TiMEx: Finds mutually exclusive groups

In cbg-ethz/TiMEx: Finding Mutually Exclusive Groups of Alterations in Cancer Datasets

Description

TiMEx is the main function of this package. It identifies all groups of mutually exclusive cancer alterations in a large binary input dataset.

Usage

TiMEx(mat, pairMu, pairPvalue, groupPvalue)

Arguments

mat	binary alteration matrix, with rows representing patients and columns representing genes
pairMu	pair-level threshold on mu (real number between 0 and 1). Default is 0.5.
pairPvalue	pair-level threshold on p-value (real number between 0 and 1). Default is 0.01.
groupPvalue	threshold for the corrected p-value, lower than which cliques are significant. Default to 0.1

Details

Dependning on the size of the dataset (both in terms of samples and alterations), TiMEx can require a reasonable time to run. For example, the approximate running time is 10 minutes for the ovarian cancer dataset, and 45 minutes for the breast cancer dataset included in this package, on a personal computer.

TiMEx displays progress messages. In a fist step, it indicates the gene which is currently being tested against the remaining genes. In a later step, it indicates the size of the clique currently being tested, and the number of cliques to test.

Value

list consisting of:

• genesSignif list of significantly mutually exclusive groups, as gene names, sorted by corrected p-value. The list contains as many elements as identified lengths of groups. For example, genesSignif[[2]] is a list containing the gene names of the significant groups of size 2. Each list of this type further has two elements, fdr and bonf, corresponding to different multiple testing correction methods. Each element is a matrix, in which rows represent gene names of significantly mutually exclusive groups.

• idxSignif list of significantly mutually exclusive groups, as indices in the input matrix, sorted by corrected p-value. The list contains as many elements as identified lengths of groups. For example, idxSignif[[2]] is a list containing the indices of the significant groups of size 2. Each list of this type further has two elements, fdr and bonf, corresponding to different multiple testing correction methods. Each element is a matrix, in which rows represent indices of significantly mutually exclusive groups.

• pvals list of corrected significant p-values corresponding to the tested cliques, ordered ascendingly. The list contains as many elements as identified lengths of significant groups. For example, pvals[[2]] is a list containing the p-values of the significant maximal cliques of size 2. Each list of this type further has two elements, fdr and bonf, corresponding to different multiple testing correction methods. Each element is a vector, of length the number of significant maximal cliques of a given size.

• posSignif list of positions of the significant groups in the input list of maximal cliques, ordered ascendingly by corrected p-value. The list contains as many elements as identified lengths of significant groups. For example, posSignif[[2]] is a list containing the positions of the significant groups of size 2. Each list of this type further has two elements, fdr and bonf, corresponding to different multiple correction methods. Each element is a vector, of length the number of significant maximal cliques of a given size.

• MusGroup list of inferred mu values corresponding to the tested cliques, ordered ascendingly by the corresponding corrected p-value. The list contains as many elements as identified lengths of significant groups. For example, MusGroup[[2]] is a list containing the mu values of the significant maximal cliques of size 2. Each list of this type further has two elements, fdr and bonf, corresponding to different multiple testing correction methods. Each element is a vector, of length the number of significant maximal cliques of a given size.

• mcStruct input structure of maximal cliques to be tested for mutual exclusivity, as returned by doMaxCliques.

• matrix input binary alteration matrix.

• groupPvalue input threshold for the corrected p-value, lower than which cliques are significant.

Author(s)

Simona Cristea, [e-mail]

References

Constantinescu et al.: TiMEx: A Waiting Time Model for Mutually Exclusive Cancer Alterations. Bioinformatics (2015).

See Also

analyzePairs for step 1 of the TiMEx procedure; doMaxCliques for step 2 of the TiMEx procedure, and findSignifCliques for step 3 of the TiMEx procedure. The data structures ovarianOutput, breastOutput, gbmDendrixOutput, and gbmMuexOutput are examples of structures resulting after running TiMEx on large cancer datasets.

Examples

# Run TiMEx on the ovarian cancer dataset with default parameters 
# (takes approximately 10 minutes)
data(ovarian)
ovarianMEGroups<-TiMEx(ovarian)

cbg-ethz/TiMEx documentation built on May 13, 2019, 1:50 p.m.