You are here

RASREC full_run vs. test_run

2 posts / 0 new
Last post
RASREC full_run vs. test_run
#1

Dear Rosetta users, 

I'm using RASREC CS-Rosetta 3.5 to pridect the structure of a membrane protein. So far I've added the following files to the setup_run: CS file (NMR data), FASTA and Fragment files. I've tried to add the Span and Lips files in the setup_run, but I didn't know the commands for these types of files; I've tried -span <spanfile>  (and ) -lips <lipsfile>, but they weren't excepted this way! So what I did is add them manually to the flags_denovo file. 

In the flags_denovo file, there's a lot of membrane flags added manually by me. I'm not sure that this is the right way to go about it. 

The probelm I'm having right now is the the difference of outcome between the test_run and the real_run. The test_run seems to go through all the stages (1-4, 7, 8). So at the end of a test_run, I get batches, centroid pool, centroid pool stage 3, centroid pool stage 4, fullatom pool, fullatom pool stage 7, fullatom pool stage 8. However, the real run only gives me batches and centroid pool. That's all, it doesn't go beyond this point! I also to get any error messages in the test.err file. 

From my readings about RASREC, I realized it quite different to abrelax method. As I've tried many times to increase the number of produces structures in RASREC (in the setup run and in the flags) but non of these attempts increase the number of output structures. In abrelax however, it very easy to set up the number of output structures in the setup_run. 

Please help me resolve this issue, as I've tried so many things before posting here and nothing worked! 

Thanks! 

Category: 
Post Situation: 
Wed, 2015-10-28 03:26
Lati

Just a little update in case someone elese is having a similar problem...

It runs out that the problem described in the post above was caused by a major hardware issue in the server I was using. The difficulty in finding it out was because there was no error messages what so ever and the test run was always complete and generated full stages. The only clue was in the core dump files that I used to get near the end of the "unsuccessful" run.

So if you're having these core files, use a different server or use your local computer. I used my local computer in the lab which has 16 cores and about 64G of ram. It took a day and a half to finish successfully (full 6 stages).

And about spcifying the number of structures, it turns out that this can't be done in RASREC protocol. The run will sample ~10,000 - 20,000 structures on its own. You can't spcify it!

I hope this is helpfull.
 

Fri, 2015-11-20 03:32
Lati