calculateSwfdr: Calculate the science-wise FDR (swfdr)

In swfdr: Science-wise false discovery rate and proportion of true null hypotheses estimation

Description

Calculate the science-wise FDR (swfdr)

Usage

calculateSwfdr(
  pValues,
  truncated,
  rounded,
  pi0 = 0.5,
  alpha = 1,
  beta = 50,
  numEmIterations = 100
)

Arguments

pValues	Numerical vector of p-values
truncated	Vector of 0s and 1s with indices corresponding to those in pValues; 1 indicates that the p-values is truncated, 0 that it is not truncated
rounded	Vector of 0s and 1s with indices corresponding to those in pValues; 1 indicates that the p-values is rounded, 0 that it is not rounded
pi0	Initial prior probability that a hypothesis is null (default is 0.5)
alpha	Initial value of parameter alpha from Beta(alpha, beta) true positive distribution (default is 1)
beta	Initial value of parameter beta from Beta(alpha, beta) true positive distribution (default is 50)
numEmIterations	The number of EM iterations (default is 100)

Value

pi0 Final value of prior probability - estimated from EM - that a hypothesis is null, i.e. estimated swfdr

alpha Final value of parameter alpha - estimated from EM - from Beta(alpha, beta) true positive distribution

beta Final value of parameter beta - estimated from EM - from Beta(alpha, beta) true positive distribution

z Vector of expected values of the indicator of whether the p-value is null or not - estimated from EM - for the non-rounded p-values (values of NA represent the rounded p-values)

n0 Expected number of rounded null p-values - estimated from EM - between certain cutpoints (0.005, 0.015, 0.025, 0.035, 0.045, 0.05)

n Number of rounded p-values between certain cutpoints (0.005, 0.015, 0.025, 0.035, 0.045, 0.05)

Examples

pVals <- runif(100)
tt <- rr <- rep(0, 100)
resSwfdr <- calculateSwfdr(pValues = pVals, truncated = tt, rounded = rr, numEmIterations=100)

swfdr documentation built on Nov. 8, 2020, 8:29 p.m.