- Posts: 11
structure distorted
- Hongtham
-
Topic Author
- Offline
- Fresh Boarder
Less
More
8 years 6 months ago #5076
by Hongtham
structure distorted was created by Hongtham
HI,
I am running Coarse Grained simulation with a membrane protein. After running about 20 ns (dt = 20 fs) with postion restraint the protein's heavy atom, the system was equilibrated, and there were some water molecules in the trainmembrane domain (the channel pore). But after removing the position restraint, the structure of protein quickly distorted, all sububits shrank and became very tight, no water molecules in the pore anymore. in the mdp file, i had some chance in the Pcoupl type)
This issue didn't happen when I ran the allatom simulation with the same system. I think the water help protein keep stable it structure. (ofcause if I run simulation with the antagonist for a while the channel may close, less water in the pore also)
I wonder this happened because there is some problem with pressure,(the lateral pressure was too big ??)
If anyone has experience about this, please help me or give me some suggestion.
Thank you so much.
Best regards,
Hongtham
The configuration for equibrating the system is below,
coulombtype = reaction field ;
rcoulomb = 1.1 ;
vdw-type = cutoff ;
rvdw-switch = 0.9 ;
rvdw = 1.1 ;
vdw-modifier = Potential-shift-verlet
epsilon_r = 15
epsilon_rf = 0
; temperature control
tcoupl = berendsen ;
tc-grps = Protein DOPC W_WF_NA_CL ;
tau-t = 1.0 1.0 1.0 ;
ref-t = 310 310 310 ;
; pressure control
Pcoupl = berendsen ;
Pcoupltype = semiisotropic ;
tau-p = 5.0 5.0 ;
compressibility = 1e-5 1e-5 ;
ref-p = 1.0 1.0 ;
The configuration for dynamics simulation is below,
Pcoupl = parrinello-rahman
Pcoupltype = semiisotropic
tau-p = 12.0 12.0
compressibility = 3e-4 3e-4
ref-p = 1.0 1.0
I am running Coarse Grained simulation with a membrane protein. After running about 20 ns (dt = 20 fs) with postion restraint the protein's heavy atom, the system was equilibrated, and there were some water molecules in the trainmembrane domain (the channel pore). But after removing the position restraint, the structure of protein quickly distorted, all sububits shrank and became very tight, no water molecules in the pore anymore. in the mdp file, i had some chance in the Pcoupl type)
This issue didn't happen when I ran the allatom simulation with the same system. I think the water help protein keep stable it structure. (ofcause if I run simulation with the antagonist for a while the channel may close, less water in the pore also)
I wonder this happened because there is some problem with pressure,(the lateral pressure was too big ??)
If anyone has experience about this, please help me or give me some suggestion.
Thank you so much.
Best regards,
Hongtham
The configuration for equibrating the system is below,
coulombtype = reaction field ;
rcoulomb = 1.1 ;
vdw-type = cutoff ;
rvdw-switch = 0.9 ;
rvdw = 1.1 ;
vdw-modifier = Potential-shift-verlet
epsilon_r = 15
epsilon_rf = 0
; temperature control
tcoupl = berendsen ;
tc-grps = Protein DOPC W_WF_NA_CL ;
tau-t = 1.0 1.0 1.0 ;
ref-t = 310 310 310 ;
; pressure control
Pcoupl = berendsen ;
Pcoupltype = semiisotropic ;
tau-p = 5.0 5.0 ;
compressibility = 1e-5 1e-5 ;
ref-p = 1.0 1.0 ;
The configuration for dynamics simulation is below,
Pcoupl = parrinello-rahman
Pcoupltype = semiisotropic
tau-p = 12.0 12.0
compressibility = 3e-4 3e-4
ref-p = 1.0 1.0
Please Log in or Create an account to join the conversation.
- mnmelo
-
- Offline
- Admin
Less
More
- Posts: 89
8 years 6 months ago #5077
by mnmelo
Replied by mnmelo on topic structure distorted
Hello Hongtham,
You don't mention which type of protein topology you're using. Is secondary structure kept only through bonded potentials? Or are you using an elastic network?
The elastic network is likely the way to go for your case. If you're already using one, I'd suggest generating a new one with an increased elastic bond cutoff and/or stronger force constant.
In addition to the above, in your .mdp you're temperature-coupling the protein on its own. Don't do it; it has too few degrees of freedom. Instead make a group of DOPC+Protein and couple that.
