Congiugate gradient gromacs

User discussions
1

I 'm running MD Simulation for a complex (protein+ligands), but I have a great difficult to use conjugate gradient minimization.
First of all, I prepared my input files using CHARMM-GUI server and then I ran energy minimization using steepest descent (1000 and 10 emtol) without problems.When I try to generate tpr using mdp file for a conjugate gradient I have this:
WARNING 1 [file topol-mod.top, line 56]:
The following macros were defined in the ‘define’ mdp field with the -D
prefix, but were not used in the topology:
FLEXIBLE
If you haven’t made a spelling error, either use the macro you defined,
or don’t define the macro

If I delete section define=-DFLEXIBLE , I can do my tpr and go on with conjugate gradient.minimization but I don’t know if it is correct delete define in the mdp file because

I see In GROMACS conjugate gradient can not be used with constraints, including the SETTLE algorithm for water as this has not been implemented. If water is present it must be of a flexible model, which can be specified in the mdp file by define = -DFLEXIBLE .(Energy Minimization — GROMACS 2021.5 documentation.

My questions are:

  1. Why the section “define=-DFLEXIBLE” is a problem for a systems using charmm-gui server?
  2. Is it correct to delete define=-DFLEXIBLE when we want use conjugate gradient?
    Best Regards
    Anna
2

CHARMM-GUI does not write water topologies that are compatible with using a flexible representation. You will find no #ifdef FLEXIBLE statement anywhere, hence you cannot turn off constraints. TIP3P was meant to be rigid, so CHARMM-GUI assumes the user will use it that way.

According to the documentation, no, you can’t do this. But in any case, I’ve never had any luck with conjugate gradients even working, let alone minimizing better than steepest descents. L-BFGS is another option that is occasionally useful.

3

thank you for your reply, I’m not having luck with congiugate gradient ,me neither! But If I want to do normal mode analysis is it necessary to do congiugate gradient for minimization or not?

4

Is it correct ,in my mdp file for congiugate gradient ,to remain the line:
define = -DPOSRES -DPOSRES_FC_BB=400.0 -DPOSRES_FC_SC=40.0 ?
It is as default of charmm-gui mdp minimization but it consideres steepest algoritm. It is the same for congiugate gradient?

5

The restraint scheme CHARMM-GUI uses assumes a rather simple minimization approach that is suitable for conventional MD that does not have such strict requirements on the maximum force. Restraints are probably going to interfere with full relaxation of the system below your target threshold. So while they might be applied in early stages with steepest descent, I don’t see how they would benefit any subsequent steps with CG.