Extract variants from imputed genotype file(s) into single BED file

For a given set of identifiers (RSID or "build37 position") extract from the UK Biobank imputed genotypes (BGEN files, field 22828) into a single BED file.

Usage

make_imputed_bed(
  in_file,
  out_bed,
  use_pos = FALSE,
  progress = TRUE,
  verbose = FALSE,
  very_verbose = FALSE
)

Arguments

in_file

A data frame or file path. Contains at least two columns: (`rsID` and `CHR`) OR (`CHR` and `POS`). Other columns are ignored.

out_bed

A string. Output BED file prefix.

use_pos

Logical. Use genomic position (CHR and POS) instead of RSID, default=FALSE

progress

Logical. Show progress through each individual file, default=TRUE

verbose

Logical. Be verbose (show individual steps), default=FALSE

very_verbose

Logical. Be very verbose (show individual steps & show terminal output from Plink etc), default=FALSE

Value

A single merged BED file (and BIM and FAM files)

Author

Luke Pilling

Examples

# Example list of variants
#    only RSID and CHR are provided
varlist <- data.frame(rsid=c("rs1800562","rs429358"), chr=c(6,19))

make_imputed_bed(in_file=varlist, out_bed="example_variants_by_rsid")

# Above example works fine if the imputed BGEN contains the provided rsid

# However if you have a list of variants defined by genomic position (CHR and POS)
# then use the `use_pos=TRUE` argument

# Example list of variants using position (build 37) not rsid
varlist_pos <- data.frame(chr=c(6,19), pos=c(26093141,45411941))

# if you provide an rsid it is ignored when `use_pos=TRUE`
make_imputed_bed(in_file=varlist_pos, out_bed="example_variants_by_pos", use_pos=TRUE)