Fit beta-uniform mixture model to a p-value distribution

Share:

Description

The function fits a beta-uniform mixture model to a given p-value distribution. The BUM method was introduced by Stan Pounds and Steve Morris to model the p-value distribution as a signal-noise decompostion. The signal component is assumed to be B(a,1)-distributed, whereas the noise component is uniform-distributed under the null hypothesis.

Usage

1
fitBumModel(x, plot = TRUE, starts=10)

Arguments

x

Numeric vector of p-values.

plot

Boolean value, whether to plot a histogram and qqplot of the p-values with the fitted model.

starts

Numeric value giving the number of starts for the optimization.

Value

Maximum likelihood estimator object for the fitted bum model. List of class fb with the following elements:

lambda

Fitted parameter lambda for the beta-uniform mixture model.

a

Fitted parameter a for the beta-uniform mixture model.

negLL

Negative log-likelihood.

pvalues

P-value vector.

Author(s)

Daniela Beisser

References

S. Pounds, S.W. Morris (2003) Estimating the occurrence of false positives and false negatives in microarray studies by approximating and partitioning the empirical distribution of p-values. Bioinformatics, 19(10): 1236-1242.

Examples

1
2
3
4
data(pvaluesExample)
pvals <- pvaluesExample[,1]
bum.mle <- fitBumModel(pvals, plot=TRUE)
bum.mle

Want to suggest features or report bugs for rdrr.io? Use the GitHub issue tracker.