METAD using Torsion angle is causing Gromacs to throw fatal LINCS errors

[distribut...@googlemail.com]

Dear PLUMED users, 

My non-plumed simulation has been running for twelve hours on my institute's cluster with no struggle at all. Unfortunately, upon using plumed 2.0.2 (I can't change the version any time soon I'm afraid) on the same Gromacs .tpr compilation I generate LINCS errors within 1100 ps of simulation time. 

I am trying to perform metadynamics on system which comprises of two proteins in a water solution. Protein A is an enzyme and protein B is its substrate. I would like protein B to rotate along its long axis (it is cylindrical in shape) in the catalytic site of enzyme A. If you wrap your hand into a fist around a finger on your other hand, and then you keep turn that finger around - that is the action I need. I am trying to isolate any primary/secondary binding sites and any region which would be unlikely to be involved in binding. 

I am not sure what kind of METAD parameters I could use, so I am testing with a torsion angle. The first two atoms belong to the substrate and the second two atoms belong to the enzyme as depicted below:

                             A4

                               | (enzyme)

                             A3 (A2 and A3 are almost parallel with one another)

--------------------------

A2 

 | (substrate)

A1

I am hoping this would encourage the two proteins to rotate around one another. I am NOT taking into account the possibility that they may drift away (translate) AND rotate at the same time. I'll leave that for another day. My input for PLUMED is:

ang: TORSION ATOMS=6603,7068,1252,1791

METAD ...

LABEL=metad

ARG=ang

PACE=500

HEIGHT=1.2

SIGMA=0.35

FILE=HILLS

GRID_MIN=-pi

GRID_MAX=pi

GRID_BIN=63

... METAD

PRINT STRIDE=10 ARG=ang,metad.bias FILE=COLVAR

Unfortunately this generates the following LINCS errors:

Not all bonded interactions have been properly assigned to the domain decomposition cells

A list of missing interactions:

               Angle of  13348 missing      3

         Proper Dih. of  19969 missing      9

               LJ-14 of  19201 missing     21

          exclusions of 188104 missing     27

Molecule type 'CollagenB'

the first 10 missing interactions, except for exclusions:

               LJ-14 atoms  261  276           global  6587  6602

               LJ-14 atoms  273  278           global  6599  6604

               LJ-14 atoms  274  276           global  6600  6602

               LJ-14 atoms  276  278           global  6602  6604

               LJ-14 atoms  276  279           global  6602  6605

               LJ-14 atoms  276  280           global  6602  6606

               LJ-14 atoms  278  281           global  6604  6607

               LJ-14 atoms  278  282           global  6604  6608

Molecule type 'CollagenC'

the first 10 missing interactions, except for exclusions:

         Proper Dih. atoms  259  261  263  264 global  7066  7068  7070  7071

               LJ-14 atoms  259  264           global  7066  7071

               Angle atoms  261  263  264      global  7068  7070  7071

               LJ-14 atoms  261  267           global  7068  7074

               LJ-14 atoms  261  271           global  7068  7078

         Proper Dih. atoms  262  261  263  264 global  7069  7068  7070  7071

               LJ-14 atoms  262  264           global  7069  7071

               Angle atoms  264  263  265      global  7071  7070  7072

               Angle atoms  264  263  269      global  7071  7070  7076

         Proper Dih. atoms  264  263  265  266 global  7071  7070  7072  7073

-------------------------------------------------------

Program mdrun_mpi_d, VERSION 4.6.5

Source code file: /home/ccspapp/Scratch/gromacs+plumed/gromacs-4.6.5/src/mdlib/domdec_top.c, line: 393

Fatal error:

60 of the 266897 bonded interactions could not be calculated because some atoms involved moved further apart than the multi-body cut-off distance (0.8 nm) or the two-body cut-off distance (0.8 nm), see option -rdd, for pairs and tabulated bonds also see option -ddcheck

For more information and tips for troubleshooting, please check the GROMACS

website at http://www.gromacs.org/Documentation/Errors

-------------------------------------------------------

I would be extremely grateful to any insight in order to make this work.

Kind regards

Anthony

[davide branduardi]

Hi, 

This may be some problems in the metadynamics parameters.

Can you try to remove the grid for the initial run and see if you have the same behaviour? If so, then try to play with metadynamics parameters (the height-the higher the more instable the system is- and width-the smaller the more instable the system is ). If not, then play with grid parameters (the GRID_BIN).

Another issue might be that the force from your TORSION is concentrated on 4 atoms only. You might want to use geometric centers of subgroup to spread the force more uniformly.

HTH

D

[distribut...@googlemail.com]

Hi Davide

Thanks for the reply. I'll give all of those a try and see what happens.

Thanks again

Anthony 

[distribut...@googlemail.com]

I've made a number of changes to the sigma parameter. As I lowered the value the simulation would run for longer before crashing with a LINCS error. My non-plummed simulation is still running, approximately 26 ns and still going.

I'm going to change the height too and drop sigma even further:

SIGMA=0.25

HEIGHT=0.55

If this fails I will try changing the individual atoms for groups. 

Thanks

Anthony

[davide branduardi]

Beware, the smaller SIGMA is the less stable your system  will be (since you will have a very narrow gaussian that can change very rapidly and add looooooots of force on your system and booom!!!! Lincs crashes!)

D

******************************************************
Davide Branduardi
davide.b...@gmail.com
Web:
https://sites.google.com/site/davidebranduardi
******************************************************

--
You received this message because you are subscribed to the Google Groups "PLUMED users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to plumed-users...@googlegroups.com.
To post to this group, send email to plumed...@googlegroups.com.
Visit this group at http://groups.google.com/group/plumed-users.
To view this discussion on the web visit https://groups.google.com/d/msgid/plumed-users/dd2c2dc7-52f1-46bc-ae1c-b5d16f70b057%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

[distribut...@googlemail.com]

Thanks Davide, I've been doing PMF calcs for a good few years now but Meta dynamics is very new to me.

Thanks again

Anthony