Pervasive Sharing of Causal Genetic Risk Factors Contributes to Clinical and Molecular Overlap between Sjögren’s Disease and Systemic Lupus Erythematosus

Figures

• 

  Figure 1. Sjögren’s Disease predicted causal genes for a distinct biological network. The predicted protein–protein interaction network from the STRING database of Sjögren’s Disease GWAS-identified putative causal genes is shown. (https://www.ndexbio.org/viewer/networks/99b77fa6-478b-11ee-aa50-005056ae23aa, accessed on 3 June 2023).

• 

  Figure 2. The Sjögren’s Disease putative causal genes network exhibits substantial overlap with that of Systemic Lupus Erythematosus. The predicted protein–protein interaction network from the STRING database of Sjögren’s disease GWAS-identified putative causal genes (teal rounded rectangles) is shown together with those of the polygenic human SLE (dark blue hexagons) and monogenic human SLE (light blue diamonds). Overlaps of SjD with either of the SLE networks are indicated with orange triangles and take priority over overlaps between monogenic and polygenic SLE (yellow circles).

• 

  Figure 3. MR analysis reveals a robust and stable causal relationship between Sjögren’s Disease and Systemic Lupus Erythematosus in a GWAS of persons with European ancestry. (A) Forest plot of SjD effect estimate on SLE in the Bentham J. et al., GWAS [40] for individual risk variants. MR Egger and inverse-variance weighted overall estimates are also shown. Data are MR effect sizes (beta ± se). Black dots and black lines indicate MR effect estimates and confidence intervals for individual variants. Red dots and red lines indicate aggregate MR effect estimates and confidence intervals as indicated for MR Egger and Inverse variance weighted analyses. (B) scatter plot of MR effect sizes (beta) ± standard error (se).

• 

  Figure 4. MR analysis reveals a robust and stable causal relationship between Systemic Lupus Erythematosus and Sjögren’s Disease in a GWAS of persons with European ancestry. (A) Forest plot of effect estimate of SLE on SjD (M13_Sjogrens) in the FinnGen GWAS [44] for individual risk variants. MR Egger and inverse-variance weighted overall estimates are also shown. Data shown are MR effect sizes (beta ± se). Black dots and black lines indicate MR effect estimates and confidence intervals for individual variants. Red dots and red lines indicate aggregate MR effect estimates and confidence intervals as indicated for MR Egger and Inverse variance weighted analyses. (B) scatter plot of MR effect sizes (beta) ± standard error (se).

• 

  Figure 5. MR analysis reveals a causal relationship or pleiotropy between Sjögren’s Disease and Systemic Lupus Erythematosus in a GWAS of persons with East Asian ancestry. (A) Forest plot of SjD on SLE in the Wang, Y.F. et al., GWAS [47] for individual risk variants. MR Egger and Inverse-variance weighted overall estimates are also shown. Data shown are MR effect sizes (beta ± se). Black dots and black lines indicate MR effect estimates and confidence intervals for individual variants. Red dots and red lines indicate aggregate MR effect estimates and confidence intervals as indicated for MR Egger and Inverse variance weighted analyses. (B) scatter plot of MR effect sizes (beta) ± standard error (se).

• 

  Figure 6. MR analysis reveals a causal relationship between Systemic Lupus Erythematosus in persons of East Asian ancestry and Sjögren’s Disease in GWAS of persons with European ancestry. (A) Forest plot of SLE from the Wang, Y.F. et al. GWAS [47] on SjD in FinnGen [44] for individual risk variants. MR Egger and Inverse-variance weighted overall estimates are also shown. Data shown are MR effect sizes (beta ± se). Black dots and black lines indicate MR effect estimates and confidence intervals for individual variants. Red dots and red lines indicate aggregate MR effect estimates and confidence intervals as indicated for MR Egger and Inverse variance weighted analyses. (B) scatter plot of MR effect sizes (beta) ± standard error (se) across the two conditions/populations.

• 

  Figure 7. Analysis scheme. [EUR] indicates European ancestry GWAS. [EAS] indicates East Asian ancestry GWAS. M13_Sjogren’s indicates the phenotype code from FinnGen used to define Sjögren’s for that GWAS. [EUR–QC] indicates genome-wide significant variants after removal of variants likely violating the relevance condition as described in the methods. Created with BioRender.com (accessed on 6 June 2023).

• 

  Figure 8. Mendelian randomization causal graph and conditions. In our study, the most likely scenario is that the chosen variant, Gk, is in linkage disequilibrium with a causal variant, Gj. In the event that this LD is perfect (r2 = 1) and the conditions of MR are met, then the association between Gj and the outcome is fully mediated by the exposure. Therefore, the ratio of marginal effects of Gj on X and Gj on Y is equal to βXY. Effects of the variant Gj through any confounder, C, or on Y are both ruled out through the Independence and Exclusion Restriction conditions. The relevance condition is that Gj (through Gk in this case) is truly associated with the effect, X. In situations of comparison across populations with different fine-scale LD patterns, the relevance assumption is likely to not be met. Created with BioRender.com (accessed on 6 June 2023).