gc_gnova_cmd.R

rm(list = ls())
options(stringsAsFactors=F)


args = commandArgs(trailingOnly=TRUE)
output_path = args[1] ## a folder to save all GNOVA output files (there will be ~2,500 files generated)
output_prefix = args[2] ## prefix of output files
sumstats = args[3] ## path to input munged summary stats
gnova_path = args[4] ## gnova path: /ysm-gpfs/pi/zhao/from_louise/yh367/GNOVA/GNOVA-master
sqpath = args[5] ## output path for simple queue task list
ref_path = args[6] ## reference table, header: Categ (source of published GWAS), 
## Trait (trait name), Curated_file (path to munged sumstats), Code (index for disease order)
## e.g. /gpfs/ysm/pi/zhao/from_louise/yh367/VA/workflow/GNOVA/sumstats/match_table_full.csv
## note this table requires updating if position of any munged sumstats changes. 
bfile_path = args[7] ## 1000 Genomes genotype data for estimating LD matrix
##e.g. /ysm-gpfs/pi/zhao/from_louise/yh367/GNOVA/genotype_1KG_eur_SNPmaf5/

mat_table_full = read.csv(ref_path, header = T) ## read 
dir.create(output_path, showWarnings = FALSE) ## mkdir if not exist
cmds = paste0('module load Python; python ', gnova_path, 'gnova.py ',sumstats,' ',mat_table_full$Curated_file,' --bfile ', bfile_path, 'eur_chr@_SNPmaf5 --out ',output_path,'/',output_prefix,'_',mat_table_full$Code,'_NS.txt;')    
writeLines(cmds, paste0(sqpath))
print(paste0('GNOVA task file written to ', sqpath))