# Position restrains in minimization and equilibration

**URL:** <https://gromacs.bioexcel.eu/t/position-restrains-in-minimization-and-equilibration/8369>\
**Category:** User discussions\
**Created:** [February 19, 2024, 11:20am UTC](https://gromacs.bioexcel.eu/t/position-restrains-in-minimization-and-equilibration/8369 "2024-02-19T11:20:56Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![Jacopo](https://dub1.discourse-cdn.com/flex017/user_avatar/gromacs.bioexcel.eu/jacopo/32/3635_2.png) [@Jacopo](https://gromacs.bioexcel.eu/u/Jacopo)\
**Post date:** [February 19, 2024, 11:20am UTC](https://gromacs.bioexcel.eu/t/position-restrains-in-minimization-and-equilibration/8369/1 "2024-02-19T11:20:56Z")

</div>

GROMACS version: all  
GROMACS modification: No

Hi everyone,

I was wondering about wether the position restraints “define = -DPOSRES” in the .mdp file have to be employed in the EM.mdp and in the two-step following equilibration of NVT.mdp and NPT.mdp before a proper MD production. I’ve seen cases in which they’re used up to the NVT step (included), but not in the NPT one (see below \*), whereas in others they are not used at all (see below \*\*). If I start from an X-ray structure, what should I do? Put them up to the NVT file, excluding the NPT one? Put them in all the three? I would be grateful for a precise answer. Moreover, can the restraints also be read by gromacs if in the EM.mdp step?

- : [1\_ASP\_simulations/tutorial\_files/sampling/input\_files · main · Gromacs constant ph / tutorials · GitLab](https://gitlab.com/gromacs-constantph/tutorials/-/tree/main/1_ASP_simulations/tutorial_files/sampling/input_files?ref_type=heads)  
\*\* : [https://zenodo.org/records/10468944](https://zenodo.org/records/10468944) (publication of January 24, so a very recent one. Look at its “1cvo\_converted” folder to understand the issue)

By the way, I ask myself why we say that the system is equilibrated after all this preparation. Indeed, there exist articles that talk of a way to minimize the G of the system starting from the very last frame of the NPT.mdp equilibration. In other words: if the system was equilibrated after the last NPT.mdp step of the equilibration, shouldn’t that mean that the average G is already at a minimum since I said the system is already equilibrated by using the last (N,P,T) ensemble step? I would expect that in a possible, subsequent (N,P,T) production I would get no lower average G.

I’ll try an answer to the very last paragraph: maybe, given the possible biases of the FF employed when using MM MD and dealing with specific chromophores such as retinal, derivatives of the p-coumaric acid (and so on), one could state that the average G value I have at the end of the equilibration is not the proper one, and that I can do better by trying a QM/MM MD, treating the chromophore at a QM level. Am I right?

Thanks for your attention, I look forward for the clarifications.

Best Regards,  
Jacopo