CoRAL accurately resolves extrachromosomal DNA structures with long-read sequencing

Abstract

Extrachromosomal DNA (ecDNA) is a prevalent mechanism of focal oncogene amplification found in approximately 20% of all cancer samples. EcDNA amplifications often appear in the most aggressive tumours, where their rapid copy-number modulations and remarkable genetic diversity contribute to faster tumour progression and drug resistance. Thus, elucidating the genomic structure of ecDNA is a critical goal for understanding tumour pathology and developing more effective therapeutic strategies. To date, the most commonly used tool for identifying ecDNA amplifications is that of paired-end short-read (Illumina) sequencing. However, short-read technologies are intrinsically limited in their ability to detect structural variants on ecDNA, phase together complex rearrangements and internal duplications, and deconvolve cell-to-cell heterogeneity of ecDNA structures. Long-read technologies, such as from Oxford Nanopore Technologies, can potentially improve inference as the longer reads are better at mapping structural variants and more likely span rearranged or duplicated regions. In this talk, I will describe CoRAL (Complete Reconstruction of Amplifications with Long reads): a computational tool for the genome-wide reconstruction of ecDNA from long reads. CoRAL reconstructs likely cyclic structures using quadratic programming that simultaneously optimizes parsimony of the reconstruction, explained copy-number, and consistency of long-read mapping. Using extensive simulations and datasets from nine previously-characterised cell lines, we demonstrate that CoRAL outperforms short-read-based approaches at ecDNA reconstruction. As long-read technologies continue to become more widespread, we anticipate that CoRAL will be a valuable tool for profiling the landscape and evolution of focal amplifications in tumours. CoRAL is publicly available at GitHub.

Biography

Matthew Jones, PhD is a postdoctoral scholar at Stanford University, USA, advised by Howard Chang, MD, PhD. His work centres on developing computational and technological approaches to better understand evolution in biological systems, especially applied to cancer. Previously, Matthew earned his PhD from University of California, San Francisco, advised by Dr. Jonathan Weissman and Dr. Nir Yosef, where he developed algorithms for reconstructing tumour lineages from CRISPR-based lineage tracers. His current research focuses on better elucidating the role of extrachromosomal DNA (ecDNA) in tumour progression and drug resistance. Matthew is currently supported by an NCI K99/R00 Early-Career Postdoctoral Fellowship.