>I will be running protein in explicit water simulations and need a timestep >of 2fs. Forces and energies in the system will be analyzed. What restraint >setup should I choose? I think no critical errors would occur if you use shake or lincs for water in your case.