David,
I'm cc'ing the galaxy-dev mailing list; let's keep this thread on the list for archival and community purposes.
I had a look through the Trinity.pl script and found that the default butterfly heap is set to 1G.
I've changed this to 2G. Would you re-run the trinity job and let me know if this helps.
Rather than mucking with the script, It's possible to change the Butterfly memory size using command line parameters:
--- Butterfly options: # # --run_butterfly :executes butterfly commands. Do not set this if you want to spawn them on a computing grid. # # --bfly_opts :parameters to pass through to butterfly (see butterfly documentation). # # --bflyHeapSpace :java heap space setting for butterfly (default: 1000M) => yields command java -Xmx1000M -jar Butterfly.jar ... $bfly_opts ---
This can be done in the Trinity wrapper at <galaxy-home>/tools/ngs_rna/trinity_all.xml ; there are already some Butterfly options set in the wrapper, so it should be easy to add more.
We see the following at the end of the Trinity logs - I don't know what it means but I thought you might... number of good components: 1 number of reads found = 1253 (from total of 1254) which came from 1252 pairs ## Read PathPair results: 1251 singletons, num pairs: 1, num pairs discarded: 0 total number of components = 1 working on subcomponent 0 method: combineSimilarPathsThatEndAtV(-1) with 1 paths method: combineSimilarPathsThatEndAtV(2) with 1 paths method: combineSimilarPathsThatEndAtV(1248) with 1 paths method: combineSimilarPathsThatEndAtV(1223) with 1 paths method: combineSimilarPathsThatEndAtV(-2) with 2 paths
Path Illustration:
==== PATH: [-1, 2, 1248, -2] === Read: [[-1, 2, 1248], []] count: 2 = Read: [[2], []] count: 1030 == Read: [[2, 1248], []] count: 23 === Read: [[2, 1248, -2], []] count: 2 = Read: [[1248], []] count: 197
Final path reported: comp12043_c0_seq1 len=1658 ~FPKM=29.7 path=[2:0-1220 1248:1221-1657] FinalPath: [-1, 2, 1248, -2] with support <1254, 1254> total number of paths reported = 1 from 1 components Done SUCCESS[53682]: java -Xmx1G -jar /gpfs/cluster/isys/galaxy/Galaxy/trinityrnaseq_r2011-08-20/Butterfly/Butterfly.jar -N 25506261 -L 300 -F 300 -C chrysalis/RawComps.6/comp12043 -V 10 --stderr SUCCESS[53680]: java -Xmx1G -jar /gpfs/cluster/isys/galaxy/Galaxy/trinityrnaseq_r2011-08-20/Butterfly/Butterfly.jar -N 25506261 -L 300 -F 300 -C chrysalis/RawComps.24/comp48425 -V 10 --stderr SUCCESS[53681]: java -Xmx1G -jar /gpfs/cluster/isys/galaxy/Galaxy/trinityrnaseq_r2011-08-20/Butterfly/Butterfly.jar -N 25506261 -L 300 -F 300 -C chrysalis/RawComps.8/comp16349 -V 10 --stderr
Sorry, 1 of 53682 jobs failed.
Overall, this looks very good - almost all the jobs succeeded.
Failed commands written to file: failed_cmds.27222.txt
Error, cmd: /gpfs/cluster/isys/galaxy/Galaxy/trinityrnaseq_r2011-08-20/util/cmd_process_forker.pl -c /gpfs/cluster/isys/galaxy/Galaxy/galaxy-dist/database/job_working_directory/679/trinity_out_dir/chrysalis/butterfly_commands.adj --CPU 4 --shuffle died with ret 256 at /gpfs/cluster/isys/galaxy/Galaxy/trinityrnaseq_r2011-08-20/Trinity.pl line 571.
This is a bit cryptic to me. You'll want to determine what error code 256 means and see whether it may be a memory issue; you may also want to email the Trinity mailing list and add for help there.
Good luck, J.