[Dock-fans] Possible problem with DOCK 3.7 post processing scripts

Trent E. Balius tbalius at aol.com
Wed Sep 26 10:00:37 PDT 2018

Thanks Corey for alerting us to your observation.  I think that we need to describe this better in the tutorials.

The rule of thumb is to make sure the energy cutoff when running extract all is less than the energy cutoff for writing to the mol2 file during docking (the mol2_score_maximum parameter).

for example, if mol2_score_maximum = -10  (this is the default value)

then run:

$DOCKBASE/analysis/extract_all.py -s -10


$DOCKBASE/analysis/extract_all_blazing_fast.py dirlist extact_all.txt -10

This should circumvent the problem of having to rerun DOCK.

I hope this helps,


Trent E. Balius, PhD
Associate Specialist, Shoichet Lab,
Dept. Pharm. Chem., UCSF
Phone: (415) 514-4253
URL: http://docking.org/~tbalius

Error message from the attracted file:

reading list of identifiers from file: extract_all.sort.uniq.txt
reading extract info from ./extract_all.txt
getting up to  1  poses for each of  341  ligands
gathering poses for molecule FRAGID_0151
reading extract info from ./extract_all.sort.uniq.txt
gathering poses for molecule FRAGID_0364
gathering poses for molecule FRAGID_0356
gathering poses for molecule FRAGID_0184
Traceback (most recent call last):
  File "/nfs/kolblab_progs/bin/DOCK-3.7.2rc1/analysis/getposes.py", line 358, in <module>
  File "/nfs/kolblab_progs/bin/DOCK-3.7.2rc1/analysis/getposes.py", line 355, in main
    bbndist=options.bbndist, bbndof=options.bbndof)
  File "/nfs/kolblab_progs/bin/DOCK-3.7.2rc1/analysis/getposes.py", line 192, in getPoses
    receptors, realRank, part, scoreToCheck, newRecScore, cache)
  File "/nfs/kolblab_progs/bin/DOCK-3.7.2rc1/analysis/getposes.py", line 111, in getOnePose
    cachedPoses = getPosesCache(file, molnum, id, rank, cache)
  File "/nfs/kolblab_progs/bin/DOCK-3.7.2rc1/analysis/getposes.py", line 92, in getPosesCache
    returnItem = marshal.loads(cacheDict[str([molnum, id, rank])])
  File "/nfs/kolblab_progs/Anaconda_x64/lib/python2.7/bsddb/__init__.py", line 270, in __getitem__
    return _DeadlockWrap(lambda: self.db[key])  # self.db[key]
  File "/nfs/kolblab_progs/Anaconda_x64/lib/python2.7/bsddb/dbutils.py", line 68, in DeadlockWrap
    return function(*_args, **_kwargs)
  File "/nfs/kolblab_progs/Anaconda_x64/lib/python2.7/bsddb/__init__.py", line 270, in <lambda>
    return _DeadlockWrap(lambda: self.db[key])  # self.db[key]
KeyError: "[334, 'FRAGID_0184', 1]"

-----Original Message-----
From: Corey Taylor <corey.taylor at uni-marburg.de>
To: dock-fans <dock-fans at docking.org>
Sent: Wed, Sep 26, 2018 9:06 am
Subject: [Dock-fans] Possible problem with DOCK 3.7 post processing scripts

Dear all,

We are using DOCK-3.7.2rc1 (cheers, John) and I believe I've bumped into
a problem with the post-processing scripts. Excuse my ignorance of the
actual mechanisms of the scripts, I'm mainly guessing.

The getposes.py script exited with the attached error. Eventually I
figured out that the pose it crashed on whilst trying to extract (in
this case, a molecule named "FRAGID_0184") was in the extraction
databases (extract_all.sort.uniq.txt, etc.) but not in test.mol2.gz so,
of course, the retrieval failed. This error was repeated whether I used
extract_all.py or extract_all_blazing_fast.py prior to running getposes_py.

Of course, the problem is fixed by simply making sure the default INDOCK
parameter 'mol2_score_maximum' (-10.0) is set high (+10.0) as per the
bkslab Wiki
(http://wiki.bkslab.org/index.php/DOCK_3.7_2018/06/05_abl1_Tutorial) and
running the dock again to ensure poses written to test.mol2.gz match
what end up in the extraction files later.

I may have missed a script parameter here so please advise if so or any
other way I haven't thought of to head-off having to re-run the entire
dock again or something similar. Mine was a very small dock so little
harm done but I dread the idea of someone else having to re-run a big
multi-day dock, hence I'm advising you guys of this in case there's a
more elegant solution possible for the scripts themselves.

Cheers guys,

Corey Taylor
Kolb Lab
Institute of Pharmaceutical Chemistry
Philipps-University Marburg
Marbacher Weg 6
35032 Marburg

Mailto: corey.taylor at uni-marburg.de
Ph: +49 6421 28 21351
Web: http://www.kolblab.org/taylor.html

Dock-fans mailing list
Dock-fans at docking.org
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.docking.org/pipermail/dock-fans/attachments/20180926/3b4b37c8/attachment.html 

More information about the Dock-fans mailing list