ldr: Estimation of posterior probabilities for the vectors of...

In repfdr: Replicability Analysis for Multiple Studies of High Dimension

Description

The function finds the posterior probabilities ofeach vector of association status for each feature, given the feature's vector of binned z-scores.

Usage

ldr(pdf.binned.z, binned.z.mat, Pi, h.vecs = NULL)

Arguments

pdf.binned.z	Same input as in repfdr. A 3-dimensional array which contains for each study (first dimension), the probability of a z-score to fall in the bin (second dimension), under each hypothesis status (third dimension). The third dimension can be of size 2 or 3, depending on the number of association states: if the association can be either null or only in one direction, the dimension is 2; if the association can be either null, or positive, or negative, the dimension is 3. Element [[1]] in the output of ztobins.
binned.z.mat	Same input as in repfdr. A matrix of the bin numbers for each of the z-scores (rows) in each study (columns). Element [[2]] in the output of ztobins.
Pi	The estimated prior probabilities for each association status vector. Can be extracted from the output of repfdr or piem, see Example section.
h.vecs	The row indices in H (see hconfigs), corresponding to the association status vectors. By default the posterior probabilities of all possible vectors of association status are computed.

Details

A subset of features (e.g most significant) can be specified as the rows in binned.z.mat, so the posterior probabilities of the vectors of association status are computed for this subset of features. See Example section.

Value

Matrix with rows that contain for each of the vectors of association status the posterior probabilities. The columns are the different feature.

See Also

repfdr, piem, hconfigs

Examples

## Not run: 
download.file('http://www.math.tau.ac.il/~ruheller/repfdr_RData/binned_zmat.RData',
  destfile = "binned_zmat.RData")
load(file = "binned_zmat.RData")
  
data(Pi)

# Fdr calculation:
output3 <- repfdr(pbz, bz, "replication",Pi.previous.result = Pi)

BayesFdr <- output3$mat[,"Fdr"]
sum(BayesFdr <= 0.05)

# The posterior probabilities for the the first five features with Bayes FDR at most 0.05:
post <- ldr(pbz,bz[which(BayesFdr <= 0.05)[1:5],],Pi)
round(post,4)

# posteriors for a subset of the association status vectors can also be reported,
# here the subset is the four first association status vectors:
post <- ldr(pbz,bz[which(BayesFdr <= 0.05)[1:5],],Pi,h.vecs= 1:4)
round(post,4)

## End(Not run)

repfdr documentation built on May 1, 2019, 9:20 p.m.