gmxtools are a series of python scripts for EVB simulations with GROMACS.
For download and updates, vizit or clone:

https://github.com/gabrieloanca/gmxtools.git or
git@github.com:gabrieloanca/gmxtools.git

If you find these tools useful, please cite the following paper:

Gabriel Oanca, Florian van der Ent, Johan Åqvist, Efficient Empirical Valence Bond Simulations with GROMACS, Journal of Chemical Theory and Computation, 2023, http://doi.org/10.1021/acs.jctc.3c00714

For suggestions, reporting bugs or for any assistance write to gabriel.oanca@icm.uu.se or to oanca.gabriel@gmail.com.

For easier use, add this folder to the PATH environment variable in the bash rc file.

The following scripts are available:

ffld2gmx.py - converts ffld parameters to OPLS-AA types for Gromacs
genposre.py - generates posre files with different constraints for the EVB and the non-EVB atoms
gmx4evb.py - builds topologies, one for each FEP frame
mapevb.py - analyses the energies and returns the EVB profile
stats.py - calculates the mean and standard deviation over several replicas, from mapevb.py output files poly.py - smoothens the EVB profiles by a 6th degree polynomial

To show this list in your terminal, type gmxtools.
To get help for any of these tools, type -h after a tool's name (e.g., gmx4evb.py -h).

• For how to write the qmatoms.dat file, follow the instruction inside examples/HOW_TO_QMATOMS file. The file qmatoms.dat can also be found inside the examples/ folder.

• A step-by-step guide for an EVB simulation in GROMACS can be found inside examples/HOW_TO file as well as in the Supporting Information of the paper above.

UPDATES APR. 11, 2024

NOTE: These updates will slightly change the protocol and the workflow presented in the Supporting Information of the paper mentioned above. If you want to follow the workflow precisely as there, then download the tools and the protocol from jctc_2023 branch (or check the JCTC2023 tag). The old protocol can also be found inside PROTOCOL(JCTC2023) folder. The new changes are as follows:

1. A new tool for mapping is introduced, mapevb.py, that will replace the QFEP tool of Q software (step III.4 in SI). By using mapevb.py, you can skip using gmx2qfep.py as well (step III.3 in SI).
2. get_ene.sh writes all the energy files in one folder named ENE (step. III.2). Execute it like get_ene.sh 15 13, where in this case '15' and '13' correspond to the numbers that will return the potential when calling the energy tool of GROMACS onto the files rerun with the first/last topolgy and with tge evbless.top, respectively.
3. The protocol changed so that the rerun is now performed after each FEP frame to avoid the requirement for large memory storage. The new protocol can be found inside the PROTOCOL folder (the old one is also saved inside PROTOCOL(JCTC2023) folder).
4. get_ene.sh writes all the energy files in one folder named ENE (step. III.2). Execute it like get_ene.sh 15 13, where in this case '15' and '13' correspond to the numbers that will return the potential when calling the energy tool of GROMACS onto the files rerun with the first/last topolgy and with evbless.top, respectively.
5. The protocol changed so that the rerun is now performed after each FEP frame to avoid the requirement for large memory storage by deleting the .trr files after the rerun. The new protocol can be found inside the PROTOCOL folder (the old one is also saved inside PROTOCOL(JCTC2023) folder).
6. gmx4evb.py now generates also the tabulated potential files together with the topologies, so you can skip the step I.17 in SI. When generating the tables, you can choose between single or double precision depending on your GROMACS installation (default: single), and the cut-off (default: 10Å).
7. For the same pair of atoms, you can now use soft-repulsions with different $\beta$ values, one for each EVB state.
   NOTE: For donor-acceptor pairs, you cannot combine soft-repulsion in one state with Lennard Jones in the other state.
8. The .ac file is now optional in ffld2gmx.py. If --resp no option is given or if the .ac file is missing, then ffld_server charges are used (step I.2 in SI). This option is not so important since charges must be written inside the qmatoms.dat file anyway, but it was annoying not to be able to skip the Gaussian calculations.