Pepspec – random runtime segfaults (rosetta 3.3 database replaced) [solved]

[Anonymous]

Hello.

When running pepspec (pepspec.linuxgccrelease) I encounter random runtime segmentation faults, occurring at different “places”/times in consecutive runs with the same input and .args files.

I’m running Rosetta 3.3 on Ubuntu 11.04 workstation (quadcore, 4GB RAM) which has (seemingly) successfully compiled.
The pepspec demo test runs without problems.
Rosetta 3.3 database is replaced with the working version.

I’ve tried several different .pdb files, processed with fbb and otherwise and several different .args files. Segmentation faults occur with all.

I thought there was a problem with my .pdb files or .args, so I tried:
-running pepspec on the demo .pdbs (generated with anchordock step) with my .args file
-running pepspec on my .pdb with demo .args file (adjusted input filenames etc.)
Neither worked – randomly timed segmentation faults occurred in both.

I’ve also run the Anchor Docking (pepspec_anchor_dock.linuxgccrelease) step on a complex, successfully, and used the output .pdbs of it as input for pepspec, with my .args file and the demo one, adjusted. Both runs ended with segmentation faults.

I moved from a virtual machine setup to a dual-boot Ubuntu workstation, in case there was a memory sharing issue, but it didn’t help.

Could this be a result of faulty compiling on my part?
(scons built the targets and demo run of pepspec works, as do some other Rosetta applications, but I have very little experience with compilers)

First time working with Rosetta, so I’m likely making a mistake somewhere. Any and all help concerning the cause of the problem is greatly appreciated.

I’ve attached the terminal output for three consecutive runs with the same input files that terminated in segfaults, as well as the contents of .args file for the run.
I can provide my input .pdb files if that would be of help.

P.S. If anyone has a test .pdb and .args file confirmed to run successfully with pepspec and would be interested in letting me test my Rosetta pepspec installation with those, I’d greatly appreciate it.

Thank you for your time and consideration.

[Anonymous]

Hi loreseeker. This error is occurring because the pepspec code was dependent on some other code in rosetta that was changed immediately before the most recent release. I’ve attached a patch for the pepspec application. Just rename this file (pepspec.cc.txt) to “pepspec.cc” and move it into the folder “rosetta_source/src/apps/public/pepspec/”, then recompile. You may want to make a backup of the original pepspec.cc file first. Email me at “dr (dot) chris (dot) king (at) gmail” if you still have problems and I can get your inputs and see what the problem is. Sorry for the confusion.

[Anonymous]

Hello, chrisk.

Thank you very much for your assistance. I’ve followed your instructions and they’ve solved the problem. Runs finish normally now.
My apologies for a delayed reply – wasn’t around a computer last week.

Thank you again for your help.

[Author]

Posts