[gmx-users] zn finger simulation

[DAVID DE SANCHO]

Dear allI am trying to run a simulation of a dimer including a zinc finger using Gromacs 4.5.3. I have gone through the documentation and searched the forums but could not find a specific answer to my problem.
In order to produce the initial topology (top) and coordinate (gro) files for the simulation I am running pdb2gmx using the PDB structure of the complex, including the Zn ions. I realized that for them to be constrained in their position, i.e. in the Zn finger, I should create bonds editing the specbond.dat file. I introduced the following lines:
CYS     SG      1       ZN      Zn      1       0.23    CYS     ZNHISE    NE2     1       ZN      Zn      1       0.21    HISE    ZN
and updated the counter from 8 to 10. Needless to say, the Zn atoms are coordinated to His and Cys residues in the first protein. The distances correspond to what I measured using VMD. However pdb2gmx did not seem to be linking the aforementioned residues. I then moved the Zn lines in the PDB to the end of the section in the PDB file corresponding to chain A, the one with the Zn fingers. Then the links were produced. So the first question is: should one move the special atoms to the chain to which they supposedly correspond?
After successfully pdb2gmx-ing, I tried the Gromacs preprocessor (grompp). It complained producing 138 errors like. 
ERROR 132 [file PDBcode_forcefield_Protein_chain_A.itp, line 13663]:  No default Proper Dih. types /   No default Angle types
By looking at the lines in the itp file for the chain A I found that the errors referred to the lack of dihedrals or angles of the Zn atoms with their partners. Could you give some suggestions on how to produce these missing parameters?
Thank you very much,

David
 		 	   		  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20110720/6e475cb4/attachment.html>