Good luck,
Manel
You don't mention which type of protein topology you're using. Is secondary structure kept only through bonded potentials? Or are you using an elastic network?
The elastic network is likely the way to go for your case. If you're already using one, I'd suggest generating a new one with an increased elastic bond cutoff and/or stronger force constant.
In addition to the above, in your .mdp you're temperature-coupling the protein on its own. Don't do it; it has too few degrees of freedom. Instead make a group of DOPC+Protein and couple that.
Good luck,
Manel
Please Log in or Create an account to join the conversation.
- Hongtham
-
Topic Author
- Offline
- Fresh Boarder
Less
More
- Posts: 11
8 years 6 months ago - 8 years 6 months ago #5078
by Hongtham
Replied by Hongtham on topic structure distorted
Hi,
I didnt use Elastic network. I have run my system in all atom for a while first, after checking the RMSD of the protein and sure the system get well equilibium, I used martinize.py script to convert the protein from the last frame of this simulation to coarse grained presentation. After that, I embedded my protein to a pre-equilibrated Coarse grained DOPC bilayer. The solvate and ionization were done later. Bythe way, I ran my simulation with Martini forcefield version 2.2 for protein, 2.0 version for lipid and ions.
I saw you recommendation about the temperature-coupling. I don't clearly understand. Can you please give me some explaination? I think we should couple seperately to get better temperature control as the Gromacs manual mentions.
Thank you so much.
Hongtham
I didnt use Elastic network. I have run my system in all atom for a while first, after checking the RMSD of the protein and sure the system get well equilibium, I used martinize.py script to convert the protein from the last frame of this simulation to coarse grained presentation. After that, I embedded my protein to a pre-equilibrated Coarse grained DOPC bilayer. The solvate and ionization were done later. Bythe way, I ran my simulation with Martini forcefield version 2.2 for protein, 2.0 version for lipid and ions.
I saw you recommendation about the temperature-coupling. I don't clearly understand. Can you please give me some explaination? I think we should couple seperately to get better temperature control as the Gromacs manual mentions.
Thank you so much.
Hongtham
Last edit: 8 years 6 months ago by Hongtham. Reason: add more information
Please Log in or Create an account to join the conversation.
- mnmelo
-
- Offline
- Admin
Less
More
- Posts: 89
8 years 6 months ago #5079
by mnmelo
Replied by mnmelo on topic structure distorted
In the martinize.py script you have several options to tune the CG representation of your protein. Try usingas an option to martinize.py (read
this paper
for more information).
As to the temperature coupling:
You are defining your protein as a separate group to thermostat to 310K. A protein has a small number of degrees-of-freedom. Thermostatting those separately will dampen their velocity fluctuations too much. The best is to always couple a solute together with a solvent, as heat will typically transfer efficiently between the two.
In your case your solute is the protein and your solvent (for this purpose) is the membrane. They should be coupled to the same bath. You do this by creating an index group that has both membrane and protein (named, for instance, DOPC_Protein), and then in your .mdp you use it alongside the water and ions:
Finally, for equilibration a compressibility of 4.5e-5 is typically used.
Good luck,
Manel
-ff elnedyn22As to the temperature coupling:
You are defining your protein as a separate group to thermostat to 310K. A protein has a small number of degrees-of-freedom. Thermostatting those separately will dampen their velocity fluctuations too much. The best is to always couple a solute together with a solvent, as heat will typically transfer efficiently between the two.
In your case your solute is the protein and your solvent (for this purpose) is the membrane. They should be coupled to the same bath. You do this by creating an index group that has both membrane and protein (named, for instance, DOPC_Protein), and then in your .mdp you use it alongside the water and ions:
tc-grps = DOPC_Protein W_WF_NA_CL ;
tau-t = 1.0 1.0 ;
ref-t = 310 310 ;Finally, for equilibration a compressibility of 4.5e-5 is typically used.
Good luck,
Manel
Please Log in or Create an account to join the conversation.
- Hongtham
-
Topic Author
- Offline
- Fresh Boarder
Less
More
- Posts: 11
8 years 6 months ago #5080
by Hongtham
Replied by Hongtham on topic structure distorted
Hi,
I got it. Thank you so much :D
Hongtham
I got it. Thank you so much :D
Hongtham
Please Log in or Create an account to join the conversation.
Time to create page: 0.108 seconds