I am trying to run MD for fluorescein, a small molecular weight green dye. I need to add a new residue (which exists in pdb file that I created by deleting the rest of the molecule and leaving fluorescein alone) to aminoacids.rtp in order to create a topology file using gromos 53a6 force field. I am having difficulty in modifying the rtp file because I need to define gromos types for bonds, angles, dihedrals and impropers. I found the needed parameters from another paper which uses Amber ff. Can I use these same charge values used in Amber , numerically they look pretty different than the ones in Gromos. And how is the charge group defined?
I define one of the oxygens (O4 in the attached jpg) in the molecule as OE and the carbons bonded to it as C(C1 and C13), however there is no bond type for C(bare carbon)-OE(ether oxygen). I define the carbons in the ring having one H as CR1( CH in ring) and the rest of the carbons as C. Although it says C is a bare carbon, I think the other definitions do not fit, because it is not aliphatic, or methane or CH2. I guess the definition and nomenclature is different, is there a rule of thumb when defining the atom types? What is bare carbon and sp3 carbon with heavy atoms mean or what are gromos names of carbons that I call as C-bare carbon and CR1?
Parametrizing a molecule with GROMOS has a lot of pros and cons. On the positive side, parameters are generally very transferable between molecules because of the group-based parametrization scheme (building blocks are parametrized, not full molecules, so that bigger molecules can be assembled). On the negative side, there is little or no relationship with QM or other straightforward calculations. The force field is highly empirical, so if no parameters are known, it can be frustrating.
The molecule at hand is mostly straightforward to parametrize. The aromatic rings are mostly CR1 (note they need explicit H), like PHE or TYR. The "bare" carbon is a carbon that has no H atoms attached to it (the carbonyls and the ones linking any of the rings). The problematic areas are the oxyanion and the ether. There are ether parameters for GROMOS that were published a couple of years back, in JCTC I think, but I don't recall the exact reference. The oxyanion will not have known parameters, so that's the biggest challenge. One could get some rough charges from QM, but the problem is target data. For GROMOS, usually neat liquid densities and free energies of solvation are the target data. For a molecule like this, likely that information is not straightforward to obtain. Charge groups are easy - 2 to 4 atoms that add up to an integral charge (usually -1, 0, or 1).
Regarding parameters from other force fields, do not mix in anything from AMBER; it's a very different parametrization scheme. Though this illustrates a good point - if the molecule is hard to parametrize under one force field, it may be more straightforward with another, which makes that other force field a better choice. CHARMM and AMBER have direct connection with QM and their parametrization protocols are well documented in the literature. In a sense, it is much easier (though more laborious) to parametrize something for AMBER and CHARMM, though both of these force fields also have generalized force fields (GAFF and CGenFF) and online servers for generating parameters (RED among others for AMBER, ParamChem for CGenFF), though manual refinement is still often necessary when the parameters are lacking.
Hi Ozge,
For parameterizing molecules for the GROMOS force field I found the Automated Topology Builder run by Alan Mark's group in Brisbane a highly valuable starting point, working surprisingly well in many cases. If you feel like checking it out, here's the URL:
http://compbio.biosci.uq.edu.au/atb/
Hope that helps a bit. Good luck.
Cheerz,
Christian
- Badges
- Science topic