normal water

  • yamahdi
  • yamahdi's Avatar Topic Author
  • Offline
  • Fresh Boarder
More
8 years 4 months ago #467 by yamahdi
water was created by yamahdi
Hi everybody
we have made a CG model of a protein and lipid bilayer and each of protein and lipid has been simulated a short MD separately. now we are solvating the system. when we use CG water the system shows infinite forces oin some water molecules and in VMD we see many water molecules are very close together and have been bonded to each other. we have tested solvating our CG system with all atom water(sec216) and in this case no problem has been shown.
another test we have done is to define a bigger simulation box for the system and then solvating the system with CG water. in this case also the system is solvating good.
I should remember that we have used the technic of changing vdw radius and it lt did not lead to better result in the case of our simulation box but in bigger simulation box it worked well.

would you please help me explaining where may be our mistake?
thanks for your kind help in advance

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

More
8 years 4 months ago #468 by xavier
Replied by xavier on topic water
Dear Yamahdi,

The problem you describe is some thing that we have also observed and the reason for the overlapping of water molecules and the inefficiency of the -vdwd option is not clear. It seems that there is a confusion in gromacs with the vdw radius of the beads themselves ...

It is good news that using a bigger box solves the problem ... it might indicate that the problem with a small box probably happens when placing the different small box, which seems to result in a an overlapping of the water boxes.

In conclusion you do not have done anything wrong and even seem to have found a solution to an old and unclear problem. Note that we had solved the problem by using boxes of equal sizes ...

XAvier.

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

  • yamahdi
  • yamahdi's Avatar Topic Author
  • Offline
  • Fresh Boarder
More
8 years 4 months ago #469 by yamahdi
Replied by yamahdi on topic water
Hi
would you please explain me how had you solved the problem using equal size boxes?
thanks

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

More
8 years 4 months ago #470 by xavier
Replied by xavier on topic water
That was by building a water box of the same size as protein was! Or the other way around actually. Note that this was actually used to solvate a protein in a membrane bilayer.

But your solution is more efficient :))

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

More
6 years 8 months ago #1141 by pimfrederix
Replied by pimfrederix on topic water
Has anybody managed to do this with polarisable water?

I used Xavier's approach of matching the protein size with the box size succesfully before with the normal water, but there are some really weird problems with genbox (as acknowledged on the Gromacs mailing list) putting molecules outside the box or having an inhomogeneous density.
Solvating with a single pol. water molecule (mentioned as a workaround of the problem) is very inefficient / doesn't work properly. Also using a bigger box doesn't work properly (water density to big in center of the box).

I hope someone can help, really struggling and we're about to submit an automated job of 8000 different systems, so manually correcting each box isn't really an option. Much appreciated!

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

More
6 years 8 months ago #1143 by xavier
Replied by xavier on topic water
I am not sure why this would not work for polarizable water!

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

More
6 years 8 months ago #1145 by pimfrederix
Replied by pimfrederix on topic water
Me neither, but in my experience it doesn't! I get a well-solvated box, but there is a small number (varying from 2-10 in my tests) of water molecules well outside the box, it's really strange.

I've now found a way to work around it, solvating using a box with only 1 PW and size of roughly 0.4x0.4x0.4 for obtaining the roughly correct density.

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

More
6 years 8 months ago #1155 by xavier
Replied by xavier on topic water

pimfrederix wrote: Me neither, but in my experience it doesn't! I get a well-solvated box, but there is a small number (varying from 2-10 in my tests) of water molecules well outside the box, it's really strange.

I've now found a way to work around it, solvating using a box with only 1 PW and size of roughly 0.4x0.4x0.4 for obtaining the roughly correct density.


That is good that this work! We had not tried that before!

