Why huge amount of SOL added during gmx genion in some cases?

User discussions
, , , ,
1

GROMACS version: 5.1.4
i have two protein structures (PDBs), with same sequence of amino acids hommology analysis was performed using different software.
when i want to run MD, during solvate and genion both have extremely different amount of solvent added to the simulation box. After checking the log files of MD run higher sol PDB mdrun was extremely slow (three times slower).
How can i solve this issue? how to control the solvent added and perform faster mdrun

protein 1
(MD run : 5.08 (ns/day)
[ molecules ]
; Compound #mols
Protein_chain_A 1
SOL 40005
NA 112
CL 134

protein 2
( MD run: 17.540 (ns/day)

[ molecules ]
; Compound #mols
Protein 1
SOL 20602
NA 60
CL 82

2

Are the box sizes similar in box cases?

Try using gmx editconf with “-d” to specify distance between protein and box boundary.

Another potential issue could be that the two homology models may have differences between the two structures. This could arise due to different template structures used for homology modelling. For example, an open and closed ion channel will have the same amino acid sequence but different structurrs. Is the RMSD between the 2 structures small enough?

Regards,
Raman

3

Indeed, the physical dimensions of the protein matter here, not the number of residues. An elongated state of the same protein vs. a compact one will drive the size of the box and how many water molecules are added to the box.

4

thank you for your explanation.

i did use this command
gmx_mpi editconf -f filename.pdb -o filename.pdb -d 1.2 gmx_mpi editconf -f filename.pdb -o filename.pdb -d 1.2

now i am receiving this error
DD load balancing is limited by minimum cell size in dimension Y

does this means my MD performance is affected or should i ignore this error. how to fix this issues．

5

Is the RMSD between the 2 structures small enough?

yes you were right. they did look similar upon super imposed few domains are quiet different.
Executive: RMSD = 4.700 (375 to 375 atoms)
yet for a small seq of 191aa, SOL 40005 did seem very peculiar.

6

thank you for clarifying. i just checked my log files and saw this , what this means?

NOTE: DLB can now turn on, when beneficial
At step 2200 the performance loss due to force load imbalance is 2.1 %
NOTE: Turning on dynamic load balancing
Writing checkpoint, step 32800 at Fri Sep 2 16:20:30 2022
DD load balancing is limited by minimum cell size in dimension Y
DD step 49999 vol min/aver 0.742! load imb.: force 52.7%