Dear ModPipe community
I managed to install modpipe in our system, but I'm struggling to test it. If I try to run the demo sequence using mode 0010 (providing the missing databases in the configuration file) I get the following error:
$ ../main/ModPipe.pl --conf_file modpipe.conf --sequence_id 97e075794f588a59e8a0fb8a945814b1MLGIKIKP --hits_mode 0010 --score_by_tsvmod OFF
MPLib::MPModules::GetProfile__M> Found profile for sequence MPLib::MPModules::GetProfile__M> Seqid : 97e075794f588a59e8a0fb8a945814b1MLGIKIKP MPLib::MPModules::GetProfile__M> Profile: /nfs/[...]/modpipe-2.2.0/demo//data/97e/97e075794f588a59e8a0fb8a945814b1MLGIKIKP/sequence/97e075794f588a59e8a0fb8a945814b1MLGIKIKP-uniprot90.prf MPLib::MPModules::GetProfile__M> Will use the above profile MPLib::MPModules::GetProfile__M> PRFUPDATE: OFF MPLib::MPUtils::CopyXPRFDB__M> No. of structure profiles in database: 39330 *PLLib::Modeller::RunModeller__E> MODELLER failed: exited with value 1** **HitsPrfPrf.pl__E> MODELLER failed on profile: 97e075794f588a59e8a0fb8a945814b1MLGIKIKP-uniprot90.prf** **MPLib::MPModules::GetHitsPP__E> Failed to calculate hits (0010): 97e075794f588a59e8a0fb8a945814b1MLGIKIKP** *MPLib::MPModules::GetHitsPP__E> /nfs/[...]/modpipe-2.2.0/src/HitsPrfPrf.pl --alignment_basename 97e075794f588a59e8a0fb8a945814b1MLGIKIKP-prfprf-0010 --set_score_statistics ON --output_alidir 97e075794f588a59e8a0fb8a945814b1MLGIKIKP-prfprf-0010-ali --hitfile_name 97e075794f588a59e8a0fb8a945814b1MLGIKIKP-prfprf-0010.hits --ccmatrix_offset -100 --profile_list /nfs/[...]/modpipe-2.2.0/dbs/pdb95/db/pdb95_prf.list --pdb_repository /nfs/[...]/modpipe-2.2.0/demo//db/pdb --target_profile 97e075794f588a59e8a0fb8a945814b1MLGIKIKP-uniprot90.prf --score_type CCMAT --gap_extend_cost -70 --custom_tag "MODELLER Prf-Prf (0010)" --profile_pssmdb /nfs/[...]/modpipe-2.2.0/dbs/pdb95/db/pdb95_prf.pssm --target_code 97e075794f588a59e8a0fb8a945814b1MLGIKIKP --summary_file 97e075794f588a59e8a0fb8a945814b1MLGIKIKP-prfprf-0010.sum --gap_open_cost -700 --e_value 1 ModPipe.pl__E> Failed to get Prf-Prf hits (0010): 97e075794f588a59e8a0fb8a945814b1MLGIKIKP
When running the modeller command from the tmp directory the following ABNORMAL TERMINATION error appears:
$ /nfs/[...]/modpipe-2.2.0/ext/mod/bin/modpy.sh /nfs/[...]/bin/python2 97e075794f588a59e8a0fb8a945814b1MLGIKIKP-prfprf.py
MODELLER 9.13, 2014/02/12, r9876
PROTEIN STRUCTURE MODELLING BY SATISFACTION OF SPATIAL RESTRAINTS
Copyright(c) 1989-2014 Andrej Sali All Rights Reserved
Written by A. Sali with help from B. Webb, M.S. Madhusudhan, M-Y. Shen, G.Q. Dong, M.A. Marti-Renom, N. Eswar, F. Alber, M. Topf, B. Oliva, A. Fiser, R. Sanchez, B. Yerkovich, A. Badretdinov, F. Melo, J.P. Overington, E. Feyfant University of California, San Francisco, USA Rockefeller University, New York, USA Harvard University, Cambridge, USA Imperial Cancer Research Fund, London, UK Birkbeck College, University of London, London, UK
Kind, OS, HostName, Kernel, Processor: 4, Linux [...] 2.6.32-358.14.1.el6.x86_64 x86_64 Date and time of compilation : 2014/02/11 22:46:42 MODELLER executable type : x86_64-intel8 Job starting time (YY/MM/DD HH:MM:SS): 2014/06/05 13:57:11
*alloc_p_215E> Dynamic memory allocation failed.** ** Routine, variable, status: ipssm 3** ***** ABNORMAL TERMINATION of Modeller. This is a bug** ** in Modeller. Please report it to the Modeller** ** developers, including any input files necessary** ** to reproduce the problem.*
Directory names and hostname have been redacted.
