UC San Diego

In recent decades, advances in sequencing technologies have led to an explosion of discoveries in cancer. While observing large chromosomal abnormalities under the microscope has demonstrated genome rearrangements can drive cancer progression, more recent technologies enabled discoveries of mutations private to single cancer patients and uncovered a broader mutation diversity. My dissertation introduces novel connections between computational methods and sequencing techniques to solve open problems in genome rearrangement research. To improve non-invasive cancer monitoring, genome rearrangements can serve as the ideal cancer biomarker for accurately monitoring tumor burden and catching relapse earlier. My approach, AmBre (Amplication of Breakpoints), characterizes a target genome rearrangement's breakpoints for use as a quantitative marker in measuring amounts of tumor DNA. For a target genome rearrangement such as CDKN2A deletion, AmBre accounts for diverse deletion breakpoints and amplies any DNA harboring the CDKN2A deletion. Since only the tumor DNA is amplied, breakpoints can be detected in tissues or blood with little tumor DNA in high background of unmutated DNA. Furthermore, AmBre relies on sequencing technologies to read the enriched DNA. For parallel detection of breakpoints across numerous samples, a geometry based rearrangement caller was developed to handle long reads generated by Pacific Biosciences sequencing instruments. In addition, I will discuss the limitations of sequencing technologies in inferring mechanisms for rearranging genomes. Specifically, sequencing data alone cannot infer a complex cancer chromosome was formed by a single shattering and repair mechanism (chromothripsis) or a series of progressive rearrangements. Lastly, genomes are diploid and genome rearrangements can appear on one or both homologous chromosomes. Detecting genome rearrangements is challenging and inferring which chromosome is affected by the rearrangement is even more difficult. Having already called genome rearrangements such as deletions, I will show how proximity-ligation sequencing can be repurposed to assign deletions to a chromosome by phasing deletions with variants. In effect, my endeavors in genome rearrangement research show the field is constantly evolving with advances being made by complementing sequencing strategies and computational methods