Modeling the temporal dynamics of the gut microbial community in adults and infants

• DOI: 10.1371/journal.pcbi.1006960
• Publication Date: 2019-06-27T17:27:31Z
• Authors: Liat Shenhav; Ori Furman; Leah Briscoe; Mike Thompson; Justin D. Silverman; Itzhak Mizrahi; Eran Halperin
• Keywords: Adult; Male; 0303 health sciences; Time Factors; QH301-705.5; Computational Biology; Infant; Models, Biological; Gastrointestinal Microbiome; 03 medical and health sciences; Databases, Genetic; Humans; Female; Biology (General); Research Article

Given the highly dynamic and complex nature of human gut microbial community, ability to identify predict time-dependent compositional patterns microbes is crucial our understanding structure functions this ecosystem. One factor that could affect such interactions, wherein community composition at a given time point affects later point. However, field has not yet settled on degree effect. Specifically, it been recently suggested only minority taxa depend in earlier times. To address issue identifying predicting temporal we developed new model, MTV-LMM (Microbial Temporal Variability Linear Mixed Model), linear mixed model for prediction dynamics. can (i.e., whose abundance be predicted based previous composition) longitudinal studies, which then used analyze trajectory microbiome over time. We evaluated performance real synthetic series datasets, found outperforms commonly methods modeling. Particularly, demonstrate effect points underestimated by least 10 when applying approaches. Using MTV-LMM, considerable portion microbiome, both infants adults, significant component points. This suggests major defining future phenomenon considerably more common than previously reported microbiome.