You are here

B1_SampleSegment1.sh from RosettaES tutorial crashes

13 posts / 0 new
Last post
B1_SampleSegment1.sh from RosettaES tutorial crashes
#1

Hi all,

I am new to Rosetta and tried to follow the RosettaES tutorial, as well as applying it to my own case. I downloaded and installed the latest weekly release (Rosetta 2017.18) for both Linux (Scientific Linux v6.9) and MacOS X (10.11.5). Running the A_PickFragments.sh script worked ok. It is only when I continue with the B1_SampleSegment1.sh script the cryptic error occurs (see below). I compiled the MacOS version using the ./scons.py -j1 option and it compiled successfully (scons:done building targets) just to see if this makes a difference. It didn't :((

I would be extremely grateful if someone could help me out and point me in the right direction!

core.pack.dunbrack.RotamerLibrary: Dunbrack 2010 library took 0.434736 seconds to load from binary

ERROR: Assertion `moving_connection.icoor().is_internal() && fixed_connection.icoor().is_internal()` failed.
ERROR:: Exit from: src/core/conformation/util.cc line: 89
BACKTRACE:
0   rosetta_scripts.static.macosclangrelease 0x000000010ee702a6 print_backtrace(char const*) + 54
1   rosetta_scripts.static.macosclangrelease 0x000000010ee70198 utility::exit(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&, int, std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&, int) + 1624
2   rosetta_scripts.static.macosclangrelease 0x000000010e96b045 core::conformation::orient_residue_for_ideal_bond(core::conformation::Residue&, core::chemical::ResidueConnection const&, core::conformation::Residue const&, core::chemical::ResidueConnection const&, core::conformation::Conformation const&, bool) + 1829
3   rosetta_scripts.static.macosclangrelease 0x000000010e93f8fc core::conformation::Conformation::append_polymer_residue_after_seqpos(core::conformation::Residue const&, unsigned long, bool) + 188
4   rosetta_scripts.static.macosclangrelease 0x000000010c500870 protocols::loop_grower::LoopGrower::single_grow(core::pose::Pose&, core::pose::Pose&, protocols::loop_grower::LoopPartialSolutionStore&, utility::vector1<std::__1::shared_ptr<core::chemical::ResidueType const>, std::__1::allocator<std::__1::shared_ptr<core::chemical::ResidueType const> > > const&, utility::vector1<std::__1::shared_ptr<core::chemical::ResidueType const>, std::__1::allocator<std::__1::shared_ptr<core::chemical::ResidueType const> > > const&, unsigned long, unsigned long, unsigned long, int, int, unsigned long, unsigned long, bool, unsigned long, bool, bool, unsigned long, int) + 1088
5   rosetta_scripts.static.macosclangrelease 0x000000010c4ece03 protocols::loop_grower::LoopGrower::apply(core::pose::Pose&) + 15411
6   rosetta_scripts.static.macosclangrelease 0x000000010c524187 protocols::loop_grower::FragmentExtension::apply(core::pose::Pose&) + 10455
7   rosetta_scripts.static.macosclangrelease 0x000000010d6db402 protocols::rosetta_scripts::ParsedProtocol::apply_mover(core::pose::Pose&, protocols::rosetta_scripts::ParsedProtocol::MoverFilterPair const&) + 1122
8   rosetta_scripts.static.macosclangrelease 0x000000010d6d663e protocols::rosetta_scripts::ParsedProtocol::apply(core::pose::Pose&) + 1246
9   rosetta_scripts.static.macosclangrelease 0x000000010d73c5c6 protocols::jd2::JobDistributor::run_one_job(std::__1::shared_ptr<protocols::moves::Mover>&, long, std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >&, std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >&, unsigned long&, unsigned long&, bool) + 6582
10  rosetta_scripts.static.macosclangrelease 0x000000010d73a0f4 protocols::jd2::JobDistributor::go_main(std::__1::shared_ptr<protocols::moves::Mover>) + 356
11  rosetta_scripts.static.macosclangrelease 0x000000010d71a432 protocols::jd2::FileSystemJobDistributor::go(std::__1::shared_ptr<protocols::moves::Mover>) + 66
12  rosetta_scripts.static.macosclangrelease 0x000000010bc6a55d main + 829
13  libdyld.dylib                       0x00007fff941965ad start + 1
14  ???                                 0x0000000000000021 0x0 + 33
protocols.rosetta_scripts.ParsedProtocol: [ ERROR ] Exception while processing procotol:
protocols.jd2.JobDistributor: [ ERROR ]

