This repository contains:

• pipelines for re-analysis of iCLIP and scRNA-Seq data for UMI-tools publication
• ipython notebook for simulations
• ipython notebooks to generate publication figures

To run the scRNA-Seq pipeline, create a new directory and copy the files in the 'data' directory to the new directory, along with the configuration files:

mkdir scRNA_seq
cp [UMI-Tool_pipelines git directory]/data/* scRNA_seq
cp [UMI-Tool_pipelines git directory]/pipeline_scRNASeq/* scRNA_seq
cd scRNA_seq

Then run:

python [UMI-Tool_pipelines git directory]/pipeline_ScRNASeq.py -v10 make full

To run the iCLIP analysis with the default configuration, create a new directory, download the datafiles, copy the configuration:

mkdir iCLIP_analysis
cd iCLIP_analysis
wget ftp://ftp.sra.ebi.ac.uk/vol1/fastq/SRR205/004/SRR2057564/SRR2057564.fastq.gz
.   .
.   .
.   .
wget ftp://ftp.sra.ebi.ac.uk/vol1/fastq/SRR205/004/SRR2057598/SRR2057598.fastq.gz
cp [UMI-Tools_pipelines git directory]/pipeline_iCLIP/SRSF_pipeline.ini pipeline.ini

For the complete analysis of the SRSF data download the following SRR records from ENA: SRR2057564, SRR2057565, SRR2057566, SRR2057567, SRR2057568, SRR2057569, SRR2057570, SRR2057571, SRR2057572, SRR2057573, SRR20 57574, SRR2057575, SRR2057576, SRR2057577, SRR2057578, SRR2057579, SRR2057580, SRR2057581, SRR2057582, SRR2057583, SRR2057584, SRR2057 585, SRR2057586, SRR2057587, SRR2057588, SRR2057589, SRR2057590, SRR2057591, SRR2057592, SRR2057593, SRR2057594, SRR2057595, SRR2057596, SRR2057597, SRR2057598

The pipeline requires a genome sequence (in fasta format) and a bowtie index, named the same way and in the same directory. The pipeline.ini file must be edited to point to these files, so for example in the ini file:

[bowtie]

# The location of the genome and indexes for use with bowtie
# The directory should contain a fasta genome with the .fa
# extension and a bowtie index with the same prefix.
# For the SRSF data set this should be an mm9 genome/index
# fo the TDP dataset this should be a hg19 genome/index
index_dir=~/genomes/bowtie/
genome=mm9

Will make the pipeline look for the genome in the ~/genomes/bowtie directory. It will use the mm9 and expect mm9.fa and mm9.1.ebwt, mm9.2.ebwt, etc to be present in that directory.

Now run the pipeline with:

python [UMI-Tool_pipelines git directory]/pipeline_iCLIP.py make full

To run the TDP analysis, copy TDP_pipeline.ini rather than SRSF_pipeline.ini, set the bowtie index to a hg19 index and run:

python [UMI-Tool_pipelines git directory]/pipeline_iCLIP.py make runNotebooks1

Pipelines will run either locally or on a cluster. The default cluster configuration is to use an SGE cluster manager with the all.q queue, the dedicated pe environment and mem_free. These defaults can be altered by changing the following settings in the [cluster] section of the pipeline.ini files:

• `queue`: The default queue to submit jobs to. Leave as NONE to have your cluster manager decide. (default: all.q)
• `parallel_environment`: The SGE parallel environment to request when submitting multi-process jobs (default: dedicated)
• `memory_resource`: The resource name to use when requesting a certian amount of memory for a job. (default: mem_free)
• `pe_queue`: If this variable is set then a different queue is used when submitting parallel jobs. (no default)
• `options`: any other options to pass to the queue manager.

The pipelines are also compatible with the SLURM and torque cluster managers. Set the manager parameter in the [cluster] section of the ini.

It is also possible to run the pipelines locally by adding --no-cluster to the command. However this will take a very long time. Running the iCLIP pipeline on our cluster with 100 parallel jobs (each possibly using multiple cores) takes around 50 hours. We estimate running locally would take many weeks.

These pipelines require the following dependencies: