Background A genetic overlap between type 2 diabetes and depression has been reported in twin studies, but the finding has not been replicated with data from genome-wide association studies. Visceral adiposity has been postulated as being on the causal pathway of the association between type 2 diabetes and depression. Since waist-to-hip ratio can be a proxy measure of intra-abdominal fat deposition, we examined its effect on the association using the polygenic scores approach in the UK Biobank. Methods Type 2 diabetes polygenic scores were constructed from the association summary statistics of the Diabetes Genetic Replication And Meta-analysis Consortium, and depression polygenic scores from the Psychiatric Genetics Consortium Major Depressive Disorders Workgroup (29 studies at seven association p-value thresholds [p=0·001 to p=0·5). Logistic regression examined the association between type 2 diabetes polygenic scores and depression case-control status and the effect of body-mass index (BMI)-adjusted waist-to-hip ratio on the association, adjusting for ancestry, centres, and genotyping batches. Findings The UK Biobank sample with genotyping data consisted of 152 551 participants. There were 10 005 cases and 19 314 controls for depression among individuals of European ancestry, with a mean age of 57·1 years (SD 7·8), BMI of 27·5 kg/m2 (4·7), and waist-to-hip ratio of 0·88 (0·09). Type 2 diabetes polygenic scores were not predictive of depression case-status at all p-value thresholds examined. The interaction between waist-to-hip ratio and type 2 diabetes polygenic scores had an effect on depression case-status (at p-value threshholds <0·2, β=0·37, p=0·02). Interpretation Our exploratory study tentatively suggests that waist-to-hip ratio might have a role in the effect of type 2 diabetes polygenic scores on depression case-status. Higher adiposity is associated with greater level of inflammation, which is in turn associated with increased risk of type 2 diabetes and depression. Further research is needed to determine the direction of causation and to replicate our finding, given the cross-sectional design and the proxy use of waist-to-hip ratio for visceral adiposity. Funding Novo Nordisk UK Research Foundation.