) server supports GROMOS96 43A1, but
the same caveat applies as far as accuracy. At least it's based on a newer
force field.
Unfortunately, "shortcuts" taken during parameterization usually leave you with
junk in the topology. Parameterizing new species is very time-consuming, if
done properly.
> calculated by prodrg is inaccurate.
> Additional question, would G53a6 is basically better than G45a3 for SDS?
>
There are a number of papers that have done simulations with SDS. You should
probably base your force field choice on what you find in the literature.
>
> DundeePRo ff45a3
> Length Potential Length Potential
> CHn-CHn 0.153 3.35E+05 0.153 7.15E+06
> C-SO1 0.143 2.51E+05 0.143 8.18E+06
> SO1-S 0.143 3.35E+05
> S-SO2 0.143 3.35E+05
> S=SO3 0.15 3.77E+05 0.15 8.37E+06
> S=SO4 0.15 3.77E+05 0.15 8.37E+06
>
> angle Potential angle Potential
> C-C-C 111 460.2 111 530
> C-C-SO1 109.5 460.2 109.5 520
> C-SO1-S 120 397.5
> SO1-S-SOn 109.5 460.2
>
>
> dihedral Potential dihedral Potential
> S-SO1-SO3-SO2 35.3 836.8 35.3 334.8
> C-C-C-C 0 5.9 3
> SO1-C-C-C 0 5.9 3
> S-SO1-C1-C2 0 3.8 3 0 1.3 3
> C1-SO1-S-SO4 0 1.3 3
> ~
> Mainly the potential is different. And I didnot find the exact value on
> the blank parts in ff45a3 categories.
>
You're comparing different force fields, so yes, there will be differences. If
you're planning to constrain all bonds, then the force constants there are not
significant and aren't used (they apply to a quartic potential). Otherwise, use
parameters that consistently originate from one force field parameter set. If
you can't find certain parameters, then you should consider your force field
choice more thoroughly.
-Justin
> File generated by Dundee:
>
> [ bonds ]
> ; ai aj fu c0, c1, ...
> 1 2 2 0.153 334720.0 0.153 334720.0 ; C12 C11
> 2 3 2 0.153 334720.0 0.153 334720.0 ; C11 C10
> 3 4 2 0.153 334720.0 0.153 334720.0 ; C10 C9
> 4 5 2 0.153 334720.0 0.153 334720.0 ; C9 C8
> 5 6 2 0.153 334720.0 0.153 334720.0 ; C8 C7
> 6 7 2 0.153 334720.0 0.153 334720.0 ; C7 C6
> 7 8 2 0.153 334720.0 0.153 334720.0 ; C6 C5
> 8 9 2 0.153 334720.0 0.153 334720.0 ; C5 C4
> 9 10 2 0.153 334720.0 0.153 334720.0 ; C4 C3
> 10 11 2 0.153 334720.0 0.153 334720.0 ; C3 C2
> 11 12 2 0.153 334720.0 0.153 334720.0 ; C2 C1
> 12 13 2 0.143 251040.0 0.143 251040.0 ; C1 SO1
> 13 14 2 0.143 334720.0 0.143 334720.0 ; SO1 S
> 14 15 2 0.143 334720.0 0.143 334720.0 ; S SO2
> 14 16 2 0.150 376560.0 0.150 376560.0 ; S SO3
> 14 17 2 0.150 376560.0 0.150 376560.0 ; S SO4
>
> [ pairs ]
> ; ai aj fu c0, c1, ...
> 1 4 1 ; C12 C9
> 2 5 1 ; C11 C8
> 3 6 1 ; C10 C7
> 4 7 1 ; C9 C6
> 5 8 1 ; C8 C5
> 6 9 1 ; C7 C4
> 7 10 1 ; C6 C3
> 8 11 1 ; C5 C2
> 9 12 1 ; C4 C1
> 10 13 1 ; C3 SO1
> 11 14 1 ; C2 S
> 12 15 1 ; C1 SO2
> 12 16 1 ; C1 SO3
> 12 17 1 ; C1 SO4
>
> [ angles ]
> ; ai aj ak fu c0, c1, ...
