If the box is periodic, it is not sufficient to merely fit the protein-ligand complex into the box; the box must be large enough to ensure that there will be no artificial electrostatic interactions between opposite sides of the protein-ligand complex during the simulation.

Martin Klvana Thank you for your answer. My box size is big enough to fit protein Ligand complex. Its 1000 amino acids protein, i gave 2 Angstrom boundary box and ran simulation for 50ns. But, after simulation I found 0.8 Angstrom is enough to fit protein ligand, so now how can i reduce the boundary box size to 0.8 and extend the simulation to 100ns ??

Yeshwanth Mahesh, as i wrote, it is not sufficient to merely fit the protein-ligand complex into the box . . . If the box is periodic, and if the cutoff for electrostatic interaction is 10 Angstroms, neither 0.8 nor 2 Angstroms is enough to prevent artificial electrostatic interactions from happening. Besides, changing the size of the box in the middle of the simulation can produce artifacts (which is why probably nobody does it -- which is why you may have a hard time finding out how to do it) -- it is a good practice to compare, if possible, systems treated in the same way (which includes the size of the simulated system). If you want to change the size of the box, do it during the preparation of the simulated system, i.e., prior the equilibrating simulation. You will not lose any real time because the analysis of the simulation data, not the simulation itself, is the limiting step in molecular dynamics studies.

the length unit in gromacs is nm, not angstrong.

when you reduce the box dimension, it means the number of particles in your simulation is decreased also so that you can not start the new simulation from what you did, it's dirty

Martin Klvana Thank you very for your suggestion