I solvate the protein system described in the martini tutorial (martini-workshop/02_protein_basics/tutorial.md at main · marrink-lab/martini-workshop · GitHub) and I get the following result:
Output configuration contains 11142 atoms in 10637 residues
Volume : 1300.18 (nm^3)
Density : 2433.53 (g/l)
Number of solvent molecules: 10254
Obviously, the density is out of range. But if you perform a simple calculation to get the density from the box volume and the number of water beads the result is:
with 72g as the mass of a water bead (4 normal individual water molecules).
Mass of water used for solvatation:
10254 X 72 /6.022 X 10^-23 = 1.226 X 10^-18 g
Volume of the box:
1300 X 10^-24 = 1.3 X 10^-21 L
Density:
(1.226 X 10^-18) / (1.3 X 10^-21)= 943 g/L, which is a reasonable value not very far from 1000 g/L considering that the protein mass is not taken into account.
It appears that there is a bug in the calculation of density. Is it important for the following operations of simulation ? In all the cases I’ve tried, the NPT equilibration is not very good with a big RMSD (usually more than 50 bar for a target pressure of 1 bar).
So what’s wrong?
Thanks for your comments