Causal effects of inflammatory protein biomarkers on inflammatory diseases

Many disease-associated protein biomarkers protect against disease development.


The PDF file includes:
Figs. S1 to S14 Table S1 Other Supplementary Material for this manuscript includes the following: Table S2 Fig. S1. Causal effect of IL-12B on psoriasis. Genetic instruments are shown with their effects on the biomarker plotted against their effects on disease. Black dots (with corresponding 95% CI) denote SNPs that GSMR identifies as being valid, while red dots (with corresponding 95% CI) are identified as statistical outliers by the HEIDI-outlier procedure. Their effect on disease is likely modified by horizontal pleiotropy and they are therefore removed prior to causal estimation. The black solid line denotes the estimate by GSMR which is adopted as the main result (corresponding raw P value is shown in figure legend), while the cyan line denotes the estimate by the inverse variance weighted method, the magenta line denotes the estimate by the weighted median method, and the green line denotes the estimate by the MR-Egger method. Note that in the last three, alternative methods, all genetic variants are used in the causal estimation. Finally, the black dashed line denotes the estimate by GSMR with the HEIDI-outlier removal procedure switched off. Genetic instruments are shown with their effects on the biomarker plotted against their effects on disease. Black dots (with corresponding 95% CI) denote SNPs that GSMR identifies as being valid, while red dots (with corresponding 95% CI) are identified as statistical outliers by the HEIDI-outlier procedure. Their effect on disease is likely modified by horizontal pleiotropy and they are therefore removed prior to causal estimation. The black solid line denotes the estimate by GSMR which is adopted as the main result (corresponding raw P value is shown in figure legend), while the cyan line denotes the estimate by the inverse variance weighted method, the magenta line denotes the estimate by the weighted median method, and the green line denotes the estimate by the MR-Egger method. Note that in the last three, alternative methods, all genetic variants are used in the causal estimation. Finally, the black dashed line denotes the estimate by GSMR with the HEIDI-outlier removal procedure switched off. Genetic instruments are shown with their effects on the biomarker plotted against their effects on disease. Black dots (with corresponding 95% CI) denote SNPs that GSMR identifies as being valid, while red dots (with corresponding 95% CI) are identified as statistical outliers by the HEIDI-outlier procedure. Their effect on disease is likely modified by horizontal pleiotropy and they are therefore removed prior to causal estimation. The black solid line denotes the estimate by GSMR which is adopted as the main result (corresponding raw P value is shown in figure legend), while the cyan line denotes the estimate by the inverse variance weighted method, the magenta line denotes the estimate by the weighted median method, and the green line denotes the estimate by the MR-Egger method. Note that in the last three, alternative methods, all genetic variants are used in the causal estimation. Finally, the black dashed line denotes the estimate by GSMR with the HEIDI-outlier removal procedure switched off. Genetic instruments are shown with their effects on the biomarker plotted against their effects on disease. Black dots (with corresponding 95% CI) denote SNPs that GSMR identifies as being valid, while red dots (with corresponding 95% CI) are identified as statistical outliers by the HEIDI-outlier procedure. Their effect on disease is likely modified by horizontal pleiotropy and they are therefore removed prior to causal estimation. The black solid line denotes the estimate by GSMR which is adopted as the main result (corresponding raw P value is shown in figure legend), while the cyan line denotes the estimate by the inverse variance weighted method, the magenta line denotes the estimate by the weighted median method, and the green line denotes the estimate by the MR-Egger method. Note that in the last three, alternative methods, all genetic variants are used in the causal estimation. Finally, the black dashed line denotes the estimate by GSMR with the HEIDI-outlier removal procedure switched off. Genetic instruments are shown with their effects on the biomarker plotted against their effects on disease. Only four SNPs were available for the MR, wherefore the HEIDI-outlier procedure in GSMR was switched off and all four instruments were assumed to be valid (black dots with corresponding 95% CI). A simple visual inspection of the data also supports this assumption. The black dashed line denotes the estimate by GSMR, with HEIDI switched off, which is adopted as the main result (corresponding raw P value is shown in figure legend), while the cyan line denotes the estimate by the inverse variance weighted method, the magenta line denotes the estimate by the weighted median method, and the green line denotes the estimate by the MR-Egger method. Note that in the last three, alternative methods, all genetic variants are used in the causal estimation. Genetic instruments are shown with their effects on the biomarker plotted against their effects on disease. Black dots (with corresponding 95% CI) denote SNPs that GSMR identifies as being valid. No statistical outliers were identified by the HEIDI-outlier procedure. The black solid line denotes the estimate by GSMR which is adopted as the main result (corresponding raw P value is shown in figure legend), while the cyan line denotes the estimate by the inverse variance weighted method, the magenta line denotes the estimate by the weighted median method, and the green line denotes the estimate by the MR-Egger method. Note that in the last three, alternative methods, all genetic variants are used in the causal estimation. Finally, the black dashed line denotes the estimate by GSMR with the HEIDI-outlier removal procedure switched off. Note that this line coincides with the black solid line, as no outliers were detected. Genetic instruments are shown with their effects on the biomarker plotted against their effects on disease. Black dots (with corresponding 95% CI) denote SNPs that GSMR identifies as being valid. No statistical outliers were identified by the HEIDI-outlier procedure. The black solid line denotes the estimate by GSMR which is adopted as the main result (corresponding raw P value is shown in figure legend), while the cyan line denotes the estimate by the inverse variance weighted method, the magenta line denotes the estimate by the weighted median method, and the green line denotes the estimate by the MR-Egger method. Note that in the last three, alternative methods, all genetic variants are used in the causal estimation. Finally, the black dashed line denotes the estimate by GSMR with the HEIDI-outlier removal procedure switched off. Note that this line coincides with the black solid line, as no outliers were detected. Genetic instruments are shown with their effects on the biomarker plotted against their effects on disease. Black dots (with corresponding 95% CI) denote SNPs that GSMR identifies as being valid, while red dots (with corresponding 95% CI) are identified as statistical outliers by the HEIDI-outlier procedure. Their effect on disease is likely modified by horizontal pleiotropy and they are therefore removed prior to causal estimation. The black solid line denotes the estimate by GSMR which is adopted as the main result (corresponding raw P value is shown in figure legend), while the cyan line denotes the estimate by the inverse variance weighted method, the magenta line denotes the estimate by the weighted median method, and the green line denotes the estimate by the MR-Egger method. Note that in the last three, alternative methods, all genetic variants are used in the causal estimation. Finally, the black dashed line denotes the estimate by GSMR with the HEIDI-outlier removal procedure switched off.
Genetic effect on biomarker Genetic effect on disease

