run_proptests {colocPropTest}

run proportional tests on extreme subset of snp pairs from two coloc style datasets. Of all functions in this package, this is the main one that should be used.

Description

run proportional tests on extreme subset of snp pairs from two coloc style datasets. Of all functions in this package, this is the main one that should be used.

Usage

run_proptests(
  S1,
  S2,
  LD,
  topsnps = "auto",
  r2.thr = 0.95,
  maxtests = 10000,
  nauto = 200,
  adjust_n = FALSE
)

Arguments

S1

coloc dataset 1

S2

coloc dataset 2

LD

LD matrix - rownames, colnames capture the snps and S1$snp[j] must be represented

topsnps

list of topsnps to be considered for testing or, if "auto", will be automatically selected

r2.thr

r2 threshold for initial tagging step - includes only one of any set of snps in mutually high LD with r2 > r2.thr

maxtests

maximum number of test pairs to consider. if more than maxtests pairs available, will select a random sample.

nauto

number of snps to use when automatically defining topsnps. only has an effect if topsnps=="auto"

adjust_n

TRUE if you want to adjust for variable sample size between snps. This is only set up for case control data at the moment (ask if you need quantitative) and requires that you supply separately the number of cases and controls at each snp in each dataset, as vector elements of the lists called n1 (cases) and n0 (controls)

Value

data.table containing the tests run

Author(s)

Chris Wallace

Examples

library(colocPropTest)
library(coloc)
data(coloc_test_data)
attach(coloc_test_data)
LD=D1$LD
dimnames(LD)=list(D1$snp,D1$snp)
results=run_proptests(D1,D2,LD=LD,topsnps=D1$snp,maxtests=100)
min(results$fdr)