MAMA (version 2.2.1)

MAP.Matches: Wrapper function for MAP-Matches method

Description

This is a wrapper function for MAP-Matches method.

Usage

MAP.Matches(data, varname, t.cutoff = "98.00%", multiple = TRUE, perm = c("both", "columns", "labels")[1], nperm = 1000, test = c("t", "t.equalvar")[1], sig.col, sig.cutoff = 0.05)

Arguments

data

Object of class MetaArray

varname

Character String - name of one column in clinical data matrices to be used as class labels

t.cutoff

Character String - quantile of T statistics to be selected, e.g. "95.00%" selects the top 5 percent of absolute values

multiple

Logical - when TRUE only paterrns with multiple '1' are used

perm

Character String - if "labels" only class labels are permuted for statistical analysis (empirical significance), if "columns" only genes in each dataset are selected randomly, if "both" both class labels and genes are permuted and two p-values returned

nperm

Numeric - number of permutations

test

Character String - if "t" then unequal variance t-test is used, if "t.equalvar" equal variance t-test is used

sig.col

Character String - which p-value is used for selection of significant patterns. Possible values are: "p.col.strong", "p.col.weak", "p.lab.strong", "p.lab.weak" , "col" refers to column permutations, "lab" to class labels, "weak" to soft match and "strong" to strong match

sig.cutoff

Numeric - p-value for selection of sigificant patterns

Value

Object of class MAP.Matches.res containing

tests

Data.frame of test statistics

bin.matrix

Binary matrix from tests, 1 means the test statistics was higer than threshold

sumarization

Sumarization of bin.matrix: number of selected genes in each dataset, genes with at least one 1 in pattern, probability of observing strong or soft match in the data

MAP

Data frame describing observed patterns: number of strong n.strong and soft n.soft matches and number of genes involved n.sig

stat.analysis

Results of statistical analysis

genes

List of genes observed with each pattern

all.genes

Names of the all genes in the analysis

References

Yang, X., Bentink, S. and Spang, R. 2005, Detecting Common Gene Expression Patterns in Multiple Cancer Outcome Entities, Biomedical Microdevices, Vol.7:3, pp. 247-251

Examples

data(ColonData)
MAP.Matches(ColonData, "MSI", nperm = 100, sig.col="p.lab.strong")