> 1 2 3 1 111.0 460.2 111.0 460.2 ; C12 C11 C10
> 2 3 4 1 111.0 460.2 111.0 460.2 ; C11 C10 C9
> 3 4 5 1 111.0 460.2 111.0 460.2 ; C10 C9 C8
> 4 5 6 1 111.0 460.2 111.0 460.2 ; C9 C8 C7
> 5 6 7 1 111.0 460.2 111.0 460.2 ; C8 C7 C6
> 6 7 8 1 111.0 460.2 111.0 460.2 ; C7 C6 C5
> 7 8 9 1 111.0 460.2 111.0 460.2 ; C6 C5 C4
> 8 9 10 1 111.0 460.2 111.0 460.2 ; C5 C4 C3
> 9 10 11 1 111.0 460.2 111.0 460.2 ; C4 C3 C2
> 10 11 12 1 111.0 460.2 111.0 460.2 ; C3 C2 C1
> 11 12 13 1 109.5 460.2 109.5 460.2 ; C2 C1 SO1
> 12 13 14 1 120.0 397.5 120.0 397.5 ; C1 SO1 S
> 13 14 15 1 109.5 460.2 109.5 460.2 ; SO1 S SO2
> 13 14 16 1 109.5 460.2 109.5 460.2 ; SO1 S SO3
> 13 14 17 1 109.5 460.2 109.5 460.2 ; SO1 S SO4
> 15 14 16 1 109.5 460.2 109.5 460.2 ; SO2 S SO3
> 15 14 17 1 109.5 460.2 109.5 460.2 ; SO2 S SO4
> 16 14 17 1 109.5 460.2 109.5 460.2 ; SO3 S SO4
>
> [ dihedrals ]
> ; ai aj ak al fu c0, c1, m, ...
> 14 13 16 15 2 35.3 836.8 35.3 836.8 ; imp S SO1
> SO3 SO2
> 4 3 2 1 1 0.0 5.9 3 0.0 5.9 3 ; dih C9 C10
> C11 C12
> 5 4 3 2 1 0.0 5.9 3 0.0 5.9 3 ; dih C8 C9
> C10 C11
> 6 5 4 3 1 0.0 5.9 3 0.0 5.9 3 ; dih C7 C8
> C9 C10
> 7 6 5 4 1 0.0 5.9 3 0.0 5.9 3 ; dih C6 C7
> C8 C9
> 8 7 6 5 1 0.0 5.9 3 0.0 5.9 3 ; dih C5 C6
> C7 C8
> 9 8 7 6 1 0.0 5.9 3 0.0 5.9 3 ; dih C4 C5
> C6 C7
> 10 9 8 7 1 0.0 5.9 3 0.0 5.9 3 ; dih C3 C4
> C5 C6
> 11 10 9 8 1 0.0 5.9 3 0.0 5.9 3 ; dih C2 C3
> C4 C5
> 12 11 10 9 1 0.0 5.9 3 0.0 5.9 3 ; dih C1 C2
> C3 C4
> 13 12 11 10 1 0.0 5.9 3 0.0 5.9 3 ; dih SO1 C1
> C2 C3
> 14 13 12 11 1 0.0 3.8 3 0.0 3.8 3 ; dih S SO1
> C1 C2
> 12 13 14 17 1 0.0 1.3 3 0.0 1.3 3 ; dih C1 SO1
> S SO4
>
>
> Thank you for your time!
>
> Best!
> Xueming
>
>
>
--
========================================
Justin A. Lemkul
Ph.D. Candidate
ICTAS Doctoral Scholar
MILES-IGERT Trainee
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin
========================================
More information about the gromacs.org_gmx-users
mailing list