Hello everyone,
I am trying to run a MD simulation for a non-standard ssDNA molecule. As it had many nonstandard residues I used ACPYPE and Antechamber to prepare GROMACS compatible topology. I divided the PDB file of the structure into 4 sections/parts having same residue IDs.
I then added H atoms to each of the different parts. Then I used YASARA AutoSmiles server to get the charges. Following this I fed the mol2 file with user defined charges options to ACPYPE. I used ff amber and the resultant gro files were all assembled into a single topology.
I did in vacuo and in-solvent minimization using both Steepest Descent and Conjugate Gradient methods. The structure did converge to tolerances of small forces (100 - 200 kcal/mol/nm). However I am unable to have a proper equilibration as it is failing on LINCS / Charge group moved too far in GROMACS.
Can anybody help with some suggestions?
Perform directly by selecting Charmm ff. Let Gromacs decide. It will work.
Dear Agnivo,
I also noticed that you asked about DNA pdb. How did you get that? If you don't have correct structure, you will end up getting error.
The best way is to generate the DNA structure through 3D-DART server and download with pdb fixed parameters and then generate topology with charmm or amber ff.
http://www.gromacs.org/Documentation/Terminology/Blowing_Up#Diagnosing_an_Unstable_System
When dealing with custom residues, isolate the problem by trying to simulate each residue separately. You probably have some inadequate parameters somewhere.
Dr. Lemkul,
I have a SOMAmer which has 6 non standard DNA residues. I calculated the partial charges through YASARA server then used ACPYPE with GAFF to generate the topologies. These were then merged to create a combined topology. However, the structure is not getting equilibrated and I believe the geometry is not right.
Do you think that I can use GAMESS to do a thorough geometery optimization and then use that structure for MD in GROMACS ?
Did you validate that the topology actually reproduces any target data, like QM energetics, dipole moments, water interactions, etc? I doubt that simply minimizing the structure will solve your problem. During MD, structures quickly move from the minimum energy surface upon being thermalized. I maintain that the most likely explanation is an inadequately parametrized model.
Thanks again for the explanation. I am very new to this area and lack background. Could you please list a few methods to correctly parametrize a new model. What should be the steps to go about this task.
If you're parametrizing for AMBER, read some of their nucleic acid papers. Any new species should be derived in a manner consistent with the existing force field.
- Badges
- Science topic