Hi Benrezkallah,

dssp takes a bit long to run, but a 200 ns simulation should be done in about half an hour? Did you remove all the water and ions first? This will drastically speed up the analysis since there will be less i/o load.

Second, instead of breaking it into two sets of 100 ns, you could either:

a) use the -dt flag on do_dssp to only calcuate the secondary structure for every 10 or 100 ps instead of every frame

b) use gmx trjconv with -skip to skip every 2/5/10 frames. This will make the trajectory smaller without losing too much data.

If neither of those work, you should simply be able to merge them with some bash commands (e.g., cat, paste, awk, head, tail, or something) after a bit of processing since they are just text files.

Cheers,

John

An alternative way to solve your problem is to use VMD. VMD also has capabilities to perform secondary structure analysis using the timeline menu. You can generate trajectories for whatsoever time point you want and simply load them into VMD.

Yes, VMD might also be slow depending upon your system size after excluding water, and non-protein molecules from the trajectory. Once you have the secondary structure from the VMD, you can save the output in your desired format. You can write a small script that would help you generate secondary structures plots and define colours for each type of secondary structure motif if you would like be a bit fancy.

Use mdtraj lib in python. The gmx tool is not very feasible.