[ERROR] Exception caught by JobDistributor for job input_0001

[ ERROR ] EXCN_utility_exit has been thrown from: src/core/conformation/util.cc line: 89
ERROR: Assertion `moving_connection.icoor().is_internal() && fixed_connection.icoor().is_internal()` failed.


protocols.jd2.JobDistributor: [ WARNING ] input_0001 reported failure and will NOT retry
protocols.jd2.JobDistributor: no more batches to process...
protocols.jd2.JobDistributor: 1 jobs considered, 1 jobs attempted in 12 seconds
Error: [ ERROR ] Exception caught by rosetta_scripts application:1 jobs failed; check output for error messages
Error: [ ERROR ]
beam file is empty exiting protcol

Category: 
Post Situation: 
Tue, 2017-07-04 08:42
ab

Might be a version problem.  I'm not sure the published protocol has been released yet (we had a release hold that was only recently lifted).  A new version of Rosetta should be posted in the next few days, so the first thing to try is that.

I'll ping the author.

Tue, 2017-07-04 11:27
smlewis

The new weekly was just posted; try with that.

Tue, 2017-07-04 15:43
smlewis

Hi smlewis, many thanks for your reply and efforts. I downloaded & installed the latest weekly and tried again. Again the A_PickFragments.sh script from the RosettaES tutorial (obtained from https://faculty.washington.edu/dimaio/files) works fine. But this time something new happened when I started the B1_SampleSegment1.sh script, the error is:

ready for parallel
ERROR: Option matching -corrections:beta_nov15 not found in command line top-level context
Did you mean:
     -corrections:beta_nov16

caught exception ERROR: Option matching -corrections:beta_nov15 not found in command line top-level context
Did you mean:
     -corrections:beta_nov16

beam file is empty exiting protcol


OK, I changed that to beta_nov16, but this time the error is:

ready for parallel
core.init: Rosetta version unknown:f08443d88de09b035118dcdb86981c843d85c77f 2017-05-25 05:19:10 -0400 from /home/benchmark/release/rosetta/git/release/rosetta.binary.linux.release.git
core.init: command: /lmb/home/aberndt/work/rosetta_bin_linux_2017.26.59567_bundle/main/source/bin/grower_prep.static.linuxgccrelease -database /lmb/home/aberndt/work/rosetta_bin_linux_2017.26.59567_bundle/main/database -parser:protocol RosettaES.xml -edensity:mapfile postprocess_job720.mrc -s bridge_ren_with_gap.pdb -default_max_cycles 200 -ignore_unrecognized_res -nstruct 1 -overwrite -parser::script_vars readbeams=False -parser::script_vars beams=NA -parser::script_vars steps=1 -parser::script_vars pcount=1 -parser::script_vars filterprevious=False -parser::script_vars filterbeams=NA -edensity:sliding_window 3 -mapreso 4 -corrections::beta_nov16 -missing_density_to_jump
core.init: 'RNG device' seed mode, using '/dev/urandom', seed=-388924545 seed_offset=0 real_seed=-388924545
core.init.random: RandomGenerator:init: Normal mode, seed=-388924545 RG_type=mt19937
beam file is empty exiting protcol

 

 

Wed, 2017-07-05 03:45
ab

Brandon's forum account is pending; in the meantime here's his suggestions:

 

Brandon Frenz [2:25 PM] 
if you want to let the user asking about ES know that they should just delete the -beta_corrections flag entirely (instead of changing it to 16) and to make sure they have the paths in the runES.sh script set correctly. Particular the -database flag which was set wrong in the version included in rosetta (I've since changed it so it will be fixed in the next weekly).

[2:25] 
Also you could suggest that if they are trying to run it on their own personal case to first try and use the test case included here: https://faculty.washington.edu/dimaio/wordpress/software/

new messages
[2:26] 
So that we can parse out if the issue is with their model or with the code.

