Diet, nutrition and our lifestyles are known to affect both the risk of type 1 and type 2 diabetes but evidence for causality of association remains uncertain for many of the proposed factors. Studies using genetic markers as proxy indicators for modifiable exposures (“Mendelian randomisation”) are increasingly used to test for causality between proposed risk factors and disease risk. This talk examines whether common nutritional factors characteristic to our modern lifestyles appear to be truly diabetogenic, and reviews genetic epidemiological evidence relating to the causal role of obesity, vitamin D, coffee and alcohol in diabetes.