Pittsburgh Supercomputing Center 

Advancing the state-of-the-art in high-performance computing,
communications and data analytics.

Trinity Example Files

 

Note: you will probably need to edit the job script for it to run succesfully.  Some things that may need to change:  your file names may be different;  your files may be stored in a different directory structure;  your job might need more time to complete; your job may require more or fewer cpus.  There may be others.

Be sure to tailor the script to suit your situation before submitting a job.  If you don't,  you may use up some of your allocation on a job that fails. If you have any questions, please contact remarks@psc.edu.

 

Run Trinity in one step

This example runs all stages of Trinity in one step.   This is only advised for small datasets — datasets that require a maximum of 16 cores or 128 Gb of memory, and that are expected to completely finish within 96 hours. 

Files include:

 

reads.left.fq.gz, input data      
reads.right.fq.gz, input data      
trinity_all.job, PBS script  View  
trinity_all.log, trinity output and error   View  
trinity_all.output, PBS standard output and error  View  

 

 

Run Trinity in multiple steps

Running Trinity in multiple steps is recommended for large datasets. If a job requires more than 96 hours to complete, PSC User Services can extend the time. To request a time extension, please contact User Services as soon as you know that it is necessary.

These jobs break the Trinity run into four steps:

  • Stage 1 - generate the kmer-catalog and runs Inchworm
  • Stage 2 - do the Chrysalis clustering of Inchworm contigs and mapping reads
  • Stage 3 - conduct the Chrysalis deBruijn graph construction
  • Stage 4 - run Butterfly and generate the final Trinity.fasta file

To run Trinity this way, submit the job for stage 1 (trinity_stage1.job). When it completes sucessfully, submit the second job, trinity_stage2.job, and so on.

Files include:

Stage 1: generate the kmer-catalog and run Inchworm

reads.left.fq.gz, input data    
reads.right.fq.gz, input data    
trinity_stage1.job, PBS script View  
trinity_stage1_example.log, trinity output and error View  
trinity_stage1.output, PBS standard output and error View  

Stage 2: do Chrysalis clustering of Inchworm contigs and mapping reads

trinity_stage2.job, PBS script View  
trinity_stage2_example.log, trinity output and error View  
trinity_stage2.output, PBS standard output and error View  

Stage 3: construct the Chrysalis deBruijn graphs

trinity_stage3.job, PBS script View  
trinity_stage3_example.log, trinity output and error View  
trinity_stage3.output, PBS standard output and error View  

Stage 4: run Butterfly and generate the final Trinity.fasta file

trinity_stage4.job, PBS script View  
trinity_stage4_example.log, trinity output and error View  
trinity_stage4.output, PBS standard output and error View  

Big assemblies scripts

This job script shows how to run Trinty with big assemblies: those with more than 200 million reads or those that use more than 64 cores.

big_assemblies.job, PBS script View  

Stay Connected

Stay Connected with PSC!

facebook 32 twitter 32 google-Plus-icon