testGroupIndependence: Multiple independence tests between reference groups and list...

View source: R/groups.R

testGroupIndependenceR Documentation

Multiple independence tests between reference groups and list of groups

Description

Test multiple contingency tables comprised by two groups (one reference group and another containing remaining elements) and provided groups.

Usage

testGroupIndependence(ref, groups, elements, pvalueAdjust = "BH")

Arguments

ref

List of character: list of groups where each element contains the identifiers of respective elements

groups

List of characters: list of groups where each element contains the identifiers of respective elements

elements

Character: all available elements (if a data frame is given, its rownames will be used)

pvalueAdjust

Character: method used to adjust p-values (see Details)

Details

The following methods for p-value adjustment are supported by using the respective string in the pvalueAdjust argument:

  • none: Do not adjust p-values

  • BH: Benjamini-Hochberg's method (false discovery rate)

  • BY: Benjamini-Yekutieli's method (false discovery rate)

  • bonferroni: Bonferroni correction (family-wise error rate)

  • holm: Holm's method (family-wise error rate)

  • hochberg: Hochberg's method (family-wise error rate)

  • hommel: Hommel's method (family-wise error rate)

Value

multiGroupIndependenceTest object, a data frame containing:

attribute

Name of the original groups compared against the reference groups

table

Contingency table used for testing

pvalue

Fisher's exact test's p-value

See Also

parseCategoricalGroups() and plotGroupIndependence()

Other functions for data grouping: createGroupByAttribute(), getGeneList(), getSampleFromSubject(), getSubjectFromSample(), groupPerElem(), plotGroupIndependence()

Examples

elements <- paste("subjects", 1:10)
ref      <- elements[5:10]
groups   <- list(race=list(asian=elements[1:3],
                           white=elements[4:7],
                           black=elements[8:10]),
                 region=list(european=elements[c(4, 5, 9)],
                             african=elements[c(6:8, 10)]))
groupTesting <- testGroupIndependence(ref, groups, elements)
# View(groupTesting)

nuno-agostinho/psichomics documentation built on Feb. 11, 2024, 11:16 p.m.