Analysis performed by Mike Gloudemans Updated 1/21/2020
This repository contains the scripts required to generate the colocalization scores described in the paper. Most figure-level analysis is outside of the scope of this repository; the aim of this repository is merely to show how the colocalization scores were generated using FINEMAP and the eCAVIAR formula, and to hopefully give an idea of how similar analyses could be run.
The top-level project directory should contain the folders bin, data, output, tmp, and scripts. All scripts must be
run from this top-level directory, or they'll be unable to locate the required files.
The file pipe.sh gives a step-by-step pipeline outlining all necessary steps.
QC and/or other analyses not directly used in the generation of colocalization results.
Scripts to launch the main colocalization portion of the analysis.
A few post-processing steps, to combine the results into one big table after running colocalization analysis.
Some formatting steps required to shape the raw QTL and GWAS data into the format required for running colocalization analysis.
The following analyses were performed using other publicly available tools. To fully complete the
analyses, you will have to install or link these tools within the bin subfolder, or modify the
pipe.sh script to include the paths to the directories where these tools are installed.
- The tools for downloading and munging the publicly available GWAS summary statistics are available at https://github.com/mikegloudemans/gwas-download/. (The UKBB, PhenomeScanner, and GWAS Catalog corpora are not explicitly included, but straightforward tools exist for downloading summary statistics in bulk from these other sources.)
- The analysis performed in this paper uses an integration of the publicly available tools FINEMAP (Benner et al. 2016) and eCAVIAR (Hormozdiari et al. 2016). My pipeline that uses these methods is available in a basic form at https://bitbucket.org/mgloud/production_coloc_pipeline/src, where you'll find further instructions for setting up the general colocalization analysis framework. An extended and easier-to-use pipeline with a greater variety of options and analyses will soon be available at https://github.com/mikegloudemans/ensemble_coloc_pipeline.
- Graphical visualization of colocalizations was performed using LocusCompare (Liu et al. 2019) but this tool is not strictly required to reproduce the results in this paper.
This project makes uses of a variety of data tables. I have linked publicly available data hosted on other sites, and provided directions below for obtaining those for which I've performed additional processing. If you have trouble accessing any of these data, please contact me directly (see Contact section below) and I'll share the relevant files directly, ASAP.
- eQTL files were obtained through the i2QTL Consortium and should be obtained directly from the owners of that project.
- GWAS summary statistics for the LocusCompare portion of the analysis are publicly available; consistently-formatted versions of these and other GWAS can be downloaded directly.
- An hg38-formatted version of the 1000 Genomes VCF is required for computing allele frequencies in a reference population.
To run all the scripts listed here, the data folder technically needs to contain all of the following files:
data/1KG: 1KG VCF for hg19, publicly available for download as described above.data/gwas: GWAS summary statistics for all traits of interest. (Download process described above)data/qtls: The QTL summary association statistics for all molecular traits measured in iPSC. This folder should additional contain subdirectories for each iPSC molecular QTL of interest, includingtrans_eqtlif that is one you wish to run.
For any questions about this colocalization processing pipeline, please contact Mike Gloudemans (mgloud@stanford.edu). I'll be glad to help you get these analyses up and running!