Dear Rosetta users
I have generated a model by fitting a crystal structure to electron density map using flexible fitting software (MDFF). This is a 864 aa long protein with two domains. However I am not satisfied with the fitting of a specific regions( a stretch of ~50 aa between two domains). This region contains a small helix surrounded by loops. I am trying to remodel this region using rosetta loop modeling and using electron density restraints. I am using following flags in the loop modeling protocol-
-database ~/path/to/rosetta_src_2015.39.58186_bundle/main/database (I am giving actual path in the file)
The loop_def.txt contains just one line
LOOP 489 541 0
When I am running this using loopmodel.linuxgccrelease, I am getting following error.
[ERROR] Exception caught by JobDistributor for job loop_model_input_0001Cannot normalize xyzVector of length() zero
protocols.jd2.JobDistributor: loop_model_input_0001 reported that its input was bad and will not retry
protocols.jd2.FileSystemJobDistributor: job failed, reporting bad input; other jobs of same input will be canceled: loop_model_input_0001
protocols.jd2.JobDistributor: no more batches to process...
protocols.jd2.JobDistributor: 5 jobs considered, 1 jobs attempted in 2778 seconds
caught exception 1 jobs failed; check output for error messages
I have following queries
1) Is my approach correct for doing this task? If yes how to solve this error. If no, then what should I be doing for modeling this region.
2) Is remodel a better option for such kind of modeling, If yes how should I go about using it?
Kindly let me know if anyother explanation is required.
Sorry for the long post
The "Cannot normalize xyzVector of length() zero" error does happen occasionally - often it's just a fluke due to random number draws. I'd recommend re-running the protocol and see if things work better with a different random number seed.
It's only if you consistently get that error with different runs with different random number seeds would I be worried about the issue.