Gromacs with coordNum colvar in protein-lig binding SMD

User discussions
1

GROMACS version: Gromacs 2021 with colvars
GROMACS modification: Yes
I am trying to run an SMD using ‘coordNum’ colvar to slowly increase number of contacts between protein and ligand. Following is my “colvar.dat” to run using “gmx mdrun -s smd.tpr -deffnm smd -colvars colvar.dat”

The simulation runs but the first step itself shows number of contacts of 400 which is not correct. I made sure in the gro file that ligand not closer to the protein at all (~20 A away).

Following is my colvar.dat content.

colvar {
name CNUM
width 1.0
outputAppliedForce on
outputVelocity on
coordNum {
cutoff 4.0
group1 { atomNumbers 141 142 144 145 147 149 150 153 154 155 158 160 172 173 175 176 178 179 180 181 184 185 187 188 191 193 }
group2 { atomNumbers 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 }
}
}

harmonic {
colvars CNUM
centers 0
targetCenters 200
forceConstant 1
targetNumSteps 500000
outputAccumulatedWork on
}

Any help in figuring out why the ligand is not moving towards the given protein atoms but the *.traj output show 400 contacts from the beginning?

Thanks,

2

Hi @jubilantsundar, in what unit is your cutoff? It looks like it may be Å, but for GROMACS it should be nm.

Does that explain the discrepancy?

Giacomo

3

Hi @giacomo.fiorin,

Thanks for pointing that out. It did explain the discrepancy. Now, I removed the cutoff value since the default value makes sense to me.
How about the units for the forceconstant? What is it? What number makes sense for protein-ligand binding SMD using 3 atoms from each in Gromacs setup?

Thanks

4

The force constant’s unit would be consistent, i.e. you have 1 kJ/mol/nm^2 that may be a bit low, but I don’t know the details of your problem. See the brief discussion here.

The only recommendation that is generally valid is to begin testing the static restraint before you move it continuously (SMD) or step-wise (umbrella sampling).

Giacomo

5

Thanks. Increasing forceconstant bring the transition expected,
I was thinking of the units in pN as I am coming from NAMD use experience.

Thanks for helping quickly.