Causal Inference in Microbiome Medicine: Principles and Applications

The causality between microbiota and diseases can be established by computational approaches. Mendelian randomization analysis is efficient in revealing the direct microbiota–disease causality, which can be evaluated by the interventionist framework. Mediation analysis and the structural equation model could reveal both direct and indirect causalities between microbiota and diseases by analyzing the mediation effect of various factors. The confirmed microbiota–disease causality and the microbiota-associated genetic variations could facilitate the implementation of personalized dietary intervention. Microorganisms that colonize the mammalian skin and cavity play critical roles in various physiological functions of the host. Numerous studies have revealed strong associations between the microbiota and multiple diseases. However, association does not mean causation. To clarify the mechanisms underlying microbiota-mediated diseases, research is moving from associative analyses to causation studies. In this article, we first introduce the principles of the computational methods for causal inference, and then discuss the applications of these methods in microbiome medicine. Furthermore, we examine the reliability of theoretically inferred causality by the interventionist framework. Finally, we show the potential of confirmed causality in microbiota-targeted therapy, especially in personalized dietary intervention. We conclude that a comprehensive understanding of the causal relationships between diets, microbiota, host targets, and diseases is critical to future microbiome medicine. Microorganisms that colonize the mammalian skin and cavity play critical roles in various physiological functions of the host. Numerous studies have revealed strong associations between the microbiota and multiple diseases. However, association does not mean causation. To clarify the mechanisms underlying microbiota-mediated diseases, research is moving from associative analyses to causation studies. In this article, we first introduce the principles of the computational methods for causal inference, and then discuss the applications of these methods in microbiome medicine. Furthermore, we examine the reliability of theoretically inferred causality by the interventionist framework. Finally, we show the potential of confirmed causality in microbiota-targeted therapy, especially in personalized dietary intervention. We conclude that a comprehensive understanding of the causal relationships between diets, microbiota, host targets, and diseases is critical to future microbiome medicine. the effect of some dietary supplements to promote the growth of Bifidobacterium. the prediction score of the age of fecal sampling according to random forest analysis. toxins that are found in the outside environment. a treatment method that transfers gut microbes in stool samples from healthy donors to patients through the gastrointestinal tract. an approach to study the relationship between host phenotype and microbiota by transplanting human microbiota to germ-free mice. the effect of mediation on the interaction between exposure and outcome. an approach to pinpoint the pathogenic microbial taxa by comparing the microbiota from mice with a different susceptibility to disease, such as healthy mice, morbid mice, and the cohoused healthy and morbid mice. the composition of a microbial community and the abundance of its members. a genome-wide association study that takes microbial traits as one of host phenotype to elucidate the associations between host genetic variations and the microbiome. the community of bacteria, archaea, protists, fungi, and viruses that live in a given ecosystem. one of the fields in biological sciences that aims to understand the mechanisms of biological phenomena by employing high-throughput technologies, including genomics, transcriptomics, proteomics, metagenomics, and metabolomics. individual nutrition guidelines that are formulated based on the combination of an individual's genetic, gut microbiota, environmental, and lifestyle factors. a commercially available probiotic cocktail of eight strains of lactic-acid-producing bacteria: Lactobacillus plantarum, Lactobacillus delbrueckii subsp. bulgaricus, Lactobacillus paracasei, Lactobacillus acidophilus, Bifidobacterium breve, Bifidobacterium longum, Bifidobacterium infantis, and Streptococcus salivarius subsp. thermophilus. the end products of microbial carbohydrate fermentation. The most common SCFAs are butyrate, propionate, and acetate.