Grant: Doctoral Research Award: Canada Graduate Scholarships

Advancements in next-generation sequencing of human pathogens have catalyzed the generation of massive amounts of data well-suited to model their evolution and transmission. The human immunodeficiency virus (HIV) epidemic is one of the best examples of how viral sequences routinely collected for drug resistant mutation surveillance have altered prioritization of public health service delivery. Comparisons of patients' HIV sequences, which are more similar when viruses are linked through recent transmission events, allow reconstruction of viruses' shared ancestry through the transmission network. In British Columbia (BC), HIV sequences are clustered according to thresholds of genetic similarity in order to identify subpopulations experiencing high transmission rates. Individuals in or associated with large or rapidly growing clusters receive prioritized treatment services, partner contact tracing, and HIV testing and pre-exposure prophylaxis for partners. While demonstrated to be an effective way to prioritize resources, clustering methods can be improved and optimized to best identify individuals with heightened transmission activity. Outside of BC, alternative clustering and transmission analysis methods have been applied, yet the comparative accuracy and sensitivity of these methods has not been explored. Therefore, I propose a systematic evaluation of HIV transmission modelling methods in regard to their thresholds, evolutionary tree models, and sensitivities to sampling biases. Using a combination of empirical data from the BC Drug Treatment Program and simulated epidemic data, we will identify and optimize the method that most accurately identifies clusters of individuals with high transmission rates. By pinpointing populations to whom public health resources should be prioritized, this research will improve the efficiency of HIV health service delivery and pave the way for prioritization schemes for other pathogens, such as hepatitis C virus, as well.