normal How to set the bond force constant?

  • Leliel
  • Leliel's Avatar Topic Author
  • Visitor
10 years 5 months ago #2529 by Leliel
How to set the bond force constant? was created by Leliel
HI,i am undergraduate student, learning Martini FF.
When i use and read the martinize.py , i found that a series of parameters :

#RES# BEADS BONDS ANGLES DIHEDRALS
# BB-SC SC-SC BB-SC-SC SC-SC-SC
"DA": [spl("SNa SNa SP1 SNa"),
[(0.120,5000),(0.220,5000),(0.320,5000),(0.420,5000),(0.520,5000),],
[(10.0, 100),(20.0, 100),(30.0, 100),(40.0, 100),(50.0, 100),(60.0, 100),(70.0, 100),(80.0, 100)],
[(10.0, 100, 10),(20.0, 100, 10),(30.0, 100, 10),(40.0, 100, 10)],
[(50.0, 100)]],

AND i think they're for nucleic acid, but I cant figure out what's the meaning of 0.120,0.220...?

can you explain each number in one [] for me? i know that 5000/100 are the force constant.

BTW, i had read the paper of MARTINI for protein, but i can't find how you set the force constant. you said the bonded sections (length,angle,dihedral angles,)are obtained by the distribution of the PDB, but I dont know how to get the force constant.

Thanks in advance!

Please Log in or Create an account to join the conversation.

More
10 years 5 months ago #2530 by jaakko
Replied by jaakko on topic How to set the bond force constant?
Hi Leliel,

I've answered this before here . In short, the script has dummy parameters to test that the script works for DNA but those should not be used for anything, they're just placeholders. (The script does state this if you try to run it with DNA input.) We are working on those DNA parameters and once they're done they'll be added to the website and a new version of the script.

As to your other question, the force constants come from fitting the widths of those distributions.

Please Log in or Create an account to join the conversation.

More
10 years 5 months ago #2534 by xavier
Replied by xavier on topic How to set the bond force constant?
Here to complement Jaakko,

Leliel wrote: HI,i am undergraduate student, learning Martini FF.
When i use and read the martinize.py , i found that a series of parameters :

#RES# BEADS BONDS ANGLES DIHEDRALS
# BB-SC SC-SC BB-SC-SC SC-SC-SC
"DA": [spl("SNa SNa SP1 SNa"),
[(0.120,5000),(0.220,5000),(0.320,5000),(0.420,5000),(0.520,5000),],
[(10.0, 100),(20.0, 100),(30.0, 100),(40.0, 100),(50.0, 100),(60.0, 100),(70.0, 100),(80.0, 100)],
[(10.0, 100, 10),(20.0, 100, 10),(30.0, 100, 10),(40.0, 100, 10)],
[(50.0, 100)]],

AND i think they're for nucleic acid, but I cant figure out what's the meaning of 0.120,0.220...?

These numbers are the reference distance for the bonds, the distance you want the bond to fluctuate around

can you explain each number in one [] for me? i know that 5000/100 are the force constant.

The numbers are listed for all the bonds defined in the bond section.

BTW, i had read the paper of MARTINI for protein, but i can't find how you set the force constant. you said the bonded sections (length,angle,dihedral angles,)are obtained by the distribution of the PDB, but I dont know how to get the force constant.

The procedure goes like this:
1- you extract the distribution of bond(s) from a source you chose, PDB structures or atomistic simulation, or experimental source.
2- you run a CG simulation with a first topology and you determine the distribution of the same bonds and compare to the "Experimental" ones.
3- you adjust your CG topology to get a better match with the "experimental" distributions. This last step is repeated until you reach the best match possible ... CG topologies are limited in resolution so you won't get all your distributions perfectly.

Thanks in advance!

Please Log in or Create an account to join the conversation.

  • Leliel
  • Leliel's Avatar Topic Author
  • Visitor
10 years 5 months ago #2550 by Leliel
Replied by Leliel on topic How to set the bond force constant?

The procedure goes like this:
1- you extract the distribution of bond(s) from a source you chose, PDB structures or atomistic simulation, or experimental source.
2- you run a CG simulation with a first topology and you determine the distribution of the same bonds and compare to the "Experimental" ones.
3- you adjust your CG topology to get a better match with the "experimental" distributions. This last step is repeated until you reach the best match possible ... CG topologies are limited in resolution so you won't get all your distributions perfectly.


So in the first procedure, you mentioned a distribution of bond(s) in atomistic simulation/PDB files, and compared them with the CG simulation ,but i think it's not the best way to adjust the constant: cause the angle A-B-C in Atomistic file cant be exactly the angle beadA-beadB-beadC in coarse grained file. So i think i think it's properly applicable to use a script to transfer the AT files --> CG files in advance and do a statistic work on the CG files.

I'm not sure if i put my point in a clear way ,and do you have the script to transfer a fine grained DNA to a coarse grained one? If you do , it's downloadable in this website just like the martinize.py??

thanks!

Please Log in or Create an account to join the conversation.

More
10 years 5 months ago #2552 by xavier
Replied by xavier on topic How to set the bond force constant?
You are correct I forgot to mention that the atomistic data should be mapped into the a CG trajectory to be able to extract the appropriate distributions.

I am not sure the DNA is already available in the current script. If so it would be in the martinize.py script. If not you'd have to be patient.

Leliel wrote:

The procedure goes like this:
1- you extract the distribution of bond(s) from a source you chose, PDB structures or atomistic simulation, or experimental source.
2- you run a CG simulation with a first topology and you determine the distribution of the same bonds and compare to the "Experimental" ones.
3- you adjust your CG topology to get a better match with the "experimental" distributions. This last step is repeated until you reach the best match possible ... CG topologies are limited in resolution so you won't get all your distributions perfectly.


So in the first procedure, you mentioned a distribution of bond(s) in atomistic simulation/PDB files, and compared them with the CG simulation ,but i think it's not the best way to adjust the constant: cause the angle A-B-C in Atomistic file cant be exactly the angle beadA-beadB-beadC in coarse grained file. So i think i think it's properly applicable to use a script to transfer the AT files --> CG files in advance and do a statistic work on the CG files.

I'm not sure if i put my point in a clear way ,and do you have the script to transfer a fine grained DNA to a coarse grained one? If you do , it's downloadable in this website just like the martinize.py??

thanks!

Please Log in or Create an account to join the conversation.

More
10 years 5 months ago #2559 by xavier
Replied by xavier on topic How to set the bond force constant?
So, here is the comment of Jakkoo on the DNA stuff in martinze.py: there are a few things related to DNA in the current martinize.py but it has been placed there some time ago in preparation of more stuff but it can NOT be used as it is ... many details are still missing.

Please Log in or Create an account to join the conversation.

More
1 year 9 months ago #9414 by yogi@martini
Replied by yogi@martini on topic How to set the bond force constant?
Is there a mathematical curve fitting technique to get bond constant and bond average bond length from the bond distribution obtained from atomistic to cg mapped data. I am asking for approximate way.

Please Log in or Create an account to join the conversation.

Time to create page: 0.100 seconds