DMPfold: Understand the run of DMPfold “run_dmpfold.sh”?

I ran dmpfold using this command:

run_dmpfold.sh sequence.fasta sequence.21c sequence.map ./dmpfold-lysin 10 500 &> dmpfold-log.log


It produced files like this:

sequence_sub_embed_1.pdb ..... sequence_sub_embed_500.pdb

And, now I see some more files like this being produced:

sequence_1.pdb .... sequence_2.pdb .....etc

Can you please let me know what is dmpfold doing here? And can I stop dmpfold in between when desired number of models are generated? And, while it is running, can I move the previously made pdbs to new location becuase of disk space issues or all pdbs have to be in the "dmpfold-lysin" as stated in the command-line above? Thanks.

From the docs:

The final model is final_1.pdb and other structures may or may not be generated as final_2.pdb to final_5.pdb if they are significantly different. Many other files are generated totalling around 100 MB - these should be deleted to save disk space if you are running DMPfold on many sequences.

These other files include intermediate CNS files (sequence_sub_embed_1.pdb etc.), output files before clustering (sequence_1.pdb etc.) and saved ensembles of previous iterations (ensemble.1.pdb etc.).

Let the run finish or you won't get the final output files. If you want to use all the generated structures, use the ensemble.n.pdb file from the last iteration (i.e. the highest number). You can also use the other ensemble files if you want to include models from previous iterations.

I would strongly recommend using default "3 50" parameters until you get things working. Don't move files during the run. You really don't get much benefit of going beyond 50 models per iteration, trust me. We tried it relatively extensively!

• Just one question: "output files before clustering" and "ensemble files": are these also 500 + 500 in number? – DanielSebas Sep 12 '19 at 17:10
• Yes, after a run has finished the sequence_n.pdb files will be 500 in number and correspond to the 500 structures in ensemble.n.pdb. Other ensemble files contain structures from previous iterations. – jgreener Sep 12 '19 at 17:18
• I noticed a behavior of the program I have: I left "run_dmpfold.sh" to run for two days without touching the system using the command line I mentioned, at that time dmpfold had produced "sequence_71.pdb" file, and after 2 days when I check, there were no more number pdb files than "sequence_71.pdb". When I logged into the system as it was on sleep mode, the number of "sequence_71.pdb" stared increasing to "sequence_72.pdb"..etc. Hence, actually 2 days were lost where dmpfold was not producing files. Can you please let me know why this might have happened? – DanielSebas Sep 15 '19 at 8:39
• It sounds like your machine went to sleep, which will have paused all running processes. This isn't a DMPfold issue, it's more an OS issue. Have a look at tools like mac-how-to.gadgethacks.com/how-to/… to sort this out. – jgreener Sep 16 '19 at 9:25
• yes, I corrected it by "aborting suspend" in my system settings. Thanks – DanielSebas Sep 16 '19 at 11:52