Dear All:
My group is a licensee of Amber14 (the most current version we've licensed) into which AmberTools 12 has been integrated (and updated via the "update_amber --upgrade" mechanism, last checked just yesterday. The update appears to have been successful.
I've been interested in trying the 'amberlite' suite as a straightforward means of generating trajectories for further MM-GB/PBSA binding energy evaluations, and have worked my way through the p38 example (Section 5.10.3.2, page 142 of the AmberTools12 Reference Manual).
Everything works fine until the last step of the tutorial. I've followed along with my own calculations, using just the initial supplied protein/ligand complex p38.pdb file and not leaning on any of the other pre-supplied files. The *leap.* files are created (all correctly symlinked to their default filenames as suggested in the tutorial).
I can determine the interaction energy in the unrefined original complex as per the tutorial, obtaining results very close (within a tenth of a kcal/mol) to the values in the tutorial. I can correctly perform the minimization and determine this interaction energy of the minimized complex, again obtaining results close to those in the tutorial. Finally, I am able to run the 100 ps MD trajectory per the example with no problems. I see the successful creation of the com.binpos trajectory file. However, upon attempting to perform the final step of the tutorial (pymdpbsa upon the trajectory) I get the error message shown at the very bottom of this message.
Looking in the P38_******.tmpdir subdirectory, I see (among some symlinks) ten P38.L.pdb.* files, which seems to make sense to me, given the provided flags " --start 10 --stop 100 --step 10 ". However I don't understand why 'getpdb' would then be looking for a file 'P38.L.pdb.20" given those flags.
I'm assuming that the "failure in molsurf" message is just a downstream consequence of the script breaking at an earlier point in time. That binary appears to be healthy (I get usage information upon running it from the command line).
I would greatly appreciate any insight or assistance anyone could provide. A search of the reflector archive locates a few other questions about failures in amberlite's pymdpbsa, those questions a few years old now, but nothing pertaining to my particular 'getpdb' error, and those prior questions appeared to have been unanswered to date.
My 'mdnab' steps generated a binary com.binpos file of 4296404 bytes. Error message at the bottom, technical details below.
OS: Ubuntu Linux 16.04 LTS
Python: 2.7.11 (Anaconda custom) (also tried system Python 2.7.12, bypassing the Anaconda install)
Sourcing amber.sh in ~/.profile
Thank you again.
With thanks and kind regards,
-Michael Bartberger
$ pymdpbsa --proj P38 --traj com.binpos --start 10 --stop 100 --step 10
getpdb: can't open file P38.L.pdb.20
failure in molsurf
Traceback (most recent call last):
File "/home/mdb/apps/amber14/bin/pymdpbsa", line 792, in <module>
opt.start, opt.stop, opt.step, gb, sa)
File "/home/mdb/apps/amber14/bin/pymdpbsa", line 95, in nrgtable_gb
etot, ebat, evdw, ecoul, egb, sasa = ffgbsa(prmtop, pdbfile, gb, sa)
File "/home/mdb/apps/amber14/bin/pymdpbsa", line 271, in ffgbsa
etot = etot + sasa
UnboundLocalError: local variable 'etot' referenced before assignment
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Sun Sep 03 2017 - 09:30:02 PDT
