Release 2.4
New in Cpipe 2.4
Installer
- Created new installer which manages a full cpipe install from scratch in as little as 30 minutes by downloading assets stored in a Swift object store (NeCTAR)
- The installer bundles most runtimes like Python, Perl, Groovy etc., ensuring Cpipe will run identically on different machines
- This also means that Cpipe requires less software to be installed on the host machine
- This process is managed by an install script in the root directory (
./install.sh) - This process is documented in the install documentation
Run Script
- Cpipe now has a central run script (
./cpipe) which can be used to run most common cpipe commands, includingrun,test, manipulate batches, designs and metadata files. - The script is documented in the commands documentation
Documentation
- Comprehensive documentation is now included in Cpipe, which is visible either in your local copies or on GitHub
- The documentation covers installing, running, configuring and various other aspects to operating Cpipe
- Have a look at the documentation index here
Docker support
- Cpipe is now able to run in a docker container
- Just clone the repo, copy in the swift_credentials.sh file (refer to the documentation for an explanation), and run
docker build .and you can generate a Cpipe docker container
ALL Design
- In line with new recommendations, we now encourage you to use the built-in ALL design instead of creating a gene list for your own analysis
- This design will cover all official UCSC genes instead of limiting the analysis. This will save having to re-analyse the data in case your gene list expands
Configuration Changes
- The bpipe.config file has moved to bpipe.config.template, which is copied to bpipe.config when your run the install. This prevents your bpipe.config settings from being overwritten when you pull never versions of Cpipe.
- The bpipe.config has been rewritten to include a template for use with an HPC queue system (e.g. torque). This means you only need to remove one comment and specify your queue system and Cpipe should work with your system
- Command names have been given to each command in the pipeline, meaning that their settings (wall time, RAM etc.) can be configured in bpipe.config using the relevant name. For example, "vep" for the VEP command.
- The
samples.txtfile previously located in each batch directory has been renamed toconfig.batch.groovyto reflect that it can be used to change any Cpipe configuration settings for the batch - New
JAVA_OPTSvariable in config.groovy/config.batch.groovy for setting flags that should always be used for Java, for instance-noverifyfor certain versions of Java that dislike Groovy bytecode.
Miscellaneous Changes
- The Cpipe directory has been restructured to place all reference data in
data, all tools and runtimes intools, and all installation tasks intasks - Some un-needed files have been removed from Cpipe
Support
2.4 will receive continued support on the 2.4 branch of Cpipe. Unlike the dev branch, which will often contain broken developments versions of Cpipe, if you remain on the 2.4 branch and pull periodically you will keep up-to-date but without any breaking changes.