- Posts: 416
Protein in water using Elnedyn
- bt12d026
-
Topic Author
- Visitor
10 years 9 months ago #1774
by bt12d026
Replied by bt12d026 on topic Protein in water using Elnedyn
Sir, How to change the side chain bead masses in Protein.itp alone? (without changing in Martini_v2.2.itp)
Please Log in or Create an account to join the conversation.
- xavier
-
- Offline
- Admin
Less
More
10 years 9 months ago #1775
by xavier
In gromacs the definition of molecule is as follows. You can add the mass of a particle after its charge.
[ atoms ]
; nr type resnr residu atom cgnr charge mass
Replied by xavier on topic Protein in water using Elnedyn
bt12d026 wrote: Sir, How to change the side chain bead masses in Protein.itp alone? (without changing in Martini_v2.2.itp)
In gromacs the definition of molecule is as follows. You can add the mass of a particle after its charge.
[ atoms ]
; nr type resnr residu atom cgnr charge mass
Please Log in or Create an account to join the conversation.
- Rajat Desikan
-
- Offline
- Expert Boarder
Less
More
- Posts: 90
10 years 6 months ago - 10 years 6 months ago #2477
by Rajat Desikan
Replied by Rajat Desikan on topic Protein in water using Elnedyn
Hi All,
I have a similar problem with my protein. I used the following command to get my topology (latest martinize.py)
./martinize.py -f 8-mer.pdb -o LE_I_elnedyn.top -x LE_I.pdb -ss 8-mer.dssp -p All -ff elnedyn22
1) HIS and TRP spring constants are too high:
628 629 1 0.19500 94000 ; HIS
74 75 1 0.25500 73000 ; TRP
Should I set both to 60000? I get the following warning from grompp
"Note4: The bond in molecule-type Protein_A between atoms 628 BB and 629 SC1 has an estimated oscillational period of 1.1e-01 ps, which is less than 10 times the time step of 2.0e-02 ps. Maybe you forgot to change the constraints mdp option."
2) I changed all the S-bead masses in martini_v2.2.itp to 72. For cholesterol, I changed the mass in martini_v2.0_cholesterol.itp back to 45 :)
3) A previous suggestion was
"Converted the constraints between backbone beads (in the section [constraints]) to bonds with a high force constant. The block of backbone-backbone constraints has been shifted to the [bonds] section and appended a column containing force constants, ~90000 kJ/mol."
Since I am interested in protein stability in a variety of membranes, I am worried about doing this step. Am I artificially stiffening the protein backbone beyond normal by doing this? Is there any way of matching this/tuning force constants with fully atomistic simulations? (I have a 100 ns AA simulation of the same protein in membrane)...
I have a similar problem with my protein. I used the following command to get my topology (latest martinize.py)
./martinize.py -f 8-mer.pdb -o LE_I_elnedyn.top -x LE_I.pdb -ss 8-mer.dssp -p All -ff elnedyn22
1) HIS and TRP spring constants are too high:
628 629 1 0.19500 94000 ; HIS
74 75 1 0.25500 73000 ; TRP
Should I set both to 60000? I get the following warning from grompp
"Note4: The bond in molecule-type Protein_A between atoms 628 BB and 629 SC1 has an estimated oscillational period of 1.1e-01 ps, which is less than 10 times the time step of 2.0e-02 ps. Maybe you forgot to change the constraints mdp option."
2) I changed all the S-bead masses in martini_v2.2.itp to 72. For cholesterol, I changed the mass in martini_v2.0_cholesterol.itp back to 45 :)
3) A previous suggestion was
"Converted the constraints between backbone beads (in the section [constraints]) to bonds with a high force constant. The block of backbone-backbone constraints has been shifted to the [bonds] section and appended a column containing force constants, ~90000 kJ/mol."
Since I am interested in protein stability in a variety of membranes, I am worried about doing this step. Am I artificially stiffening the protein backbone beyond normal by doing this? Is there any way of matching this/tuning force constants with fully atomistic simulations? (I have a 100 ns AA simulation of the same protein in membrane)...
Last edit: 10 years 6 months ago by Rajat Desikan. Reason: Small error in typing
Please Log in or Create an account to join the conversation.
