How to apply COM pull on molecules of the same type?

GROMACS version: 2021.3
GROMACS modification: No

There are n molecules of type SOL in a certain GRO file. These n molecules are randomly distributed in the box.

I want to apply a constant force on all molecules of the same type, I am following this but it requires setting an index or so, and appears to be too complicated.

I just want it as simple as:

pull-name = SOL

then it pulls the Center of Mass of all the SOL molecules

How to achieve this?

You can use accelerate groups instead:
accelerate = SOL

Hi Hess, Thank you for the response, I checked the accelerate parameter but the unit is in nm/ps^2 which seems to be extremely high? Does it accept 1E-14 input for mixed precision gromacs?

nm/ps^2 is not extremely high. 1e-14 is extremely low. You will not see any effect with such a low value, as this lead to forces that are are a factor 10^16 smaller than the typical forces between atoms.

I tried converting it to m/s^2 and its about: 1.0e15 m / (s^2), its quite high for experimenting neutral species. There are several for charged particle though, but these are from Particle Accelerators. Apologies for citing Wiki :’(

Anyway, does gromacs accept input for accelerate as low as 1e-14?

Hi organicchemistry_01,

unless you are going to simulate for seconds (not in terms of CPU time), you are never going to see anything move when applying an acceleration of just 1e-14 nm/ps^2. That is the order of magnitude of gravity, and as you may know gravity is most of the times neglected when studying nanoscopic systems because all other forces are orders of magnitude larger.

I suggest to first ask yourself: how much (or how fast) do I want the COM to move in the span of my simulation time (realistically measured in nanoseconds)? And then choose the acceleration field accordingly. Comparison to typical macroscopic acceleration values is meaningless.

I see, so COM pull will be like FAST FORWARD, so I can use it for scaling time or process duration vs force too?

Would you be kind to share a reference that does force/time scaling?

I would eventually use this to compare the gromacs result with an actual force pull.

I am not 100% sure about what you mean by FAST FORWARD, but it seems to me we have two different things in mind.

My observation was mainly addressing your concerns about using an acceleration on the order of nm/ps^2.

I would eventually use this to compare the gromacs result with an actual force pull.

Can you expand on that? What are you simulating in practice?

I am simulating a force applied, so we can use the accelerate parameter for this. Example would be with AFM, SEM, and Ultra-centrifuge. For now, I am interested in force created by ultracentrifuge to specific molecules/polymer chain, I wish to see “sedimentation” happening via gromacs

I think the timescales you are thinking of a are far off. I assume sedimentation would happen on the order of seconds or (much) more, whereas MD simulations can reach microseconds. Thus you need to apply a many orders of magnitude larger force to see something happen in an MD simulation.