Development and Deployment of a Genomic Cancer Data Extraction Pipeline on the Cloud

Cancer is a disease that affects millions of people worldwide and although treatments have evolved over the past few decades, it remains one of the most challenging diseases to treat effectively. Enormous datasets from human cancer cells including whole-genome sequences have recently become available. These datasets hold undiscovered clues to the origins, spread, and response to treatments of cancer. Telomeres, which form the ends of human chromosomes, are critical for genome stability and play key roles in cancer development and progression. Cancer-associated genomic DNA changes occurring in telomere and telomere-adjacent genome regions remain understudied due to their complex and repetitive DNA sequence organization; we are developing new telomere and subtelomere feature-based metrics to detect and quantify them using available genome datasets. Here, we present a computational pipeline that enables the selective extraction of key telomere-associated data from large genomic datasets on a high-performance computing cluster or a cloud environment. Our pipeline consists of three main workflows used to extract the telomeric, the subtelomeric repeat, and the subtelomere-adjacent single-copy read data, respectively. We employed our pipeline on the AWS HealthOmics cloud infrastructure and tested it on a metastatic cancer dataset of 101 patients, extracting these telomere-associated reads in an automatic, reproducible and timely manner with the use of Nextflow. These well-defined smaller datasets are being used to detect and quantify changed subtelomeric and telomeric features during cancer development and progression.