- Posts: 8
Simulation involving PI34
- cchan2242
-
Topic Author
- Offline
- Fresh Boarder
Less
More
9 years 6 months ago #4118
by cchan2242
Simulation involving PI34 was created by cchan2242
Hi all,
Whenever I do simulation of a lipid bilayer involving PI34 (force-field parameters in martini_v2.0_glycolipids.itp), the system could not even get through the minimization. However when I change the composition to maybe DOPC, DOPE and POPS, in which POPS is also a charged lipid, the system works better (refer to md.chem.rug.nl/cgmartini/index.php/forum...9-segmentation-fault ).
Ultimately I would like to run simulations for protein-above-membrane systems. I will greatly appreciate that if anyone could provide similar experience.
Following is the error message in minimization.log:
Fatal error:
15 of the 25382 bonded interactions could not be calculated because some atoms involved moved further apart than the multi-body cut-off distance (1.4 nm) or the two-body cut-off distance (1.4 nm), see option -rdd, for pairs and tabulated bonds also see option -ddcheck
Following is the minimization parameters:
integrator = steep
dt = 0.01
nsteps = 5000
nstxout = 0
nstvout = 0
nstlog = 100
nstxtcout = 100
xtc-precision = 10
rlist = 1.4
coulombtype = shift
rcoulomb = 1.2
epsilon_r = 15
vdw-type = shift
rvdw-switch = 0.9
rvdw = 1.2
tcoupl = v-rescale
tc-grps = Protein non-Protein
tau-t = 1.0 1.0
ref-t = 300 300
Whenever I do simulation of a lipid bilayer involving PI34 (force-field parameters in martini_v2.0_glycolipids.itp), the system could not even get through the minimization. However when I change the composition to maybe DOPC, DOPE and POPS, in which POPS is also a charged lipid, the system works better (refer to md.chem.rug.nl/cgmartini/index.php/forum...9-segmentation-fault ).
Ultimately I would like to run simulations for protein-above-membrane systems. I will greatly appreciate that if anyone could provide similar experience.
Following is the error message in minimization.log:
Fatal error:
15 of the 25382 bonded interactions could not be calculated because some atoms involved moved further apart than the multi-body cut-off distance (1.4 nm) or the two-body cut-off distance (1.4 nm), see option -rdd, for pairs and tabulated bonds also see option -ddcheck
Following is the minimization parameters:
integrator = steep
dt = 0.01
nsteps = 5000
nstxout = 0
nstvout = 0
nstlog = 100
nstxtcout = 100
xtc-precision = 10
rlist = 1.4
coulombtype = shift
rcoulomb = 1.2
epsilon_r = 15
vdw-type = shift
rvdw-switch = 0.9
rvdw = 1.2
tcoupl = v-rescale
tc-grps = Protein non-Protein
tau-t = 1.0 1.0
ref-t = 300 300
Please Log in or Create an account to join the conversation.
- Clement
-
- Offline
- Admin
Less
More
- Posts: 211
9 years 6 months ago #4119
by Clement
Replied by Clement on topic Simulation involving PI34
You're using too much processors for your minimization, and some bonded interactions are spread over more than one voxel of your domain decomposition. For now you can increase the -rdd/-rcon values to 1.6 and/or decrease the number of CPUs with -nt.
Please Log in or Create an account to join the conversation.
- cchan2242
-
Topic Author
- Offline
- Fresh Boarder
Less
More
- Posts: 8
9 years 6 months ago - 9 years 6 months ago #4121
by cchan2242
Replied by cchan2242 on topic Simulation involving PI34
I have adjusted the no. of processors to 1 and/or -rdd to 1.6 however for this time errors are as follows:
Fatal error:
Too many LINCS warnings (1303)
If you know what you are doing you can adjust the lincs warning threshold in your mdp file
or set the environment variable GMX_MAXCONSTRWARN to -1,
but normally it is better to fix the problem
I have also tried another simulation of only membrane in water and it returns the same error for minimization on LINCS warnings.
Also I am curious why it seems much more "unstable" with the PI34 lipids? Is it because it is highly charged?
Fatal error:
Too many LINCS warnings (1303)
If you know what you are doing you can adjust the lincs warning threshold in your mdp file
or set the environment variable GMX_MAXCONSTRWARN to -1,
but normally it is better to fix the problem
I have also tried another simulation of only membrane in water and it returns the same error for minimization on LINCS warnings.
Also I am curious why it seems much more "unstable" with the PI34 lipids? Is it because it is highly charged?
Last edit: 9 years 6 months ago by cchan2242.
Please Log in or Create an account to join the conversation.
