DISCOVER supplementary methods

The notebooks below contain all the code required to reproduce the figures and results of the paper A novel independence test for somatic alterations in cancer shows that biology drives mutual exclusivity but chance explains co-occurrence.

Simulated data analyses

• Pairwise analyses of simulated data

  Compares the Binomial, Fisher's exact and DISCOVER tests on simulated data.

• Group test

  Compares the DISCOVER group test to six alternative methods (CoMEt, MEGSA, MEMo, muex, mutex, and TiMEx) on simulated data.

Pan-cancer analyses

• Download PanCan12 data

  Downloads the mutation and copy number data for the TCGA PANCAN12 studies.

• Gene selection

  Selects the genes for use in the pairwise analyses.

• Pairwise analysis

  Performs pairwise co-occurrence and mutual exclusivity analyses.

• Within-chromosome co-occurrence analysis

  Tests for co-occurrences between genes located on the same chromosome, in order to assess whether the DISCOVER test will detect these 'positive controls'.

• STRING enrichment

  Determines the overlap of mutually exclusive gene pairs with the STRING functional interaction network.

• MSigDb group tests

  Identifies significantly mutually exclusive gene sets based on predefined gene sets extracted from MSigDb.

• De novo gene set identification

  Detects de novo mutually exclusive gene sets based on correlation clustering of pairwise mutual exclusivities.