The configuration file that has been used it's identical as the one provided in the modpipe website as an example, with the addition of the XPRF_LIST and XPRF_PSSMDB variables.
Thanks for your help, Marco
On 6/5/14, 6:03 AM, Marco Galardini wrote: > *alloc_p_215E> Dynamic memory allocation failed.** > ** Routine, variable, status: ipssm 3**
This happens if you give Modeller an empty PSSM database (it tries to allocate 0 bytes of memory to store it). It's already "fixed" in the upcoming 9.14 release, in that it'll give you a more intelligible error message, but the underlying cause is that something is wrong with your PSSM file.
Ben
Hi Ben,
thanks for your prompt reply: the problem is that the PSSM file should be ok.
the conf file has these values:
XPRF_LIST /nfs/[...]/modpipe-2.2.0/dbs/pdb95/db/pdb95_prf.list XPRF_PSSMDB /nfs/[...]/modpipe-2.2.0/dbs/pdb95/db/pdb95_prf.pssm
the files in that location are there and not empty:
$ ls -lh total 1.3G -rw-r--r-- 1 [...] [...] 3.7M Jun 5 09:53 pdb95_prf.list -rw-r--r-- 1 [...] [...] 1.1G Jun 5 09:55 pdb95_prf.pssm
$ head -n 1 pdb95_prf.list /nfs/[...]/modpipe-2.2.0/dbs/pdb95/data/2a/12asA/12asA-uniprot90.prf
$ head pdb95_prf.pssm /nfs/[...]/modpipe-2.2.0/dbs/pdb95/data/2a/12asA/12asA-uniprot90.prf 327 583 405 409 421 336 436 371 355 421 355 386 450 386 421 386 646 471 379 286 336 336 336 286 318 690 286 492 409 318 409 409 318 268 353 318 336 336 386 518 764 357 347 256 280 392 216 261 700 412 520 507 267 266 289 292 311 357 567 260 349 468 258 588 583 268 280 622 310 348 496 335 356 287 588 396 509 394 339 224 306 363 298 346 416 286 286 340 334 550 458 383 381 329 668 495 430 646 354 281 314 374 274 434 566 282 321 420 294 470 365 421 420 371 760 466 421 374 332 319 369 377 221 403 602 226 303 322 379 573 343 327 337 288 721 516 342 452 408 223 274 663 318 353 580 255 372 455 285 458 294 322 331 320 602 443 400 404 334 235 291
Any other parameter or file I should look for?
Many thanks, Marco
On 05/06/2014 19:55, Ben Webb wrote: > On 6/5/14, 6:03 AM, Marco Galardini wrote: >> *alloc_p_215E> Dynamic memory allocation failed.** >> ** Routine, variable, status: ipssm 3** > > This happens if you give Modeller an empty PSSM database (it tries to > allocate 0 bytes of memory to store it). It's already "fixed" in the > upcoming 9.14 release, in that it'll give you a more intelligible > error message, but the underlying cause is that something is wrong > with your PSSM file. > > Ben
On 6/6/14, 5:29 AM, Marco Galardini wrote: > thanks for your prompt reply: the problem is that the PSSM file should > be ok.
Looks OK. If you can send me the Modeller input file and whatever file it's trying to read (presumably the .pssm) then I can take a look and see if I can reproduce your problem here. (If you can cut down the .pssm to a smaller size and still reproduce the problem, that would help ;) - look for the // separators.)
Ben
Dear Ben,
turned out that the problem was indeed the "//" separator between the profiles: when I applied 'sed' to change the path in the pssm and list files the double slash was removed, resulting in this quite obscure error.
Sorry for using your time, Marco
On 11/06/2014 21:27, Ben Webb wrote: > On 6/6/14, 5:29 AM, Marco Galardini wrote: >> thanks for your prompt reply: the problem is that the PSSM file should >> be ok. > > Looks OK. If you can send me the Modeller input file and whatever file > it's trying to read (presumably the .pssm) then I can take a look and > see if I can reproduce your problem here. (If you can cut down the > .pssm to a smaller size and still reproduce the problem, that would > help ;) - look for the // separators.) > > Ben
Hi Ben,
sorry to bother you, have you had any time to look into our case? The pssm profiles seem ok to me, any other thing we should look at? Should the pdb directory contain uncompressed structures?
Kind regards, Marco
On 05/06/14 19:55, Ben Webb wrote: > On 6/5/14, 6:03 AM, Marco Galardini wrote: >> *alloc_p_215E> Dynamic memory allocation failed.** >> ** Routine, variable, status: ipssm 3** > > This happens if you give Modeller an empty PSSM database (it tries to > allocate 0 bytes of memory to store it). It's already "fixed" in the > upcoming 9.14 release, in that it'll give you a more intelligible > error message, but the underlying cause is that something is wrong > with your PSSM file. > > Ben
-
Ben Webb -
Marco Galardini