in MM-PBSA analysis, you are removing all of the explicit water during the
postprocessing. So as long as the original simulations were reasonable,
there is no need to try to match up the exact # water molecules or even the
solvate command cutoff distance.
if the numerical values that you listed (2 columns) are the rmsd values, I
suspect this is an imaging issue. In your periodic system, molecules get
"wrapped" in the periodic box, but you want to put them back together
before calculating rmsd of a complex. Try using the cpptraj autoimage
command prior to the rms command.
On Mon, Apr 1, 2024 at 11:36 AM Yang Wei via AMBER <amber.ambermd.org>
wrote:
> Dear AMBER Community,
>
> I am currently employing MMPBSA with a multiple trajectory approach to
> compute the binding free energy of two systems. The first involves the
> interaction between a protein (123 amino acids/2,047 atoms) and a single
> molecule (132 atoms), while the second entails a protein-protein complex
> (123 amino acids/2,047 atoms and 163 amino acids/2,765 atoms,
> respectively).
>
> 1. For the 1st system, 10 angstrom is used for the solvatebox of the
> complex (protein-molecule), so that 8,938 water (or 26,814 atoms) was
> added. The input for the MD simulation is written as follows:
>
> NVT production 200 ns
> &cntrl
> imin=0, irest=0,
> nstlim=200000000, dt=0.002, ntx=1,
> ntpr=1000, ntwx=1000, ntwr=50000,
> cut=12, ntb=1,
> ntc=2, ntf=2,
> ntt=3, gamma_ln=2.0,
> tempi=300.0, temp0=300.0,
> ioutfm=1, ntwv=-1, ig=-1, iwrap=1
> &end
> &wt
> type='END'
> &end
>
> Currently, I have maintained an equal number of water molecules for the
> complex (protein-molecule), receptor (protein), and ligand (molecule) to
> preserve concentration. However, utilizing 8,938 water molecules for ligand
> (a single molecule) appears excessive and computationally wasteful. I did a
> test by employing 3000 water molecules for ligands and the resulting
> binding energy difference is only around ~1 kcal/mol. My question is that
> whether it is necessary to employ the same number of water molecules for
> all components, and if the implicit model necessitates an equivalent box
> size as the explicit model.
>
> 2. For the second system, a 10 angstrom solvate box was employed for the
> protein-protein complex, resulting in the addition of 19,629 water
> molecules (or 58,887 atoms). Apart from nstlim=400000000, the input for the
> MD simulation remains identical to the first system.
>
> Prior to the MMPBSA analysis, an RMSD analysis was conducted:
>
> parm ../../Ternary_VHL_solv.prmtop
> trajin ../../Ternary_VHL_solv_nvt_prod.netcdf
> reference ../../Ternary_VHL_solv.inpcrd
>
> rms :1-285.N,CA,C reference out rmsd.agr
>
> quit
>
> Upon visualizing the frames (e.g., Frame 92253 and Frame 92254), the
> fluctuations were likely caused by the size of the box.
> .....
> 92251.000 2.9623
> 92252.000 3.0159
> 92253.000 2.9695
> 92254.000 31.2806
> 92255.000 31.2697
> 92256.000 31.3480
> 92257.000 31.2952
> 92258.000 31.3905
> 92259.000 31.4305
> 92260.000 3.0645
> 92261.000 3.0589
> 92262.000 3.0633
> .....
>
> I'm concerned about the potential impact of these fluctuations on the
> MMPBSA analysis and whether enlarging the size of the water box would be
> necessary. This could lead to longer simulation times and require more
> frames for the entropy analysis.
>
> Thank you in advance.
>
> Best,
>
> Yang
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Mon Apr 01 2024 - 10:00:02 PDT
