Inference of Causal Interaction Networks of Gut Microbiota Using Transfer Entropy

Abstract

Understanding the complex dynamics of gut microbiota interactions is essential for unraveling their influence on human health. In this study, we employed transfer entropy analysis to construct a causal interaction network among gut microbiota genera from the time-series data of bacterial abundances. Based on the longitudinal microbiome data from two subjects, we found that the constructed gut microbiota regulatory networks exhibited power-law degree distribution, intermediate modularity, and enrichment of feedback loops. Interestingly, the networks of the two subjects displayed differential enrichment of feedback loops, which may be associated with the differential recovery dynamics of the two subjects. In summary, the transfer entropy-based network construction provides us with valuable insights into the ecosystem of gut microbiota and allows us to identify key microbial hubs that play pivotal roles in shaping the microbial balances.