Hi all, I'm using Modeller 9v3 (Linux) and 9v1 (WindowsXP). I perform a rudimentary automodel with the script below. On both systems I get 200 output models, all of which appear to have an identical number of residues and atoms, however the script just hangs when it executes the cluster command.
---------------- # Homology modeling by the automodel class from modeller import * from modeller.automodel import *
log.verbose() env = environ()
# directories for input atom files env.io.atom_files_directory = './'
a = automodel(env, alnfile = 'sequences3.pir', knowns = ('prot1','prot2','prot3'), sequence = 'protx', assess_methods = (assess.GA341, assess.DOPE)) a.starting_model= 1 a.ending_model = 200
a.make()
a.cluster() -----------------
I've then used the clustering script found at http://salilab.org/modeller/wiki/Cluster_PDBs with the 200 models, which uses transfer_xyz. This time the script ends with a segmentation fault (Linux) or closes unexpectedly (Windows XP). There are are no errors reported in the log file in either instance, it simply executes fine up to the point where clustering is requested and gets no further.
As this hasn't been mentioned before I assume it's not a bug. Has anyone else had this problem and overcome it?
Thanks! James
James Irving wrote: > I'm using Modeller 9v3 (Linux) and 9v1 (WindowsXP). I perform a > rudimentary automodel with the script below. On both systems I get 200 > output models, all of which appear to have an identical number of > residues and atoms, however the script just hangs when it executes the > cluster command.
How long did you leave the script running? Many parts of Modeller are computationally expensive, and can run for half an hour or more without producing any output.
> I've then used the clustering script found at > http://salilab.org/modeller/wiki/Cluster_PDBs with the 200 models, which > uses transfer_xyz. This time the script ends with a segmentation fault > (Linux) or closes unexpectedly (Windows XP). There are are no errors > reported in the log file in either instance, it simply executes fine up > to the point where clustering is requested and gets no further. > > As this hasn't been mentioned before I assume it's not a bug.
On the contrary, a segmentation fault is *always* a bug. Thanks for bringing it to my attention. Could you send me (not the list) your full input files (e.g. alignments, PDBs, etc.) so that I can reproduce and fix your two problems?
Ben Webb, Modeller Caretaker