You are here

RosettaCM: An internal error has occurred when "Running the Hybridize mover"

5 posts / 0 new
Last post
RosettaCM: An internal error has occurred when "Running the Hybridize mover"
#1

Hello,

I am trying to replicate the tutorial for RosettaCM according to the following URL:

https://www.rosettacommons.org/docs/latest/application_documentation/structure_prediction/RosettaCM

The step 1 (using setup_RosettaCM.py) is fine where I can obtain the thread pdb file and flags and rosetta_cm.xml.

But when I tried to run step 2 (Run the Hybridize mover). It crashed.

The command I used is listed below:

# step 1

python2 /sysapps/cluster/software/Rosetta/3.11/rosetta_bin_linux_2019.35.60890_bundle/tools/protein_tools/scripts/setup_RosettaCM.py \
--fasta IN.fasta \
--alignment aln.fasta \
--alignment_format fasta \
--templates 6u8qA.pdb \
--rosetta_bin /sysapps/cluster/software/Rosetta/3.11/rosetta_bin_linux_2019.35.60890_bundle/main/source/bin \
--platform linux \
--compiler gcc \
--build default

# step 2

cd rosetta_cm

/sysapps/cluster/software/Rosetta/3.11/rosetta_bin_linux_2019.35.60890_bundle/main/source/bin/rosetta_scripts.default.linuxgccrelease \
@flags \
-database /sysapps/cluster/software/Rosetta/3.11/rosetta_bin_linux_2019.35.60890_bundle/main/database \
-nstruct 20

The input files I used:

>IN
FLDGIDKAQEEHEKYHSNWRAMASDFNLPPVVAKEIVASCDKCQLKGEAMHGQVDCSPGIWQLDCTHLEGKVILVAVHVA
SGYIEAEVIPAETGQETAYFLLKLAGRWPVKTVHTDNGSNFTSTTVKAACWWAGIKQEFGIPYNPQSQGVIESMNKELKK
IIGQVRDQAEHLKTAVQMAVFIHNFKRKGGIGGYSAGERIVDIIATDIQTKELQKQITKIQNFRVYYRDSRDPVWKGPAK
LLWKGEGAVVIQDNSDIKVVPRRKAKIIRDYGKQMAGDDCVASRQDED

and the template is the 6u8qA.pdb (chain A based on the PDB id: 6u8q)

The crashed log file is uploaded.

Please help to figure out what is the reason for this crash?

Another question is, in this protocol, do we need the Fragment Library? (I did not see any thing about Fragment Library in this protocol)

Thanks.

-Ming

AttachmentSize
ROSETTA_CRASH.log5.14 KB
Post Situation: 
Mon, 2020-06-15 05:56
haom

This is a segfault in carbohydrates code (although backtraces become suspect when segfaults occur). 

The general advice with segfaults is to compile the non-release build (remove mode=release when compiling - this will take a while) and run it again.  Debug mode failures often give more interpretable error messages.

Do you intentionally have carbohydrates in your model, and if so are they important or can you remove them?  (If they are the source of the crash, and you don't care about them, then removing them from the input is a fast fix.)

Yes, hybridize wants fragments for one of the middle phases of its modeling.

Mon, 2020-06-15 16:07
smlewis

Hi smlewis,

Thanks for your reply.

Do you intentionally have carbohydrates in your model, and if so are they important or can you remove them?

-- no. I just want to follow the protocol but use my own target sequence and the selected template.

Yes, hybridize wants fragments for one of the middle phases of its modeling.

-- but in the following files, I do not see where is the Fragment.

The flags is listed below:

# i/o
-in:file:fasta /nethome/haom/IN/IN.fasta
-nstruct 20
-parser:protocol rosetta_cm.xml

# relax options
-relax:dualspace
-relax:jump_move true
-default_max_cycles 200
-beta_cart
-hybridize:stage1_probability 1.0
 

and rosetta_cm.xml is listed below:

 

<ROSETTASCRIPTS>
    <TASKOPERATIONS>
    </TASKOPERATIONS>
    <SCOREFXNS>
        <ScoreFunction name="stage1" weights="score3" symmetric="0">
            <Reweight scoretype="atom_pair_constraint" weight="0.1"/>
        </ScoreFunction>
        <ScoreFunction name="stage2" weights="score4_smooth_cart" symmetric="0">
            <Reweight scoretype="atom_pair_constraint" weight="0.1"/>
        </ScoreFunction>
        <ScoreFunction name="fullatom" weights="beta_cart" symmetric="0">
            <Reweight scoretype="atom_pair_constraint" weight="0.1"/>
        </ScoreFunction>
    </SCOREFXNS>
    <FILTERS>
    </FILTERS>
    <MOVERS>
        <Hybridize name="hybridize" stage1_scorefxn="stage1" stage2_scorefxn="stage2" fa_scorefxn="fullatom" batch="1" stage1_increase_cycles="1.0" stage2_increase_cycles="1.0">
            <Template pdb="/nethome/haom/IN/rosetta_cm/6u8qA_thread.pdb" cst_file="AUTO" weight="1.0" />
        </Hybridize>
    </MOVERS>
    <APPLY_TO_POSE>
    </APPLY_TO_POSE>
    <PROTOCOLS>
        <Add mover="hybridize"/>
    </PROTOCOLS>
</ROSETTASCRIPTS>
 

In this two files, I do not see where is Fragment File (rosetta_scripts.default.linuxgccrelease will automatically call Fragment from database?)? How can I add Fragment into rosetta_cm.xml or flags?

The command I used according to the tutorial and the prompt from setup_RosettaCM.py

/sysapps/cluster/software/Rosetta/3.11/rosetta_bin_linux_2019.35.60890_bundle/main/source/bin/rosetta_scripts.default.linuxgccrelease \
@flags \
-database /sysapps/cluster/software/Rosetta/3.11/rosetta_bin_linux_2019.35.60890_bundle/main/database \
-nstruct 20

Here this step crashed.

 

Is there other details for succuessfully run this protocol?

Thanks,

Ming

 

 

Mon, 2020-06-15 16:21
haom

You can add an element like this to your hybridize element in the XML (so it is at the same level as your template element, a sub element of the hybridize element):

 

<Fragments three_mers="inputs/1bbh.3mers_trim5" nine_mers="inputs/1bbh.9mers_trim5"/>
 

 see also https://www.rosettacommons.org/demos/latest/tutorials/rosetta_cm/rosetta_cm_tutorial.html

 

 

Mon, 2020-06-15 20:49
smlewis

Thank you, smlewis.

I will try that one.

 

Tue, 2020-06-16 05:55
haom