- helgi
-
- Offline
- Administrator
Less
More
- Posts: 54
9 years 6 months ago #4122
by helgi
Replied by helgi on topic Simulation involving PI34
Hello, hello,
So this could indeed be the PI34 lipid, at 10fs it should be stable but maybe depending on the PI34 concentration, how long you equilibrated (with an even lower time step) and other things in your system etc. You can verify if this is the PI34 and then what atoms in the PI34 by checking in your log file what atoms are giving the LINCS warnings.
If this only happens rather rarely you can try changing the .mdp options:
lincs_order = 4
lincs_warnangle = 30
to order of 6 or 8 (will slow down your simulation a tiny bit) and but the warnangle at 90.
If this is still a problem and your simulation is not too sensitive to the orientation of the PI34 head I would recommend removing the dihedral between the head the the lipid tail, this is a weak orientational preference that came from limited testing of atomistic PIPs and will undoubtedly depend on the lipid mixture, PIP concentration etc. etc. so its exact value is most likely not correct in your system. This dihedral makes the PIPs rather computationally unstable so removing it you will have a more stable system (should even work with a time step of 20fs). This was done in the new Plasma Membrane simulations we did - you can find the .itp's in cgmartini.nl/cgmartini/index.php/example...ons2/lipid-membranes
Hope this helps
Cheers,
- Helgi
So this could indeed be the PI34 lipid, at 10fs it should be stable but maybe depending on the PI34 concentration, how long you equilibrated (with an even lower time step) and other things in your system etc. You can verify if this is the PI34 and then what atoms in the PI34 by checking in your log file what atoms are giving the LINCS warnings.
If this only happens rather rarely you can try changing the .mdp options:
lincs_order = 4
lincs_warnangle = 30
to order of 6 or 8 (will slow down your simulation a tiny bit) and but the warnangle at 90.
If this is still a problem and your simulation is not too sensitive to the orientation of the PI34 head I would recommend removing the dihedral between the head the the lipid tail, this is a weak orientational preference that came from limited testing of atomistic PIPs and will undoubtedly depend on the lipid mixture, PIP concentration etc. etc. so its exact value is most likely not correct in your system. This dihedral makes the PIPs rather computationally unstable so removing it you will have a more stable system (should even work with a time step of 20fs). This was done in the new Plasma Membrane simulations we did - you can find the .itp's in cgmartini.nl/cgmartini/index.php/example...ons2/lipid-membranes
Hope this helps
Cheers,
- Helgi
Please Log in or Create an account to join the conversation.
- cchan2242
-
Topic Author
- Offline
- Fresh Boarder
Less
More
- Posts: 8
9 years 6 months ago #4126
by cchan2242
Replied by cchan2242 on topic Simulation involving PI34
Unfortunately the problem is not yet solved. I have studied the .itp's for the Plasma Membrane Simulation and make the following adjustments:
1. I have commented our the dihedral in PI34 but the LINCS warnings remains and all the atom numbers reported point to the PI34 lipids.
2. As POP2 was used in the Plasma Membrane Simulation which only differs from PI34 with a C3A particle to D3A in the lipid tail, I tried simulation with POP2 replacing PI34 for both "membrane in water" and "protein+membrane in water" using the itp for the Plasma Membrane Simulation and both of them failed with LINCS warnings (pointing to POP2 instead).
Looking more close to the LINCS warnings, they all points to particles in the head of PI34 or POP2. What is really wrong with the heads? Do we actually need some special treatments for the highly charged heads?
1. I have commented our the dihedral in PI34 but the LINCS warnings remains and all the atom numbers reported point to the PI34 lipids.
2. As POP2 was used in the Plasma Membrane Simulation which only differs from PI34 with a C3A particle to D3A in the lipid tail, I tried simulation with POP2 replacing PI34 for both "membrane in water" and "protein+membrane in water" using the itp for the Plasma Membrane Simulation and both of them failed with LINCS warnings (pointing to POP2 instead).
Looking more close to the LINCS warnings, they all points to particles in the head of PI34 or POP2. What is really wrong with the heads? Do we actually need some special treatments for the highly charged heads?
Please Log in or Create an account to join the conversation.
