[gmx-users] Distance restraint problem with 4.0.5 version

[Mark Abraham]

Alexandre Suman de Araujo wrote:
> Hi Gmxers
> 
> Some months ago I performed some simulations, with GMX 3.3.3 version, 
> where I apply distance restraint between water oxygen atoms and a static 
> dummy site located at the center of a sphere to keep waters inside this 
> sphere. After performed an upgrade to GMX 4.0.5 version this distance 
> restraint scheme stopped to work.
> I read distance restraint material in 4.0.5 paper manual and I can't 
> find any modification reported there.
> 
> Looking at gmxdump output from tpr's generated with grompp from 3.3.3 
> and 4.0.5 versions and the same .mdp, .gro and .top files, I had this:
> 
> 3.3.3 version: grep -i disre gmxdump3.txt |more
> 
>   disre                = Simple
>   disre_weighting      = Conservative
>   disre_mixed          = FALSE
>   nstdisreout          = 100
>         functype[468]=DISRES, label=   0, type=1, low= 0.00000000e+00, 
> up1= 3.00000000e+00, up2= 9.39999962e+00, fac= 1.00000000e+00)
>         functype[469]=DISRES, label=   1, type=1, low= 0.00000000e+00, 
> up1= 3.00000000e+00, up2= 9.39999962e+00, fac= 1.00000000e+00)
>         functype[470]=DISRES, label=   2, type=1, low= 0.00000000e+00, 
> up1= 3.00000000e+00, up2= 9.39999962e+00, fac= 1.00000000e+00)
>         functype[471]=DISRES, label=   3, type=1, low= 0.00000000e+00, 
> up1= 3.00000000e+00, up2= 9.39999962e+00, fac= 1.00000000e+00)
>         functype[472]=DISRES, label=   4, type=1, low= 0.00000000e+00, 
> up1= 3.00000000e+00, up2= 9.39999962e+00, fac= 1.00000000e+00)
>         functype[473]=DISRES, label=   5, type=1, low= 0.00000000e+00, 
> up1= 3.00000000e+00, up2= 9.39999962e+00, fac= 1.00000000e+00)
>         .
>         .
>         .
>         functype[724]=DISRES, label= 256, type=1, low= 0.00000000e+00, 
> up1= 3.00000000e+00, up2= 9.39999962e+00, fac= 1.00000000e+00)
>         functype[725]=DISRES, label= 257, type=1, low= 0.00000000e+00, 
> up1= 3.00000000e+00, up2= 9.39999962e+00, fac= 1.00000000e+00)
>         functype[726]=DISRES, label= 258, type=1, low= 0.00000000e+00, 
> up1= 3.00000000e+00, up2= 9.39999962e+00, fac= 1.00000000e+00)
>        0 type=468 (DISRES) 389 392
>        1 type=469 (DISRES) 393 396
>        2 type=470 (DISRES) 397 400
>        3 type=471 (DISRES) 401 404
>        4 type=472 (DISRES) 405 408
>        5 type=473 (DISRES) 409 412
>        .
>        .
>        .
>        256 type=724 (DISRES) 1413 1416
>        257 type=725 (DISRES) 1417 1420
>        258 type=726 (DISRES) 1421 1424
> 
> 4.0.5 version: grep -i disre gmxdump4.txt
> 
>   disre                = Simple
>   disre_weighting      = Conservative
>   disre_mixed          = FALSE
>   nstdisreout          = 100
>         functype[470]=DISRES, label=   0, type=1, low= 0.00000000e+00, 
> up1= 3.00000000e+00, up2= 9.39999962e+00, fac= 1.00000000e+00)
>            0 type=470 (DISRES) 0 3
> 
> 
> It is clear that 3.3.3 version is applying the constraints to all waters 
> (what is correct) and 4.0.5 to just one. However, I can't find a way to 
> fix this.

Hmm, the above indicates that 3.3.3 wasn't folding the function type 
definitions correctly, and apparently 4.0.5 is folding the types as well 
as the references to them. If you're sure you've used the same 
.mdp/.top, then please submit a bugzilla (http://bugzilla.gromacs.org/) 
and attach a .tar of the .top, .mdp, .tpr and .gro files.

Also have a look at grep -i disre -C 5 gmxdump4.txt to see if there's 
anything interesting in the local contexts of the differences.

Mark

> Does anyone could help me with this issue? Are there modifications in 
> distance restraint algorithm in 4.0.5 version compared with 3.3.3 
> version? Or this can be a bug?
> 
> Thanks
>