VMD-L Mailing List
From: Mayne, Christopher G (cmayne2_at_illinois.edu)
Date: Wed Sep 23 2015 - 09:36:08 CDT
It is generally best to refine only the atoms that will be affected by the modification to the sidechain. Here, I would definitely leave the backbone alone (there are highly optimized CMAP terms in there). Since the electronic structure of the sidechain is fairly isolated (i.e., no units of unsaturation), you can likely leave CB and CG as-is. Since CD is directly bonded to the functional unit that you have modified, it should probably be included in the optimization.
On Sep 22, 2015, at 11:26 AM, Amy Rice wrote:
Thank you for the help! I have gone back and included the protein and CGen files when creating the initial parameter file as suggested, and that made things much easier. This time, I was able to complete the charge optimization without encountering the error I had before. I did have a question about this step though. It appears from the top_all36_prot.rtf that the backbone charges (and even some of the side chains, such as CB and CG in the residues with longer side chains) are all standardized. I excluded all the aliphatic hydrogens from the optimization, as explained in the screencast tutorials. I decided also to exclude the backbone atoms and use the partial charges as given in the arginine residue. Would you recommend excluding more of the atoms (CB and CG, maybe CD), or including everything in the charge optimization?
I'm not quite sure what's going on here. It seems like it's looking for an atom type that is a concatenation of several atoms, and I'm not sure why it would do that since you're PSF seems OK. I can come back to this problem, but I see a couple of other issues here.
Based on the contents of your initial PAR file I can tell that you didn't include the par_all36_prot.prm file as an "associated parameter file". Many of the parameters that you have in your PAR file are defined by the standard CHARMM force field for proteins. Including this file when generating the initial parameter file will significantly reduce the number of missing parameters that you will have to work on.
You should check out the CGenFF topology file, specifically how residue GUA changes to MGUA. Comparing these two entries will be helpful in assessing the changes associated with methylating the nitrogens of guanidinium, e.g., how the partial atomic charges change, and it will give you proper LJ parameters for the terminal methyl. Many of the missing parameters should be defined, albeit in CGenFF terms (i.e., using CGenFF atom types rather than CHARMM protein atom types). It may be a little tedious, but you should be able to translate these.
Anything that remains missing won't required parameterizing the full molecule that you've constructed. At worst you can truncate the system to an N-methyl and N-propyl guanidinium.
On Sep 17, 2015, at 2:39 PM, Amy Rice wrote:
I am attempting to parameterize a modified arginine residue following the ffTK screencasts, using VMD 1.9.2. I've been able to complete the first 4 steps without issue, but I'm running into some errors during the charge optimization, in both downhill mode and simulated annealing. The error message reads:
"can't read "ljPars(HA2 HA2 HA2 H)": no such element in array"
I noticed a similar problem was discussed previously (http://www.ks.uiuc.edu/Research/vmd/mailing_list/vmd-l/22158.html Thank you for the help,
Illinois Institute of Technology
Illinois Institute of Technology
Thank you for the help,