- helgi
-
- Offline
- Administrator
Less
More
- Posts: 54
9 years 6 months ago #4129
by helgi
Replied by helgi on topic Simulation involving PI34
In addition to the removed dihedral there is also the bond (atoms 5 6) that was changed from a constraint to a bond (to increase stability), the C to D bead in the tail would have no effect on the stability, just changes the type of lipid tail you use. When you do energy minimization it is better to make all the constraint into bonds but then change it back for the actual simulations. With these plasma membrane PIPs I have not had any problems at 20fs and have simulated for hundreds of microsecond, but of course this can depend on your system (other components, initial minimization, PIP concentration etc) so without actually seeing your system it's hard to say what it going wrong. As suggested before try changing the LINCS options, running at a lower time step should also work, and hopefully after good equilibration you can increase it back up to 20fs, else you will have to figure out which atoms are exploding (PIP head beads, but why, what nocks into them to make them move so fast that the LINCS can't resolve the constraint at a 20fs time step).
Hope this helps,
Good luck,
- Helgi
Hope this helps,
Good luck,
- Helgi
Please Log in or Create an account to join the conversation.
- cchan2242
-
Topic Author
- Offline
- Fresh Boarder
Less
More
- Posts: 8
9 years 6 months ago #4131
by cchan2242
Replied by cchan2242 on topic Simulation involving PI34
I might provide more information about my system for more advice: Currently the system is simply a lipid bilayer simulation in water with ions. If the membrane simulation works, I would like to add a protein above the membrane (so the water box will increase in size). The membrane is composed of DOPC, DOPE, POPS and PI34 (4:3:2:1) and ion concentration is 0.18M. The topology is as follows:
#include "martini_v2.2.itp"
#include "martini_v2.0_lipids.itp"
#include "martini_v2.0_glycolipids.itp"
#include "martini_v2.0_ions.itp"
[ system ]
COMPLEX BILAYER
[ molecules ]
; X: 28.000 (36 lipids) Y: 11.000 (14 lipids)
; 504 lipids in upper leaflet, 504 lipids in lower leaflet
; NDX Solute 1 0
; Charge of protein: 0.000000
; NDX Membrane 1 14056
; Charge of membrane: -700.000000
; Total charge: -700.000000
; NDX Solvent 14057 35493
; NDX System 1 35493
DOPC 201
DOPE 151
POPS 100
PI34 50
DOPC 201
DOPE 151
POPS 100
PI34 50
W 20737
NA+ 700
Another system, in which PI34 is replaced by POP2 (with similar topology), is also mentioned in the above replies. All the systems are constructed using insane.py.
I an wondering why the POP2 membrane does not work either, its topology was directly copied from the plasma membrane simulation so should not it work?
#include "martini_v2.2.itp"
#include "martini_v2.0_lipids.itp"
#include "martini_v2.0_glycolipids.itp"
#include "martini_v2.0_ions.itp"
[ system ]
COMPLEX BILAYER
[ molecules ]
; X: 28.000 (36 lipids) Y: 11.000 (14 lipids)
; 504 lipids in upper leaflet, 504 lipids in lower leaflet
; NDX Solute 1 0
; Charge of protein: 0.000000
; NDX Membrane 1 14056
; Charge of membrane: -700.000000
; Total charge: -700.000000
; NDX Solvent 14057 35493
; NDX System 1 35493
DOPC 201
DOPE 151
POPS 100
PI34 50
DOPC 201
DOPE 151
POPS 100
PI34 50
W 20737
NA+ 700
Another system, in which PI34 is replaced by POP2 (with similar topology), is also mentioned in the above replies. All the systems are constructed using insane.py.
I an wondering why the POP2 membrane does not work either, its topology was directly copied from the plasma membrane simulation so should not it work?
Please Log in or Create an account to join the conversation.
- chaos
-
- Visitor
9 years 4 months ago #4225
by chaos
Replied by chaos on topic Simulation involving PI34
Hi cchan2242,
A colleague just helped me to sort this out. Here is what to do.
Before the minimization, i.e. before creating the tpr, do the following temporary modification in the martini_v2.0_glycolipids.itp file (just go straight to the PI34 definition if this is the only lipid from this file you are interested in):
Initial martini_v2.0_glycolipids.itp:
; DIPALMITOYLPHOSPHATIDILINOSITOL 3-4 BI-PHOSPHATE ( PIP2(3,4) )
[ moleculetype ]
PI34 1
[ atoms ]
1 P1 1 PI34 C1 1 0 72.000
2 Na 1 PI34 C2 2 0 72.000
...
[ bonds ]
; 1 2 1 0.4 30000
; 1 3 1 0.4 30000
; 2 3 1 0.4 30000
1 4 1 0.35 1250
4 7 1 0.47 1250
7 8 1 0.37 1250
7 13 1 0.47 1250
13 14 1 0.47 1250
14 15 1 0.47 1250
15 16 1 0.47 1250
8 9 1 0.47 1250
9 10 1 0.47 1250
10 11 1 0.47 1250
11 12 1 0.47 1250
; 2 5 1 0.3 30000
; 2 6 1 0.35 30000
; 1 5 1 0.40 25000
; 3 6 1 0.31 30000
; 5 6 1 0.60 25000
....
