Prepares summary statistics Z Score Files from IMPG and runs a finemapping analysis using Paintor and Caviarbf.

If you use this tool please ensure you cite the primary papers for the methods, these are listed below.

###Paintor

https://github.com/gkichaev/PAINTOR_FineMapping/

Kichaev, Gleb, et al. "Integrating functional data to prioritize causal variants in statistical fine-mapping studies." (2014): e1004722.

###Caviarbf

https://bitbucket.org/Wenan/caviarbf

Chen, Wenan, et al. "Fine Mapping Causal Variants with an Approximate Bayesian Method Using Marginal Test Statistics." Genetics (2015): genetics-115.

Run the following commands to generate the files

- numpy
- scipy
- GEMINI framework https://github.com/arq5x/gemini/
- PAINTOR https://github.com/gkichaev/PAINTOR_FineMapping/
- caviarbf https://bitbucket.org/Wenan/caviarbf
- tabix http://www.htslib.org/doc/tabix.html

Before installation it is important you change the file fine_mapping_pipeline/config/__init__.py to either a local installation or the 1000 genomes FTP server. The {0} is replaced internall for each chromosome. The following string is an example for the EBI 1000 genomes FTP.

__1000_genomes_template__="""ftp://ftp.1000genomes.ebi.ac.uk/vol1/ftp/release/20130502/ALL.chr{0}.phase3_shapeit2_mvncall_integrated_v5a.20130502.genotypes.vcf.gz"""

python setup.py install
bash install_paintor.sh

Basically the pipeline takes a list of either rsids, or chromosome"\t"position, and a folder containing ZScores statistics that have been output using the ImpG program, or any program as long as it conforms to the output we expect.

https://github.com/huwenboshi/ImpG

The pipeline looks at the 1st two columns for chromosome and position, and the 5th column for the ZScore.

Example ZScore input files are found in the tests/test_data/zscores/ folder. Example RSID file for the Zscores is found in the tests/test_data/rsids.txt

To get familiar with the pipeline run the following commands on the test dataset.

First you need to prepare you ImpG data for running the pipeline.

    fine_mapping_pipeline prepare -s tests/test_data/rsids.txt -z tests/test_data/zscores/ -f 25000 -b hg19 -o tests/test_output

Then follow up with either paintor and/or caviarbf with you parameters of interest, more customisation to be added soon.

Run these commands to get a feel.

    fine_mapping_pipeline paintor -d tests/test_output -o tests/test_output/paintor -c 1
    fine_mapping_pipeline caviarbf -d tests/test_output -o tests/test_output/caviarbf -c 1

This will place the analysis results from paintor and caviarbf in the tests/test_output/paintor and tests/test_output/caviarbf directories.

In the future, I plan to add plotting tools of the results data.

Also, currently PAINTOR uses a bonferonni correction for the number of annotations used, to restrict the number of annotations to say the genome segmentations use the argument --no-dhs to the paintor sub-command of the fine-mapping pipeline.

Tox is used to perform Unit testing of the Pipeline, make sure it is installed and accessible on your command-line.

tox

or alteratively you can use py.test directly.

py.test