Streamlined tophat/bowtie python pipeline

[Gus]

Hello everyone,

I am not sure if this is the correct place for this but I welcome this thread being moved if there is a better place.

I have developed a streamlined tophat/cufflinks pipeline that can run all steps for multiple sample groups/experiments with a single command if so desired. Log files record settings and output messages for easy tracking and methods writing months after you actually ran your data through the pipe. Email notifications are available and a single stage failure does not cause the whole run to fail.

It has a feature for running these analyses in a an SGE administered cluster environment as well.

I have not encoded ALL options from every program yet but welcome fellow programmers adding options/features that they use/need and pushing their changes to github for inclusion in the next release.

The project is quite functional for me but it is still considered below a v1.0 release at the present moment.

I have written the docs using sphinx and they are hosted on github as well.

I have tried to make the code as 'self documenting' as possible with liberal use of docstrings as well.

My main point in posting this here is to get feed back and attract collaboration on developing 'blacktie' to the point that people with a minimal command line comfort level will feel confident using these valuable tools. There may be other pipelines out there in the wild but I did not find any that were marketed past a single lab's needs.

Please try it out and give me feed back or feature requests if you feel the call.


Links:
github project repo: https://github.com/xguse/blacktie
docs: http://xguse.github.com/blacktie/

for up to the hour builds pull from the 'develop' branch


blacktie is also listed on the PyPI so v0.1-dev is install-able via:

$ [sudo] pip install blacktie

Thank you for your time and help,

Gus

[Gus]

Bumped version to 0.1.1

Hi everyone. I just bumped the PyPI version to 0.1.1 after adding some option checking and making sure that the CufflinksCall objects properly dealt with --GTF vs --GTF-guide options.

I also updated the docs.

If you installed via pip you should:

$ [sudo] pip install -U blacktie

Gus

[Gus]

I have released a new version of blacktie (0.2.0rc1) which has a number of minor fixes and new features.

A major gotcha will be that the yaml config file has been changed slightly to facilitate the inclusion of biological replicate data. So please take a look at the demo config file in the docs posted at: http://xguse.github.io/blacktie/tuto...iguration-file

Thanks and please let me know what you think. I need your input to make improvements.

Gus

News
====

0.2.0rc1
--------

*Release date: 2013-04-19*

* git tag: 'v0.2.0rc1'
* email_notification is now adjustable for other email servers
* Added support for handling biological replicates in cuffdiff runs.
* Major changes to yaml config:
    * condition_queue[index].group_id --> condition_queue[index].experiment_id
    * addition of condition_queue[index].replicate_id to track replicate data
    * condition_queue[index].name now represents description of an 'experimental condition' and will be shared by replicates.

[Gus]

This is an update.

I have cummerbund working inside of rpy2 and expect a new release containing automation of the basic functionality of cummerbund very soon.

Gus

[Gus]

I have a new blacktie release that now integrates automated cummeRbund plots.

Blacktie v0.2.1 with screenshots!

[Gus]

BUGFIX release: 0.2.1.2

0.2.1.2
-----------
*Release date: 2013-07-17*

* version number system adopted to conform to PEP386
* This is a hot fix to squash a bug described in github issue: https://github.com/xguse/blacktie/issues/10
* **in short:** on a Mac ``pprocess`` complained "AttributeError: 'module' object has no attribute 'poll'" when trying to set up a queue.
* the quick fix is to look for that exception and continue without ``pprocess`` if encountered to avoid the fatal exception.

Attention:
- It is recommended that all users update to 0.2.1.2 to avoid this issue.
- Anyone with knowledge of this issue is encouraged to comment on the `issues thread <https://github.com/xguse/blacktie/issues/10>`_ above or in the Comments section below.