find.a0 {siggenes}R Documentation

Computation of the Fudge Factor

Description

Provides the required information for obtaining the optimal choice of the fudge factor in the Empirical Bayes Analysis of Microarrays that uses the modified t statistics.

Usage

    find.a0(data,x,y,paired=FALSE,mat.samp=NULL,B=100,balanced=FALSE,na.rm=FALSE,delta=0.9,
    alpha=(0:9)/10,include.0=TRUE,p0=NA,stable=TRUE,number.int=139,rand=NA,plot.legend=TRUE)

Arguments

data the data set that should be analyzed. Every row of this data set must correspond to a gene.
x vector of the columns of data that correspond to the treatment group. In the paired case, (x[i],y[i]) build a pair. If, e.g., the first n1 columns contain the gene expression values of the treatment group,x=1:n1.
y vector of the columns of data that correspond to the control group. In the paired case, (x[i], y[i]) are an observation pair.
paired paired (TRUE) or unpaired (FALSE) data. Default is FALSE.
mat.samp a permutation matrix. If specified, this matrix will be used, even if rand and B are specified.
B number of permutations used in the calculation of the null density.
balanced if TRUE, only balanced permutations will be used. Default is FALSE.
na.rm if FALSE (default), the expression score of genes with one or more missing values will be set to NA. If TRUE, the missing values will be replaced by the genewise mean of the non-missing values.
delta a gene will be called differentially expressed, if its posterior probability of being differentially expressed is large than or equal to delta.
alpha a vector of possible values for the fudge factor a0 in terms of quantiles of the standard deviations of the genes.
include.0 if TRUE (default), a0=0 will also be a possible choice of the fudge factor.
p0 the prior probability that a gene is differentially expressed. If not specified, it will automatically be computed.
stable if TRUE (default), p0 will be computed by the algorithm of Storey and Tibshirani (2003). If FALSE, the (unstable) estimate that ensures that the posterior probability of being differentially expressed is always non-negative is computed.
number.int number of equally spaced intervals between the minimum and and the maximum of the expression scores z that are used in the logistic regression for estimating the ratio of the null density to the mixture density.
rand if specified, the random number generator will be put in a reproducible state.
plot.legend if TRUE (default), a legend will be added to the plot of the expression scores vs. their logit-transformed posterior probability.

Value

a list of the numbers of genes called differentially expressed by the EBAM analysis for several choices of a0, and the plot of the expression scores vs. their corresponding logit-transformed posterior probability of being significant.

sig.a0 vector containing the number of differentially expressed genes for the specified set of values for a0.
a0 the optimal choice of the fudge factor using the criterion of Efron et al. (2001) that the a0 should be used which leads to the most differentially expressed genes.

Note

The results of find.a0 must be assigned to an object for the further analysis with ebam.

Author(s)

Holger Schwender, holger.schw@gmx.de

References

Efron, B., Tibshirani, R., Storey, J.D., and Tusher, V. (2001). Empirical Bayes Analysis of a Microarray Experiment, JASA, 96, 1151-1160.

Storey, J.D., and Tibshirani, R. (2003). Statistical significance for genome-wide experiments, Technical Report, Department of Statistics, Stanford University.

Schwender, H. (2003). Assessing the false discovery rate in a statistical analysis of gene expression data, Chapter 7, Diploma thesis, Department of Statistics, University of Dortmund, http://de.geocities.com/holgerschw/thesis.pdf.

See Also

ebam ebam.wilc


[Package Contents]