- xavier
-
- Offline
- Admin
Less
More
- Posts: 416
10 years 6 months ago #2509
by xavier
Replied by xavier on topic Protein in water using Elnedyn
The original ElNeDyn has been parameterised to reproduce the internal dynamics of proteins and this led to parameters among which a 150.000 kJ/mol/nm^2 force constant between backbone beads. This make BB-BB bonds stiff, which they are in an atomistic simulation. The value of this force constant does not matter ... the force constant in the elastic network is much more important.
In the latest versions of martinize we have tried to made ElNeDyn better but we introduced a couple of modifications that together make the model much less stable.
We now advise to switch the BB-BB constrains back into conventional bonds with a 150.000 kJ/mol/nm^2 force constant.
You should not need to change the force constant for the TRP and HIS. They become unstable when constrains are used on the backbone.
The masses need to be 72 for the S type particles.
In the latest versions of martinize we have tried to made ElNeDyn better but we introduced a couple of modifications that together make the model much less stable.
We now advise to switch the BB-BB constrains back into conventional bonds with a 150.000 kJ/mol/nm^2 force constant.
You should not need to change the force constant for the TRP and HIS. They become unstable when constrains are used on the backbone.
The masses need to be 72 for the S type particles.
Rajat Desikan wrote: Hi All,
I have a similar problem with my protein. I used the following command to get my topology (latest martinize.py)
./martinize.py -f 8-mer.pdb -o LE_I_elnedyn.top -x LE_I.pdb -ss 8-mer.dssp -p All -ff elnedyn22
1) HIS and TRP spring constants are too high:
628 629 1 0.19500 94000 ; HIS
74 75 1 0.25500 73000 ; TRP
Should I set both to 60000? I get the following warning from grompp
"Note4: The bond in molecule-type Protein_A between atoms 628 BB and 629 SC1 has an estimated oscillational period of 1.1e-01 ps, which is less than 10 times the time step of 2.0e-02 ps. Maybe you forgot to change the constraints mdp option."
2) I changed all the S-bead masses in martini_v2.2.itp to 72. For cholesterol, I changed the mass in martini_v2.0_cholesterol.itp back to 45 :)
3) A previous suggestion was
"Converted the constraints between backbone beads (in the section [constraints]) to bonds with a high force constant. The block of backbone-backbone constraints has been shifted to the [bonds] section and appended a column containing force constants, ~90000 kJ/mol."
Since I am interested in protein stability in a variety of membranes, I am worried about doing this step. Am I artificially stiffening the protein backbone beyond normal by doing this? Is there any way of matching this/tuning force constants with fully atomistic simulations? (I have a 100 ns AA simulation of the same protein in membrane)...
Please Log in or Create an account to join the conversation.
- Rajat Desikan
-
- Offline
- Expert Boarder
Less
More
- Posts: 90
10 years 6 months ago #2514
by Rajat Desikan
Replied by Rajat Desikan on topic Protein in water using Elnedyn
Thank you, Xavier.
"We now advise to switch the BB-BB constraints back into conventional bonds with a 150.000 kJ/mol/nm^2 force constant. You should not need to change the force constant for the TRP and HIS. They become unstable when constrains are used on the backbone."
If I understood this right, I should manually copy all the BB-BB constraints into the bonds section with a force constant of 150000...
I have a few doubts.
1) In the bonds section, I have the following header
[ bonds ]
#ifdef RUBBER_BANDS
#ifndef RUBBER_FC
#define RUBBER_FC 500.000000
#endif
Should I change this to:
[ bonds ]
#ifdef RUBBER_BANDS
#ifndef RUBBER_FC
#define RUBBER_FC 150000.000000
#endif
2) I should not tinker with this at all, right?
; Sidechain bonds
1 2 1 0.31000 2800 ; MET
...
3) In constraints, there seem to be two types.
Type I
[ constraints ]
1 3 1 0.37319 ; MET(C)-ALA(C)
3 4 1 0.37378 ; ALA(C)-GLU(1)
4 6 1 0.37691 ; GLU(1)-ILE(1)
...
Are these the constraints that I should copy to [bonds]?
