6. Available data

REVOLVER comes with usable multi-region data from several cohorts, often in the form of its own outputs; these data can be used to re-run fit with different parameters, or just test the method. We also provide survival data for some single-sample cohorts, which one can use as readout to validate REVOLVER's groups.

Most of the datasets are embedded in REVOLVER's R package, and can be loaded via the usual data(...) call; datasets that are too large to include in a package are available as RData at the Github repository caravagn/revolver.misc, and referenced here.

We provide also vignettes for some of these case studies.

Data to fit models via REVOLVER.

• n = 19 Colorectal adenomas and carcinomas: Cross et al, Nat Eco Evo; to appear.

  • original input data (dataframe): data(CRC);
  • REVOLVER object (cohort and fits): data(CRC.cohort).

• n = 99 Non-small cell lung cancers (from TRACERx): Jamal-Hanjani et al; NEJM 2017.

  • original input data (dataframe): data(TRACERx);
  • REVOLVER output (fits, clusters and jackknife; ~100MB): caravagn/revolver.misc.

• n = 50 Primary and Metastatic breast cancers: Yates et al; Nat Med 2015

  • REVOLVER output (fits, clusters and jackknife): data(Breast.fit)

• n = 10 Primary and Metastatic renal carcinomas: Gerlinger et al; NEJM 2012

  • available upon request.

Data for survival analysis. All these datasets have been downloaded from the cBio portal at MSKCC.

• Non-small cell lung cancers: n = 230 from TCGA LUAD 2014, plus n = 182 from Broad Institute LUSC 2012, plus from n = 177 from TCGA LUSC 2012

  • Curated RData: TRACERx-survival.RData.

• Breast cancers: n = 1318 from METABRIC 2016, plus n = 434 from TCGA BRCA 2012.

  • Curated RData: Breast-cancer-survival.RData.

• n = 300 TCGA KIRC provisional 2012 kidney cancers

  • available upon request.