Pipeline Execution

The pipeline excution relies on the JIP pipeline system. It allows local execution and also provides HPC clusters integration.

The main GRAPE workflow can be run with the grape run command and is composed by two blocks:

  • pre-processing step to generate all the required input files for running the pipeline (one for all the samples in the project)
  • processing of the RNA-seq data (one for each samples)

The first block can also be run independently with the grape run setup command. The setup block check all the prerequisites for any configured pipeline and run all the necessary steps to reach a valid initial state from which the pipeline can be run. At present it consits of the following steps:

  • GEM genome index
  • GEM transcriptome index

In the second block the following steps are performed:

  • GEMtools RNA mapping pipeline run
  • GEMtools mappings filtering
  • isorform expression estimation with the Flux Capacitor

The pipeline can be run for a subset of the project dataset. Assuming that foo is a valid id for a sample you could run:

$ grape run foo

to run the pipeline on the ‘foo’ sample data.

To run the pipeline on a HPC cluster you need to configure the JIP. Please refer to the JIP documentation for more information about this topic. With a valid JIP cluster configuration the grape submit command can be used to run jobs on the cluster.

Note

A specific JIP database file for each GRAPE project is used and can be found at <project>/.grape/grape_jip.db.

Default Pipeline

At the moment a Default Pipeline is configured, which includes the following modules:

  • CRGtools, a set of in-house scripts and binaries from the Computational Biology of RNA Processing group at CRG Barcelona
  • SAMtools, utilities to manipulate sam files
  • GEMTools, utilities and pipelines for RNA-seq mapping
  • FluxCapacitor, a program for isoform quantification estimation