Type II: constraints on individual amino acids
8 9 1 0.20000 ; VAL
11 12 1 0.25500 ; ASP
16 17 1 0.19500 ; THR
...
And these should remain in the constraints section, right?
I am sorry for being naive. I just do not want to get anything wrong. Thanks for your time and patience!
"We now advise to switch the BB-BB constraints back into conventional bonds with a 150.000 kJ/mol/nm^2 force constant. You should not need to change the force constant for the TRP and HIS. They become unstable when constrains are used on the backbone."
If I understood this right, I should manually copy all the BB-BB constraints into the bonds section with a force constant of 150000...
I have a few doubts.
1) In the bonds section, I have the following header
[ bonds ]
#ifdef RUBBER_BANDS
#ifndef RUBBER_FC
#define RUBBER_FC 500.000000
#endif
Should I change this to:
[ bonds ]
#ifdef RUBBER_BANDS
#ifndef RUBBER_FC
#define RUBBER_FC 150000.000000
#endif
2) I should not tinker with this at all, right?
; Sidechain bonds
1 2 1 0.31000 2800 ; MET
...
3) In constraints, there seem to be two types.
Type I
[ constraints ]
1 3 1 0.37319 ; MET(C)-ALA(C)
3 4 1 0.37378 ; ALA(C)-GLU(1)
4 6 1 0.37691 ; GLU(1)-ILE(1)
...
Are these the constraints that I should copy to [bonds]?
Type II: constraints on individual amino acids
8 9 1 0.20000 ; VAL
11 12 1 0.25500 ; ASP
16 17 1 0.19500 ; THR
...
And these should remain in the constraints section, right?
I am sorry for being naive. I just do not want to get anything wrong. Thanks for your time and patience!
Please Log in or Create an account to join the conversation.
- xavier
-
- Offline
- Admin
Less
More
- Posts: 416
10 years 6 months ago #2517
by xavier
Replied by xavier on topic Protein in water using Elnedyn
Ok. I was not clear.
You need to change the backbone-backbone constrains into bonds with a force constant value of 150000 kJ/mol/nm^2.
You do have to do it by hand. Go to the constrain section. The type I you find is the one. Backbone-backbone distance are around 0.38 nm. You need to copy these and move them to [bonds] section. Use the same format of other bonds.
Your point 1) relates to the elastic network used to control the conformation of the protein. Do NOT modify this section.
Your point 2) it is about side chains ... you need to modify backbon-backbone distance ... so you are correct do NOT modify them.
Your point 3) type I is the one you need to modify.
In addition, you may have trouble with HIS backbone-side chain bond. That happens. You'd need to turn the bond into a constrain for more stability.
You need to change the backbone-backbone constrains into bonds with a force constant value of 150000 kJ/mol/nm^2.
You do have to do it by hand. Go to the constrain section. The type I you find is the one. Backbone-backbone distance are around 0.38 nm. You need to copy these and move them to [bonds] section. Use the same format of other bonds.
Your point 1) relates to the elastic network used to control the conformation of the protein. Do NOT modify this section.
Your point 2) it is about side chains ... you need to modify backbon-backbone distance ... so you are correct do NOT modify them.
Your point 3) type I is the one you need to modify.
In addition, you may have trouble with HIS backbone-side chain bond. That happens. You'd need to turn the bond into a constrain for more stability.
Rajat Desikan wrote: Thank you, Xavier.
"We now advise to switch the BB-BB constraints back into conventional bonds with a 150.000 kJ/mol/nm^2 force constant. You should not need to change the force constant for the TRP and HIS. They become unstable when constrains are used on the backbone."
If I understood this right, I should manually copy all the BB-BB constraints into the bonds section with a force constant of 150000...
I have a few doubts.
1) In the bonds section, I have the following header
[ bonds ]
#ifdef RUBBER_BANDS
#ifndef RUBBER_FC
#define RUBBER_FC 500.000000
#endif
Should I change this to:
[ bonds ]
#ifdef RUBBER_BANDS
#ifndef RUBBER_FC
#define RUBBER_FC 150000.000000
#endif
2) I should not tinker with this at all, right?