Wed, 2017-07-05 11:31
smlewis

Thanks a lot smlewis and Brandon!!!

I am running now the tutorial and so far so good ;o)) !!!!!!!!! As suggested I deleted the -beta_corrections flag entirely (instead of changing it to 16) in the runES.sh. I noted however that there is a weights="beta_nov15" tag in the RosettaES.xml which I did NOT remove (should I have???)

Thanks again, guys!

Thu, 2017-07-06 03:34
ab

beta_nov15 is the same as REF2015, which is the current default scorefunction.  (The release hold was in place while we mopped up bugs exposed by the switch).  So I think the XML is ok.

Thu, 2017-07-06 07:12
smlewis

thanks!

Fri, 2017-07-07 02:48
ab

Yep the xml should be fine. 

Thu, 2017-07-06 13:30
brandon.frenz

Hi Brandon, many thanks for your help!!! I successfully ran the B1_SampleSegments1.sh script now. I provided a map (~4 A) segment (prepared in Chimera) with the pdb fit into it. deleted 31 residues in my pdb to see what RosettaES would rebuild. It did a good job rebuilding 2 helices, however I am a bit surprised that the sequence of the pdb outputs (32 pdb were written out in the loop_1 dir) in that deleted region is the same as in the fasta input. My motivation was to get the register right. Did I overlook something?

Many thanks!

Fri, 2017-07-07 02:47
ab

Sorry I am not sure I understand the question. The fasta file should match the the pdbfile + any additional residues you would like to add to it. Is that not what happened? If it's just that the models aren't accurate (be sure to check all of them, or at least relax them all and check the top scoring), you can run a second round with taboo (which uses the old rounds results to guide the sampling to new locations) and/or increase the beamwidth parameter to something higher, perhaps 64 or even 256+ if you have enough CPU time for it, although that probably isn't necessary. 

For easy benchmarking you may want to consider adding the flag -in::file::native *native pdbfile* which will cause rosetta to report the RMS of the output files in their name. It will also produce a file named "rank.txt" which will show the best RMS to the native at every step so that you can track where it lost the native path.

Fri, 2017-07-07 16:03
brandon.frenz

Hi Brandon, many thanks for your suggestions which I will try out!!!

Sorry about my maybe naive approach: The pdb and the according fasta file I am providing are most likely to be wrong in the amino acid register (with Gly's and Val's in bulky protrusions of the 4 A density which cannot be true) and I would like to use RosettaES or any other program to build in my segmented map with that in mind. However, it builds 'only' those residues in the density that are present in the fasta and deleted in the pdb. Ideally I would like Rosetta to go over the whole fasta sequence and score which residues would best fit in that particular segment and then build in it. Is there such an option???

Also I noticed that the B1.2_InspectIntermediates.sh script does no do much in addition to the B1_SampleSegment1.sh, is that right? Because the latter script wrote out already 32 pdb (my gap in the pdb was 32 residues, so it is this number + 1 ?) files named after_filter_32__4.46717e-316_-405.082.pdb. Which number is the score here btw?405?

You suggested a second run with taboo to guide the sampling to new locations. How would I set that up?

Many thanks for your help again!!!

Sat, 2017-07-08 03:07
ab

So the RosettaES tool basically has 1 job. Take an input structure, and a fasta file, and find the best possible way in which the residues missing from the PDB will fit into the density. If I understand you correctly you don't actually trust the input structure and therefore would like to essentially rebuild the whole thing de novo. In that case you want to go back one step and use the de novo density application, for which you can find a tutorial on the DiMaio lab website: http://dimaiolab.ipd.uw.edu/software/. That protocol has an option to pass a calpha trace and where you can pass in your full pdb structure and that may help improve sampling. If the de novo density app is successfull it will produce an incomplete model fit into the density. If you remove any portions of that model you are unhappy with you can then put it into RosettaES which will add all the residues still missing. Alternatively if you there is some portion of your original model you are confident is correct you can remove everything except those segments and pass it into ES which will fill in the missing gaps.

Sat, 2017-07-08 13:49
brandon.frenz