# Negative force from pulling simulation **URL:** \ **Category:** User discussions\ **Tags:** pull-code\ **Created:** [November 10, 2022, 6:47pm UTC](https://gromacs.bioexcel.eu/t/negative-force-from-pulling-simulation/5227 "2022-11-10T18:47:28Z")\ **Posts on this page:** 6\ **Page:** 1
**Author:** ![anna.stephens](https://avatars.discourse-cdn.com/v4/letter/a/e274bd/32.png) [@anna.stephens](https://gromacs.bioexcel.eu/u/anna.stephens)\ **Post date:** [November 10, 2022, 6:47pm UTC](https://gromacs.bioexcel.eu/t/negative-force-from-pulling-simulation/5227/1 "2022-11-10T18:47:28Z")
GROMACS version: 2020 GROMACS modification: No I’m conducting a coarse-grained molecular dynamics simulation where I’m pulling my peptide through to the centre of mass of my bilayer system. The distance between the protein’s COM and membrane is ~ 3nm, the time is 50ns, so my pulling rate is about 0.0007 nm/ps with a force constant of 1,000. When I plot my force evolving with time it begins from 0 then tends to negative kJ/mol/nm, could anyone help with why this is? I feel like my system is unrealistic but my pulling force feels pretty reasonable. Thanks in advance! Anna ---
**Author:** ![anna.stephens](https://avatars.discourse-cdn.com/v4/letter/a/e274bd/32.png) [@anna.stephens](https://gromacs.bioexcel.eu/u/anna.stephens)\ **Post date:** [November 11, 2022, 9:22am UTC](https://gromacs.bioexcel.eu/t/negative-force-from-pulling-simulation/5227/2 "2022-11-11T09:22:27Z")
I should add, when the peptide first reaches the membrane I get my maximum on the PMF free energy which I assume is my transition state, I’m guessing my protein has just gone further than the membrane’s COM so the force is now acting in the opposite direction to restore it towards the COM but not sure how to avoid this situation? ---
**Author:** ![hess](https://dub1.discourse-cdn.com/flex017/user_avatar/gromacs.bioexcel.eu/hess/32/416_2.png) [@hess](https://gromacs.bioexcel.eu/u/hess)\ **Post date:** [November 15, 2022, 8:30am UTC](https://gromacs.bioexcel.eu/t/negative-force-from-pulling-simulation/5227/3 "2022-11-15T08:30:13Z")
It is very hard to conclude anything from a non-equilibrium simulation. You should umbrella potentials at fixed locations, or AWH which is more efficient and easier to set up. PS are you using pull-geometry cylinder? ---
**Author:** ![anna.stephens](https://avatars.discourse-cdn.com/v4/letter/a/e274bd/32.png) [@anna.stephens](https://gromacs.bioexcel.eu/u/anna.stephens)\ **Post date:** [November 15, 2022, 8:49am UTC](https://gromacs.bioexcel.eu/t/negative-force-from-pulling-simulation/5227/4 "2022-11-15T08:49:54Z")
Hey, I’m using distance between the COM of the protein and the membrane. Is this the starting step for the umbrella potentials at fixed locations or is it something I would put in the input .mdp file? ; Pull code pull = yes pull\_ncoords = 1 ; only one reaction coordinate pull\_ngroups = 2 ; two groups defining one reaction coordinate pull\_group1\_name = Protein pull\_group2\_name = Membrane pull\_coord1\_type = umbrella ; harmonic potential pull\_coord1\_geometry = distance ; simple distance increase pull\_coord1\_dim = N N Y pull\_coord1\_groups = 1 2 pull\_coord1\_start = yes ; define initial COM distance \> 0 pull\_coord1\_rate = -0.00007 nm per ns pull\_coord1\_k = 1000 ; kJ mol^-1 nm^-2 Thanks so much! Anna ---
**Author:** ![anna.stephens](https://avatars.discourse-cdn.com/v4/letter/a/e274bd/32.png) [@anna.stephens](https://gromacs.bioexcel.eu/u/anna.stephens)\ **Post date:** [November 15, 2022, 10:02am UTC](https://gromacs.bioexcel.eu/t/negative-force-from-pulling-simulation/5227/5 "2022-11-15T10:02:03Z")
![pulling_force](https://europe1.discourse-cdn.com/flex017/uploads/bioexcel1/original/2X/d/d2f8e37db1066e1084deb855c4e43c68e1aaa02e.png) ---
**Author:** ![hess](https://dub1.discourse-cdn.com/flex017/user_avatar/gromacs.bioexcel.eu/hess/32/416_2.png) [@hess](https://gromacs.bioexcel.eu/u/hess)\ **Post date:** [November 16, 2022, 7:32am UTC](https://gromacs.bioexcel.eu/t/negative-force-from-pulling-simulation/5227/6 "2022-11-16T07:32:09Z")
You should not use the distance if you want to pull along the z-axis. Precisely for this case we have pull-geometry cylinder. But if your molecule is a whole protein, it will be very difficult to get converged results.