GWAS rsID annotation and LiftOver

Summary

This repository contains workflow / scripts for:

(1) annotating GWAS with rsIDs (e.g. in regenie output, when only CHR and GENPOS columns are present)

(2) performing liftover of genome build (38 <--> 37)

Annotation and liftover can be done as a part of one workflow or completely independently.

Workflow

Please note:

• remember to make all .sh files executable (chmod 777 file.sh)
• it is recommended to do this on a server, as it might be very slow in a local environment

Obtaining genome reference data

Both datasets will be downloaded into refdata/ folder in the working directory (i.e. this cloned repo).

Please note: if you are in Exeter, the data download step can be skipped - simply create a symbolic link to data in /slade/projects/Public_Ref_Datasets/dbsnp/annotate_rsids_and_liftover/refdata/ (ln -s)

Data for rsID annotation

Download dbSNP v157 data in VCF format (for the specified genome build 37/38), and subset it to only required for annotation columns in .txt format.

To run: (recommended to run with nohup or screen as it might take a while)

./00_download_ref_data_annotation.sh 38 # specify build 38 or 37

Output: refdata/dbSNP157_refdata_build38.txt

Data for LiftOver

Download the chain files required for performing liftover; must specify the starting genome build (38 or 37).

To run:

./00_download_ref_data_liftover.sh 38 # specify build 38 or 37

Output: refdata/hg38ToHg19.over.chain or refdata/hg19ToHg38.over.chain

(1) Genome annotation

Script for performing rsID annotation; The main .sh script internally calls .R script:

├── 01_annotate_GWAS_with_rsids.sh
    └── 01_annotate_helper.R

Assumptions:

• The input GWAS file is in the regenie format (i.e. the specific column order and names; modify those in your file before running the script if needed): CHROM GENPOS ID ALLELE0 ALLELE1 A1FREQ INFO N TEST BETA SE CHISQ LOG10P EXTRA

• The input GWAS file name does not contain dots in the file name; only to separate the file extension:

  • my_GWAS.txt.gz - ok
  • my.GWAS.txt.gz - not ok

• If running on a server, make sure to load the necessary R modules: (e.g. for Exeter, uncomment line 42 in 01_annotate_GWAS_with_rsids.sh)

• The script uses the dbSNP annotation data (refdata/dbSNP151_refdata_build38.txt) generated earlier using 00_download_ref_data_annotation.sh, so it assumes that the input GWAS data is in build 38.

To run:

Provide full paths to annotaion file and input GWAS file:

./01_annotate_GWAS_with_rsids.sh /path/to/refdata/dbSNP157_refdata_build38.txt /path/to/your/GWAS.txt.gz

Output: /path/to/your/GWAS_rsids.txt.gz

(2) Genome liftover

Script for performing liftover; The main .sh script internally calls .R script:

├── 02_liftover_GWAS.sh
    └── 02_lifover_helper.R

Assumptions:

• The same assumptions as for (1) apply here;
• If running on a server, make sure you load the necessary R module: (e.g. for Exeter, uncomment line 22 in 02_liftover_GWAS.sh )

To run:

Provide full paths to input GWAS file (output from (1) or a separate file in regenie format), path to folder containing the downloaded chain files, and the starting genome build (i.e. if need b38 -> b37, add '38'):

./02_liftover_GWAS.sh /path/to/your/GWAS_rsids.txt.gz /path/to/refdata/ 38

Output: /path/to/your/GWAS_rsids_b37.txt.gz