important solution to the problem of analyzing heterogeneous samples. Massively parallel sequencing enables the analysis of independent, clonal, DNA molecules [5,6] and has been used early on to digitally measure the presence of low prevalence mutations in complex DNA mixtures [7,8] or in EGFR exons of heterogeneous tumor DNA samples [9]. Constant improvements of the massively parallel sequencing technology offer the opportunity to revise the balance between breadth and depth of such assays and identify a wide variety of potentially actionable DNA changes in a patient's tumor. Broad assays like whole genome and whole exome sequencing have been used to discover new cancer mutations [10,11], or study clonal selection in breast cancer [12]. However, their performance on heterogeneous clinical samples has not been demonstrated and the significance of the vast majority of the mutations identified is not clear; therefore, such broad sequencing approaches currently have limited clinical utility for personalized cancer treatment. In contrast, a more targeted sequencing approach assaying all clinically actionable genes, but no extraneous regions, allows for the depth of sequencing to be maximized for a more accurate analysis of heterogeneous clinical samples.
