Background: Type 1 diabetes (T1D) is a common complex metabolic disease characterized by the autoimmune destruction of insulin-producing β cells of the pancreatic islets. Due to a drastic loss of β cells, resulting in no to low insulin production, persons with T1D require life- long insulin injections. Genetic factors play a significant role in the aetiology of T1D, and the heritability of T1D is estimated to be between 85-88%. Many of the known susceptibility genes are involved in immune processes, which is consistent with its aetiology as an autoimmune disorder in which auto-antigens attack the pancreatic islets cells. However, the main causes of morbidity and mortality today are the complications that arise from even mild hyperglycaemia in T1D cases. Glycemic control is often measured using an assay called haemoglobin A1c (HbA1c), that provides an estimate of average blood glucose in individuals over an ~ 3month period. Objectives: In this thesis, I used phenotype and genetic data from the Diabetes Complications and Control Trial (DCCT), a large (n=1441) randomized clinical trial designed to assess the impact of two different (conventional and intensive) approaches to glycemic control, to examine the relationship between imputed gene expression and mean HbA1c. Hypothesis: I hypothesis that variation in levels of imputed gene expression in blood and/or other tissues will be associated variation in glycemic control. Secondarily, I hypothesis that this association may differ between treatment arms, because of the effect of treatment on HbA1c. Methods: Gene expression or transcriptome imputation is a relatively new method in genetic epidemiology in which training sets from different tissues are used to identify the set of single nucleotide polymorphisms (SNPs) within a given window (typically 2Mb) surrounding a gene that best predict RNA abundance in that tissue using regression regularization or a similar statistical approach. Then, the set of SNPs that predict gene expression in each tissue can be used to impute the transcriptome in an independent genotyped dataset. Here, I use the open-source PredictDB data repository, the DCCT genotype data set and the PrediXcan workflow described by Gamazon et al to test for associations between imputed gene expression and mean HbA1c in ten different training sets (tissues). Results: I find weakly suggestive associations between imputed gene expression and mean HbA1c for nine genes, and I find a treatment*gene interaction for two of these genes. Three of the associated genes mapped to chromosomal regions with known genome wide association study (GWAS) hits, and one of the genes (MAPKI81) has been associated with glycemic control in individuals with Type 2 diabetes. Conclusions and future work: Collectively, this suggests that employing transcriptome imputation may help identify non- coding variants that influence (disease) phenotypes by dysregulating gene expression. In future work, this workflow will be applied to multiple datasets and then a meta-analysis will be performed, in the hopes that we wil be powered to find statistically significant associations between HbA1c and predicted gene expression. The ultimate goal of this research is to tease apart the genetic determinants of hyperglycemia in T1D and discover new therapeutic approaches to controlling blood sugar.