[gmx-users] Re: Error during CNP simulation

[Bharath K. Srikanth]

Hi Chris

My EM works just fine if I use a cubical box of water, so I'm baffled as
to what's going wrong in this case.

I tried the entire process again, this time removing position restraints.
Let me describe my entire procedure on this occasion, once again:

1) I first assembled a bilayer of 128 DOPC lipids, in a box of dimension
7.5 x 7.5 x 7.5 nm3.

2) I then removed the water molecules, resized the box length by changing
the bottom line of the coordinate file, and inserted a CNP to an
appropriate location, about 7 nm above the bilayer.

3) Then, I performed an energy minimization, and the CNP did not move
about. So far, so good.

4) I solvated the box with water using genbox, and then attempted an
energy minimization (using steepest descents). This is where the problem
arises.

These are the coordinates from the input coordinate file, before the
energy minimization:
  129F16    C01 1793  10.600   3.474   3.330
  129F16    C02 1794  10.438   3.758   3.294
  129F16    C03 1795  10.381   3.588   3.947
  129F16    C04 1796  10.795   3.703   3.387
  129F16    C05 1797  10.564   3.308   3.584
  129F16    C06 1798  10.289   3.506   3.415
  129F16    C07 1799  10.322   3.939   3.528
  129F16    C08 1800  10.655   3.430   3.872
  129F16    C09 1801  10.840   3.479   3.620
  129F16    C10 1802  10.641   3.958   3.453
  129F16    C11 1803  10.342   3.876   3.850
  129F16    C12 1804  10.610   3.991   3.756
  129F16    C13 1805  10.675   3.736   3.950
  129F16    C14 1806  10.307   3.397   3.721
  129F16    C15 1807  10.862   3.787   3.699
  129F16    C16 1808  10.172   3.686   3.649

And these are the coordinates from the output coordinate file, within just
TEN steps of energy minimization:
 4849F16    C01 6513   1.015   0.049   5.350
 4849F16    C02 6514   1.294  -3.135   4.063
 4849F16    C03 6515   0.663  -1.872   3.948
 4849F16    C04 6516   1.521  -1.365   3.815
 4849F16    C05 6517   2.111  -0.464   4.030
 4849F16    C06 6518   0.546  -2.205   6.177
 4849F16    C07 6519  -0.418  -2.289   6.608
 4849F16    C08 6520  -0.301  -1.054   4.991
 4849F16    C09 6521  -0.154  -1.757   5.881
 4849F16    C10 6522   0.595  -0.223   5.592
 4849F16    C11 6523   1.450  -0.980   7.148
 4849F16    C12 6524   0.055  -1.027   5.487
 4849F16    C13 6525   0.020  -0.473   5.753
 4849F16    C14 6526   1.492  -0.255   4.474
 4849F16    C15 6527   0.926  -0.762   4.489
 4849F16    C16 6528   0.147  -2.022   3.748

I have no idea why this is happening- that the molecule could move so much
seems ridiculous. Besides the em.mdp file, is something wrong with, for
example, my method of resizing the box?

Thanks!
Bharath


--original message--

Sorry Bharath, I simply can not believe that a macromolecule drifted 3 nm
during
energy minimization.
What happened to the intervening solvent? I think that you have some
problem with
your technique here.

Why did you use 120 constraints? Is this for bonds within the CNP?

You need to do 2 things:

a) figure out what is going wrong with your EM... what you describe seems
impossible
to me.
b) simplify your system to debug the problem. So you get the same warnings
about too
many constraints
if you omit the position restraints?

Chris.