You are here

Relax is resulting in a Segmentation Fault

4 posts / 0 new
Last post
Relax is resulting in a Segmentation Fault
#1

First off, I hate to be asking about a "Segmentation Fault" here because I know it can be any number of things; but I'm completely stuck and need to ask for help.

I am preparing a structure for downstream modeling, and I am running a very simple relax following the steps outlined here (the short protocol):

relax.linuxgccrelease \
    -database /databases/rosetta/ \
    -relax:constrain_relax_to_start_coords \
    -relax:coord_constrain_sidechains \
    -relax:ramp_constraints false \
    -s /inputs/3BBT_cleaned.pdb \
    -ex1 \
    -ex2 \
    -use_input_sc \
    -flip_HNQ \
    -no_optH false

 

And almost immedately run into a Segmentation Fault error. To me, it looks like all inputs are ok. The only user-defined input is really the PDB, which I can run through clean_pdb.py, and upon inspection, looks alright (I attached it anyhow).

What might be the cause of the segmentation fault? Are there any additional verbose logging options that I could get more information from?

Here is the full output of the relax run:

core.init: Rosetta version unknown:cbe8723f7038f0b9e5d24fca9c3728b2fc952a37 2016-08-02 10:58:29 -0400 from /scratch/local-benchmark/release/rosetta/git/release/rosetta.binary.linux.release.git
core.init: command: /rosetta/bin/relax.static.linuxgccrelease -database /databases/rosetta/ -relax:constrain_relax_to_start_coords -relax:coord_constrain_sidechains -relax:ramp_constraints false -s /inputs/3BBT_cleaned.pdb -ex1 -ex2 -use_input_sc -flip_HNQ -no_optH false
core.init: 'RNG device' seed mode, using '/dev/urandom', seed=1401820833 seed_offset=0 real_seed=1401820833
core.init.random: RandomGenerator:init: Normal mode, seed=1401820833 RG_type=mt19937
core.scoring.ScoreFunctionFactory: SCOREFUNCTION: talaris2014
core.scoring.etable: Starting energy table calculation
core.scoring.etable: smooth_etable: changing atr/rep split to bottom of energy well
core.scoring.etable: smooth_etable: spline smoothing lj etables (maxdis = 6)
core.scoring.etable: smooth_etable: spline smoothing solvation etables (max_dis = 6)
core.scoring.etable: Finished calculating energy tables.
basic.io.database: Database file opened: scoring/score_functions/hbonds/sp2_elec_params/HBPoly1D.csv
basic.io.database: Database file opened: scoring/score_functions/hbonds/sp2_elec_params/HBFadeIntervals.csv
basic.io.database: Database file opened: scoring/score_functions/hbonds/sp2_elec_params/HBEval.csv
basic.io.database: Database file opened: scoring/score_functions/rama/Rama_smooth_dyn.dat_ss_6.4
basic.io.database: Database file opened: scoring/score_functions/P_AA_pp/P_AA
basic.io.database: Database file opened: scoring/score_functions/P_AA_pp/P_AA_n
basic.io.database: Database file opened: scoring/score_functions/P_AA_pp/P_AA_pp
protocols.relax.FastRelax: ================== Using default script - no constraint ramping ==================
protocols.jd2.PDBJobInputter: Instantiate PDBJobInputter
protocols.jd2.PDBJobInputter: PDBJobInputter::fill_jobs
protocols.jd2.PDBJobInputter: pushed /inputs/3BBT_cleaned.pdb nstruct index 1
protocols.evaluation.ChiWellRmsdEvaluatorCreator: Evaluation Creator active ... 
protocols.jd2.PDBJobInputter: PDBJobInputter::pose_from_job
protocols.jd2.PDBJobInputter: filling pose from PDB /inputs/3BBT_cleaned.pdb
core.chemical.ResidueTypeSet: For ResidueTypeSet fa_standard there is no shadow_list.txt file to list known PDB ids.
core.chemical.ResidueTypeSet:     This will turn off PDB component loading for ResidueTypeSet fa_standard
core.chemical.ResidueTypeSet:     Expected file: /databases/rosetta/chemical/residue_type_sets/fa_standard/shadow_list.txt
core.chemical: Warning: protein residue GLY is not explicitly listed as either L or D in its params file.
core.chemical: To avoid seeing this warning in the future, add "L_AA", "D_AA", or "ACHIRAL_BACKBONE" to the "PROPERTIES" line of the params file.
core.chemical: Warning: protein residue XXX is not explicitly listed as either L or D in its params file.
core.chemical: To avoid seeing this warning in the future, add "L_AA", "D_AA", or "ACHIRAL_BACKBONE" to the "PROPERTIES" line of the params file.
core.chemical: Warning: protein residue V03 is not explicitly listed as either L or D in its params file.
core.chemical: To avoid seeing this warning in the future, add "L_AA", "D_AA", or "ACHIRAL_BACKBONE" to the "PROPERTIES" line of the params file.
core.chemical: Warning: protein residue V04 is not explicitly listed as either L or D in its params file.
core.chemical: To avoid seeing this warning in the future, add "L_AA", "D_AA", or "ACHIRAL_BACKBONE" to the "PROPERTIES" line of the params file.
core.chemical: Warning: protein residue HTY is not explicitly listed as either L or D in its params file.
core.chemical: To avoid seeing this warning in the future, add "L_AA", "D_AA", or "ACHIRAL_BACKBONE" to the "PROPERTIES" line of the params file.
core.chemical: Warning: protein residue  BZ is not explicitly listed as either L or D in its params file.
core.chemical: To avoid seeing this warning in the future, add "L_AA", "D_AA", or "ACHIRAL_BACKBONE" to the "PROPERTIES" line of the params file.
core.chemical: Warning: protein residue SAL is not explicitly listed as either L or D in its params file.
core.chemical: To avoid seeing this warning in the future, add "L_AA", "D_AA", or "ACHIRAL_BACKBONE" to the "PROPERTIES" line of the params file.
core.chemical: Warning: protein residue BCS is not explicitly listed as either L or D in its params file.
core.chemical: To avoid seeing this warning in the future, add "L_AA", "D_AA", or "ACHIRAL_BACKBONE" to the "PROPERTIES" line of the params file.
Segmentation fault

