Galaxy

From UABgrid Documentation
Revision as of 00:10, 16 September 2011 by Curtish@uab.edu (Talk | contribs)

Jump to: navigation, search

Contents


Overview

The UAB Galaxy platform for experimental biology and comparative genomics designed to help you analyze multiple alignments, compare genomic annotations, profile metagenomic samples and more from your web browser. This platform is built on Galaxy, backed by the Cheaha compute cluster, and powered by UABgrid. Documentation on the UAB installation can be found on the UAB Galaxy wiki.

Using Galaxy / Tutorials

There are several UAB tutorials on NGS Analysis with Galaxy, created for HPC Boot Camp 2011 and a nice talk by Jeremy Goecks during Reserach Computing Day 2011.

There are also numerous general tutorials online at the Penn State public Galaxy site that are worth looking at.


Galaxy@UAB

The UAB Galaxy instance can be accessed at http://galaxy.uabgrid.uab.edu using BlazerID credentials. The https/ssl access will be available soon. The UAB Galaxy instance is using revision 50e249442c5a from the upstream galaxy repository.

Temporary Protocol for moving large sequence files (>2GB) to UAB's galaxy instance (or very large numbers of files).

Hardware

Behind the scenes the Galaxy server at UAB is powered by Cheaha cluster.

Available Tools

Following is a list of tools available through Galaxy platform right now. More description will be added soon.

Software Version Information
bwa 0.5.9 Further information
bowtie 0.12.7 Further Information
lastz 1.02.00 Further information
samtools 0.1.12a Further information
Legacy blast (megablast) 2.2.25 Further information
srma 0.1.15 Further information
velvet 1.1.03 Further information
Top Hat 1.2.0 Further information
Cuff Links 1.0.1 Further information
Lift Over 26-Apr-2011 18:26 2.6M Further information
R R-2.13.0 Further information
RPy 1.0.3 Further information
ps2pdf  ?? Further information
MACS 1.4.0rc2 Further information
taxonomy2tree r3 Further information
sputnik NA Further information
beam2 Unknown Further information
addscores NA Further information
clustalw 2.1 Further information
gmaj NA Further information
gpass NA Further information
HYPHY 2.0020110330 beta Further information
laj NA Further information
pass2 NA Further information
twoBitToFa NA Further information
Perl revision 5 version 8 subversion 8 Further information
perM 3.3 Further information
GNUPlot 4.4.3 Further information
Numpy 1.6.0 Further information
numexpr 1.4.2 Further information
hdf5 1.8.7 Further information
Cython 0.14.1 Further information
Python Tables (tables) 2.2.1 Further information
FastX Toolkit 0.0.13 Further information


Adding Novel Datasets

Prerequisites

You should have checked out your own galaxy instance and run it from git as described in http://projects.uabgrid.uab.edu/galaxy/wiki/GalaxyDevelopment

Introduction

In order to add a new data set, a series of dependent files must be created and configured on cheaha.uabgrid.uab.edu. The configuration files are located in or under:

  • /share/apps/galaxy/galaxy-latest

The dependent files should be located in or under:

  • /lustre/project/public_datasets

Some of the older data sets are still located in or under:

  • /lustre/project/galaxy/public_dataset

I describe here setting up a basic genome and include only a description of how to set up 3 critical pieces:

  • bwa
  • bowtie
  • samtools

Obviously you will need to have an account on cheaha but you will also need to be in the galaxy-admin group.

FASTA File

Download your FASTA file (doing any conversions needed) and place the file in: /lustre/project/public_datasets/primary/MY_GENOME/MY_GENOME.fa

You should following the naming conventions in tool-data/shared/ucsc/builds.txt as shown below.

  • sacCer2 S. cerevisiae June 2008 (SGD/sacCer2) (sacCer2)

For instance, if there is already an entry for your build of S. cerevisiae, then use the dbkey (leftmost column) in builds.txt to name MY_GENOME. In this case it would be sacCer2. In some cases (tree shrew, obscure chimeric mouse genomes you construct yourself) there will be no entry. You will need to edit and create one yourself and update builds.txt.

Make sure the extension is .fa and don't worry if there are multiple files, they can be concatenated together as shown below.

Directory Creation (example for sacSer2)

mkdir /lustre/project/public_datasets/primary/sacCer2
cd /lustre/project/public_datasets/primary/sacCer2
wget http://hgdownload.cse.ucsc.edu/goldenPath/sacCer2/bigZips/chromFa.tar.gz
tar xzvf chromFa.tar.gz
cat chr*.fa 2micron.fa > sacCer2.fa


Index Creation

A not fully tested script has been written to index the genomes, it can be run instead of commands below to index genomes.

