[gmx-users] Re:: Specified frame doesn't exist or file not seekable

[Mark Abraham]

sudheer babu wrote:
> 
> 
> I am having the evidence that I have drawn RMSD plot for 7.25ns simulation, in rmsd.xvg file showed the values of the time(ps), mentioned in *trjconv command 

That's evidence if you used the .xtc file for producing that RMSD plot 
and the full system was used for the RMSD plot and there's data for 
those times you wanted. The original error message from trjconv implied 
that at least one of those three was not true.

> 
> Im pasting my .mdp file
> 
> title               =  dpt_prod
> 
> constraints         =  hbonds
> integrator          =  md
> dt                  =  0.002    
> nsteps              =  500000
> nstcomm             =  1
> 
> ;output control statements
> nstxout             =  100
> nstlog              =  100
> nstenergy           =  100
> nstlist             =  10

Well hopefully there was more to it than this, but you've certainly left 
off the most pertinent option... nstxtcout!

> My idea of doing this trjconv command is by mistaken I have deleted 7thns .trr but I have that file in .xtc format, so I am trying to get 7ns file in .trr format for continue my simualation

No, you don't have "that file in .xtc format". Whether you produced this 
XTC format trajectory with mdrun, or with trjconv on mdrun's .trr, then 
it contains a (possibly whole) subset of your system, with coordinates 
only, and at a reduced precision. Since there are no velocities, and 
only approximate coordinates, you have no chance of an 
ensemble-preserving restart. You'll need to start again and take more 
care of valuable output :-)

Mark