Despite the use of potent immunosuppressive agents, injury secondary to the alloimmune response principally directed against mismatched HLA antigens, remains a significant cause of renal transplant dysfunction and loss. One of the principal ways to limit the detrimental effect of immune alloreactivity towards the graft is through minimisation of donor-recipient HLA mismatches, but this has not changed conceptually since the early period of clinical transplantation. The central hypothesis of the research described in this thesis is that assessment of histocompatibility based on evaluation of structural differences between donor and recipient HLA molecules has the potential to improve graft outcomes after kidney transplantation. Amino acid sequence analyses have been central to gaining an understanding of the relative immunogenicity of a mismatched donor HLA antigen in the context of a recipient's HLA repertoire. While these methods offer advantages over basic enumeration of HLA mismatches, regardless of how sophisticated they may be, they are unable to fully describe the intricate nature of the B cell epitope--BCR interaction that initiates a humoral response. This interaction requires, but is not limited to, structural and electrostatic potential complementarity. The aim of this thesis is to describe a fully structural description of HLA immunogenicity through analysis of the three-dimensional physicochemical environment at the surface of the HLA molecule. First, I describe the bioinformatic techniques that integrate HLA sequence and X-ray crystallographic data to produce accurate models of HLA molecular structure and, subsequently, calculate the electrostatic potential in the three-dimensional area surrounding the molecule. Through quantitative comparison of the three-dimensional electrostatic potential on the surface of each HLA molecule, Electrostatic Mismatch Score 3D (EMS-3D) is derived. EMS-3D represents the electrostatic disparity between a group of HLA molecules. Subsequently, I demonstrate, using examples of HLA mutagenesis studies described in the literature, that patterns of antigenicity are better explained through analysis of surface electrostatic potential than by analysis of sequence. Of greater clinical interest, the ability of EMS-3D to predict the development of \textit{de novo} alloantibody responses was assessed in a cohort of patients who received a standardised immunisation and had their antibody profile characterised using Luminex single-antigen beads. This analysis validated EMS-3D as a valuable predictor of the development and magnitude of an alloantibody response, and therefore as a potentially useful clinical tool in the allocation of donor organs. The clinical utility of EMS-3D was assessed in a national cohort of kidney transplants, which found EMS-3D to better predict transplant outcome than conventional, currently-employed, measures of histocompatibility. HLA polymorphisms are implicated in many disease processes, particularly in autoimmunity. In the last chapter, the methodology developed in this thesis is employed to analyse the association between particular HLA polymorphisms and Inflammatory Bowel Disease (IBD). Analyses of the electrostatic potential within the HLA peptide binding groove found that particular patterns of electrostatic potential within the HLA peptide binding groove were associated with disease phenotype. Future work will examine how the examination of multiple discrete regions, ostensibly epitopes, can further improve the prediction of the HLA alloimmune response.