The genomic landscape of somatic viral integration in cancer

Resumen

Viruses cause 10-15% of human cancers worldwide and some oncogenic viruses nucleic acid may become integrated into the human genome. This process can induce tumours, a mechanism called insertional mutagenesis. The present work aims to shed light on viral integrations in cancer, detecting and characterising somatic Hepatitis B virus (HBV) integrations in hepatocellular carcinoma (HCC) samples from the Pan-Cancer Analysis of Whole Genomes (PCAWG) dataset. In addition, a new tool for the detection of viral integrations in both, Next Generation and Third Generation Sequencing data, is presented here. This pipeline was used for the analysis of human cancer samples HBV and Human Papillomavirus (HPV) positive and cell lines Kaposis Sarcoma Herpesvirus (KSHV) positive. These analyses revealed that non-canonical HBV insertions occur in association with megabase-size deletions that remove telomeric regions of chromosomes and can cause the loss of tumour suppressor genes. Furthermore, we found that HBV somatic insertions in HCC are typically acquired early in tumour evolution. In addition, VIRUNS (VIral-Reference UNions Search) was developed. It was benchmarked using in silico genomes and compared with three short-read integration detection tools, displaying the best balance between precision and recall. VIRUNS successfully characterised HPV integrations in the PCAWG dataset, which are usually coupled with local sequencing coverage changes. Finally, KSHV-human DNA junctions, not described until the moment, were detected by VIRUNS in two primary effusion lymphoma cell lines.