UCSF

Over the past two decades, high-throughput DNA sequencing has revolutionized our understanding of epigenetics. The development of quantitative assays that use sequencing to measure chromatin accessibility and methylation state have provided key insights into oncogenesis and cancer metastasis. Recently developed 3rd generation technologies offer significant improvements over current sequencing platforms, including kilobase-scale read lengths and native detection of epigenetic marks on single molecules. Profiling techniques built on these platforms can therefore capture the epigenetic states of single chromatin fibers with unprecedented resolution. However, the inherently higher input requirements for these assays and sequencing platforms have limited their general applicability in medical settings, where sample material is constrained. In this dissertation, we address this problem by developing a new generalizable strategy for preparing native 3rd generation sequencing libraries from low-input samples using a hyperactive transposase. In Chapter 1, we motivate and contextualize this work by discussing how the epigenome is dysregulated in cancer progression and the current tools we use to study it. Then, in Chapter 2 we discuss how our transposase-mediated method enables the study of chromatin fibers in a range of clinically relevant samples including cancer cell lines and patient derived xenograft models. In Chapter 3, we present further methodological improvements that enable direct library preparation from cells and nuclei, collectively lowering input requirements ~20X and making native single molecule profiling studies competitive with existing epigenomic assays. Finally, we conclude in Chapter 4 by considering how our method is a general tool for using 3rd generation sequencing as a read out for biological assays. We demonstrate this for two specific cases: high-depth profiling of targeted genomic regions and resolving chromatin states in single cells.