Gut Microbiota and Psychiatric Disorders: A Two-Sample Mendelian Randomization Study

Evidence supports the observational associations of gut microbiota with a variety of psychiatric disorders, but the causal nature of such associations remains obscure. Aiming to comprehensively investigate their causal relationship and to identify specific causal microbe taxa for psychiatric diseases, we conducted a two-sample Mendelian randomization (MR) analysis of gut microbiome with 15 psychiatric diseases. Specifically, the microbiome genome-wide association study (GWAS) in 18,473 individuals from the MiBioGen study was used as exposure sample, and the GWAS for 15 psychiatric diseases was used as outcome samples. One-hundred ninety bacterial taxa from six levels were available for analysis. At a multiple-testing corrected significance level (phylum P < 5.56 x 10(-3), class P < 3.33 x 10(-3), order P < 2.63 x 10(-3), family P < 1.67 x 10(-3), genus P < 4.90 x 10(-4), and species P < 3.33 x 10(-3)), the following eight causal associations from seven bacterial features (one phylum + three classes + one order + one family + one species) were identified: family Prevotellaceae with autism spectrum disorder (P = 5.31 x 10(-4)), class Betaproteobacteria with bipolar disorder (P = 1.53 x 10(-3)), class Actinobacteria with schizophrenia (P = 1.33 x 10(-3)), class Bacteroidia and order Bacteroidales with Tourette syndrome (P = 2.51 x 10(-3) and 2.51 x 10(-3)), phylum Actinobacteria and class Actinobacteria with extroversion (P = 8.22 x 10(-4) and 1.09 x 10(-3)), and species Clostridium innocuum with neuroticism (P = 8.92 x 10(-4)). Sensitivity analysis showed no evidence of reverse causality, pleiotropy, and heterogeneity. Our findings offered novel insights into the gut microbiota-mediated development mechanism of psychiatric disorders.