- Posts: 211
Making the itp file in New Molecule Parameterizing
- yipy0005
-
Topic Author
- Visitor
9 years 9 months ago #3985
by yipy0005
Making the itp file in New Molecule Parameterizing was created by yipy0005
In the tutorial for parameterizing novel molecules, I noticed that the [ bonds ], [ angles ] & [ dihedrals ] for toluene are not labelled with anything. Instead the [ constraints ] are labelled with some bond lengths. Therefore, I have a few question to ask:
1. What's the difference between [ bonds ] & [ constraints ]?
2. I know the functional form for Martini for [ bonds ] is 1, how about [ angles ] and [ dihedrals ]?
Thanks, any help is really appreciated. =)
1. What's the difference between [ bonds ] & [ constraints ]?
2. I know the functional form for Martini for [ bonds ] is 1, how about [ angles ] and [ dihedrals ]?
Thanks, any help is really appreciated. =)
Please Log in or Create an account to join the conversation.
- Clement
-
- Offline
- Admin
Less
More
9 years 9 months ago #3986
by Clement
Replied by Clement on topic Making the itp file in New Molecule Parameterizing
Bonds = harmonic (or something else) link between two particles: requires a length and a force constant
Constraint = fixed distance between two particles: requires only a length
I would argue that your second question is more GROMACS oriented, and would point at the manual ("Bonded interactions", pp. 73 of the GROMACS 4.6.4 manual) to avoid any mistakes I could make explaining that...
If nothing is defined in [ bonds ], [ angles ], etc. it just means you don't need any specific descriptors to respect the correct geometry (phase space) visited by the molecule...
Constraint = fixed distance between two particles: requires only a length
I would argue that your second question is more GROMACS oriented, and would point at the manual ("Bonded interactions", pp. 73 of the GROMACS 4.6.4 manual) to avoid any mistakes I could make explaining that...
If nothing is defined in [ bonds ], [ angles ], etc. it just means you don't need any specific descriptors to respect the correct geometry (phase space) visited by the molecule...
Please Log in or Create an account to join the conversation.
- yipy0005
-
Topic Author
- Visitor
9 years 9 months ago #3987
by yipy0005
Replied by yipy0005 on topic Making the itp file in New Molecule Parameterizing
Okie, then I wish to ask would you recommend that the [ angles ] & [ dihedrals ] be unspecified if I'm not familiar or unsure with the angles of the beads in my CG model?
Please Log in or Create an account to join the conversation.
- Clement
-
- Offline
- Admin
Less
More
- Posts: 211
9 years 9 months ago #3988
by Clement
Replied by Clement on topic Making the itp file in New Molecule Parameterizing
I'm not really sure of the meaning of your question... Confusing.
1) Are you asking if you should leave the empty [ bonds ], [ angles ] and [ dihedrals ] sections in the topology? Nothing is defined there, it's perfectly useless. The relevant part of the topology for toluene is:
Rest is unimportant. Leave it or delete it, won't change a thing.
2) If you're asking if you should *not* define any value if you don't know the equilibrium length/angle, I'd say no. This is the trial and error part we mention in the tutorial. For instance, define an angle if you think the conformation occupied by three beads should be constrained to a certain phase space (if the distribution of values extracted from your atomistic simulation is limited to a certain range), fill the required fields with the equilibrium value (the center of your distribution), and a first rough guess at a force constant (check other topologies to have an idea of the magnitude of this latter). Run a simulation with your CG model, extract the distribution of angular conformations visited by the corresponding beads and compared it to your atomistic one. If it matches without modifying other bonds/dihedrals/etc. defined in the molecule, you're done; if not, change force constant and do that last part again. Remember you can *try* things to have a feel at how your CG model behaves; once you understand how defining an angle in a part of your molecule affect the rest, you'll get how to continue your parameterization.
3) General comment: try to be slightly more precise in your questions please... And make sure the matter is Martini related and not some generalities about GROMACS syntax.
1) Are you asking if you should leave the empty [ bonds ], [ angles ] and [ dihedrals ] sections in the topology? Nothing is defined there, it's perfectly useless. The relevant part of the topology for toluene is:
[ moleculetype ]
TOL 1
[ atoms ]
1 C4 1 TOL RI1 1 0
2 SC4 1 TOL RI2 2 0
3 SC4 1 TOL RI3 3 0
[constraints]
1 2 1 0.248
2 3 1 0.191
3 1 1 0.248 Rest is unimportant. Leave it or delete it, won't change a thing.
2) If you're asking if you should *not* define any value if you don't know the equilibrium length/angle, I'd say no. This is the trial and error part we mention in the tutorial. For instance, define an angle if you think the conformation occupied by three beads should be constrained to a certain phase space (if the distribution of values extracted from your atomistic simulation is limited to a certain range), fill the required fields with the equilibrium value (the center of your distribution), and a first rough guess at a force constant (check other topologies to have an idea of the magnitude of this latter). Run a simulation with your CG model, extract the distribution of angular conformations visited by the corresponding beads and compared it to your atomistic one. If it matches without modifying other bonds/dihedrals/etc. defined in the molecule, you're done; if not, change force constant and do that last part again. Remember you can *try* things to have a feel at how your CG model behaves; once you understand how defining an angle in a part of your molecule affect the rest, you'll get how to continue your parameterization.
3) General comment: try to be slightly more precise in your questions please... And make sure the matter is Martini related and not some generalities about GROMACS syntax.
Please Log in or Create an account to join the conversation.
- yipy0005
-
Topic Author
- Visitor
9 years 9 months ago #3989
by yipy0005
Replied by yipy0005 on topic Making the itp file in New Molecule Parameterizing
Oops, sorry sorry.! I apologize for raising vague questions that are seemingly more related to GROMACS.! Nonetheless, your 2nd point is the answer I'm looking for. Parameterizing a novel molecule is really hard work.! I totally salute you all for coming up with the MARTINI concept as well as the parameterization for various molecules. =)
Please Log in or Create an account to join the conversation.
Time to create page: 0.102 seconds