Advanced Rank Product Analysis

Description

The function performs the Rank Product method to identify differentially expressed genes. It is possible to do either a one-class or two-class analysis. It is also possible to combine data from different studies (e.g. datasets generated by different laboratories. This function has been kept only to guarantee backward compatibility, in fact the same results can be obtained by RankProducts.

Usage

1
2
RPadvance(data, cl, origin, num.perm = 100, logged = TRUE, na.rm = TRUE, 
gene.names = NULL, plot = FALSE, rand = NULL, huge = FALSE)

Arguments

data

the data set that should be analyzed. Every row of this dataset must correspond to a gene

cl

a vector containing the class labels of the samples. In the two class unpaired case, the label of a sample is either 0 (e.g., control group) or 1 (e.g., case group). For one class data, the label for each sample should be 1

origin

a vector containing the origin labels of the samples. The label is the same for samples within one lab and different for samples from different labs.

num.perm

in this version of the package, this parameter is not used any more, but it is kept for backward compatibility

logged

if "TRUE" data have been previously log transformed. Otherwise it should be set as "FALSE"

na.rm

if "FALSE", the NA value will not be used in computing rank. If "TRUE" (default), the missing values will be replaced by the genewise median of the non-missing values. Gene with a number of missing values greater than 50% are still not considered in the analysis

gene.names

if "NULL", no gene name will be attached to the outputs, otherwise it contains the vector of gene names

plot

if "TRUE", plot the estimated pfp vs the rank of each gene

rand

if specified, the random number generator will be put in a reproducible state

huge

if "TRUE" not all the outputs are evaluated in order to save space

Value

A summary of the results obtained by the Rank Product method.

pfp

estimated percentage of false positive predictions (pfp), both considering upregulated an downregulated genes

pval

estimated pvalues per each gene being up- and down-regulated

RPs

the Rank Product statistics evaluated per each gene

RPrank

rank of the Rank Product of each gene in ascending order

Orirank

ranks obtained when considering each possible pairing. In this version of the package, this is not used to compute Rank Product (or Rank Sum), but it is kept for backward compatibility

AveFC

fold changes of average expressions (class1/class2). log fold-change if data has been log transformed, original fold change otherwise

allrank1

fold change of class 1/class 2 under each origin. log fold-change if data has been log transformed, original fold change otherwise

allrank2

fold change of class 2/class 1 under each origin. log fold-change if data has been log transformed, original fold change otherwise

nrep

total number of replicates

groups

vector of labels (as cl)

Note

Percentage of false prediction (pfp), in theory, is equivalent of false discovery rate (FDR), and it is possible to be large than 1.

The function looks for up- and down- regulated genes in two seperate steps, thus two pfps are computed and used to identify gene that belong to each group. The function is able to replace function RP in the same library. it is a more general version, as it is able to handle data from differnt origins.

Author(s)

Francesco Del Carratore, francesco.delcarratore@postgrad.manchester.ac.uk
Andris Janckevics, andris.jankevics@gmail.com

References

Breitling, R., Armengaud, P., Amtmann, A., and Herzyk, P.(2004) Rank Products: A simple, yet powerful, new method to detect differentially regulated genes in replicated microarray experiments, FEBS Letter, 57383-92

See Also

topGene RP RSadvance plotRP RP.advance RankProducts

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
# Load the data of Golub et al. (1999). data(golub) 
# contains a 3051x38 gene expression
# matrix called golub, a vector of length called golub.cl 
# that consists of the 38 class labels,
# and a matrix called golub.gnames whose third column 
# contains the gene names.
data(golub)

##For data with single origin
subset <- c(1:4,28:30)
origin <- rep(1,7)
#identify genes 
RP.out <- RPadvance(golub[,subset],golub.cl[subset],
        origin,plot=FALSE,rand=123)
      
#For data from multiple origins
      
#Load the data arab in the package, which contains 
# the expression of 22,081 genes
# of control and treatment group from the experiments 
#indenpently conducted at two 
#laboratories.
data(arab)
arab.origin #1 1 1 1 1 1 2 2 2 2
arab.cl #0 0 0 1 1 1 0 0 1 1
RP.adv.out <- RPadvance(arab,arab.cl,arab.origin,
                num.perm=100,gene.names=arab.gnames,logged=TRUE,rand=123)

attributes(RP.adv.out)
head(RP.adv.out$pfp)
head(RP.adv.out$RPs)
head(RP.adv.out$AveFC)