[gmx-users] "Too many LINCS warnings" in a minimization after solvation with coarse-grained waters

[ZHANG Cheng]

Dear Fotis and Peter,


Thank you very much for the help.


Fotis, Can I modify the mdp file to use "soft" potential modifier, how to do that?


I think my problem is not the first reason (i.e. something wrong with the system structure or topology), because the potential is decreasing for the minimization step after inserting 10 protein molecules
https://github.com/lanselibai/martini/blob/master/20190121_LINCS/minimization%20after%20insert%2010%20proteins.jpg
$ Steepest Descents converged to machine precision in 4251 steps,
$ but did not reach the requested Fmax < 1.
$ Potential Energy  = -2.4130119e+05
$ Maximum force     =  9.1535597e+00 on atom 2335
$ Norm of force     =  7.1063030e-01





Peter, how to replace all constraints for stiff bonds?





------------------ Original ------------------
From:  "ZHANG Cheng"<272699575 at qq.com>;
Date:  Tue, Jan 22, 2019 05:58 AM
To:  "gromacs.org_gmx-users"<gromacs.org_gmx-users at maillist.sys.kth.se>;

Subject:  Re: "Too many LINCS warnings" in a minimization after solvation with coarse-grained waters



I also tried to reduce the "emtol" gradually in the mdp file, i.e. from 1000 to 100 to 10. It passed the "emtol = 1000" but it stopped again at "emtol = 100", i.e. outputting dozens of pdb files before the "LINCS warnings".


Then I looked at the edr files.
The potential in "emtol = 1000" and "emtol = 100" runs were actually converging
https://github.com/lanselibai/martini/blob/master/20190121_LINCS/emtol%201000%20then%20100.png


My understanding for the "LINCS warnings" is, the system is not stable. But why the potential is still converging?


Do I need to adjust the "lincs warning threshold", or "set the environment variable GMX_MAXCONSTRWARN to -1"? How to do that?


Is there a "standard" mdp file for minimization for a coarse-grained system with 10 proteins in water?
I am using this, but I do not know how to modify it.
https://github.com/lanselibai/martini/blob/master/20190121_LINCS/minimization_solvate.mdp




------------------ Original ------------------
From:  "ZHANG Cheng"<272699575 at qq.com>;
Date:  Tue, Jan 22, 2019 00:12 AM
To:  "gromacs.org_gmx-users"<gromacs.org_gmx-users at maillist.sys.kth.se>;

Subject:  "Too many LINCS warnings" in a minimization after solvation with coarse-grained waters



I am doing coarse-grained (CG) modelling for 10 proteins in a box. I was told "Too many LINCS warnings" in the minimization after solvation with coarse-grained waters.


I try to diagnose the problems based on 
http://manual.gromacs.org/documentation/2018/user-guide/terminology.html#blowing-up


My procedure is


1) A single CG-protein was firstly minimized in vacuum, no problem


2) Then 10 of this protein were inserted to a box, followed by a minimization. It "stopped because the algorithm tried to make a new step whose size was too small, or there was no change in the energy since last step." So I think this minimization is also successful.


3) The system was then solvated by
$ gmx solvate -cp 10_noW_minimized.gro -cs water-box-CG_303K-1bar.gro -radius 0.21 -o system-solvated.gro -p system.top


4) Then the solvated system is minimized by
$ gmx grompp -f minimization_solvate.mdp -c system-solvated.gro -p system.top -o system-min-solvent.tpr
$ gmx mdrun -deffnm system-min-solvent -v -c system-min-solvent.gro

PDB structures were outputted from step 327 to step 710, and it stopped due to the "LINCS warnings".


The "minimization_solvate.mdp" is here
https://github.com/lanselibai/martini/blob/master/20190121_LINCS/minimization_solvate.mdp


The "system-min-solvent.log" is here
https://github.com/lanselibai/martini/blob/master/20190121_LINCS/system-min-solvent.log


So I think the system with 10 proteins in vacuum is okay (right?). But when CG-water is added, it got problem? How to modify my system? Let me know if you need other information. Thank you.


Cheng