This repository has been archived by the owner on Apr 11, 2023. It is now read-only.
Cluster Flow v0.4
This is a huge release that got lost on it's way. A great number of things have changed since the
v0.3 release, so please be prepared for a bit of tinkering if you're updating.
I'll try to make releases smaller and more frequently in the future to avoid this happening again!
Download
Download Cluster Flow v0.4 (.zip)
Download Cluster Flow v0.4 (.tar.gz)
Changelog
- Warning: Break of backwards compatability
- The way that genome references are handled has been rewritten.
- Genome references are no longer tied to specific types, they are now agnostic.
to the type of reference, making it far easier to whatever type of reference you need.
Additionally, the wizard to add genome paths has been written and is now largely automated,
making it super fast to add new genomes. - A consequence of this change is any
genomes.config
files written before v0.3 of
Cluster Flow will no longer work. Thankfully the fix is easy! Replace@bowtie_path
with@reference bowtie
.@gtf_path
changes to@reference gtf
and so on.
@genome_path
changes to@reference fasta
. - If you have any custom pipelines these will also need to be updated.
@require_bowtie
changes to@require_reference bowtie
and so on. See updated example module files
for examples on how to update custom modules. - Apologies for any inconvenience that this change incurs. Feel free to get in touch
if you have any problems.
- Genome references are no longer tied to specific types, they are now agnostic.
~/clusterflow/
directory moved to~/.clusterflow/
to reduce home directory clutter.- Cluster Flow won't find your old config file - run
mv ~/clusterflow/ ~/.clusterflow/
to fix.
- Cluster Flow won't find your old config file - run
- The way that genome references are handled has been rewritten.
- New Stuff
- New website! http://clusterflow.io
- New home of cluster flow documentation
- Docs are now versioned along with CF version.
- Online interactive demo and other cool stuff. Code can be found at https://github.com/ewels/clusterflow-website
- You can now run Cluster Flow locally (new
@cluster_environment
local
)- Tested on Mac OSX and Linux. Includes
--qstat
and--qdel
functionality - Allows easy testing and use of pipelines for those without access to a HPC cluster.
- New
--environment
command line option allows you to set this at run time.
- Tested on Mac OSX and Linux. Includes
- The
--make_config
wizard has been renamed to--setup
and does a lot more stuff- Should make first-run of Cluster Flow much easier - just download and run
cf --setup
- Should make first-run of Cluster Flow much easier - just download and run
- Support for STAR RNA-seq aligner (thanks to @stu2)
- New ChIP-Seq analysis pipeline (thanks to @orzechoj)
- Includes six new modules:
bedToNrf
,bedtools_bamToBed
,deeptools_bamCoverage
,deeptools_bamFingerprint
,phantompeaktools_runSpp
andpicard_dedup
- Includes six new modules:
- Modules are given more information via the run file to help
decide the amount of memory and cores they bid for (eg. number of files, reference) - Running job stats (
cf --qstat
) now has a summary at the bottom, and counts running and completed jobs. - New config option
@time_multiplier
to globally adjust time requests for slow clusters- To stop things getting silly, there's also
@max_time
to put a top limit on what's possible..
- To stop things getting silly, there's also
- Added a special case value
null
for the@environment_module_alias
config option, to prevent Cluster Flow from loading that specific module. - All perl scripts now have
env perl
in shebang to increase portability - Modules can now have file extensions, as long as they have
.cfmod
at the end of the basename- This helps editing tools with syntax highlighting, amongst other things
- Python comes to Cluster Flow! The first Python module is up and running, along with a
Helpers.py
module file- See the
example_module.py
file for help in writing your own modules in Python - The basic Perl module helpers are now available in the Python packages as well, more translation to follow
- See the
- Now using GRIDEngine
h_vmem
memory option instead ofvf
- Gives a hard memory limit instead of a request limit at job submission time
- Thanks to @stu2 and @s-andrews
- Support for explicit GRIDEngine queue nomination on the command line
- Modules now print their software versions to the log where possible.
- New
--merge
(command line) and@merge_regex
(config file) options to automatically merge input files.- This is implemented using a new module,
cf_merge_files
, which can also be used in pipelines - If the supplied regexes only match single files, the module can be used to simply rename files
- This is implemented using a new module,
- New
--runfile_prefix
option to help avoid potential filename clashes. - New
@cluster_project
config option to specify project for cluster jobs. - Added compatability with GRIDEngine
~/.sge_request files
(by ignoring them).
Thanks to @s-andrews - New tophat module called
tophat
which introduces a workaround for buggy MAPQ
reporting by tophat whilst keeping unique alignments. Thanks to @FelixKrueger.- The previous tophat module is still available if you're not interested in MAPQ scores and
would like slightly faster processing. This is now calledtophat_broken_MAPQ.cfmod
.
- The previous tophat module is still available if you're not interested in MAPQ scores and
- Pipeline completion e-mails are now written to disk as well (HTML and plain text)
- New log file containing the job submission commands as well as the output received from the cluster at submission (usually numeric job identifiers)
- Removed the
--qstatcols
command line option and added the@colourful
config option to replace it- The config wizard is also updated to add this to your personal config
- Added checks to make sure that we have at least one config file, and that the cluster environment is set
- Added new
@environment_module_always
config option to always load certain environment modules at run time. - Added new
@require_python_package
pipeline option to check that a Python pacakge is installed before pipeline launch
- New website! http://clusterflow.io
- Bugs Squashed
- Fixed output filename problem in tophat with output cleaning
- Fixed bugs causing minimum memory allocation regardless of availability
- Fixed bug causing Bowtie2 to fail if Bowtie1 index absent
- Cleaned up some unrecognised output that always made it into the log file
- Allow multiple merge regexes