 

No real warnings or errors to speak of. I'm lost here, so any pointers that anyone has will go a long ways.

Thanks!

AttachmentSize
3BBT_cleaned.pdb169.81 KB
Post Situation: 
Thu, 2017-08-24 18:46
brspurri

I see your Rosetta version is a year old.

Here are some guesses:

1) mismatched database and code, if you have more than one copy of each on your machine

2) Try compiling and running the debug build (remove mode=release when compiling), that often gives better error messages for seg faults (particularly if the error is a vector overrrun; those are bounds checked in debug mode).  This is a fast error so it won't be painful.

3) Try the same command line with a different PDB - that will help debug issue 1 above, and give us data on whether it's related to the actual PDB or something else is wrong

4) just try recompiling, in case it's bitrot of some sort (or a corrupted bit of hard disk or something)

5) try a more recent (or just different) rosetta

6) check your input for colinear atoms - 3 colinear atoms will cause problems because their torsions are undefined, although usually later than this.  That's usually an issue with PDBs created programattically; your file name implies it's from the PDB so this seems unlikely.

7) wait for Rocco to come along and guess right

Speaking for Rocco as well - we REALLY APPRECIATE that you note ahead of time "I know seg fault is a useless error to report" - saves us a ton of stress worrying about your understanding of the problem!

Fri, 2017-08-25 08:38
smlewis

The prevelence of "Warning: protein residue XXX is not explicitly listed as either L or D in its params file." strongly indicates to me that it's a database/code version mis-match issue.

Fri, 2017-08-25 09:39
rmoretti

That was exactly it! Thank you. Apparently, I somehow mixed data/versions when downloading new bundles. Thanks all.

Sat, 2017-08-26 07:08
brspurri