Skip to content

Latest commit

 

History

History
100 lines (65 loc) · 3.2 KB

SCC_README.md

File metadata and controls

100 lines (65 loc) · 3.2 KB

Setting Up Dependencies

The following workflow eliminates the direct use of most modules. Basically you'll be setting up your own custom installs of each dependency. To use this setup it's advised to not use the modules that you're replacing, especially Python. To see what modules you have loaded, use the list command:

module list

To remove a package, use the unload command:

module unload python2.7

Before we proceed, you'll need the linga-proxy module loaded if you're working on scc4.bu.edu:

module load linga-proxy

The python version you'll need to use is called anaconda:

module load anaconda/2.1.0

This provides access to the conda tool, which is used to install packages where ever you want. To setup your own install environment, use the following command:

conda create -p ./demo_env --copy python==2.7.3 pip --yes

This will create an install in the directory ./demo_env, which means it's setup relative to where you are currently working. It will install python (currently pinned to the version used by the pipeline, but that can change in the future) and pip, which you can use to install packages that conda is not aware of. Basically try conda install whatever and if no package is available, then try pip install whatever.

Once you've created the environment, then you can "activate" it, which means make it the current one to use (this process is complementary to the module system):

source activate ./demo_env

Now install all the other dependencies:

conda install --copy --yes \
   -c file:///restricted/projectnb/montilab-p/conda_channel \
   samtools bowtie2 cufflinks cutadapt fastqc\
   htseq pysam r==3.0.0 subread tophat

Traditionally conda is ment for hard to install python packages (e.g. numpy, scipy and matplotlib), and conda is aware of an online repository (hense the 'linga-proxy' requirement above). The montilab-p project has a "channel" containing all the packages that the pipeline uses, some that use R, some that use Python, all ready to be installed in the same manor as a python package. The above command installs these programs for you in your ./demo_env directory.

Currently the pipeline is not configured to be installed in the same manor, but that will change in the near future. When that happens, you can install it along with all the other packages (in fact, you'll need to just install the pipeline, and conda will automagically install all the dependencies).

Using conda-based setup

If you haven't already done this, activate your conda environment:

module load anaconda
source activate path/to/your/demo_env/can/be/relative

Now just run the pipeline:

python path/to/CBMgithub/tools/RNASeq_pipeline/RNASeq_pipeline.py -p param.txt

If you ever want to stop using a conda environment, then 'deactivate' it:

source deactivate

This will bring back the default anaconda install.

Using Git Repo

To get latest version: git fetch origin

To add a file: git add file (or *)

To push all the changes to the github repository: git commit (filename or -a) git push origin master

#nice and simple slide set that walks you through git + github http://rogerdudler.github.io/git-guide/

Also there is a graphical interface: gitk