gogoGPCR is a framework for performing burden testing on UK Biobank Whole-Exome Sequencing (WES) data. This repo contains a series of notebooks for pre-processing and quality-controlling UKB VCF files, mainly with Hail, a Dockerfile for pre-processing phenotypes with PHESANT, and a series of Shell scripts for performing burden testing, with regenie.
Despite the name, gogoGPCR can, in principle, be run for any gene (or set hereof).
For information on running individual notebooks, see notebooks/README.md. Likewise, see scripts/README.md for running individual scripts.
WARNING: gogoGPCR will probably (99% confidence) not work out of the box and requires tweaking for your particular setup. In the very least, it may serve as inspiration for your own analyses.
Notebooks and scripts should generally be run in numerical order. Notebook/scripts starting with 0*_ e.g. notebooks/02_covariates.py, need only be run once and the generated file can be re-used for further analyses. Files wigth *a,b,c_ e.g. 03c_metabolic_phenotypes_BT.py represent different versions of the same step.
├── README.md
├── config.toml # Majority of configuration
├── notebooks # Pre-processing and QC Jupyter notebooks
│ ├── 00_prescriptions.py
│ ├── 01_sample_hard_filter.py
│ ├── ...
│ ├── 4_forest_plots.py
│ └── README.md
├── docker # Dockerfile for running PHESANT
│ └── phesant
│ ├── install_packages.R
│ └── Dockerfile
├── scripts # Shell scripts for running regenie step 1 and 2
│ ├── 00_merge_files.sh
│ ├── ...
│ ├── 1a_regenie_step1_BT.sh
│ ├── 2a_regenie_step2_BT.sh
│ └── README.md
├── src # QC and utility functions
│ ├── matrixtables.py
│ ├── ...
│ └── utils.py
├── data
│ ├── examples # Example regenie output
│ │ └── *.regenie
│ └── misc # Miscellaneous data files for mapping purposes
│ ├── Data_Dictionary_Custom.csv
│ ├── ...
│ └── xgen_plus_spikein.b38.bed
└───thesis # Please don't cite
└── MSc_Thesis.pdf
Run jupytext --from py:percent --to notebook notebooks/*.py