Trinity genome guided

270 views
Skip to first unread message

Eric Wafula

unread,
May 5, 2015, 11:43:11 AM5/5/15
to trinityrn...@googlegroups.com, Brian Haas
Hi Brain,

Might you have an idea why genome guided trinity assembly fails with the error below?
Seem do to samtools sort, but cannot figure out that the cause is.

[bam_sort_core] merging from 210 files...

bash: line 1:   839 Exit 9                  gsnap -d gsnap_target.gmap -D . -A sam --nofails -N 1 -t 30 -w 10000 -n 20 /scratch/projects/Colletotrichum/combined_R1.fq /scratch/projects/Colletotrichum/combined_R2.fq

       840 Aborted                 | samtools view -bt gsnap_target.fa.fai -

       841 Done                    | samtools sort -o - - > gsnap.coordSorted.bam

Error, cmd: bash -c "set -o pipefail; gsnap -d gsnap_target.gmap -D . -A sam --nofails -N 1 -t 30 -w 10000 -n 20 /scratch/projects/Colletotrichum/combined_R1.fq /scratch/projects/Colletotrichum/combined_R2.fq | samtools view -bt gsnap_target.fa.fai - | samtools sort -o - - > gsnap.coordSorted.bam " died with ret 34304 at /scratch/biotools/software/assembly/trinityrnaseq_r20140717/Trinity line 1990.


Command:

/scratch/biotools/software/assembly/trinityrnaseq_r20140717/Trinity --seqType fq --JM 180G --left combined_R1.fq --right combined_R2.fq --SS_lib_type FR --CPU 30 --no_cleanup --genome genome.new.fasta --genome_guided_max_intron 10000 --genome_guided_sort_buffer 20G --genome_guided_CPU 30 --GMAP_CPU 30


Approx 800,000,000 reads

Eric

Brian Haas

unread,
May 5, 2015, 6:13:20 PM5/5/15
to Eric Wafula, trinityrn...@googlegroups.com
Hi Eric,

It looks like something went wrong during the gsnap alignment phase.  Note, we have removed support for running gsnap as part of Trinity-GG in newer releases.  I suggest to upgrade to the latest Trinity, and to run GSNAP (or whatever your favorite short read spliced alignment tool is) separately to generate your coordinate-sorted BAM file, then give the coord-sorted bam file to the new Trinity software for assembly.

best,

~brian

--
--
Brian J. Haas
The Broad Institute
http://broadinstitute.org/~bhaas

 

Eric Wafula

unread,
May 8, 2015, 9:51:54 AM5/8/15
to Brian Haas, trinityrn...@googlegroups.com
Hi Brian,

I am not able to install any of the Trinity v2 releases presumably because of our gcc compiler (v4.1.2 on centos 5.9). I've spent a week trying to update gcc to newer versions with/without environment modules unsuccessfully. Upgrading the to centos 6.x which might have compatible gcc to install Trinity v2  is currently not an option for other reasons. 
Is there any modification that can be made on the Trinity install procedure to have it install with an older gcc like ours?

See compilation error below.

Thanks.

Eric

#####################################################

make[1]: Entering directory `/scratch/biotools/software/assembly/trinityrnaseq-2.0.6/Inchworm'

Making install in src

make[2]: Entering directory `/scratch/biotools/software/assembly/trinityrnaseq-2.0.6/Inchworm/src'

g++ -DHAVE_CONFIG_H -I. -I..    -std=c++0x -pedantic -fopenmp -Wall -Wextra -Wno-deprecated -m64 -g -O2 -MT Fasta_entry.o -MD -MP -MF .deps/Fasta_entry.Tpo -c -o Fasta_entry.o Fasta_entry.cpp

cc1plus: error: unrecognized command line option "-std=c++0x"

make[2]: *** [Fasta_entry.o] Error 1

make[2]: Leaving directory `/scratch/biotools/software/assembly/trinityrnaseq-2.0.6/Inchworm/src'

make[1]: *** [install-recursive] Error 1

make[1]: Leaving directory `/scratch/biotools/software/assembly/trinityrnaseq-2.0.6/Inchworm'

make: *** [inchworm_target] Error 2

[ekw10@comandra trinityrnaseq-2.0.6]$

#################################################################


Brian Haas

unread,
May 8, 2015, 6:42:12 PM5/8/15
to Eric Wafula, trinityrn...@googlegroups.com
Hi Eric,

I'm actually not sure how to get it to install with older versions of GCC.  I'd continue to try to get gcc 4.9 or higher installed and go with that.  GCC should build from source (though take a long time - several hours usually).  You can always stick with an older version of Trinity if needed. We try to keep certain things backwards compatible as we move forward w/ Trinity development.

If you're in a bind, you can always use our Galaxy site for generating your Trinity assembly:

and then do all the downstream analyses locally (not requiring a Trinity compilation - just the plugins).

best,

~brian

Reply all
Reply to author
Forward
0 new messages