The prevalence of the mutations detected in complex DNA mixture has traditionally been limited to approximately 20% using Sanger sequencing [21,22]. The development of specific mutation enrichment or detection strategies has greatly increased this sensitivity [3,23,24], but impaired the breadth of the assay. The UDT-Seq approach presented here offers a streamlined method to implement in clinical care massively parallel sequencing of cancer mutational hotspots in heterogeneous samples. The simultaneous sequencing of a calibration sample enhances the robustness of the assay and therefore the reliability of the results. We have shown that this approach can comprehensively detect low prevalence mutations by screening 71,081 DNA positions located in cancer mutational hotspots. The sensitivity of the assay down to mutations present at 5% prevalence permits detection of mutations in heterogeneous or poor quality samples with rare mutated clones, low cellularity, or contamination with stroma or immune cell infiltration, all of which are commonly seen in clinical samples. Importantly, our data suggest that in order to increase the reliability and identify mutations present at less than 5% prevalence, the accuracy of the next generation
