GROMACS version: 2024.06
GROMACS modification: Yes/No
Hello, everyone! I have conduct a simulation of pulling a protein in the specific direction, and my pull seeting of mdp file is as below. But after a short time of simulation, the program ends:
Distance between pull groups 1 and 2 (6.633696 nm) is larger than 0.49 times
the box size (6.633633).
You might want to consider using “pull-geometry = direction-periodic” instead.
I visualize the end structure, found the protein is extremely far away from the starting location, which didn’t obey my expectation. And i check the pullx file and found the pull rate was almost 8 nm/628 ps, which was bigger than my settings.
So why would this happen? Dose the setting of pull-coord1-rate have no influence on the pulling process？Are there any methods to control the pull rate?
In addition, how to set a virtual point as the reference group?
It’s difficult for me to guess what went wrong here. But your setup of pulling along a particular direction in 3D space looks very strange. This might cause issues, maybe in combination with periodic boundary conditions.
And your force constant is very small for pulling a protein.
Thank you for your reply! The pulling direction was setted by measuring the helix axial vector of the protein. Is there any fault? What’s the suitable parameter of the force constant? Thank you for your help in advance!
My protein molecule contains 52 residues, which doesn’t seem very large to me. During the simulation, the proteins do not rotate, but will shift sideways perpendicular to helix axial. How should I choose the reference group? I have tried smaller k values such as 250, 750, etc. In this case, the protein will rotate. When k is set to 1000, no rotation will occur, but a lateral shift will occur. How to solve this shift? Thank you very much for your reply
It’s not about the size of the protein, but of the COM-groups.
You can (should) not choose a vector in 3D-space, unless you fix the orientation of your molecule(s).
A force constant of 250 is likely too small, 750 or 1000 is the same. The standard deviation of the thermal spread you get in the RC is sqrt(kT/fc). This is 0.1 nm for 250. That might be ok, but the spring will extend a lot when pulling.
My simultaion system contains a complex (including protein, colored in green) and another small protein (about 52 residues, colored in red), the small protein has a typical helix and its another partial region has uncovalent interaction with the complex in the interface. So my research goal is to pull the small protein along the helix direction, and then elucidate the positional and conformational change of the region of complex which interacts with the small protein.
If any information is required, please inform me! Thank you for your reply in advance!
The best is always if you can define a local COM in the large protein to define a distance from.
Use a vector is problematic. If you do this with respect to the COM of the protein, you complex might start to rotate. If you do it with an absolute reference (use pull group 0), you whole complex might move along.
I suppose you currently have issues because the distance vectors between the two COMs is nearly perpendicular to the pull direction.
1）When conducting the pulling force, the small protein moves, but the protein’s helix distorts and becomes the loop. I am not sure whether the k value is too big or not. Or any other reasons?
2）How sould i define the reference group? If i define the reference group within the big protein, the pulling direction is not parallel with the distance vector, so may i choose another reference group such as an virtual atom site?
The force constant is often not so critical, but shouldn’t be too small, as you will then extend the spring instead of moving along the reaction coordinate. So larger is generally better (up to a limit).
The reference group should be chosen such that the distance between the groups is somewhat similar to the direction you want to move in. It should not be too small, as then there is a risk of local deformation.
Thank you for your suggestion! I take the distance as the pulling method and choose one particular amino acid as the reference group. And the mdp settings are as below:
After a period of simulation, i analyze the gro file and find the small protein which was pulled deformed (the red is the original structure and the purple is the final structure of the small protein). How should i set the parameters to avoid the deformation of protein? Thank you for your reply!
Reaction coordinates can be difficult. You can apply restraints or guide the system more, but the more you guide, the more the answer will be what you put in. You will have to find a balance here. I would suggest to search for literature on similar projects.
