[Mendelian randomization of the causal relationship between ω-3 polyunsaturated fatty acids and major depression]

A two-sample Mendelian randomization method was used to explore the causal relationship between ω-3 polyunsaturated fatty acids and the risk of major depressive disorder.

Using a large-scale genome-wide association study database, a series of quality control steps were taken to select qualified SNPs that were strongly associated with exposure. To make the conclusion more robust and reliable, inverse variance weighting, MR-Egger regression, weighted median, robustly adjusted profile scores, and pleiotropic residual and outlier tests were used based on different assumptions of two-sample Mendelian randomization five method for causal estimation. The causal relationship between ω-3 PUFA and the risk of MDD was studied with OR value and 95%CI as evaluation indicators. To assess horizontal pleiotropy, heterogeneity, and stability, MR-Egger intercept tests, Cochran' Q and I~2 statistics, and "leave-one-out" sensitivity analyzes were performed. In addition, other datasets are used to check the result of the original dataset.

A total of 66 SNP sites were included in the original data set as instrumental variables. The result of MR-PRESSO showed that there was no causal association between ω-3 PUFA and RA(OR=0.973, 95%CI 0.920-1.028, P=0.332). And the random effect IVW(OR=0.973, 95% CI 0.920-1.028, P=0.329), MR-Egger(OR=0.995, 95% CI 0.899-1.100, P=0.920), weighted median(OR=0.988, 95% CI 0.918-1.063, P=0.248) and MR-RAPS(OR=0.972, 95% CI 0.927-1.020, P=0.247) the causal estimation result of the four method tended to be consistent with the MR-PRESSO method, the result were robust. The validation set was consistent with the original data set in the estimation of the causal effect of ω-3 PUFA and MDD, and there was no causal relationship between the two. MR-Egger intercept test, original data set(intercept=-0.002, P=0.599) and validation set(intercept=0.012, P=0.628) indicated that the causal analysis result were less likely to be influenced by gene-level pleiotropy big.

The result of two-sample Mendelian randomization analysis showed that there was no causal relationship between ω-3 PUFA and the risk of MDD.