[gmx-users] CONSTRAINED SIMULATIONS

[chris.neale at utoronto.ca]

Dear Giuseppe:

OK, those are the forces. They seem pretty huge and massively  
fluctuating, although I use umbrella sampling myself so this might be  
quite normal for constraint sampling (something for you to look into).

The only thing that comes to mind is that your running up against a  
pbc problem in determining either the center of mass of each protein  
prior to solving the constraint or that your constraint is oscillating  
about different "shortest distances" over different images.

The gmx3 manual also says this:

Another limitation is connected with pressure scaling. When pressure  
scaling is used, the shortest distance between the pull groups and the  
reference group should be within in the central unit-cell, not to  
periodic images of the pull groups. Otherwise the pressure scaling  
will affect the reference distances and thus also the forces. The  
exception is constraint pulling with constraint distance? > 0, this  
will work correctly with pressure scaling under all conditions.

So here's what I recommend:

1. create a system that has no water so that it runs faster.

2. Use the sd integrator so that you get some random noise.
integrator          =  sd
tc_grps             =  System
tau_t               =  1.0
ld_seed             =  -1
ref_t               =  300
gen_temp            =  300
gen_vel             =  yes
unconstrained_start =  no
gen_seed            =  -1

3. see if you can reproduce the problem.

4. Make the box much bigger and center your solutes (trjconv -center  
-pbc mol) so that you need not worry about pbc problems.

5. see if you can reproduce the problem.

Chris

-- original message --

Got it!

Ok, first lines:

# CONSTRAINT    3.0
# Component selection: 1 1 1
# nSkip 1
# Ref. Group 'protein_b'
# Nr. of pull groups 1
# Group 1 'protein_a'  Pos. 3.099513 2.982803 3.377482
#####
0.000000                -787.501479
0.002000                -271.159600
0.004000                -1147.351351
0.006000                -108.722797
0.008000                -97.813955
0.010000                58.592018
0.012000                428.034442
0.014000                1022.187858
0.016000                -60.867559
0.018000                334.483373
0.020000                1328.818408
0.022000                940.307012
0.024000                368.857413
0.026000                1680.598490
0.028000                857.879861
0.030000                362.426529
0.032000                1717.334105
0.034000                810.192387
0.036000                105.655760
0.038000                1048.031839
0.040000                484.148742
0.042000                1312.769160