Restore missing or old rsID column for a GWAS summary statistics file
- python 3.7+
- pandas 1.3+ (1.2.3 has a bug with sorting)
- bash 4+
- clone this repo
- Download the latest full dbSNP file in vcf.gz format for the corresponding build (e.g.
dbSNP151_GRCh37.vcf.gz,dbSNP154_GRCh38.vcf.gz) - Run
create_config.shto create a config file for the program:bash create_config.sh diabetes_GWAS_ss.tsvor. create_config.sh diabetes_GWAS_ss.tsv
- Edit the generated config file
diabetes_GWAS_ss.tsv.jsonto make sure you have:- indices of columns for chromosome and base pair position in the config file corresponding to these columns in your summary statistics file
diabetes_GWAS_ss.tsv - path to the dbSNP file
- filenames for intermediate and the resulting file not conflicting with your existing files
- indices of columns for chromosome and base pair position in the config file corresponding to these columns in your summary statistics file
- Run
kukrsID, passing config file:python3 kukrsID.py diabetes_GWAS_ss.tsv.json
- GWAS summary statistics file is missing "rsID" column
- GWAS summary statistics file is old and has old rsIDs (that were merged into the newer ones)
$ python3 ~/kukrsID/add_missing_rsID_column.py continuous-20508-both_sexes.tsv.json
=== Step 1/3: Sort GWAS SS file by Chr and BP ===
--- Step 1/3 finished in 394.9851267337799 seconds ---
=== Step 2/3: Generate rsID column and save it to a file ===
input GWAS SS file: "/home/kukubuntu/data/continuous-20508-both_sexes_asc.tsv"
successfully restored 28940544 rsIDs based on provided dbSNP file: '/media/kukubuntu/Samsung_T5/dbSNP151_GRCh37.vcf.gz'
--- Step 2/3 finished in 3006.899871110916 seconds ---
=== Step 3/3: Concatenate columns of the sorted GWAS SS file and rsID file ===
concatenated GWAS SS file with the file with restored rsID at: "/home/kukubuntu/data/continuous-20508-both_sexes_asc_with-kukrsIDs.tsv"
--- Step 3/3 finished in 348.7267150878906 seconds ---
see resulting file at:
"/home/kukubuntu/data/continuous-20508-both_sexes_asc_with-kukrsIDs.tsv"