Chromoscope: Interactive Visualization for Clinical Interpretation of Whole Cancer Genomes

A detailed analysis of whole DNA sequences holds the promise of identifying novel ways to cure and diagnose cancer. However, the size and complexity of the data, as well as the difficulty in obtaining highly sensitive and specific mutation calls, prevent physicians and scientists from interpreting complete DNA profiles. To facilitate the clinical interpretation of cancer genomes, we built Chromoscope, a web application for visualization of mutations in cancer, especially for structural alterations of chromosomes. The Chromoscope platform (https://chromoscope.bio/) allows the user to load their samples and view the mutations at multiple scales, from the scale of chromosomes, through genes, to individual sequencing reads. We will extend Chromoscope to provide an interactive overview of each cancer case and support a comparative analysis of groups of cancers. We will prepare interactive visualizations of cancer genome datasets from thousands of patients from newly published cohorts, as a resource for reference and discovery. The platform will facilitate genomic education for the public and clinicians and ultimately enable the discovery of clinically relevant features in cancer genome data and accelerate the use of whole cancer genomes in clinical applications.

Final Report:

The Chromoscope web application makes cancer genome analysis more accessible to researchers who do not code and prefer Google Maps–style interactions with data. The app offers a visual interface for institutions generating cancer genome data and seeking to open interpretation to a wider audience. Thanks to funding from ICI, we conducted user testing that led to a redesign of the graphical interface, with emphasis on easily accessible documentation, video demonstrations, and features for highlighting clinically relevant variants. Furthermore, Chromoscope’s documentation defines data formats for structural and other variant types compatible with the platform, facilitating data exchange and the construction of even larger cohorts.

Chromoscope is now the go-to website for rapid assessment of structural variant landscapes across cancer types and for evaluating the prevalence of structural variants associated with homologous recombination deficiency, replication stress, or focal amplifications — all of which are linked to emerging cancer therapies that are either approved or in clinical trials. The Chromoscope app provides easy access to data from the Pan-Cancer Analysis of Whole Genomes and MSK Spectrum cohorts. Both datasets are accessible directly via chromoscope.bio, as well as through cbioportal.org; the latter integration additionally enables exploration of large cohorts and selection of cases of interest for detailed analysis in Chromoscope. We are currently working to bring additional large cancer genome cohorts to Chromoscope by integrating with the Genomics Data Commons, Genomics England, and Cancer Cell Line Encyclopedia data infrastructures.

Observations made using Chromoscope regarding the co-occurrence of structural variants and mutations have led to the redefinition of mutational signatures — an important tool for understanding mechanisms of cancer initiation and a biomarker for existing and future cancer therapies. Our manuscript on this topic is currently under review.

We are planning to submit follow-up funding applications to support maintenance of the application, extension of AI-assisted interpretation of structural variants, and expanded visualization of clinical biomarkers.

Learn More About Their Work:

Nature Methods
Sehi L’Yi, Dominika Maziec, Victoria Stevens, Trevor Manz, Alexander Veit, Michele Berselli, Peter J. Park*, Dominik Głodzik* & Nils Gehlenborg*. Chromoscope: interactive multiscale visualization for structural variation in human genomes

Link:  Chromoscope: interactive multiscale visualization for structural variation in human genomes | Nature Methods

Poster presented at the Dana Farber/ Harvard Cancer Center symposium 2024

NIH Pub Med: Mechanisms of tandem duplication in the cancer genomeScully, Ralph, Dominik Glodzik, Francesca Menghi, Edison T. Liu, and Cheng-Zhong Zhang. "Mechanisms of tandem duplication in the cancer genome." DNA repair (2024): 103802.

Video Presentation:

YouTube Video on how the duo of cBioportal + Chromoscope work