Page Comparison

Trinity, developed at the Broad Institute and the Hebrew University of Jerusalem, performs de novo reconstruction of transcriptomes from RNA-seq data. If you run this on Artemis, you will achieve best performance if you specify the --workdir option to use local node storage instead of /project, /scratch, or /home. This is because large assemblies generate millions of small files that degrade the performance of Artemis's Lustre filesystem. Node local storage performs much better than Lustre for this workload.

The way to tell Trinity to write to node local storage is to You can use the PBS variable "$TMPDIR" . $TMPDIR contains the path to the compute node's to tell Trinity to write to local node storage. You Remember that you need to additionally specify an a working directory with the word "trinity" in it for the job to run. To set up the Therefore, to make a Trinity working directory, add the following two lines to your PBS script before you run Trinity:

...

You can then add the option --workdir ${TRINITY_WORKDIR} to your Trinity command. An example PBS script to run the sample job provided with Trinity version 2.6.6 is provided below. In this job, I copied /usr/local/trinity/2.6.6/sample_data/test_Trinity_Assembly to /project/Project/abcd1234/trinity/input/test_Trinity_Assembly before submitting the job. You will need to project used in the below script to match your Artemis project.

#!/bin/bash
#PBS -P Project
#PBS -l select=1:ncpus=4:mem=16gb
#PBS -l walltime=1:00:00
#PBS -j oe

module load python/2.7.13
module load bowtie2/2.3.3.1
module load samtools/1.8
module load jellyfish/2.2.6
module load salmon/0.11.0
module load trinity/2.6.6

export TRINITY_WORKDIR=${TMPDIR}/trinity_workdir
export OUTDIR=/project/Project/abcd1234/trinity/trinity_output
mkdir -p ${TRINITY_WORKDIR}

cd /project/Project/abcd1234/trinity/input/test_Trinity_Assembly

#######################################################
##  Run Trinity to Generate Transcriptome Assemblies ##
#######################################################

Trinity --seqType fq --max_memory 4G \
              --left reads.left.fq.gz \
              --right reads.right.fq.gz \
              --output ${OUTDIR} \
              --workdir ${TRINITY_WORKDIR} \
              --SS_lib_type RF \
              --CPU 4

...