[gmx-users] Re: Simulation in the high temperature conditions

[Justin A. Lemkul]

James Starlight wrote:
> Justin,
> 
> 
> I've applied disres on each backbone atom of my potein within cutoff 
> distance of 1nm ( Rc=1.0 nm). I've selected this value for cutoff to 
> decrease overall ammount of the restains in my itp file. Also such value 
> ( 1nm) was selected because of the relatively tight packing of the alpha 
> helices in the TM buddle of membrane protein.
> 

I still don't see how you're defining this cutoff anywhere in genrestr or any 
other step.  Do you have some special index group that you're using in your 
genrestr command?  It may not be relevant, I'm just very confused.

> The selected values for disre_dist, disre_up2 and disre_frac were 1, 1.2 
> and 0.5 nm respectually . Also I've made disres with 1 and 0 nm values 
> but I have not noticed any difference in the resulted behaviour of the 
> constrained system. As the consequence I have not clearly realise what 
> exactly is the disre_frac and in what exactly cases this option could be 
> helpfull.
> 

The -disre_dist option sets the tolerance for defining the 'lo' value.  By 
default, it is a fixed distance below the actual distance contained in the 
coordinate file.  The default value is 0.1, so if your distance is 0.5, the 'lo' 
value is set to 0.4 (i.e. 0.5 - 0.1).  When using the -disre_frac value, the 
value specified in -disre_dist is ignored, so be aware that when you combine the 
two, you override -disre_dist (as stated in genrestr -h).  So if you're using 
-disre_frac of 0.5, your 'lo' value for a 0.5-nm distance is 0.25.

> 
> This is an example of my output itp file
> 
> [ distance_restraints ]
> ;   i     j ? label      funct         lo        up1        up2     weight
>     1     5 1     0          1          0    1.64731    2.64731          1
>     1    10 1     1          1          0     1.7326     2.7326          1
>     1    12 1     2          1          0    1.83886    2.83886          1
>     1    14 1     3          1          0    1.96079    2.96079          1
>     1    15 1     4          1          0    2.03503    3.03503          1
>     1    17 1     5          1          0    1.99007    2.99007          1
>     1    19 1     6          1          0    2.08879    3.08879          1
>     1    27 1     7          1          0    2.13916    3.13916          1
>     1    29 1     8          1          0    2.27147    3.27147          1
>     1    30 1     9          1          0    2.35793    3.35793          1
> 
> 
> I've made 10 ns simulation of such system and observe rapid shrinking of 
> my protein starting with first 100ps like the distances that I chose 
> were too small and forces shrink my protein. But when I've tried larger 
> distance value for disre_dist my protein denatured rapidly as no disres 
> were presented.
> 

The problem here is that you're allowing all distances between 0 and 'up1' to 
experience zero restraining force.  That gives the system incredibly wide 
latitude and basically makes the restraints useless.  What you're calling either 
shrinking or denaturing are likely just two possible outcomes from the structure 
being destabilized in some way.  Try running genrestr with default options to 
see if you get a better result.

> So the main question is How I could specify this disre values if I want 
> to restrain the motion of the helixes of my protein within disre value ( 
> e.g 10 A) relatively current confrmation ?
> 

Are you wanting the restraints to be maintained within a 1-nm distance of their 
current value?  That's a very wide latitude.  If that's what you want, then use 
the -disre_dist option without -disre_frac.

-Justin

-- 
========================================

Justin A. Lemkul
Ph.D. Candidate
ICTAS Doctoral Scholar
MILES-IGERT Trainee
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

========================================