performSOGL: Function to perform analysis using Similarity of Ordered Gene...

In MAMA: Meta-Analysis of MicroArray

Description

This a wrapper function to perform all steps designed in Similarity of Ordered Gene Lists.

Usage

performSOGL(data, varname, test, B, which = c("score", "empirical"), min.weight = 1e-05, two.sided = TRUE, percent = 0.95)

Arguments

data	MetaArray object, the rownames in the gene expression data matrices must be equal
varname	A string indicating which column of clinical data matrices should be used to compute test statistic. Same column is used in all datasets.
test	"FCH" for fold change (function fold.change) or "T" for T-test (function meta.test)
B	Number of permuatations
which	if "empirical" then empirical confidence intervals of number of overlapping genes are also provided, if "score" only random and subsampled scores necessary for tunning alpha parameter are calculated
min.weight	Minimal weight for score calculation
two.sided	if TRUE both top and bottom of the ordered gene lists are considered, if FALSE only top ones
percent	Percentage (Numeric between 0 and 1) of the score for genes selection

Value

Object of class SOGLresult, it is a list containig:

ordering	Ordered Gene Lists as a data.frame where columns refer to datasets
alpha.selected	Selected value of alpha parameter
alpha.considered	Vector of alpha considered for selection
pAUC	pAUC values related to all alphas considered
random	Random scores (permutations of class labels)
subsample	Scores after subsampling from each class and dataset
emp.ci	Empirical confidence intervals for number of overlapping genes
common.genes	Vector of number of overlapping genes
score	Observed similarity score
significance	Significance of the observed score in form of p-value
genes	Genes that account for observed similarity score
all.genes	Names of the all genes in the analysis

Author(s)

Ivana Ihnatova

References

Yang, X., Bentink, S., Scheid, S. Spang, R., Similarities of ordered gene lists, 2005

MAMA documentation built on Jan. 15, 2017, 3:05 p.m.