Table 3. Results from the GSMR analysis with more stringent thresholds for the SNP correlations (R2) and SNP significance levels (P).
Causal effect estimates with FDR < 0.05 in any of the sensitivity analyses with stricter parameter settings are included.
| Thresholds | R2 = 0.6; P = 1 × 10−6 (main analysis) | R2 = 0.2; P = 1 × 10−6 | R2 = 0.6; P = 5 × 10−8 | R2 = 0.2; P = 5 × 10−8 | |||||||||||||
| Biomarker | Disease |
No. of
IVs* |
OR
(95% CI) |
P | FDR† |
No. of
IVs* |
OR
(95% CI) |
P | FDR† |
No. of
IVs* |
OR
(95% CI) |
P | FDR† |
No. of
IVs* |
OR
(95% CI) |
P | FDR† |
| CD40 | Psoriasis | 22 | 0.95 (0.92–0.99) |
6.6 × 10−3 |
1.8 × 10−1 |
11 | 0.98 (0.93–1.02) |
3.1 × 10−1 |
8.6 × 10−1 |
15 | 0.92 (0.88–0.96) |
2.3 × 10−4 |
1.9 × 10−2 |
9 | 0.96 (0.92–1.01) |
1.1 × 10−1 |
8.4 × 10−1 |
| IL-12B | Psoriatic arthropathy |
19 | 0.79 (0.72–0.87) |
4.5 × 10−6 |
3.4 × 10−4 |
15 | 0.82 (0.74–0.91) |
2.3 × 10−4 |
1.4 × 10−2 |
10 | 0.73 (0.64–0.83) |
1.4 × 10−6 |
2.2 × 10−4 |
8 | 0.63 (0.54–0.73) |
5.2 × 10−10 |
1.7 × 10−7 |
| Psoriasis | 16 | 0.84 (0.80–0.88) |
2.7 × 10−13 |
1.0 × 10−10 |
13 | 0.89 (0.84–0.93) |
5.1 × 10−7 |
9.6 × 10−5 |
8 | 0.82 (0.77–0.87) |
1.1 × 10−11 |
3.6 × 10−9 |
6 | 0.86 (0.81–0.92) |
1.0 × 10−5 |
1.6 × 10−3 |
|
| IL-18R1 | Asthma | 78 | 1.01 (1.00–1.01) |
1.6 × 10−1 |
7.5 × 10−1 |
47 | 1.02 (1.01–1.02) |
5.0 × 10−4 |
2.0 × 10−2 |
48 | 1.01 (1.00–1.02) |
4.0 × 10−2 |
5.7 × 10−1 |
35 | 1.02 (1.01–1.03) |
1.1 × 10−4 |
7.1 × 10−3 |
| Eczema | 93 | 1.02 (1.01–1.03) |
7.0 × 10−5 |
4.4 × 10−3 |
49 | 1.03 (1.01–1.05) |
1.9 × 10−4 |
1.4 × 10−2 |
63 | 1.03 (1.01–1.04) |
2.6 × 10−5 |
2.8 × 10−3 |
38 | 1.04 (1.02–1.06) |
1.5 × 10−5 |
1.6 × 10−3 |
|
| Hay fever | 74 | 1.02 (1.01–1.03) |
5.3 × 10−4 |
2.8 × 10−2 |
43 | 1.03 (1.01–1.04) |
3.1 × 10−5 |
3.0 × 10−3 |
49 | 1.01 (1.00–1.02) |
4.2 × 10−2 |
5.7 × 10−1 |
31 | 1.03 (1.02–1.05) |
2.4 × 10−5 |
1.9 × 10−3 |
|
| Allergy | 81 | 1.02 (1.01–1.02) |
1.4 × 10−3 |
4.8 × 10−2 |
45 | 1.02 (1.01–1.03) |
5.2 × 10−4 |
2.0 × 10−2 |
52 | 1.02 (1.01–1.03) |
7.7 × 10−4 |
4.0 × 10−2 |
33 | 1.02 (1.01–1.04) |
2.7 × 10−4 |
1.5 × 10−2 |
|
| Celiac disease | 92 | 1.05 (1.02–1.08) |
3.2 × 10−3 |
1.0 × 10−1 |
49 | 1.05 (1.01–1.10) |
2.3 × 10−2 |
4.1 × 10−1 |
63 | 1.06 (1.02–1.09) |
6.4 × 10−4 |
4.0 × 10−2 |
38 | 1.05 (1.00–1.10) |
6.6 × 10−2 |
8.2 × 10−1 |
|
