In the past few years, interferon-induced transmembrane (IFITM) proteins have been identified as important antiviral factors. The current understanding of IFITMs suggests that they localise within distinct cellular compartments from where they exert their broad antiviral role. For example, IFITM1 localises to the plasma membrane and restricts viruses that do not require endocytosis to infect host cells. In contrast, IFITM2 and IFITM3 are found in the early and late endosomes, respectively, and are potent inhibitors of viruses that depend on endosomal pathways for infection.

I begin this dissertation by providing some background on the biology and function of IFITM proteins, including details of in vitro assays that have helped elucidate IFITMs role as antiviral factors. I also describe some early candidate-gene association studies that have attempted to correlate genetic variation within these genes with variation in viral restriction. I also describe how genetic association studies have been used more broadly to understand the biology underlying both infectious and non-communicable diseases.

Evidence from in vitro, and in vivo work has demonstrated the IFITMs role as potent antiviral factors, however, no genome-wide association study has reported any significant associations to genetic variant in or around these genes. In Chapter 2, I explore reasons why this may be the case and calculate the coverage of IFITM genes by commercially available genotyping arrays. I show that IFITM2 and IFITM3 are amongst the 7% of all protein coding genes with less than 25% common variant (minor allele frequency > 5%) coverage across all arrays. Poor coverage of genetic variation is therefore one explanation for the lack of IFITM associations in GWAS.

The lack of coverage in the genotyping arrays led me to explore other tools to capture variation in the IFITM region. I employ a targeted sequencing method using two different sequencing technologies: short-read sequencing (Illumina MiSeq) and single molecule, real-time sequencing (PacBio RS). Conventional pulldown protocols for targeted sequencing have not been designed for single molecule, real-time sequencing at the time, thus in Chapter 3, I provide some details of the optimisation work required to adapt the targeted method for PacBio sequencing. I then assess the performance of the method for both Illumina and PacBio sequencing.

In Chapter 4, I apply the targeted sequencing method described in Chapter 3 to test genetic variants in and around IFITM1, IFITM2 and IFITM3 for association with rapid disease progression in HIV. I also explore the contribution of rare genetic variants (MAF < 1%) to this phenotype by testing for a differential enrichment between cases and controls across each of the three genes.

Studies in vitro have also reported that IFITM proteins are potent restrictors of dengue virus infection. In Chapter 5, I use genotype data across a cohort of 2,008 Vietnamese children diagnosed with dengue haemorrhagic fever (DHF) and 2,018 cord blood controls to test if common variants are associated with the disease.