I did a molecular dynamics simulation of 50 molecules of Tween 80 using GAFF to determine the density for 100 ns.
But, after MD production, the plot of volume, density, Box-X, and Box-Y data against time (ps) continues to decrease or shrink. It is also seen in the trajectory that the simulation box continues to shrink so that the molecules became closer together.
The system is evolving in response to the topology and initial conditions. Your molecules want to condense, either because they were initially too far apart or because the attractions within and between the molecules cause them to collapse inward. The target in this case should be whatever the actual density is, and if it is incorrect after a sufficiently long simulation (one in which the volume/density actually reach a stable value), it suggests that your topology is not correct and you will have to refine it.
Thanks a lot Mr.Lemkul for the explanation.
For more details, What part I must refine in the topology?
What comparison should be used to ensure that the topology is correct?
Is there any software or method that could confirm my topology?
Please help me Sir…
For your information, topol.top as the topology file is generated by AmberTools21 and Acpype. When I saw it, it is like there is no something wrong in there.
Here is my topology file topol.top (72.2 KB)
Still falling short, now you need to tell us how exactly you created your topology in AMBER. Was it via tleap? Which tleap commands? How did you obtain the charges? Which force field(s)?
It’s really hard to help without critical informations. Never ever assume we are familiar with your system, we are familiar with our tools, if much!
Funny thing is that looking at your topol it seems you used only GAFF (or GAFF2?). And if it’s only GAFF, why didn’t you use acpype -i ...?
You didn’t say what are your molecules. Are they lipids? You didn’t show how did you run in Gromacs. Did you add water? How does your system looks after 100 ns simulation?
Well, just saying, but the more information you give, the greater is your chance to get help.
Thank you for the detailed question, Sir…
Yes, I created the topology with AnteChamber21 via tleap with AMBER99SB protein force field (leaprc.ff99SB).
Here is the tleap commands:
I have tried to use acpype -i, but I got a lot of error messages which told me can’t to do so.
So, I tried to created .prmtop file first used AnteChamber21, passed it through acpype, and the topology is created.
As far as I know, Tween 80 is emulsifier or surfactant derived from polyethoxylated sorbitan and oleic acid (fatty acid) which appeared as liquid. So, technically is kind of lipid. I want to see the density of it as early stage. So, I could confirm that my simulation is represent the experiment. For do that, I did not add water.
and here is my system looks like after 100 ns simulation
According to that snapshot, 50 molecule of Tween 80 divided into 2 groups on the box edges which is unrealistic since Tween 80 is a liquid not a gas. Also, the box size became smaller than before.
Is the information I have provided sufficient?
Please correct me if I’m wrong, Sir…thank you for helping me…
You observed the formation of a droplet of Tween80 in the gas phase. There are not two groups; you are simply seeing the effects of periodic boundary conditions.
Since you are essentially throwing a few molecules into the gas phase, naturally you will see compression of the box under the influence of a barostat. The initial conditions are nowhere close to the environment in which Tween80 normally exists (a liquid or an aqueous solution).
Regarding the topology generation, trying to assign parameters to such a large molecule is an inappropriate approach. I suspect you will get different charges assigned to equivalent groups and you will essentially have a geometry-dependent solution. Large molecules need to be broken down into simple model compounds that represent the constituent chemical moieties. You can also get a general idea of the quality of the topology by comparing against the same (or at least similar) groups that are already parametrized for the force field. There’s nothing terribly unusual about the chemical composition of Tween80 that would make such a comparison difficult.
Thank you for your detailed explanation, Sir.
but, I still confused about something.
How Tween 80 could be in the gas phase since I want to observe it as a droplet in liquid phase?
Is it a coincidence or did something cause it?
and How to break down the Tween 80 into simple model?
Look at your initial condition - several molecules scattered in empty space. That’s essentially modeling the molecules floating around in vacuum until they find each other. If you want to study the liquid phase, you need a far more concentrated starting point. If you want to form a droplet in water, well, you need to add water :)
Look up some force field parametrization papers - what do the authors use as models for amino acid sidechains, phospholipids, etc? This will give you the right approach. We don’t parametrize massive molecules, we parametrize pieces and link them together.
Thank you so much for your explanation Mr. Lemkul. Now I have a view on what I will do in the next try. If there is something I confuse, maybe I will ask a question again to you. Again, thank you, Sir…
Thank you so much Mr. Wilter for the solution. I will try those immediately. If I confuse about something regarding to simulation setup or other things, maybe I will ask a question again to you. Once more, thank you, Sir…
Actually, your and @jalemkul’s answer made my day and keep me up since this project for my bachelor thesis…thank you very much…thank you…
Tween-80 topology parameters are already reported in literature. There are a few all-atom simulations looking at Tween-80 micellization or surfactant properties. Those may be a good point to start.
Hi @alanwilter , I have seen this thread before, but I cannot find it anymore. Was it removed from the thread?
Besides, was this problem solved? I’m facing the same issues and I realized that the semiempirical SQM calculation takes too long to finish for big molecules, such as an oligomer. Is that the source of the problem? Is that any way of circumventing it?