; Sidechain bonds
1 2 1 0.31000 2800 ; MET
...
3) In constraints, there seem to be two types.
Type I
[ constraints ]
1 3 1 0.37319 ; MET(C)-ALA(C)
3 4 1 0.37378 ; ALA(C)-GLU(1)
4 6 1 0.37691 ; GLU(1)-ILE(1)
...
Are these the constraints that I should copy to [bonds]?
Type II: constraints on individual amino acids
8 9 1 0.20000 ; VAL
11 12 1 0.25500 ; ASP
16 17 1 0.19500 ; THR
...
And these should remain in the constraints section, right?
I am sorry for being naive. I just do not want to get anything wrong. Thanks for your time and patience!
Please Log in or Create an account to join the conversation.
- Rajat Desikan
-
- Offline
- Expert Boarder
Less
More
- Posts: 90
10 years 5 months ago #2600
by Rajat Desikan
Replied by Rajat Desikan on topic Protein in water using Elnedyn
Thank you so much, Xavier. Superbly detailed reply.
Please Log in or Create an account to join the conversation.
- Rajat Desikan
-
- Offline
- Expert Boarder
Less
More
- Posts: 90
10 years 5 months ago - 10 years 5 months ago #2625
by Rajat Desikan
Replied by Rajat Desikan on topic Protein in water using Elnedyn
Hi,
Following up on the previous suggestions, I changed the BB-BB constraints to bonds as below:
;FROM CONSTRAINTS
1 3 1 0.37319 150000; MET(C)-ALA(C)
3 4 1 0.37378 150000; ALA(C)-GLU(1)
4 6 1 0.37691 150000; GLU(1)-ILE(1)
But I get the following warning from grompp:
"The bond in molecule-type Protein_A between atoms 1 BB and 3 BB has an estimated oscillational period of 9.7e-02 ps, which is less than 5 times the time step of 2.0e-02 ps. Maybe you forgot to change the constraints mdp option."
Should I decrease the timestep/decrease the spring constant? Or do I ignore the warning?
Thanks,
Following up on the previous suggestions, I changed the BB-BB constraints to bonds as below:
;FROM CONSTRAINTS
1 3 1 0.37319 150000; MET(C)-ALA(C)
3 4 1 0.37378 150000; ALA(C)-GLU(1)
4 6 1 0.37691 150000; GLU(1)-ILE(1)
But I get the following warning from grompp:
"The bond in molecule-type Protein_A between atoms 1 BB and 3 BB has an estimated oscillational period of 9.7e-02 ps, which is less than 5 times the time step of 2.0e-02 ps. Maybe you forgot to change the constraints mdp option."
Should I decrease the timestep/decrease the spring constant? Or do I ignore the warning?
Thanks,
Last edit: 10 years 5 months ago by Rajat Desikan.
Please Log in or Create an account to join the conversation.
- xavier
-
- Offline
- Admin
Less
More
- Posts: 416
10 years 5 months ago #2626
by xavier
Replied by xavier on topic Protein in water using Elnedyn
THis fine ... just ignore that warning ... you are on the good track.
Rajat Desikan wrote: Hi,
Following up on the previous suggestions, I changed the BB-BB constraints to bonds as below:
;FROM CONSTRAINTS
1 3 1 0.37319 150000; MET(C)-ALA(C)
3 4 1 0.37378 150000; ALA(C)-GLU(1)
4 6 1 0.37691 150000; GLU(1)-ILE(1)
But I get the following warning from grompp:
"The bond in molecule-type Protein_A between atoms 1 BB and 3 BB has an estimated oscillational period of 9.7e-02 ps, which is less than 5 times the time step of 2.0e-02 ps. Maybe you forgot to change the constraints mdp option."
Should I decrease the timestep/decrease the spring constant? Or do I ignore the warning?
Thanks,
Please Log in or Create an account to join the conversation.
- Rajat Desikan
-
- Offline
- Expert Boarder
Less
More
- Posts: 90
10 years 5 months ago #2627
by Rajat Desikan
Replied by Rajat Desikan on topic Protein in water using Elnedyn
Thanks, Xavier!
Please Log in or Create an account to join the conversation.
Time to create page: 0.106 seconds