#!/bin/bash
# Argument #1 = GENOME_NAME - Should be of the form ce6, sacCer2, etc..
# This scripts builds the indices for a new genome. It assumes that:
# /lustre/project/public_datasets/primary/$1 and /lustre/project/public_datasets/primary/$1/$1.fa exist
mkdir /lustre/project/public_datasets/derived/$1
cd /lustre/project/public_datasets/derived/$1
mkdir bowtie
mkdir bowtie/color
mkdir bowtie/base
mkdir bwa
mkdir perm
mkdir sam
#Bowtie
cd /lustre/project/public_datasets/derived/$1/bowtie/base
cat /lustre/project/public_datasets/derived/sacCer2/bowtie/base/reindex_bowtie_sacCer2.bsh | perl -pe "s/sacCer2/$1/g;" > reindex_bowtie_$1.bsh
cat /lustre/project/public_datasets/derived/sacCer2/bowtie/base/submit_index_job  | perl -pe "s/sacCer2/$1/g;" > submit_index_job
chmod +x submit_index_job
chmod +x reindex_bowtie_$1.bsh
/lustre/project/public_datasets/derived/$1/bwa/submit_index_job
#BWA
cd /lustre/project/public_datasets/derived/$1/bwa
cat /lustre/project/public_datasets/derived/sacCer2/bwa/reindex_bwa_sacCer2.bsh | perl -pe "s/sacCer2/$1/g;" > reindex_bwa_$1.bsh
cat /lustre/project/public_datasets/derived/sacCer2/bwa/submit_index_job  | perl -pe "s/sacCer2/$1/g;" > submit_index_job
ln -s /lustre/project/public_datasets/primary/$1/$1.fa
chmod +x submit_index_job
chmod +x reindex_bwa_$1.bsh
/lustre/project/public_datasets/derived/$1/bwa/submit_index_job
#SAM and SRMA
cd /lustre/project/public_datasets/derived/$1/sam
cat /lustre/project/public_datasets/derived/sacCer2/sam/reindex_sam_sacCer2.bsh | perl -pe "s/sacCer2/$1/g;" > reindex_sam_$1.bsh
cat /lustre/project/public_datasets/derived/sacCer2/sam/submit_index_job  | perl -pe "s/sacCer2/$1/g;" > submit_index_job
ln -s /lustre/project/public_datasets/primary/$1/$1.fa
chmod +x submit_index_job
chmod +x reindex_sam_$1.bsh
/lustre/project/public_datasets/derived/$1/sam/submit_index_job

Configuration File Update

Follow the detail instructions that Shantanu has posted for creating a git branch for your changes, they can be found here: http://projects.uabgrid.uab.edu/galaxy/wiki/GalaxyDevelopment

I have written a small script to update the bowtie, bwa, samtools and srma indices called update_indices.bsh as shown below:

#!/bin/bash
#Argument #1 GENOME_NAME (ex. ce6)
#Argument #2 through 8 GENOME DESCRIPTION (C. elegans May 2008 (WS190/ce6) (ce6))
# Try to use description from tool-data/shared/ucsc/builds.txt if available
echo "$1        $1      $2 $3 $4 $5 $6 $7 $8    /lustre/project/public_datasets/derived/$1/bowtie/base/$1" >> /home/ozborn/projects/galaxy/galaxy/tool-data/bowtie_indices.loc
echo "$1        $1      $2 $3 $4 $5 $6 $7 $8    /lustre/project/public_datasets/derived/$1/bowtie/base/$1" >> /home/ozborn/projects/galaxy/galaxy/tool-data/bowtie_indices_color.loc
echo "$1        $1      $2 $3 $4 $5 $6 $7 $8    /lustre/project/public_datasets/derived/$1/bwa/$1.fa" >> /home/ozborn/projects/galaxy/galaxy/tool-data/bwa_index.loc
echo "$1        $1      $2 $3 $4 $5 $6 $7 $8    /lustre/project/public_datasets/derived/$1/bwa/$1.fa" >> /home/ozborn/projects/galaxy/galaxy/tool-data/bwa_index_color.loc
echo "index     $1      /lustre/project/public_datasets/derived/$1/sam/$1.fa" >> /home/ozborn/projects/galaxy/galaxy/tool-data/sam_fa_indices.loc
echo "$1        $1      $2 $3 $4 $5 $6 $7 $8    /lustre/project/public_datasets/derived/$1/sam/$1.fa" >> /home/ozborn/projects/galaxy/galaxy/tool-data/srma_index.loc

This will be committed and others can add to it to update blast databases, perM, and other indices as needed.


Final Steps

Log in to your local galaxy and see if you can run your job. If it all works out, contact Shantanu and push the changes to production. It is best to send a patch or set up your directory so he can pull from it.

Available datasets

Genome Downloaded Blast Database BWA Index Bowtie Index PerM Index Sam Index SRMA Dict
hg19 (by chromosome) Yes Yes Yes Yes Yes Yes Yes
Mouse (mm9) Yes Yes Yes Yes Yes Yes Yes
Vaccinia Western Reserve Yes No Yes Yes Yes Yes Yes
Mycoplasma pneumonniae (M129) Yes Yes Yes Yes Yes Yes Yes
Mycoplasma pneumonniae (FH) Yes Yes Yes Yes Yes Yes Yes
Chromosome 11 Mouse Contigs Yes No Yes Yes Yes Yes Yes
S. cerevisae (sacCer2) Yes No Yes Yes No Yes Yes
C. elegans (ce6 and WS226) Yes No Yes Yes No Yes Yes
Tree Shrew 62 Yes No Yes Yes No Yes Yes

Public instance

A public instance of Galaxy maintained by Penn State University is at http://usegalaxy.org/

Support

In order to facilitate interaction among UAB Galaxy users, share experience, and provide peer-support we have established a galaxy-users group. To join this group and participate in email discussions please subscribe to the galaxy-user group. On-line archives of these discussions are available here. Please note, the email discussions are a public forum. You are advised to only post information you are authorized to share and comfortable with being public.

References

Personal tools
Namespaces

Variants
Actions
Navigation
Toolbox