Installing witsGWAS

Getting a copy of witsGWAS:

For WITS cluster users:

1. Copy the witsGWAS folder at /opt/exp_soft/bioinf/witsGWAS/ into your home directory.
2. Take a look at the README.txt for quick setup info

For Independent users:

witsGWAS software requirements

Software	Version
Admixture	>= 1.21
Affymetrix Power Tools	1.15.2
Eigenstrat
Emmax	>= 23.1.1
Perl	>= 5.10.1
PLINK	>= 1.07
PLINK	>= 1.9
Python	>= 2.7.2
R	>= 3.2.1
Rubra	0.1.5
Ruffus	2.2
pyfiglet	0.7.4
Graphviz (a.k.a Dot program)	2.26.0

Download the witsGWAS codebase

witsGWAS-0.1.0

Installing Rubra

• git clone https://github.com/bjpop/rubra
• cd rubra
• python setup.py install

Note: Installing Rubra automatically takes care of the installation of Ruffus 2.2

Customizing Rubra for your cluster

• Update the names of the job queues in Rubra's cluster_job.py with ones relevant to your cluster. See the example below (lines 96 -111) replacing Rubra's job queues (terri and terri-smp) with WITS cluster specific queues (WitsLong and medium).

A comparison of the differences between Rubra's cluster_job.py (left) and Rubra's cluster_job.py edited for witsGWAS (right).

Zoom in on the web browser to get a better view of the image

Tip: Rubra's cluster_job.py can be customized to run on an SGE cluster instead of PBS torque. See lines 140 - 152 in cluster_job.py

2. Changes that need to be made to PBS Torque

witsGWAS produces flagFiles (.Success) to track successful completion of jobs. To facilitate this, the keep-completed attribute in PBS Torque is set to 15 secs.

Notes: TORQUE provides the ability to report on the status of completed jobs for a configurable duration after the job has completed. This can be enabled by setting the keep_completed attribute on the job execution queue. This should be set to the number of seconds that jobs should be held in the queue. Completed jobs will be reported in the C state and the exit status is seen in the exit_status job attribute (Submitting and managing jobs, section 2.5, TORQUE Administrator Guide, version 4.1.4).

3. Setting -up the environment

witsGWAS uses a module called gwaspipe as referred to in pipeline_name_stages_config.py files. Create a module entitled: gwaspipe that sets up paths to provide access to bioinformatics software. See the image below as a guide:

An image showing the contents of the gwaspipe module

#%Module 1.0
#
#  bioinf module for use with 'environment-modules' package:
#  sets up path to provide access to bioinformatics software
# 


setenv LANG "en_GB.UTF-8"
setenv LC_ALL "en_GB.UTF-8"
prepend-path PATH /opt/exp_soft/bioinf/bin
prepend-path PATH /opt/exp_soft/apt/bin
prepend-path PATH /opt/exp_soft/python27/bin
prepend-path PERL5LIB /opt/exp_soft/bioinf/lib/perl5/site_perl/
prepend-path R_LIBS /opt/exp_soft/bioinf/R
prepend-path LD_LIBRARY_PATH /opt/exp_soft/lib64
prepend-path LD_LIBRARY_PATH /opt/exp_soft/bioinf/lib64
prepend-path LD_LIBRARY_PATH /opt/exp_soft/bioinf/lib
prepend-path  PYTHONPATH /opt/exp_soft/bioinf/lib64/python/

=== Note to WITS cluster users:

• The gwaspipe module can be viewed by logging into worker node no.1

ssh n01.core.wits.ac.za
more /etc/modulefiles/gwaspipe

1. The software listed in the witsGWAS software requirements above is already available on the WITS cluster through the gwaspipe module
2. WITS cluster users need NOT apply any changes to Rubra's cluster_job.py
3. WITS cluster users do NOT need to create the environment module gwaspipe as it is already available on the cluster

Examples

For each pipeline in witsGWAS, an example dataset and corresponding results have been included in the example_datasets sub-directory.

Pipeline	Dataset
Affymetrix QC	affyqc
	affyqc_no-pheno-info
PLINK QC	plinkqc
	plinkqc_no-sex-info
Association testing	assoc_testing
	assoc_testing_no-sex-info