Tumour virology in the era of high-throughput genomics

With the advent of massively parallel sequencing, oncogenic viruses in tumours can now be detected in an unbiased and comprehensive manner. Additionally, new viruses or strains can be discovered based on sequence similarity with known viruses. Using this approach, the causative agent for Merkel cell carcinoma was identified. Subsequent studies using data from large collections of tumours have confirmed models built during decades of hypothesis-driven and low-throughput research, and a more detailed and comprehensive description of virus–tumour associations have emerged. Notably, large cohorts and high sequencing depth, in combination with newly developed bioinformatical techniques, have made it possible to rule out several suggested virus–tumour associations with a high degree of confidence. In this review we discuss possibilities, limitations and insights gained from using massively parallel sequencing to characterize tumours with viral content, with emphasis on detection of viral sequences and genomic integration events. This article is part of the themed issue ‘Human oncogenic viruses’.