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

More
6 years 8 months ago #1156 by pimfrederix
Replied by pimfrederix on topic water
I'm afraid I was a bit premature with that conclusions as well, as it's not a stable solution: I still have the same problem for some tries, but it works in a few occasions (if the minimization takes it). Submitting the exact same job sometimes works, sometimes I get the message about the minimization converging to machine precision, but not reaching the desired force, indicating the box is exploding due to overlapping waters. It's quite frustrating :S.

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

More
6 years 8 months ago #1157 by xavier
Replied by xavier on topic water

pimfrederix wrote: I'm afraid I was a bit premature with that conclusions as well, as it's not a stable solution: I still have the same problem for some tries, but it works in a few occasions (if the minimization takes it). Submitting the exact same job sometimes works, sometimes I get the message about the minimization converging to machine precision, but not reaching the desired force, indicating the box is exploding due to overlapping waters. It's quite frustrating :S.


What has been working for us is to build a box of water of the final size desired (box size of the protein solvated) and then solvate the protein in it using the genbox command.

Or use a box of water that is bigger than the final box. You also need to use the option -vdwd 0.178.

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

More
6 years 8 months ago #1158 by pimfrederix
Replied by pimfrederix on topic water
Yes, this is what I did before with the normal water, but I encountered small problems for polarizable water as mentioned on 9/9. I used vdwd 0.2 or 0.21 though, I'll see if 0.178 works then!

Thanks for your support.

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

More
6 years 7 months ago #1177 by pimfrederix
Replied by pimfrederix on topic water

pimfrederix wrote: Yes, this is what I did before with the normal water, but I encountered small problems for polarizable water as mentioned on 9/9. I used vdwd 0.2 or 0.21 though, I'll see if 0.178 works then!

Thanks for your support.


I now found out why it doesn't work for polarizable water, but I've not figured out yet how to fix it:
Because of the 3 sites per water in the pol. water model, sometimes the WP and WM sites have slightly negative coordinates. For some reason, genbox goes mad when this happens and it creates a few extra waters with the same residue number, e.g. for residue 5877 below
5875PW W16474 1.262 10.610 2.654
5875PW WP16475 1.151 10.524 2.652
5875PW WM16476 1.231 10.592 2.789
5876PW W16477 0.792 6.774 11.373
5876PW WP16478 0.690 6.808 11.463
5876PW WM16479 0.694 6.856 11.428
5877PW W16480 5.855 0.087 8.248
5877PW WP16481 5.749 0.178 8.254
5877PW WM16482 5.879 -0.047 8.283
5877PW W16483 7.169 -4.209 10.118
5877PW WP16484 7.236 -4.100 10.063
5877PW WM16485 7.204 -4.343 10.140
5877PW W16486 3.097 -11.120 9.398
5877PW WP16487 3.206 -11.124 9.310
5877PW WM16488 3.013 -11.148 9.506
5878PW W16489 4.342 9.475 8.814
5878PW WP16490 4.442 9.377 8.821
5878PW WM16491 4.326 9.505 8.950
5879PW W16492 9.463 4.353 10.592
5879PW WP16493 9.566 4.283 10.530
5879PW WM16494 9.469 4.491 10.564
5880PW W16495 5.273 8.615 8.556
5880PW WP16496 5.323 8.611 8.425
5880PW WM16497 5.309 8.493 8.615

Sometimes these beads are outside the box, sometimes they overlap with other waters, which both causes problems. Don't know if it's better to write a script that mimics the genbox procedure but does it correctly for my purposes or a script that postprocesses the gro file and removes the extra waters? I'm not a great programmer so this may take me a while. If I have a solution I'll post it here.

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

More
6 years 7 months ago #1179 by djurre
Replied by djurre on topic water
Great you are looking in to this. You could consider submitting a bug report to the gromacs redmine server: redmine.gromacs.org/ Maybe someone in the gromacs development team has a easy solution.

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

More
2 years 11 months ago #5654 by Pim
Replied by Pim on topic water
4 years later: Gromacs 5.1 and newer have a fixed solvate code so this all goes better!

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

Time to create page: 0.110 seconds