[ constraints ]
1 2 1 0.4
1 3 1 0.4
2 3 1 0.4
2 5 1 0.3
2 6 1 0.35
1 5 1 0.40
3 6 1 0.31
5 6 1 0.60
Modified for minimization martini_v2.0_glycolipids.itp:
; DIPALMITOYLPHOSPHATIDILINOSITOL 3-4 BI-PHOSPHATE ( PIP2(3,4) )
[ moleculetype ]
PI34 1
[ atoms ]
1 P1 1 PI34 C1 1 0 72.000
2 Na 1 PI34 C2 2 0 72.000
...
[ bonds ]
1 2 1 0.4 30000
1 3 1 0.4 30000
2 3 1 0.4 30000
1 4 1 0.35 1250
4 7 1 0.47 1250
7 8 1 0.37 1250
7 13 1 0.47 1250
13 14 1 0.47 1250
14 15 1 0.47 1250
15 16 1 0.47 1250
8 9 1 0.47 1250
9 10 1 0.47 1250
10 11 1 0.47 1250
11 12 1 0.47 1250
2 5 1 0.3 30000
2 6 1 0.35 30000
1 5 1 0.40 25000
3 6 1 0.31 30000
5 6 1 0.60 25000
....
[ constraints ]
; 1 2 1 0.4
; 1 3 1 0.4
; 2 3 1 0.4
; 2 5 1 0.3
; 2 6 1 0.35
; 1 5 1 0.40
; 3 6 1 0.31
; 5 6 1 0.60
Hope this helps
Maria
A colleague just helped me to sort this out. Here is what to do.
Before the minimization, i.e. before creating the tpr, do the following temporary modification in the martini_v2.0_glycolipids.itp file (just go straight to the PI34 definition if this is the only lipid from this file you are interested in):
Initial martini_v2.0_glycolipids.itp:
; DIPALMITOYLPHOSPHATIDILINOSITOL 3-4 BI-PHOSPHATE ( PIP2(3,4) )
[ moleculetype ]
PI34 1
[ atoms ]
1 P1 1 PI34 C1 1 0 72.000
2 Na 1 PI34 C2 2 0 72.000
...
[ bonds ]
; 1 2 1 0.4 30000
; 1 3 1 0.4 30000
; 2 3 1 0.4 30000
1 4 1 0.35 1250
4 7 1 0.47 1250
7 8 1 0.37 1250
7 13 1 0.47 1250
13 14 1 0.47 1250
14 15 1 0.47 1250
15 16 1 0.47 1250
8 9 1 0.47 1250
9 10 1 0.47 1250
10 11 1 0.47 1250
11 12 1 0.47 1250
; 2 5 1 0.3 30000
; 2 6 1 0.35 30000
; 1 5 1 0.40 25000
; 3 6 1 0.31 30000
; 5 6 1 0.60 25000
....
[ constraints ]
1 2 1 0.4
1 3 1 0.4
2 3 1 0.4
2 5 1 0.3
2 6 1 0.35
1 5 1 0.40
3 6 1 0.31
5 6 1 0.60
Modified for minimization martini_v2.0_glycolipids.itp:
; DIPALMITOYLPHOSPHATIDILINOSITOL 3-4 BI-PHOSPHATE ( PIP2(3,4) )
[ moleculetype ]
PI34 1
[ atoms ]
1 P1 1 PI34 C1 1 0 72.000
2 Na 1 PI34 C2 2 0 72.000
...
[ bonds ]
1 2 1 0.4 30000
1 3 1 0.4 30000
2 3 1 0.4 30000
1 4 1 0.35 1250
4 7 1 0.47 1250
7 8 1 0.37 1250
7 13 1 0.47 1250
13 14 1 0.47 1250
14 15 1 0.47 1250
15 16 1 0.47 1250
8 9 1 0.47 1250
9 10 1 0.47 1250
10 11 1 0.47 1250
11 12 1 0.47 1250
2 5 1 0.3 30000
2 6 1 0.35 30000
1 5 1 0.40 25000
3 6 1 0.31 30000
5 6 1 0.60 25000
....
[ constraints ]
; 1 2 1 0.4
; 1 3 1 0.4
; 2 3 1 0.4
; 2 5 1 0.3
; 2 6 1 0.35
; 1 5 1 0.40
; 3 6 1 0.31
; 5 6 1 0.60
Hope this helps
Maria
Please Log in or Create an account to join the conversation.
Time to create page: 0.104 seconds