Assessing intratumor heterogeneity and tracking longitudinal and spatial clonal evolutionary history by next-generation sequencing

Cancer is a disease driven by rounds of genetic and epigenetic mutations that follow Darwinian evolution. The tumor for a given patient is often a mixture of multiple genotypically and phenotypically distinct cell populations. This contributes to failures of targeted therapies and to drug resistance, and thus it is important to study intratumor heterogeneity. Here, we propose Canopy, a statistical framework to reconstruct tumor phylogeny by next-generation sequencing data from temporally and/or spatially separated tumor resections from the same patient. We show that such analyses lead to the identification of potentially useful prognostic/diagnostic biomarkers and successfully recover the tumor's evolutionary history, validated by single-cell sequencing. Canopy provides a rigorous foundation for statistical analysis of repeated sequencing data from evolving populations.