Errors regarding force field when using unknown residues #130
Comments
|
Can you provide here the corresponding .pdb file, please? What was the source of your .itp file? Your .itp file is not complete, it has some missing values that are absent from the original OPLS-AA force field. I will give more explanations after providing the .pdb file, it will be easier.
The CHARMM/CGenFF force field requires a third (.prm) file, in addition to .itp and .pdb. |
Since I cannot attach the file itself, I have uploaded it here: https://filebin.net/wf9jxtjbrd0zt1em
I used topolgen-1.0 (found here: http://www.gromacs.org/Downloads/User_contributions/Other_software) to generate the itp file. I know the values produced are likely far off what they should be, but I thought it would just lead to inaccurate results, not total failure. There are so many sites that claim to generate itp files from pdb files, are there any you would recommend, or is it very situational? EDIT: In the next comment I use another source, which looks like it contains more information...
Ah alright, good to know. Many thanks for the guidance. |
|
I found an itp generator that looks more reliable (PRODRG - I don't think this is really suitable for all atoms force fields though, since it excludes hydrogens on carbon) and I see what you mean about the missing values before, however some of those look a bit too large.. and when I compare it to the itp files from your SAMPL6 data, it looks a bit excessive.. Unfortunately, I still see errors, less of them, but it seems one of the atom types is still not recognised: I thought I'd post this, as a slightly closer attempt, with 1 atom type error, but still with the qtot error. |
|
I believe I may have solved this issue, using the LigParGen web server, I will close the thread when sure. |
|
LigParGen should give you a correct input file. I paste below an example of a compound that we have used for the SAMPL7 Physical Properties challenge. I will write you about your molecule later, in a separate post, as the answer requires more time. |
|
Thanks, I get 1 warning, which I am going to temporarily allow to see what happens:
|
|
You should not neglect this warning. It means that the order of atoms is not the same in the input .pdb and .itp files. Reordering the atoms in .pdb according to the order from .itp should fix this. |
|
Attached are the .pdb and .itp for your structure, using the original OPLS-AA atom types. You will note that, compared with your original .pdb file, the atoms were reordered according to the order from the .itp file. Remarks:
I'll let you to explore all these and come back if you have questions. Please report if the files attached (remove the .txt extension, which was necessary to be able to attach the files here) work correctly with MDPOW. |
Thank you for these explanations, they are very helpful.
I am currently running the simulations using the corrected .itp file from LigParGen, but after this I will be sure to run yours (which looks like it came from a personal script) and report back. It will be interesting to see how the two compare, especially vs other LogP predictions such as those from QSAR methods. Thanks again. |
Yes, it was generated by a script called MOL2FF that we have developed in house. But given that there are quite many parameters that are absent from the OPLS-AA force field, it requires very often manual (expert) intervention on the output .itp file and therefore is not adapted for public usage. |
So I changed all the residue names to DIA (from PRO) to stop it confusing it with the proline residue (admittedly I should have done this with the pdb file first before creation of the itp file). This allows the mdpow-equiilibrium to proceed to completion. However, when I try to run I don't quite know what to infer from the errors in the log files: Thank you again for your guidance. |
|
Try to run manually the command provided at the end of the error message: Maybe you should increase the |
This outputs the following error: When I use the full directory path (putting /Users/mark/Desktop/diaz/diaz/Equilibrium/water/MD_NPT/ infront of md.mdp), the command does works, which makes me think my hierarchy structure is wrong.. even though I was able to run mdpow-equilbirum. My directory hierarchy looks correct, I think, especially considering the FEP folder which is created before the error is in the right place, but I will double check. Many thanks. |
|
So I thought there might have been a problem with the data gathered from the 50ns mdpow-equilibrium data, because when I deleted it(moved into backup directory) and did a very short mdpow-equilbirum simulation, the data outputted from this short run worked with the fep step... I find this all very bizarre, because, like I said, when I put /Users/mark/Desktop/diaz/diaz/Equilibrium/water/MD_NPT/ infront of md.mdp in the following command and run this separately, the 50ns data does not error, which makes me think my first point is invalid:
Any ideas on what might be going wrong? Note: This is referring to simulations using LipParGen topology, I will use your topology in a run soon. |
I looked back to your last messages, and I'm surprised by this error. DIA should be in big caps, not in small caps. Could you please check in the .pdb, .itp, .yml files that DIA is in big caps everywhere? We need to find from where it comes this small caps 'dia' in your error messages (there are several occurrences). |
The command does work, but it's not the right md.mdp file. These command should be executed in the FEP folder, in |
The presence of md_prev.cpt is normal for longer simulations. Running the commands from two different terminals (but not simultaneously) should not be a problem.
As I said in my previous message, this is the wrong md.mdp file, you should use the one from the FEP folder. Have you tried with
Let's see if my suggestions bring any new information helping to understand the origin of the problem. Using LipParGen topology should not be a problem. |
I realise now what has happened:
I thought this was the case, but thanks for confirming.
I had tried this, but I guess since it was a Fatal error it didn't make a difference. Thanks again for the guidance, it's a shame it took me so long to realise my blunder. |
|
No problem. We all make mistakes and we are here to learn. Please rerun the whole protocol with short simulation times (e.g. 1 ns for equilibrium and 100 ps for each FEP window) to confirm that everything runs smoothly. The resulting numbers will not be reliable as the simulations will not be converged, but this is only for checking that the whole protocol works correctly now in your hands. Please also open an issue for checking that the ligand code is exactly the same in all input files (.pdb, .itp and .yml). As you said, when a user realizes a possible mistake at the FEP stage, it is already too late. |
The whole protocol (up to the failing analysis stage of md-pow, as previously mentioned) runs smoothly.
I will do this now. Thanks again. |
I'm trying to run the package on quite an unusual molecule and I thought that my .itp file was correct, however, I am starting to doubt this.
When I run:
I was under the impression that mymolecule.itp was incorporated automatically by the package to the top of the topology file with #include.
Or am I required to update the force field files myself aswell? If so, are these the ones in the MDPOW directory, or in the gromacs installation in /opt/gromacs/? (Apologies if this is more of a GROMACS question- I thought I'd ask here first because I'm not sure if I would be required to edit files within the gromacs installation or the MDPOW directory)
I get error messages implying that the force field has not been adapted to the new "residue":
grompp_qtot() failed. See warning and screen output for clues.I assume grumpp_qtot is failing because of these errors, or should I change my qtot to an integer, I didn't think this was actually essential.
Here is my .ipt and .yml file
my .ipt file:
My .yml file:
Another quick question, what is the prm parameter in the setup section of the yml file (it is not present in the benzene tutorial example)?
Thank you very much for any assistance, it is greatly appreciated.
The text was updated successfully, but these errors were encountered: