Whenever I try to use the gmx editconf command, I am confused about the -d option (minimum distance between protein surface and box edge).
How should we decide the correct value for -d? Do we need to select it manually (e.g., 1.0 nm, 1.2 nm, 1.5 nm), or does GROMACS automatically choose it?
Earlier, my supervisor mentioned that GROMACS would show an error if the -d value was too small, and then we would manually increase it until no overlap occurred. However, in the recent versions, no such error appears even when I try smaller or larger -d values.
Currently, I am preparing a box that consists of tetramers, but I am confused about what -d value would be appropriate to use.
What is the recommended -d value for a system containing tetramers of specific protein segments (to avoid periodic image interactions but without making the box unnecessarily large)?
Usually i keep distance as 1 , that would be enough to make sure no clashes happen, you can always check the box.gro file in vmd and check whether its clashing or not. If it clashes change it to 1.2 in the command. Increasing the box size will ultimately increase the amount of water molecules it can accumulate, hence it slows the simulation.
On what basis we determine the shape of box i.e cubic , triclinic, dodecahedron as I have read somewhere that by seeing the gro file i.e the less no.of atom present we considered that shape is it true
The -d flag introduces some spacing so that what you put in the box is not too close to its own periodic image. A (very rough) rule of thumb is to use at least the electrostatics of the cut-off of the force field you are using, generally between 1.0 and 1.2 nm. However, the structure you are inserting in the box might be quite far from the structure during dynamics, especially if you have flexible loops and regions that can extend and diffuse around a lot. So the choice of -d largely depends on your system, it can be easily much larger than the cut-off of the force field.
After creating box, it will show the volume of it in the last line. Usually from what i did, i find dodecahedron is best as it takes much lesser volumes compared to cubic. But it differs from protein to protein. You can check different shapes and see the volume it makes and pick the best one from there