GenCHORD: A Pipeline for Genomic Rearrangement Analysis

We present GenCHORD, a specialised step-detection algorithm suitable for detecting large-scale copy number variations from the coverage data of a sequenced, aligned genome.This algorithm is specifically tuned for detecting Structural Variations which arise during chromothripsis-induced cancer, and so permits a high degree of data reduction, whilst preserving pertinent features for cancer classification.We describe the underlying statistical and algorithmic model, demonstrate the ability of the to preserve, compress and encode information such that a simple Machine Learning model can classify and identify subtypes of Devil Facial Tumour Disease, a cancer affecting Sarcophilus harrisii, the Tasmanian Devil, and discuss how this tool may be used in future.