NEW YORK – A team from Weill Cornell Medicine, the Broad Institute, Dana-Farber Cancer Institute, and elsewhere has identified a handful of new complex genome rearrangement groups in cancer genomes using a junction balance-based analysis focused on thousands of available tumor genome sequences.
"Clustering of tumors according to genome-graph-derived features identified subgroups associated with DNA repair defects and poor prognosis," senior author Marcin Imielinski, a pathology and medicine, precision medicine, and computational biomedicine researcher affiliated with Weill Cornell Medicine and the New York Genome Center, and his colleagues wrote.
For a paper published in Cell on Thursday, the researchers came up with a computational strategy known as the junction balance analysis, or JaBbA, algorithm to search for structural variants based on "junction copy number" (JCN) genome-graph profiles parsed from almost 2,800 new and previously-reported tumor whole-genome sequences spanning 31 cancer types.
"We hypothesized that the topology and [copy number] of both vertices and edges of junction-balanced genome graphs might reveal novel classes of complex [structural variant] events and mutational processes," the authors explained.
Indeed, the team's analyses unearthed three main structural variant types that they dubbed "rigma," "pyrgo," and "tyfona" rearrangements — alterations marked by fragile site deletions; towering duplications involving super-enhancers; or scattered, amplified fold-back inversions, respectively, that were over-represented in certain cancer types.
The alterations occurred in a wide range of cancer types and were nothing if not complex, impacting large swathes of the affected tumor genomes, Imielinski explained in a statement.
"If I took a piece of chapter three and I duplicated it five times, and then I stuck it into the middle of chapter two, and then shredded that whole thing and distributed the pieces around the last ten chapters of the book, these are the kinds of transformations [being identified]," he said.
In the gastrointestinal tumors, for example, the researchers saw enhanced representation from the fragile site deletion chasms, which appeared to have occurred early on in the evolution of the tumors. On the other hand, breast and ovarian cancers appeared more prone to the pyrgo alterations, while tyfonas turned up more often in acral melanomas.
In a small cell lung cancer cell line known for harboring the tyfona type alteration, the team's follow-up analyses based on additional short read sequence, optical mapping, Hi-C three-dimensional interaction profiling, fluorescence in situ hybridization, RNA sequence, and other data provided a more refined look at these structural changes and the mutations that co-occurred with them.
The researchers are reportedly applying their structural variant detection strategy to cancer cases and expect the approach to become more widespread in clinics and in research labs in the future.
"Larger short-read meta-analyses spanning [more than] 10,000 cancer genome graphs, which may be imminently possible through the combination of our approach and dataset with [data from other projects and precision medicine efforts] … will help link novel features of genome-graph topology with clinical variables, including therapeutic response," the authors conclude. "This includes future investigations linking the presence of tyfonas to the efficacy of immune checkpoint inhibition in acral melanoma or other tumor types in which these events are found."