Member Site › Forums › Rosetta 3 › Rosetta 3 – General › Multi-state design with the error “unable to locate database file”
- This topic has 1 reply, 2 voices, and was last updated 5 years, 7 months ago by Anonymous.
-
AuthorPosts
-
-
April 1, 2019 at 1:30 am #3103Anonymous
Hi,
I am a new Rosetta user and learning to do the multi-state Ab design by following the RECON protocal. When I run it, I got the error said “unable to locate database file” even though I already listed the database path in .options. Could anyone help me to fix this problem? Here is the error message I got:
protocols.jd2.JobDistributor: (3) Parser is present. Input mover will be overwritten with whatever the parser creates.
protocols.jd2.MPIWorkPoolJobDistributor: (3) Slave Node 3: Requesting new job id from master
protocols.jd2.MPIWorkPoolJobDistributor: (3) Slave Node 3: Received job id 2 from master.
protocols.jd2.PDBJobInputter: (3) PDBJobInputter::pose_from_job
protocols.jd2.PDBJobInputter: (3) filling pose from PDB 4HKX_relax.pdb
Warning: (1) Unable to locate database file chemical/residue_type_sets/fa_standard/residue_types//root/Rosetta/rosetta_src_2018.09.60072_bundle/main/database/chemical/residue_type_sets/fa_standard/TYPE_SET_MODE full_atom
Error: (1)ERROR: Exception caught by JobDistributor while trying to get pose from job ‘4HKX_relax_test_0003’
Error: (1)
[ERROR] EXCN_utility_exit has been thrown from: src/core/chemical/residue_io.cc line: 167
ERROR: Cannot find file ‘/root/Rosetta/rosetta_src_2018.09.60072_bundle/main/database/chemical/residue_type_sets/fa_standard/TYPE_SET_MODE full_atom or /root/Rosetta/rosetta_src_2018.09.60072_bundle/main/database/chemical/residue_type_sets/fa_standard/residue_types//root/Rosetta/rosetta_src_2018.09.60072_bundle/main/database/chemical/residue_type_sets/fa_standard/TYPE_SET_MODE full_atom’
And here is my .options file:
-database /root/Rosetta/rosetta_src_2018.09.60072_bundle/main/database/
-linmem_ig 100 # limit memory to store rotamer pair interactions
-use_input_sc # Include rotamers from the input structure
-ex1
-nstruct 10 # Generate 10 models
-s 4HKX_relax.pdb 3UBQ_relax.pdb # Use structure 4HKX as input
-mute protocols.simple_moves.GenericMonteCarloMover # Mute the output from the GenericMonteCarloMover
-run:msd_job_dist # Use the multistate design job distributor
-run:msd_randomize # Randomize order of input PDBs in multistate design
The error only showed when I run the mpi version of Rosetta, the normal single core version worked fine. And I already checked all the database files listed in the error message and I’m sure they are all there but Rosetta just couldn’t find them.
Thanks!
Yifei
-
April 3, 2019 at 4:31 pm #14643Anonymous
I’m a little concerned about the presence of the literal “TYPE_SET_MODE” in the output — that shouldn’t be used for the name of any files, which makes me think that there might be a issue somewhere. I’m also curious as to why the string “/root/Rosetta/rosetta_src_2018.09.60072_bundle/main/database/” is showing up twice in the message.
My guess is that you have a typo somewhere in your inputs or database. Try running `grep -Rh ‘TYPE_SET_MODE’ /root/Rosetta/rosetta_src_2018.09.60072_bundle/main/database/` on the command line. All the lines printed should be things like “TYPE_SET_MODE full_atom”, without any references to path. If not, I would re-download a fresh copy of the database.
I’d also check your input files, and see if there’s any other place (besides the -database option) where you might have accidentally put the database path.
You might also want to try running things in a clean directory, carefully copying over just those files you need (carefully examining each one). Rosetta can sometimes prefer files in the local running directory rather than the database, and depending on how that’s set up, you can get issues and confusion.
-
-
AuthorPosts
- You must be logged in to reply to this topic.