[gmx-users] Re: Umbrella sampling simulations using make_edi with -restrain and -harmonic

[Poker Chen]

> I think you should add -tau 0 when using flooding as harmonic restraint
> (since you do not want the flooding potential to change).
Okay, adding -tau 0 does turn it properly into a harmonic potential, 
according to edsam.xvg. However, the average position of the resulting 
(gaussian) is near but not at the expected value, given the reference 
position in -ori.
> Can you check whether make_edi -ori wrote the correct position on the 1st eigenvector in your
> .edi file? Scroll down to a line called
>
> # NUMBER OF EIGENVECTORS + COMPONENTS GROUP 7
>
> (Group 7 means the flooding vectors), if you have a single flooding vector, the next
> line should read "1" and in the following line you will find three entries that describe
>
> <nr> <spring constant> <reference projection> <ref. proj. change per time step>
>
> which I guess read in your case
>   1   1    9.99   <no 4th entry>
>
> so the reference projection should be 9.99, as calculated from your .ori structure.
> If there is no 3rd entry, you can simply put it there manually (and leave away
> the -ori option to make_edi).
constrain.edi with the -tau 0 line and -ori does not give the reference 
position on the eigenvector. -ori seems to only write the new origin 
coordinates. So the output looks like this:

...
# NUMBER OF EIGENVECTORS + COMPONENTS GROUP 7
  1
        1   0.052626   #<-- lack of reference position after vector and 
spring-constant.
-0.04593 -0.02711 -0.01672
-0.04696 -0.01819 -0.00113
...
#NTARGET, XTARGET
  0
#NORIGIN, XORIGIN
  574
7  8.208000  9.368000  4.211000  #<-- coordinates change with -ori argument
...

I'll put the values in manually and run another set.

  - Po-chia.