Figure 2. Scatter plot of genetic associations with atrial fibrillation against associations with inflammatory bowel disease, with causal estimates (β coefficients) of inflammatory bowel disease on atrial fibrillation estimated by inverse-variance weighted (dashed line), and MR-Egger (solid line) methods. The straight lines should be the change in the log odds of atrial fibrillation per unit increase of the log odds of inflammatory bowel disease.