13. AmpliconSuite: Analyzing focal amplifications in cancer genomes

Focal amplifications in cancer genomes, particularly extrachromosomal DNA (ecDNA) amplifications, are pivotal in cancer progression, enabling high amplification of oncogenes. Distinguishing the different mechanisms with whole-genome sequencing (WGS) data is challenging due to their complex profiles of copy number (CN) and structural variation (SV). We present AmpliconSuite, a collection of tools enabling robust identification of focal amplifications from WGS data.

At the core are the AmpliconArchitect (AA) and AmpliconClassifier (AC) methods, which detect and analyze SVs and CNs using WGS data to produce robust predictions of focal amplification types, including ecDNA and breakage fusion-bridge (BFB) cycles. We combined these tools into a single, reproducible workflow, AmpliconSuite-pipeline, which is available through Nextflow, GenePattern and Bioconda. AmpliconSuite-pipeline also incorporates other upstream tools into the workflow to standardize inputs and improve filtering of inputs. It introduces our latest classification methods, such as the ecContext method within AC for classifying types of ecDNA based on their mechanisms of formation (chromothripsis, excision, etc.).

To foster collaboration, we also introduce a companion website, AmpliconRepository.org. This community-editable platform allows researchers to publicly share calls generated by AmpliconSuite. Notably, AmpliconRepository.org provides ecDNA predictions on 2,525 tumor samples from TCGA, PCAWG, and CCLE. The ongoing goal of this repository is to become the largest resource for focal amplifications in cancer, driven primarily by ecDNA but also including other mechanisms like BFB. AmpliconSuite makes identification of focal amplifications reproducible and simple, and empowers users to share analyses publicly, representing a valuable resource to investigate the mechanisms of oncogene amplification.