normal Philosophy of the martinilize

  • Leliel
  • Leliel's Avatar Topic Author
  • Visitor
10 years 8 months ago #1804 by Leliel
Philosophy of the martinilize was created by Leliel
Hi,i am vary interested in martini FF and want to martinilise the DNA in my own philosophy.but i cant know how to get info in the atomic pdb files and change them into the cg .gro and .top files.So who can tell me how to do this? not a script but the core method.

Thanks a lot!

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

More
10 years 7 months ago #1806 by xavier
Replied by xavier on topic Philosophy of the martinilize
In Maritini we often use the center of mass of the atoms that are enclosed in the CG bead as the position of the CG bead.

Please let us know if you get a reliable model of DNA based on Martini.

Leliel wrote: Hi,i am vary interested in martini FF and want to martinilise the DNA in my own philosophy.but i cant know how to get info in the atomic pdb files and change them into the cg .gro and .top files.So who can tell me how to do this? not a script but the core method.

Thanks a lot!

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

  • Leliel
  • Leliel's Avatar Topic Author
  • Visitor
10 years 7 months ago #1807 by Leliel
Replied by Leliel on topic Philosophy of the martinilize
Thanks a lot!
And is it right that once you settle down the position of every CG bead ,just use them to calculate the bond distance,angle,dihedral angle ...these parameters?

then how to parameterize the connection constant?like some k ?

If we get some progress use this model,i will surely inform you about this.

Again Thanks!

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

More
10 years 7 months ago #1808 by xavier
Replied by xavier on topic Philosophy of the martinilize
You may want to read one of the papers describing the construction of a model!

You would run a atomistic model simulation to extract the force constant of the bonds/angle/dihedrals ...

Leliel wrote: Thanks a lot!
And is it right that once you settle down the position of every CG bead ,just use them to calculate the bond distance,angle,dihedral angle ...these parameters?

then how to parameterize the connection constant?like some k ?

If we get some progress use this model,i will surely inform you about this.

Again Thanks!

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

  • Leliel
  • Leliel's Avatar Topic Author
  • Visitor
10 years 7 months ago #1809 by Leliel
Replied by Leliel on topic Philosophy of the martinilize
Well,I have read most of your published paper,and on that it said that the nonbond interaction are parameterized with some thermodynamical datas and bond interaction are parameterized with the AT MD structure,but the papers shows little about the details of how to do this...

I am an undergraduate student and these questions may be stupid ,but can you tell the procedure of constructing a martini model??

Thanks

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

More
10 years 7 months ago #1810 by xavier
Replied by xavier on topic Philosophy of the martinilize
There are two book chapters out there where the procedure is explained.

You have to use an iterative procedure in which you will refine your topology at each step. You first need to collect all the experimental data you can on the molecule you want to parameterise. Partitioning data (between different environments) is fundamental. You then want (if possible) to run an atomistic simulation of the extract bonded terms. From this all you can build your first CG topology and run an MD simulation from which you can extract data to compare to the data you collected ... it is simple after you go through the process a few times :))

good luck.

Leliel wrote: Well,I have read most of your published paper,and on that it said that the nonbond interaction are parameterized with some thermodynamical datas and bond interaction are parameterized with the AT MD structure,but the papers shows little about the details of how to do this...

I am an undergraduate student and these questions may be stupid ,but can you tell the procedure of constructing a martini model??

Thanks

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

  • Leliel
  • Leliel's Avatar Topic Author
  • Visitor
10 years 7 months ago #1811 by Leliel
Replied by Leliel on topic Philosophy of the martinilize
Am i right to explain the procedure in this way?
1.define the position of CG beads(use the collected beads' center of mass)
2.use the position of CG beads to calculate the distance of bonds,angles,dihedral angels.
3.compare these parameters to the experimental datas(or AT MD structure) using the iterative way to refine them.

that may be similar in the nonbond-interaction paramaters ?

And pls the name of book or chapter (or the website of the book)so i can get it and learn by myself.thanks in advance!

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

More
10 years 7 months ago #1812 by xavier
Replied by xavier on topic Philosophy of the martinilize
What you describe for the bonded parameters is indeed correct. It would be similar for the non-bonded but the data would be partitioning data.

The book chapters I mention are:
- S.J. Marrink, M. Fuhrmans, H.J. Risselada, X. Periole. The MARTINI force field. In "Coarse graining of condensed phase and biomolecular systems", G. Voth ed., CRC press, Chapter 2, 2008.
- X. Periole, S.J. Marrink. The Martini coarse-grained force field. In "Methods in molecular biology", Vol 924, L. Monticelli & E. Salonen Eds., Springer, 2013, pp 533-565.

You can find a reprint on the martini webpage (cgmartini.nl) in the publication section.

Leliel wrote: Am i right to explain the procedure in this way?
1.define the position of CG beads(use the collected beads' center of mass)
2.use the position of CG beads to calculate the distance of bonds,angles,dihedral angels.
3.compare these parameters to the experimental datas(or AT MD structure) using the iterative way to refine them.

that may be similar in the nonbond-interaction paramaters ?

And pls the name of book or chapter (or the website of the book)so i can get it and learn by myself.thanks in advance!

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

  • Leliel
  • Leliel's Avatar Topic Author
  • Visitor
10 years 7 months ago #1817 by Leliel
Replied by Leliel on topic Philosophy of the martinilize
Thanks!i have already read these two chapters and get to know the whole 3step-recipe of building a new topology.But i still couldn't find out by what programs or tools you make these steps. Like which script or tool you will use to calculate the center of mass position or the distribution ?

thx!

