Reconstruct 2D pmf

User discussions
1

GROMACS version: 2021.6-Colvars-2022-12-16-dev
GROMACS modification: No
Here post your question
Hi all,
I have performed a metadynamics simulation with the Colvars module; I decided to stop it before it got to the final step. In the simulation I biased three CV’s and I have the pmf file for each variable: I was wondering if there is a tool to “reconstruct” a 2d pmf having the 1d files.

Thanks in advance
Stefano

2

Hi Stefano, how did you bias the variables? If you have three separate PMF files it is likely that you ran three concurrent 1D metadynamics simulations. Did you add all three variables’s names to the bias?
https://colvars.github.io/colvars-refman-gromacs/colvars-refman-gromacs.html#sec:colvarbias

If on the other hand you already have a multi-dimensional PMF in this format:
https://colvars.github.io/colvars-refman-gromacs/colvars-refman-gromacs.html#sec:colvar_multicolumn_grid
you will find that there is a colvars_grid Python module in the colvartools folder of the Colvars repository.

Giacomo

3

Hi Giacomo,
I applied three 1d bias: actually I defined three separate metadynamics blocks. Does it make sense to try and get a bidimensional pmf? And if yes, is there a way?

Thanks again
Stefano

4

There is a metadynamics technique that makes use of multiple 1D metadynamics biases, but I am not familiar with it (I’m sure the paper can be found fairly quickly).

For a multi-dimensional PMF simply list multiple variables in the colvars keyword of the single metadynamics block. For reference, this is very similar to what you would to in PLUMED as well.

By the way, you mention 2D PMF but are biasing three CVs: what about the third one?

Giacomo

5

Thanks for all the information.

Regarding the third CV, I was just trying to get “pairwise” pmf for clarity sake.

Thanks again
stefano