Fig. S9. Causal effect of IL-18R1 on allergy.
Genetic instruments are shown with their effects on the biomarker plotted against their effects on disease. Black dots (with corresponding 95% CI) denote SNPs that GSMR identifies as being valid, while red dots (with corresponding 95% CI) are identified as statistical outliers by the HEIDI-outlier procedure. Their effect on disease is likely modified by horizontal pleiotropy and they are therefore removed prior to causal estimation. The black solid line denotes the estimate by GSMR which is adopted as the main result (corresponding raw P value is shown in figure legend), while the cyan line denotes the estimate by the inverse variance weighted method, the magenta line denotes the estimate by the weighted median method, and the green line denotes the estimate by the MR-Egger method. Note that in the last three, alternative methods, all genetic variants are used in the causal estimation. Finally, the black dashed line denotes the estimate by GSMR with the HEIDI-outlier removal procedure switched off.
Genetic effect on biomarker Genetic effect on disease  Black dots (with corresponding 95% CI) denote SNPs that GSMR identifies as being valid, while red dots (with corresponding 95% CI) are identified as statistical outliers by the HEIDI-outlier procedure. Their effect on disease is likely modified by horizontal pleiotropy and they are therefore removed prior to causal estimation. The black solid line denotes the estimate by GSMR which is adopted as the main result (corresponding raw P value is shown in figure legend), while the cyan line denotes the estimate by the inverse variance weighted method, the magenta line denotes the estimate by the weighted median method, and the green line denotes the estimate by the MR-Egger method. Note that in the last three, alternative methods, all genetic variants are used in the causal estimation. Finally, the black dashed line denotes the estimate by GSMR with the HEIDI-outlier removal procedure switched off. Half-colored squares denote estimates with P<0.05, similar to full-colored squares, but, e.g., due to the removal of outlier SNPs, the results are based on less than 4 instruments, which makes the estimates more uncertain. If only four SNPs met the stricter parameter settings, results are indicated by squares with gray upper-left corners. If less than four SNPs met the stricter thresholds, results are indicated by half-grayed squares. In both these situations (i.e.,≤ 4 SNPs), the HEIDI-outlier removal procedure was switched off.
A D A C C L 2 3 C C L 2 5 C D 4 0 C D 6 C D C P 1 C S T 5 C X C L 5 C X C L 6 C X C L 1 0   S13. Causal effect of each biomarker on each inflammatory disease with stricter parameter settings P=5´10 -8 and R 2 =0.6. Results from the GSMR analysis with stricter parameters settings. The P value threshold is set to P=5´10 -8 while the LD threshold is set to R 2 =0.6. If only one SNP was left after filtering, GSMR was unable to perform a causal estimation, which is indicated by "NA". An asterisk (*) indicates significant effect (FDR<0.05), blue shades denote a protective effect with various significance (raw P Half-colored squares denote estimates with P<0.05, similar to full-colored squares, but, e.g., due to the removal of outlier SNPs, the results are based on less than 4 instruments, which makes the estimates more uncertain. If only four SNPs met the stricter parameter settings, results are indicated by squares with gray upper-left corners. If less than four SNPs met the stricter thresholds, results are indicated by half-grayed squares. In both these situations (i.e., ≤4 SNPs), the HEIDI-outlier removal procedure was switched off.
A D A C C L 2 3 C C L 2 5 C D 4 0 C D 6 C D C P 1 C S T 5 C X C L 5 C X C L 6 C X C L 1 0  Half-colored squares denote estimates with P<0.05, similar to full-colored squares, but, e.g., due to the removal of outlier SNPs, the results are based on less than 4 instruments, which makes the estimates more uncertain. If only four SNPs met the stricter parameter settings, results are indicated by squares with gray upper-left corners. If less than four SNPs met the stricter thresholds, results are indicated by half-grayed squares. In both these situations (i.e., ≤4 SNPs), the HEIDI-outlier removal procedure was switched off.
A D A C C L 2 3 C C L 2 5 C D 4 0 C D 6 C D C P 1 C S T 5 C X C L 5 C X C L 6 C X C L 1 0