LD_prune: Perform LD pruning on SNP data

In Osmahmoud/SlopeHunter: Slope-Hunter: A ROBUST METHOD FOR COLLIDER BIAS CORRECTION IN CONDITIONAL GENOME-WIDE ASSOCIATION STUDIES

Perform LD pruning on SNP data

Description

Uses PLINK clumping method ('–clump' command), where a greedy search algorithm is implemented to randomly select a variant (or the variant with the lowest p-value, if a user wish to), referred to as the index SNP, and remove all variants within a certain kb distance in linkage disequilibrium with the index SNP, based on an r-squared threshold from the 1000 Genomes reference panel phase 3 data. Then repeats until no variants are left.

Usage

LD_prune(
  dat,
  clump_kb = 250,
  clump_r2 = 0.1,
  Random = TRUE,
  clump_p1 = 1,
  local = FALSE,
  ref_pop = "EUR",
  ref_bfile,
  seed = 77777
)

Arguments

dat	Output from harmonise_effects. Must have a SNP name column (SNP).
clump_kb	Clumping window, default is 250.
clump_r2	Clumping r-squared threshold, default is 0.1.
Random	Logical, if TRUE (the default), SNPs will be randomly pruned. Otherwise, based on p-values.
clump_p1	Clumping sig level for index SNPs, default is 1.
local	Logical, if FALSE (the default), the MRC-IEU API 'http://gwas-api.mrcieu.ac.uk/' will be used for clumping. Otherwise, your local machine will be used for clumping given that you provide a bed/bim/fam LD reference dataset.
ref_pop	Super-population to use as reference panel at the API (when local is FALSE). Default = "EUR".
ref_bfile	Path to the bed/bim/fam LD reference (e.g. "1kg.v3/EUR" for local 1000 EUR ref. population file). If local=TRUE, then this should be provided.
seed	Random number seed for random pruning

Value

Data frame

Osmahmoud/SlopeHunter documentation built on Oct. 7, 2022, 4:38 p.m.