HNF1A-MODY is the most common form of monogenic diabetes, but due to low awareness among clinicians and limited availability of genetic testing in many countries, the majority of cases are misdiagnosed as type 1 or type 2 diabetes. In clinical practice, it is apparent that some young adults with non-autoimmune diabetes share several features of MODY and type 2 diabetes. Hence, I aimed to understand the range of genetic variation in HNF1A in this group and whether we could use biomarkers to help identify those with HNF1A alleles likely to be influencing their diabetes status. I performed HNF1A sequencing in 1058 subjects with non-autoimmune diabetes diagnosed below 45 years of age and identified 26 rare non-synonymous HNF1A alleles in 30 probands and 8 relatives. A systematic assessment assigned a likely damaging functional effect to 50% of these HNF1A alleles. The assessment of HNF1A alleles and their experimental characterisation unveiled the complexity of the genotype-phenotype relationship and continuity of the functional spectrum of HNF1A alleles from benign through likely benign, likely damaging to damaging effect. To clarify the performance of C-reactive protein (CRP) and N-glycan profile as biomarkers of HNF1A-MODY, I evaluated them in the above unselected sample of individuals who also underwent HNF1A sequencing. I have shown that 3-antennary fucosylated N-glycans and CRP were significantly lower in subjects with likely damaging HNF1A alleles in this group when compared to participants without HNF1A variants. Both biomarkers distinguished individuals with likely damaging HNF1A alleles from those without HNF1A variants with a C-statistic of 0.82-0.92. Plasma levels of these biomarkers negatively correlated with the predicted pathogenicity of HNF1A alleles. The expression of several steroidogenesis enzymes had been reported to be reduced in the liver cells of Hnf1a knock-out mice. Hence, I assessed the urinary steroid profile in individuals with HNF1A-MODY and age- and BMI-matched non-diabetic subjects and those with type 2 diabetes. The activity of 11β-hydroxysteroid dehydrogenase type 2 was reduced and the activity of 5β-steroid reductase increased in individuals with HNF1A-MODY. Their performance in selecting subjects with HNF1A-MODY was inferior to that of CRP and 3-antenary fucosylated glycans. In summary, my thesis has shown the complexity of the interpretation of genetic variation which is crucial for an accurate diagnosis and the assessment of potential biomarkers for genetic disease. Careful evaluation of the functional effect of HNF1A alleles identified on sequencing informs diagnosis in individuals with early-onset diabetes. Furthermore, the combination of functional assessment and clinical phenotype can guide management in those without canonical monogenic phenotypes found to have variant HNF1A alleles.