EBexpo-class {RPointCloud}

The EBexpo Class

Description

The EBexpo object represents the results of an Empirical Bayes approach to estimate a distribution as a mixture of a (more or less) known exponential distribution along with a completely unknown "interesting" distribution. The basic method was described by Efron and Tibshirani with an application to differential expression in microarray data.

Usage

EBexpo(edata, resn = 200)
cutoff(target, prior, object)
## S4 method for signature 'EBexpo'
hist(x, xlab="", ylab="Prob(Interesting | X)", main="", ...)
## S4 method for signature 'EBexpo,missing'
plot(x, prior, post = c(0.5, 0.8, 0.9), ...)

Arguments

edata

A numeric vector; the observed data that we think comes mainly from an exponential distribution.

resn

A numeric vector; the resolution used to estimate a histogream.

x

An EBexpo object.

xlab

A character vector; the label for the x-axis.

ylab

A character vector; the label for the y-axis.

main

A charcter vector; the plot title.

...

The usual set of graphical parameters.

prior

A numeric vector of length 1; the prior probability of an observed data point coming from the known exponential distribution.

post

The posterior probabilities to display in the plot.

target

The target posterior probability.

object

An EBexpo object.

Value

The EBexpo function constructs and returns an object of the EBexpo class

The plot and hist methods return (invisibly) the EBexpo object that was their first argument.

Slots

expo:

An ExpoFit object.

theoretical.pdf:

A numerical vector representing the density funciton of the putative exponential distribution component of the mixture..

unravel:

A numeric vector; the observed empirical distribution of the complete mixture.

Methods

plot(x, prior, post = c(0.5, 0.8, 0.9), ...):

Produce a plot of a EBexpo object.

hist(x, xlab="", ylab="Prob(Interesting | X)", main="", ...):

Produce a histogram of teh observed distibution, with overlays.

Author(s)

Kevin R. Coombes <[e-mail]>

References

Efron B, Tibshirani R. Empirical bayes methods and false discovery rates for microarrays. Genet Epidemiol. 2002 Jun;23(1):70-86. doi: 10.1002/gepi.1124.

Examples

data(cytof)
diag <- AML10.node287.rips[["diagram"]]
persistence <- diag[, "Death"] - diag[, "Birth"]
d1 <- persistence[diag[, "dimension"] == 1]
eb <- EBexpo(d1, 200)
hist(eb)
plot(eb, prior = 0.56)
cutoff(0.8, 0.56, eb)