Next generation molecular diagnostics using ultrasensitive sequencing

Abstract

Massively parallel sequencing enables the exploration of the genetic heterogeneity within microbial, viral and tumor cell populations. Detecting circulating tumor DNA in blood and other body fluids has the potential to revolutionize molecular diagnostics. However, these liquid biopsies typically contain only minute amounts of highly degraded DNA and standard sequencing approaches lack the resolution to detect rare genetic variants. The overall goal of this thesis was to develop an ultrasensitive sequencing approach with single molecule resolution that requires only minimal amounts of material. To this end, we developed the simple multiplexed PCR-based barcoding of DNA for ultrasensitive mutation detection by next-generation sequencing protocol (SiMSen-Seq). SiMSen-Seq achieves ultrasensitive detection of nucleotide variants by attaching unique molecular identifiers to target DNA molecules using PCR primers. SiMSen-Seq is enabled by highly optimized reaction conditions and the use of a stem-loop structure that prevents the UMI from forming non-specific PCR products. We showed that ultrasensitive variant detection is attained mainly by using UMI, while gains in sensitivity from using high-fidelity polymerases were minor. We also demonstrated that oligonucleotide quality is essential in numerous molecular applications, including SiMSen-Seq. Next generation diagnostics tools also demand optimized preanalytical conditions to achieve the necessary variant detection sensitivity, while remaining fast, simple, and cost efficient. Therefore, we established a workflow for cell-free DNA analysis and developed quantitative PCR-based quality controls to evaluate each experimental step. We also developed a bioinformatics pipeline for processing any type of targeted sequencing data containing unique molecular identifiers, including barcode clustering, error correction, variant calling, and visualization. Next, we used SiMSen-Seq in applications requiring ultrasensitive mutant detection. We first employed SiMSen-Seq to experimentally confirm that UV light rapidly induces highly recurrent mutations within a specific promotor motif. These mutations remained sub-clonal even after weeks of cell culture, arguing against a tumor-driving role. Our results highlight the importance of sequence context for the interpretation of somatic variants in cancer. We also showed that ctDNA can be used as a clinical biomarker for tumor burden and to monitor treatment efficacy in uveal melanoma. Patients with high ctDNA levels had worse overall survival, demonstrating the clinical utility of circulating tumor-DNA-based liquid biopsy analysis. In conclusion, we showed that SiMSen-Seq is a simple, flexible, low-DNA input protocol that enables rare variant detection to address a multitude of clinical and basic research questions.