February 2001 - modeller_usage

Error using MALIGN3D
by Michael Buck 24 Jan '02

24 Jan '02

MALIGN3D Top files which worked well using Modeler 4 are giving me this error : fndatmi_284E> Beginning atom index for a residue out of bounds: 0 0 recover____E> ERROR_STATUS >= STOP_ON_ERROR: 1 1 I am using the following script # MALIGN3D READ_ALIGNMENT FILE = 'pair.aa' SET ATOM_FILES = '1.pdb' '2.pdb' SET FIT_ATOMS = 'CA' MALIGN3D GAP_PENALTIES_3D= 0 2.0, WRITE_FIT = on, WRITE_WHOLE_PDB = off If anybody has seen this problem, please help. Thank you … [View More]

2 1

problem!
by Paola D'Alessio 15 Jun '01

15 Jun '01

Dear users, I am trying to produce a chimeric model for my protein. The first thing I want to do is to check the alignment I have. When I tell Modeller to execute the "check_alignment" command, I get the following message: rdpir___E> alignment sequence not found in PDB file: 1 ./pXXXX.pdb ####################################################################################################################################### The alignment I have is: >P1;pXXXX structureX:… [View More]

2 1

Modeller4 on IRIX 6.5: additional versions available?
by Allen Smith 03 Feb '01

03 Feb '01

Hi. We're about to start using modeller4 on various IRIX 6.5 machines. These include r5000 (O2s and Indys) and r12000 (an Octane) processors. We are wondering whether the following could be made available: A. A version that will run properly on r12000 machines. These are using a later version of the r10000 processor, with both mips4 and 64-bit capabilities. The existing IRIX 6.5 r10000 versions will run only with a modification to the startup scripts of the LD_LIBRARY64_PATH … [View More]variable adding /usr/lib64/mips4 to the start of that path. The reason this is necessary is that the -r10000 flag evidently used in compiling the r10000 versions of Modeller adds /usr/lib64/mips4/r10000 to the path of directories searched for libraries. This directory contains libfastm.so, which won't run on r12000 machines (I suspect it has assembler optimizations for r10000). Forcing rld (the runtime loader) to look at /usr/lib64/mips4 first causes it to instead use the more generalized mips4 version of libfastm.so. While this modification enables the tests to run properly, it qualifies as inconvenient and prone to bugs (especially if dealing with other programs such as InsightII that also set LD_LIBRARY64_PATH) and prone to bugs. A generalized mips4, 64-bit version of modeller4 should be compilable via adding: IRIX64 ) FC=f77; FFLAGS="-automatic -L/usr/lib64/mips4 -mips4 -64 -r8 -O2 -GCM:aggressive_speculation=ON -OPT:roundoff=3"; LDFLAGS="-L/usr/lib64/mips4 -Wl,-rpath,/usr/lib64/mips4 -lfastm" ;; \ to Makefile.include1. If it will work properly, f90 might be used instead of f77. -LNO and -IPA should enable better performance. B. A version that will run well on r5000 machines. These are 32-bit machines but will run mips4 code, which should improve modeller4's performance relative to the mips3 code which the existing IRIX n32-bit versions use. A generalized mips4, n32-bit version of modeller4 should be compilable via adding: mips4 ) FC=f77; FFLAGS="-automatic -L/usr/lib32/mips4 -mips4 -n32 -r8 -O2 -GCM:aggressive_speculation=ON -OPT:roundoff=3"; LDFLAGS="-L/usr/lib32/mips4 -Wl,-rpath,/usr/lib32/mips4 -lfastm" ;; \ to Makefile.include1. Again, if it will work properly, f90 might be used instead of f77, and -LNO and -IPA should enable better performance. In both of the above, specialized r5000 and r12000 versions would probably work even better. I don't know if you have either of these types of machines available to you. We are capable of such compilation ourselves but (obviously) don't have the modeller4 source code available to us. Thanks, -Allen -- Allen Smith easmith(a)beatrice.rutgers.edu [View Less]

1 0