MR analyses of causal effects of 4 gut microbial genera on DLBCL
| Exposure . | Method . | SNP . | β (95% CI) . | P value . | OR (95% CI) . | Power . |
|---|---|---|---|---|---|---|
| Genus Alistipes | IVW | 12 | −0.566 (−1.115 to −0.017) | .043 | 0.568 (0.328-0.982) | 0.4 |
| Maximum likelihood | 12 | −0.570 (−1.128 to −0.012) | .045 | 0.565 (0.324-0.988) | ||
| MR-PRESSO | 12 | −0.441 (−0.878 to −0.004) | .071 | 0.643 (0.415-0.996) | ||
| WM | 12 | −0.545 (−1.247 to 0.157) | .128 | 0.58 (0.287-1.170) | ||
| MR-Egger | 12 | 1.235 (−1.376 to 3.846) | .376 | 3.439 (0.253-46.812) | ||
| Genus Coprobacter | IVW | 11 | 0.341 (0.008-0.674) | .045 | 1.407 (1.008-1.963) | 0.82 |
| Maximum likelihood | 11 | 0.355 (0.018-0.692) | .039 | 1.426 (1.018-1.997) | ||
| MR-PRESSO | 11 | 0.346 (0.062-0.63) | .033 | 1.426 (1.063-1.878) | ||
| WM | 11 | 0.388 (−0.061 to 0.837) | .091 | 1.474 (0.941-2.309) | ||
| MR-Egger | 11 | 0.62 (−0.734 to 1.974) | .393 | 1.859 (0.48-7.205) | ||
| Genus Ruminococcaceae UCG-002 | IVW | 20 | 0.356 (0.011-0.701) | .043 | 1.427 (1.011-2.015) | 0.75 |
| Maximum likelihood | 20 | 0.367 (0.016-0.718) | .040 | 1.443 (1.016-2.050) | ||
| MR-PRESSO | 20 | 0.298 (0.078-0.518) | .014 | 1.443 (1.083-1.677) | ||
| WM | 20 | 0.358 (−0.132 to 0.848) | .152 | 1.43 (0.876-2.335) | ||
| MR-Egger | 20 | 0.503 (−0.391 to 1.397) | .285 | 1.653 (0.676-4.044) | ||
| Genus Turicibacter | IVW | 9 | −0.504 (−0.97 to −0.038) | .034 | 0.604 (0.379-0.964) | 0.66 |
| Maximum likelihood | 9 | −0.532 (−0.964 to −0.100) | .016 | 0.587 (0.381-0.905) | ||
| MR-PRESSO | 9 | −0.226 (−0.692 to 0.24) | .36 | 0.587 (0.501-1.271) | ||
| WM | 9 | −0.516 (−1.088 to 0.056) | .078 | 0.597 (0.337-1.059) | ||
| MR-Egger | 9 | −0.359 (−2.38 to 1.662) | .738 | 0.698 (0.093-5.264) |
| Exposure . | Method . | SNP . | β (95% CI) . | P value . | OR (95% CI) . | Power . |
|---|---|---|---|---|---|---|
| Genus Alistipes | IVW | 12 | −0.566 (−1.115 to −0.017) | .043 | 0.568 (0.328-0.982) | 0.4 |
| Maximum likelihood | 12 | −0.570 (−1.128 to −0.012) | .045 | 0.565 (0.324-0.988) | ||
| MR-PRESSO | 12 | −0.441 (−0.878 to −0.004) | .071 | 0.643 (0.415-0.996) | ||
| WM | 12 | −0.545 (−1.247 to 0.157) | .128 | 0.58 (0.287-1.170) | ||
| MR-Egger | 12 | 1.235 (−1.376 to 3.846) | .376 | 3.439 (0.253-46.812) | ||
| Genus Coprobacter | IVW | 11 | 0.341 (0.008-0.674) | .045 | 1.407 (1.008-1.963) | 0.82 |
| Maximum likelihood | 11 | 0.355 (0.018-0.692) | .039 | 1.426 (1.018-1.997) | ||
| MR-PRESSO | 11 | 0.346 (0.062-0.63) | .033 | 1.426 (1.063-1.878) | ||
| WM | 11 | 0.388 (−0.061 to 0.837) | .091 | 1.474 (0.941-2.309) | ||
| MR-Egger | 11 | 0.62 (−0.734 to 1.974) | .393 | 1.859 (0.48-7.205) | ||
| Genus Ruminococcaceae UCG-002 | IVW | 20 | 0.356 (0.011-0.701) | .043 | 1.427 (1.011-2.015) | 0.75 |
| Maximum likelihood | 20 | 0.367 (0.016-0.718) | .040 | 1.443 (1.016-2.050) | ||
| MR-PRESSO | 20 | 0.298 (0.078-0.518) | .014 | 1.443 (1.083-1.677) | ||
| WM | 20 | 0.358 (−0.132 to 0.848) | .152 | 1.43 (0.876-2.335) | ||
| MR-Egger | 20 | 0.503 (−0.391 to 1.397) | .285 | 1.653 (0.676-4.044) | ||
| Genus Turicibacter | IVW | 9 | −0.504 (−0.97 to −0.038) | .034 | 0.604 (0.379-0.964) | 0.66 |
| Maximum likelihood | 9 | −0.532 (−0.964 to −0.100) | .016 | 0.587 (0.381-0.905) | ||
| MR-PRESSO | 9 | −0.226 (−0.692 to 0.24) | .36 | 0.587 (0.501-1.271) | ||
| WM | 9 | −0.516 (−1.088 to 0.056) | .078 | 0.597 (0.337-1.059) | ||
| MR-Egger | 9 | −0.359 (−2.38 to 1.662) | .738 | 0.698 (0.093-5.264) |
IVW was used as the primary method, synthesizing SNP effects by prioritizing those with higher precision under the assumption of no horizontal pleiotropy. Complementing this, the maximum likelihood method was used to optimize model fit with normally distributed effects. MR-PRESSO was applied to detect and correct for pleiotropic biases from SNPs. The WM provided robust estimates against invalid instrumental variables, focusing on the median of their weighted effects. Lastly, MR-Egger was used to identify directional pleiotropy, essential for unbiased estimation under the Instrument Strength Independent of Direct Effect (InSIDE) assumption.
MR-PRESSO, Mendelian randomization pleiotropy residual sum and outlier; WM, weighted median.