exp_weights: Exponential p-value weights

In pweight: P-Value Weighting

Description

Computes the exponential p-value weights for multiple testing. Given estimated means mu of test statistics T, the p-value weights are proportional to exp(beta*mu), for a tilt parameter beta. In addition, the large weights are truncated at a maximum value UB (upper bound), and the remaining weight is re-distributed among the rest of the statistics.

Usage

exp_weights(mu, beta = 2, UB = Inf)

Arguments

mu	the estimated means of the test statistics
beta	(optional) weights are proportional to exp(mu*beta), default beta=2
UB	(optional) upper bound on the weights (default UB = Inf)

Details

Specifically, it is assumed that T are Gaussian with mean mu. One-sided tests of mu>=0 against mu<0 are conducted using the test statistics T. To optimize power, different levels are allocated to different tests. For more details, see the paper "Optimal Multiple Testing Under a Gaussian Prior on the Effect Sizes", by Dobriban, Fortney, Kim and Owen, http://arxiv.org/abs/1504.02935

Value

The exponential weights.

See Also

bayes_weights for Bayes, spjotvoll_weights for Spjotvoll weights, and exp_weights for exponential weights

Other p.value.weighting: bayes_weights; iGWAS; spjotvoll_weights

Examples

J <- 100
mu <- rnorm(J)
beta <- 2
UB <- 20
w <- exp_weights(mu, beta, UB)

pweight documentation built on May 30, 2017, 2:54 a.m.