| LAP-TGF-β-1 | Osteoarthritis | 5 | 0.94 (0.91–0.98) |
9.0 × 10−4 |
3.4 × 10−2 |
3‡ | 0.94 (0.90–0.97) |
3.6 × 10−4 |
1.9 × 10−2 |
1‡ | NA§ | NA§ | NA§ | 1‡ | NA§ | NA§ | NA§ |
| LT-α║ | Type 1 diabetes |
4 | 0.60 (0.51–0.70) |
2.4 × 10−10 |
4.5 × 10−8 |
4‡ | 0.94 (0.87–1.02) |
1.5 × 10−1 |
7.3 × 10−1 |
3‡ | 0.99 (0.90–1.08) |
7.9 × 10−1 |
9.6 × 10−1 |
2‡ | 0.93 (0.85–1.03) |
1.5 × 10−1 |
8.4 × 10−1 |
| Rheumatoid arthritis |
5 | 0.84 (0.78–0.90) |
4.1 × 10−6 |
3.4 × 10−4 |
4‡ | 0.97 (0.92–1.01) |
1.8 × 10−1 |
7.4 × 10−1 |
3‡ | 1.00 (0.95–1.05) |
9.9 × 10−1 |
1.0 × 10−0 |
2‡ | 0.98 (0.93–1.03) |
3.8 × 10−1 |
9.1 × 10−1 |
|
| Psoriasis | 2 | 1.06 (1.01–1.11) |
1.5 × 10−2 |
2.9 × 10−1 |
4‡ | 0.83 (0.77-0.89) |
3.3 × 10−7 |
9.6 × 10−5 |
3‡ | 1.00 (0.96–1.05) |
9.2 × 10−1 |
1.0 × 10−0 |
2‡ | 1.06 (1.01–1.11) |
1.5 × 10−2 |
4.0 × 10−1 |
|
| Celiac disease | 2 | 1.55 (1.34–1.79) |
5.7 × 10−9 |
7.1 × 10−7 |
4‡ | 1.33 (1.18–1.49) |
2.7 × 10−6 |
3.4 × 10−4 |
3‡ | 0.98 (0.87–1.10) |
7.3 × 10−1 |
9.3 × 10−1 |
2‡ | 1.15 (1.01–1.31) |
3.0 × 10−2 |
5.6 × 10−1 |
|
| TWEAK | Asthma | 4‡ | 0.92 (0.88–0.97) |
6.0 × 10−4 |
2.8 × 10−2 |
2‡ | 0.92 (0.88–0.96) |
4.9 × 10−4 |
2.0 × 10−2 |
1‡ | NA§ | NA§ | NA§ | 1‡ | NA§ | NA§ | NA§ |
| VEGF-A | Ulcerative colitis |
5 | 0.86 (0.79–0.94) |
8.7 × 10−4 |
3.4 × 10−2 |
2‡ | 0.85 (0.78–0.93) |
7.5 × 10−4 |
2.6 × 10−2 |
5 | 0.86 (0.79–0.94) |
8.7 × 10−4 |
4.0 × 10−2 |
2‡ | 0.85 (0.78–0.93) |
7.5 × 10−4 |
3.5 × 10−2 |
*The number of IVs that remained after HEIDI-outliers had been removed.
†P value adjusted for multiple testing (FDR). Note that the region of raw P values corresponding to FDR < 0.05 differs slightly between the four different analyses.
‡After nonsignificant and/or correlated SNPs were removed, the number of remaining SNPs was too few (<5) for the HEIDI-outlier procedure to work properly, wherefore HEIDI was switched off in these analyses.
§After nonsignificant and/or correlated SNPs were removed, only a single SNP remained, wherefore GSMR could not estimate a causal effect.
║As highlighted in the main analyses, some of the instruments for LT-α might be invalid. In these sensitivity analyses, the HEIDI-outlier procedure had to be switched off due to the low number of instruments, and the potentially invalid instruments are therefore included in the causal estimation, in particular, rs2229092.