Skip to content

Pipeline to implement a "TAD_Pathways" analysis. Discover candidate genes based on association signals in TADs

License

Notifications You must be signed in to change notification settings

greenelab/tad_pathways_pipeline

Repository files navigation

TAD_Pathways

Leveraging TADs to identify candidate genes at GWAS signals

Gregory P. Way, Casey S. Greene, and Struan F.A. Grant - 2017

DOI

Summary

The repository contains data and instructions to implement a "TAD_Pathways" analysis for over 300 different trait/disease GWAS or custom SNP lists.

TAD_Pathways uses the principles of topologically association domains (TADs) to define where an association signal (typically a GWAS signal) can most likely impact gene function. We use TAD boundaries as defined by Dixon et al. 2012 and hg19 Gencode genes to identify which genes may be implicated. We then perform an overrepresentation pathway analysis to identify significantly associated pathways implicated by the input TAD-defined geneset.

For more specific details about our method, refer to our short report at the European Journal of Human Genetics.

We also present a 6 minute video introducing the method and discussing the experimental validation at EJHG-tube.

Setup

First, clone the repository and navigate into the top directory:

git clone git@github.com:greenelab/tad_pathways_pipeline.git
cd tad_pathways_pipeline

Before you begin, download the necessary TAD based index files and GWAS curation files and setup python environment:

bash initialize.sh

# Using conda version 4.4.11
conda activate tad_pathways

Now, a TAD_Pathways analysis can proceed. Follow an example pipeline to work from an existing GWAS or the custom pipeline example for insight on how to run TAD_Pathways on user curated SNPs.

Examples

We provide three different examples for a TAD pathways analysis pipeline. To run each of the analyses:

source activate tad_pathways

# Example using Bone Mineral Density GWAS
bash example_pipeline_bmd.sh

# Example using Type 2 Diabetes GWAS
bash example_pipeline_t2d.sh

# Example using custom input SNPs
bash example_pipeline_custom.sh

General Usage

There are two ways to implement a TAD_Pathways analysis:

  1. GWAS
  2. Custom

GWAS

To perform a TAD_Pathways analysis on publicly available GWAS results, simply browse the data/gwas_catalog/ directory to select a valid GWAS file. These files contain a curation of all significant SNPs mapped to specific traits as distributed by the NHGRI-EBI GWAS Catalog.

Each file in this directory is a tab separated text file of genome-wide significant SNPs and their genomic location along with their reported nearest gene and associated PUBMED id. For complete information on how these files were constructed, refer to https://github.com/greenelab/tad_pathways.

Each GWAS has 3 associated files, including files in data/gwas_catalog/. The other files are located in data/gwas_tad_snps/ and data/gwas_tad_genes/. All files are important for performing a TAD_Pathways analysis. See the GWAS example files for instructions on how to implement the necessary scripts.

Custom

To perform a TAD_Pathways analysis on a list of custom SNPs, generate a comma separated text file. The first row of the text file should have group names and subsequent rows should list the rs numbers of interest. There can be many columns with variable length rows.

E.g.: custom_example.csv

Group 1 Group 2
rs12345 rs67891
rs19876 rs54321
... ...

Then, perform the following steps:

source activate tad_pathways

# Map custom SNPs to genomic locations
Rscript --vanilla scripts/build_snp_list.R \
        --snp_file "custom_example.csv" \
        --output_file "mapped_results.tsv"

# Build TAD based genelists for each group
python scripts/build_custom_TAD_genelist.py \
       --snp_data_file "mapped_results.tsv" \
       --output_file "custom_tad_genelist.tsv"

The output of these steps are Group specific text files with all genes in TADs harboring an input SNP. See example_pipeline_custom.sh for more details.

Contact

For all questions and bug reporting please file a GitHub issue

For all other questions contact Casey Greene at csgreene@mail.med.upenn.edu or Struan Grant at grants@email.chop.edu