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

More
10 years 7 months ago #1821 by xavier
Replied by xavier on topic Philosophy of the martinilize
The resolution transformation is done using the reverse transformation implemented in gromacs ...

You have to get the tutorial and go through it. It will show you how to build atomistic and CG topologies to use in combination to go from an atomistic to a CG trajectory. This way you can extract the bonded terms (bonds, angles etc).

Leliel wrote: Thanks!i have already read these two chapters and get to know the whole 3step-recipe of building a new topology.But i still couldn't find out by what programs or tools you make these steps. Like which script or tool you will use to calculate the center of mass position or the distribution ?

thx!

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

  • Leliel
  • Leliel's Avatar Topic Author
  • Visitor
10 years 7 months ago #1822 by Leliel
Replied by Leliel on topic Philosophy of the martinilize
Sooooory that i come again...
i think you mention the "adaptive resolution scheme"section in Gromacs tutorial,right?
So i will install a "votca"package and use iterative boltzmann inversion method to get the parameters?

i will try to do it,thanks!

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

More
10 years 7 months ago #1823 by Clement
Replied by Clement on topic Philosophy of the martinilize
Hi!

To get the value of your bonded potentials (bond lengths, angles, dihedrals), you need to:
  1. run an atomistic (AA) simulation of the whole and/or some parts of your molecule, long enough to have converged values for the next step (for you to test; depending on the size of your molecule, can be nanoseconds to hundreds of nanoseconds).
  2. convert your AA trajectory into a CG trajectory; to be able to do that, you need a AA > CG mapping of your molecule, i.e. which AA atoms go in which CG bead (you need to write that by yourself; but it's the fun part of the parameterization). The reprocessing of the trajectory is done using this code: http://md.chem.rug.nl/cgmartini/index.php/downloads/tools/113-rt , following this tutorial: http://md.chem.rug.nl/cgmartini/index.php/tutorial/reverse-transformation . You'll find gazillions of posts concerning this reverse transformation on this Forum.
  3. extract the values you need: with the tools (g_bond, g_angle, etc.) provided by GROMACS (that's what we usually use, the rest being done with the same simulation package), you can measure distributions of bond length, angle, etc.; this will give you the first set of parameters for your trial topology.
The rest is "just" iteratively tuning the CG topology by running trial simulations, adapting force constants (you 'll need to compare the distributions of bonded parameters visited by your molecule using your iteratively updated trial topology and the ones extracted from your AA > CG mapped trajectory) and beads type (you'll need to compare partition coefficient of your compound(s) to experimental value(s), if existing; you can quickly calculate that with thermodynamic integration).

In case you didn't see it already, these steps are explained in more details here: http://md.chem.rug.nl/cgmartini/index.php/user-platform/faq#create topos .

Good luck!

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

  • Leliel
  • Leliel's Avatar Topic Author
  • Visitor
10 years 7 months ago #1837 by Leliel
Replied by Leliel on topic Philosophy of the martinilize
Hi,
i have read through the tutorials of 1e. and 3. both mentions the resolution transformation.but it says that i the package is based upon gromacs3.3.1 ,but the version of gromacs on my computer is 4.6.2. so is it OK to install two version of gromacs? or the 4.6.2 can afford the g_fg2cg function??

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

More
10 years 7 months ago #1838 by Clement
Replied by Clement on topic Philosophy of the martinilize
Nope, you need to compile the modified version of GROMACS 3.3.1 downloadable on this website.

No problem to have multiple versions of GROMACS on your machine; you need to use the different executables (mdrun, g_fg2cg, etc.) giving the full path (i.e. /wherever/is/your/gromacs-3.3.1/bin/mdrun), or source it in your current bash session (i.e. source /wherever/is/your/gromacs-3.3.1/bin/GMXRC).

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

  • Leliel
  • Leliel's Avatar Topic Author
  • Visitor
10 years 7 months ago #1839 by Leliel
Replied by Leliel on topic Philosophy of the martinilize
I have installed the modified version of GROMACS3.3.1 and use the g_fg2cg to follow the 1e. tutorial.
And now i wanna extract some bond parameters using this tool.At hand i have the cg.gro file of an DNA,using a cpp file written by myself, and a .top and .itp files written artificially.

I summarize the steps of this recipe, could pls help me to validate this?
1) do a long-time scale-fine grained MD of this DNA,long enough to extract credible parameters(100ns or somewhat?)
2) use the fg2cg to extract a coarse grained trajectory of this MD.
3) make a index using the make_ndx ,mentioned which i would like to count.
4) use g_angle to extract the gold-standard angle and dihedral as a ref.
5) use my artificially .gro to run a MD ,and also extract some angle/dihedral ,and alter the parameters until finely accord to the gold-standard.

am i right??

as well i have some questions about this recipe's details. how can i put this 100ns MD? in the water(solution) or just do a MD only with this DNA?

thanks in advance,for helping me do a coarse-grained recipe step by step, enforce my knowledge.

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

More
10 years 7 months ago #1840 by djurre
Replied by djurre on topic Philosophy of the martinilize
Yes, this sounds like the right steps to take. The 100ns should normally be done in water.

However, I would also like to make a warning: I won't be as easy as it sounds here. This protocol works very well for small molecules, but is much harder for larger biopolymers. For example the structures of proteins is not stable without an elastic network. A similar thing you might encounter for DNA. Second, the bases of DNA are very tightly packed. This might not be possible with regular of S-type Martini beads. These are only some of the problems you will face and have to solve.

In our lab, someone is working on DNA, and even thought he is one of the four smartest PhD students we have, he is still breaking his head on it. So be warned!

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

Time to create page: 0.113 seconds