You are here

Comparative modeling stops with "no more batches to process... " / no output

7 posts / 0 new
Last post
Comparative modeling stops with "no more batches to process... " / no output
#1

I have been trying to built a model using Rosetta 3.4 but it fails with "no more batches to process... ". No output is produced or any other errors (Except a warning about missing OXT residue at the termini). I have attached the log file, part of the .aln file and the flag file that I have been using.
As I can see from log, it recognizes the input pdb, the fragment libraries.. I might have a problem with the alignment file as the input format was not exactly clear to me. Other then that I have no clue what could be my mistake. I have checked the manual and a tutorial when I constructed the flags file.

Any help is appreciated.

Thanks

AttachmentSize
log.txt5.38 KB
flags.txt553 bytes
aln__.txt3.76 KB
Post Situation: 
Tue, 2012-08-28 12:13
eskici

Your argument to -out:file:silent is a directory, not a file name. It needs a file name. I am guessing that is the error. A non-existent file name is fine, but I think it's probably freaking out about getting something that does exist that it can't read/write.

Failing that (I wrote this part first but then the tip above seemed most likely to fix it):

I think from your post you've seen "rosetta_demos/public/homology_modeling_threading_basic", yes? Unfortunately that demo is for 3.3, not 3.4...I would hope it works but can't tell (I'll try giving it a test).

Try telling it it's the other type of alignment file, 'general', whatever that means.

Try passing the fragments to -frag3 and -frag9 as well?

If you already have a silent file present, try -overwrite.

Tue, 2012-08-28 12:36
smlewis

My big concern is the "0 jobs considered" part of the log, which indicates that Rosetta isn't even trying to run through the protocol. The most common cause of this issue is that Rosetta finds the output corresponding to the job, so it thinks it's already done all the jobs and doesn't bother to repeat them. This is why it suggests trying the -overwrite flag (which will force re-running). What Steven suggests is that it's getting confused with -out:file:silent being a directory, which isn't throwing the appropriate error, but instead silently thinks the output of the job already exists.

If what Steven suggests doesn't work, it would help to see if the jobs are being generated appropriately. Try re-running with "-out:level 400" on the command line. If things are working appropriately, you should see lines like "Instantiate ThreadingJobInputter" "ThreadingJobInputter::fill_jobs" and "creating job for alignment ... on template ..."

Tue, 2012-08-28 13:52
rmoretti

"What Steven suggests is that it's getting confused with -out:file:silent being a directory, which isn't throwing the appropriate error, but instead silently thinks the output of the job already exists. "

For PDB stuff, jd2 creates all the jobs, then checks as it tries to run them if they're already on disk or not. SilentFileJobOutputter has a tougher comparison to make, since it has to parse a silent file instead of just checking for the existence of files on disk. Even so, I think it ought not report trying to run 0 jobs...but it's the first thing to check.

Frankly, despite being involved in the creation of jd2, I have no idea how half of it works anymore - people have hacked in all sorts of crazy things, including a lot of things that explicitly were never supposed to happen.

Tue, 2012-08-28 14:16
smlewis

Yes I have seen the demos for 3.3 ...
Hmm I though I had a filename typed there but I don't. I will run it with presenting a file name instead and see if it makes any difference. If it doesn't then I will try changing the other arguments.
Now I remember I had one error before this "no more jobs.." The error was that it cannot find/read/... atom_descriptions.txt But the error disappeared when I typed in the full path for the database. It is probably not related but might worth mentioning.

Wed, 2012-08-29 01:40
eskici

I had run it with -out level 400 and it stops write after the setup for the output file.. I don't have any files with the same name, but in case I used the -overwrite flag (I used it in the command line right after the executable name instead of the flag file dunno if it makes a difference). Nothing is created, not even an empty file.

protocols.jd2.FileJobOutputter: FileJobOutputter ctor
protocols.jd2.SilentFileJobOutputter: SilentFileJobOutputter setup for file Gabam.silent
protocols.jd2.JobDistributor: no more batches to process...
protocols.jd2.JobDistributor: 0 jobs considered, 0 jobs attempted in 0 seconds
protocols.jd2.JobDistributor: no jobs were attempted, did you forget to pass -overwrite?
protocols.jd2.SilentFileJobOutputter: writing 0 structs.
protocols.jd2.SilentFileJobOutputter: currently have 0 structs.

Wed, 2012-08-29 02:00
eskici

I should stop replying in every 2-3 min. :)
I went through the log and realized that it didn't read any alignments from alignment file but it didn't raised any errors. Changing the alignment type to general worked! It runs now.
Thanks!

Wed, 2012-08-29 02:04
eskici