TiMEx-package: The main usages of TiMEx

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

Description

The main usages of TiMEx, a package for finding groups of mutually exclusive alterations in large cancer datasets.

Overview

The most important function in this package is TiMEx, which identifies all mutually exclusive groups in a binary dataset. TiMEx is a procedure implementing three steps: first, all pairs in the input dataset are tested for mutual exclusivity. Second, maximal cliques are identified on the basis of a selected number of pairs. Third, the resulting cliques are tested for mutual exclusivity. Additional inputs to TiMEx include thresholds on the significance and intensity of mutually exclusive pairs (pairMu and pairPvalue) and q-value cutoff on mutually exclusive groups (groupPvalue). Unless otherwise specified, TiMEx will use default values of these inputs.

Alternatively, the three steps of the TiMEx procedure can be run separately via the three functions analyzePairs, doMaxCliques, and findSignifCliques (in this order).

Preprocessing and postprocessing

This package also provides functions to preprocess the input data (doMetagene, removeLowFreqs), as well as to postprocess the identified mutually exclusive groups (produceTablesSignifGroups, subsampleAnalysis, plotGroupByName, recoverAllNamesGroups).

Datasets

Multiple datasets are available within this package. breast and ovarian are datasets downloaded from cBioPortal (TCGA) in July 2014, and preprocessed as described in Constantinescu et. al: TiMEx: A Waiting Time Model for Mutually Exclusive Cancer Alterations. Bioinformatics (2015). gbmDendrix is a glioblastoma dataset used in Leiserson et. al: Simultaneous identification of multiple driver pathways in cancer. Plos Computational Biology (2013). Additionally, this package also includes the dataset gbmMuex, used and preprocessed as described in Szczurek et. al: Modeling mutual exclusivity of cancer mutations. Research in Computational Molecular Biology (2014).

For each of these four datasets, the identified significantly mutually exclusive groups are available as separate datasets (breastOutput, ovarianOutput, gbmDendrixOutput, and gbmMuexOutput). Similarly, results of a subsampling analysis ran with 100 repetitions on the identified groups are available as separate datasets (breastSubsampling, ovarianSubsampling, gbmDendrixSubsampling, and gbmMuexSubsampling).

For breast cancer and ovarian cancer, the metagroups of genes in the original datasets (produced with the function doMetagene) are available as separate datasets (breastGroups and ovarianGroups).

Finally, the binary input matrices corresponding to the four breast cancer subtypes LuminalA, LuminalB, Her2, and Basal are available in the dataset breastSubtypes, and the significantly mutually exclusive groups identified in each of these four subtypes are available in the dataset breastSubtypesOutput.

Simulations

Datasets can be generated from the TiMEx model using the function simulateGenes.

More

For more in-depth explanations of the TiMEx package and model, including examples, please see the corresponding paper below.

References

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

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