The global burden of herpes simplex virus 1 (HSV1) is ~3.7 billion, while HSV2 afflicts ~400 million. In adults, these lifelong infections typically cause epithelial lesions, which recur whenever the virus reactivates from its lifelong latent reservoir in neurons. In newborns the outcomes are more dire, with approximately half of all HSV infections leading to invasive viral spread into the central nervous system (CNS), or viral dissemination into organs such as the lungs or liver. Rates of mortality and lifelong morbidity are significantly higher for the invasive CNS and disseminated forms of neonatal infection, than for infections that remain limited to the body surface. The contribution of HSV genetic variation to these different clinical outcomes is as yet unknown. In a recent pilot analysis of neonatal HSV samples, we found patterns of viral genetic variation that correlated with invasive spread phenotypes. We now propose to extend our genomic and phenotypic analyses of viral variation to a larger neonatal dataset and to incorporate in vivo models of virulence. The combination of viral comparative genomics, cell-based phenotyping, in vivo models of pathogenesis, and de-identified clinical data will lay the foundation for a future genome-wide association study (GWAS) for neonatal HSV. Using these data, we will probe connections between viral genetic variation and clinical outcomes such as invasive CNS vs. skin disease, severity of neurologic impairment, and response to antivirals. In Aim 1, we will use high-throughput sequencing (HTSeq) and comparative genomics to dissect viral genetic variation between and within individuals with HSV1 or HSV2 neonatal HSV disease. We will quantify differences between-hosts at the level of the overall consensus genomes, and within-host by examining minor variants in samples from distinct niches in the body. In Aim 2 we will determine the in vitro phenotypic profile of each cultured HSV isolate, by examining rates of viral replication, cell-to-cell spread, plaque morphology, and viral protein expression and localization in a panel of cell types, including neurons. In Aim 3 we will determine the rate of spread and neuroinvasion in vivo for neonatal HSV1 and HSV2 isolates, using murine models of either disseminated or CNS infection. These data will allow us to link differences in viral genetics (from Aim 1) and cell-based phenotypes (from Aim 2) with levels of neuroinvasion and virulence observed in vivo (Aim 3) or via de-identified clinical data for these neonatal isolates. Our ultimate goal is to find measures than can predict the risk of highly-invasive disease for newly infected newborns, so that individuals at highest risk can be identified and targeted for intervention(s) to limit viral invasion and improve clinical outcomes.