UCSF

The emergence of high-throughput genomic technologies has markedly accelerated virus detection, virus discovery, and viral metagenomics. DNA viral microarrays and high-throughput sequencing platforms allow for the unbiased detection of all pathogens in parallel. In this dissertation, the capacity for a novel or unexpected viral etiology in a number of idiopathic diseases is investigated. More specifically, serum from individuals with acute liver failure, ocular fluids from patients with uveitis, and bronchoalveolar lavage from individuals experiencing an acute exacerbation of idiopathic pulmonary fibrosis are analyzed for all viruses.

An alternative to the syndromic approach to virus discovery is also taken in tandem, where a hypothesized gap in virus phylogeny is targeted specifically. Overlaying viral hosts onto a phylogenetic tree of a conserved herpesvirus gene reveals a clade of herpesviruses found in a number of primates, with a distinct gap suggesting a homologous herpesvirus in human hosts. Attributes of this gap, such as sequence similarity, viral lifestyle, and tropism, are used to design a set of discovery projects. Saliva samples from patients with full-blown AIDS were collected and banked in an era before antiretroviral therapy, and tonsils, a lymphocyte rich tissue, from adolescents with recurrent tonsillitis was collected and treated with a chemical known to induce the herpesvirus lytic gene cascade. These samples were analyzed with the pan-viral microarray and high-throughput sequencing for the presence of a novel human herpesvirus.

This method of targeted analysis of a hypothesized virus gap is then generalized to all viruses in an effort to precipitate hypotheses of discovery targets, or gaps, in virus trees when they are overlaid onto their corresponding host trees. The propensity for viruses to coevolve with their hosts motivates a framework for translating host homology to homology between known viruses and their suggested orthologs (gap-recognition). These proposed gaps present insight into viral diversity and evolution, and can also be used to motivate targeted microarray-based and high throughput sequencing-based virus discovery efforts. With the rapid advances in massively parallel genomic platforms, there is a clear demand for more informed and targeted molecular and bioinformatic strategies of detection.