REIDS (version 0.1.0)

FIRMAScores: "FIRMAScores"

Description

The FIRMAScores function performs an analyais on the FIRMA scores of the sanples. If the groups are not paired, the Firma all sample score will be the minimum value of group 1. A test statistic is perforrmed on the grouping of interest to see if there is a significant difference between them. If the data is paired, the mean paired differences are obtained and tested versus zero.

Usage

FIRMAScores(Data, InformativeExons = NULL, groups = list(group1 =
  list(group1a = NULL, group1b = NULL), group2 = NULL), paired = FALSE,
  significancelevel = NULL)

Arguments

Data

The output of the FIRMA model

InformativeExons

A character vector of exon IDs. As for the REIDS model probesets are filtered out by I/NI calls model and later on exon score, the remaining exons can be specified here. Only these shall be considered in the FIRMA analysis to make the results between REIDS and FIRMA more comparable

groups

A list with two elements speficifing the columns of the data of group 1 in group1 and those of group 2 in group2. Here is the possibility to specifify multiple subgroups in group 1 for the calculation of the all FIRMA sample score.

paired

Logical. Are the groups paired? If TRUE the mean paired differences are calculated and tested whether these are significantly different from zero or not.

significancelevel

If specified, filtering is conducted on the p-values.

Value

A data frame with one line per exon. The columns contain the gene ID, the exon ID, the test statistic, a p-value and an adjusted p-value. If the groups are paired also the mean paired difference is given. Only the probesets with high enough exon scores and a significant test are kept in the data frame.

Details

The input to this function is the output of the FIRMA model. On the FIRMA scores, a t-test is performed. This is either between groups if the data is not paired and on the mean paired differences if the data is paired. If no pairing is present, an all sample FIRMA score is computed as the maximum of the minimum values of the of the subgroups in group 1. The returned p-value is adjusted for multiplicity. If a vector is given in InformativeExons, only these exon IDS are considered in the calculations.

Examples

# NOT RUN {
data(ExampleFirmaOutput)
FIRMATest=FIRMAScores(Data=ExampleFirmaOutput,InformativeExons=NULL,groups=list(group1=
list(group1a=c(1,2,3),group1b=NULL),group2=c(4,5,6)),paired=FALSE,significancelevel=NULL) 
# }