Creates weight files in PredictDB format from QTL data

Source: R/ctwas_weights.R

Creates weight files in PredictDB format from QTL data

create_predictdb_from_QTLs(
  weight_table,
  gene_table = NULL,
  cov_table = NULL,
  use_top_QTL = TRUE,
  outputdir = getwd(),
  outname
)

Arguments

weight_table

a data frame of the genes, QTLs and weights, with columns: "gene", "rsid", "varID", "ref_allele", "eff_allele", "weight". If you want to use multiple eQTLs per gene, you can set use_top_QTL=FALSE. But we assume the weights of the eQTLs are learned from multiple regression (instead of marginal effect sizes).

gene_table

a data frame (optional) with information of the genes in weight_table ("gene","genename","gene_type", etc.). If NULL, create a simple gene_table based on the weight_table

cov_table

a data frame of covariances between variants, with columns: "GENE","RSID1","RSID2", "VALUE". If NULL, do not create covariance files (.txg.gz), unless use_top_QTL=TRUE.

use_top_QTL

If TRUE, only keep the top QTL with the largest abs(weight) for each gene (molecular trait), and create a simple cov_table with covariance set to 1.

outputdir

output directory

outname

name of the output weight file