wAFd_combine: Directed wAF Combination
In songbiostat/wAF: Weighted Adaptive Fisher Method

Description Usage Arguments Value Examples

View source: R/combine.R

This function combines P-values of one-tailed and two-tailed tests using directed weighted adaptive Fisher (wAFd) method.

1	wAFd_combine(pvs_2sided, pvs_left, log = FALSE, weight = NULL, n0 = 1)

`pvs_2sided`	P-values of two-tailed tests. A matrix with dimenstions K by N. If an object of P-values from perm_score is used, K is the number of SNVs, and N is the number of permutations plus 1.
`pvs_left`	Left-side P-values of one-tailed tests. Dimensions should be the same as pvs_2sided.
`log`	Indicator of whether P-values are on the log scale.
`weight`	Weights given to the P-values. A vector with dimension K. Flat weight is used if it is not specified.
`n0`	Tuning parameter. Discard the first n0-1 P-values of each column.

A list object.

pv: P-value of wAF dtest.
stat: Observed test statitic of wAFd test.
loci_combined: Variants which are combined into the test statistic. The index of included variants are returned in the ascending order of their weighted P-values.
stat_all: wAFd statistics for all permuted samples.
pv_all: P-values of wAFd statistics for all permuted samples.

# Combine P-values of normally distributed test statistics
U <- matrix(rnorm(10 * 100), ncol=100)
p <- 2 * (1 - pnorm(abs(U)))
p.l <- pnorm(abs(U))
wt <- (1:10)/55
test <- wAFd_combine(p, p.l, weight = wt)

Y <- SNV_sparse$trait
X <- SNV_sparse$SNV[, - SNV_sparse$zero_var]
result <- perm_score(Y, X, nperm = 100)
test <- wAFd_combine(result$pvs, result$pvs_left)