marioernestovaldes
commented
7 months ago I went and took a look at the code to see how difficult was to implement the feature you are requesting. Although implementing the atom mask selection method is possible, it should take a little bit of time and testing. For the time being, there is a relatively easy way to exclude some atoms from the masks.
To update gmx_MMMPBSA from the repo, please type:
python -m pip install git+https://github.com/Valdes-Tresanco-MS/gmx_MMPBSA -U
I added a new variable that you can use in the input file named exclude_backbone. This variable controls whether to exclude backbone atoms from residues selected to treat with QM. It accepts two values:
0: Consider backbone atoms in residues selected to treat with QM
1: Do not consider backbone atoms in residues selected to treat with QMA typical QMMMGBSA input file would look like this:
Sample input file for QM/MMGBSA calculation
#This input file is meant to show only that gmx_MMPBSA works. Although, we tried to use the input files as recommended in the
#Amber manual, some parameters have been changed to perform more expensive calculations in a reasonable amount of time. Feel free to change the parameters
#according to what is better for your system.
&general
sys_name="QM/MMGBSA",
startframe=5,
endframe=14,
PBRadii=2,
forcefields="oldff/leaprc.ff99SB,leaprc.gaff"
/
#make sure to include at least one residue from both the receptor
#and ligand in the qm_residues mask when using 'ifqnt'.
#this requirement is automatically fulfilled when using the within keyword
#https://groups.google.com/g/gmx_mmpbsa/c/GNb4q4YGCH8
&gb
igb=1, saltcon=0.150,
ifqnt=1, qm_theory=PM3,
# qm_residues selection with "within" keyword (new in v1.5.0)
qm_residues="within 5"
# Exclude backbone atoms
exclude_backbone=1,
/For the time being, if you want to exclude other atoms different than those of the backbone, you could modify the mask in the createinput.py file
Please, let us know if you encounter any issues...
camattelaer
In which tool?
gmx_MMPBSA
New Feature
Currently qmmask selection is based on full residues, either directly via residue number or within specific distance of the ligand.
Additional option to set qmmask directly for selected atoms would be beneficial to prevent SCF issues if the MD was run using only an MM forcefield.
Description
It would be nice if there was an option to set the QM atom mask other than only based on residue numbers.
Currently creating QM-MM boundary through residues numbers 'cuts' through amide bonds, which is a poor choice of QM-MM boundary.
With MMPBSA.py this is currently only possible using -make-mdins and -use-mdins flags: 1) generate inputs by running MMPBSA.py with -mdins-flag 2) the user modifies the qmmask in the respective input files to not include amide backbone atoms, eg 'cutting' through Calpha-Cbeta bond 3) run MMGBSA.py using modified input files using -use-mdins flag
Currently running some tests and a simulation where gmx_MMPBSA was having troubles with SCF convergence had much better results using the MMPBSA.py approach removing the backbone atoms
I do not expect this to be quickly implemented, but I do believe it would be an added value to gmx_MMPBSA
(relevance: in my opinion low since only applies to people interested in QM-MMGBSA approach, unfortunately i am one of those people) (difficulty: in my opinion medium, perhaps even high, because it does require some additional intermediary file manipulation to run MMPBSA)
Relevance
low
Difficulty to implement
medium