Installing STRetch

Quick Start Installation

This is an example of installing STRetch using the install script. It assumes Git and Java are already installed.

The install script also downloads the hg19 genome and annotation files ~8 GB. It takes approximately 20 minutes (depending on your download speed).

Download STRetch and run the install script:

git clone [email protected]:Oshlack/STRetch.git STRetch
cd STRetch
./install.sh

Download the test data

mkdir test
cd test/
wget -O testdata.zip https://ndownloader.figshare.com/articles/4762489?private_link=cc7347f4637d9a7fe22d
unzip testdata.zip 
rm testdata.zip

Run the test data

../tools/bin/bpipe run -p EXOME_TARGET="SCA8_region.bed" ../pipelines/STRetch_exome_pipeline.groovy *.fastq.gz

(The test data only includes reads from one locus, so runs very quickly, but may produce warning messages about there being no data from many chromosomes - this is fine)

Notes:

• If you would like to use the included PCR WGS reference set as controls, you will need to edit the "CONTROL" line in pipeline_config.groovy.
• To use system versions of any software, simply edit pipeline_config.groovy to point to the location of the executables. Be aware that using different versions of the tools can results in errors.
• If using STRetch on a cluster, you can use bpipe to send jobs to the queue. You will need to create STRetch/pipelines/bpipe.config. See http://docs.bpipe.org/Guides/ResourceManagers/ for instructions, as they will likely differ by cluster.

Requirements

Operating system: Linux

The installation script requires the following to run (most of of these will already be installed on many systems). The remaining dependancies can be installed automatically.

• Git
• Java (tested on version 1.8)
• bzip2
• unzip

Bpipe - this pipeline software runs the STRetch pipelines

Python 3

• See environment.yml for specific packages
• Easiest to install using Conda: conda env create -f environment.yml
• Then activate the environment with source activate STR.
• For some cluster environments you may need to point the pipeline to the specific version of Python, which would look something like this [where you installed conda]/miniconda3/envs/STR/bin/python

These commandline tools are required, but will already be installed by environment.yml as part of the python dependencies

• BedTools
• goleft
• mosdepth

Other command line tools

• BWA (requires mem algorithm - tested on version 0.7.12)
• Samtools
• Picard
• Bazam - for extracting reads if starting from bam/cram files

Reference data

Quick start for human data: You can download a bundle of the pre-indexed hg19 genome with STR decoys and corresponding bed files here.

Required reference files:

• Reference genome with STR decoy chromosomes (fasta)
• BWA indices of reference
• .genome file of reference
• STR decoy bed file
• STR positions in genome annotated bed file

All chromosomes names in these files must be in the same sort order.

STRetch requires a reference genome that includes STR decoy chromosomes. You can generate this by adding concatenating STRdecoys.fasta to the end of any reference genome in fasta format, and then indexing the result. For example:

$ cat ucsc.hg19.fasta STRdecoys.fasta > ucsc.hg19.STRdecoys.fasta
$ bwa index ucsc.hg19.STRdecoys.fasta

STRetch also requires a bed file specifying the position of all STRs in the reference genome, with two additional columns containing the repeat unit/motif and the number of repeat units in the reference.

For example:

chr1 10000 10468 TAACCC 77.2

This file is produced by extracting the appropriate columns from Tandem Repeats Finder output.

Configuring STRetch

install.sh will create a text file STRetch/pipelines/pipeline_config.groovy (note this file must be in the same directory as the pipelines). This will contain system-specific settings, especially paths to the locations of software and reference files. There is a template, and several examples provided: pipelines/config-examples/pipeline_config_[cluster-name].groovy.

Troubleshooting ?

See the Troubleshooting wiki page.

old STRetch version 0.1.0 - R requirements

R 3 (tested on R version 3.3.1)

• install.packages(c('optparse','plyr','dplyr','tidyr','reshape2'))

Install R packages:

R
> install.packages(c('optparse','plyr','dplyr','tidyr','reshape2'))
> q()