Predicting gut microbiota dynamics in obese individuals from cross-sectional data

Abstract

Introduction: Obesity affects approximately 39% of adults worldwide. While gut microbiota has been linked to obesity, most research has focused on static taxonomic composition rather than the dynamic interactions between microbial taxa. Methods: We applied BEEM-Static, a generalized Lotka-Volterra model, to cross-sectional 16S rRNA gut microbiome data from six public datasets, comprising 2,435 profiles from lean and obese individuals. Results: A total of 57 significant microbial interactions were identified in obese individuals (79% negative), compared to 37 in lean individuals (92% negative). For example, Bacteroidetes showed a stronger inhibitory effect on Firmicutes in obese individuals (−0.41) than in lean ones (−0.26). Firmicutes and Proteobacteria exhibited consistently higher carrying capacities in obese populations. Discussion: These findings suggest that microbial interaction networks—not just taxonomic abundance—play a key role in obesity-related dysbiosis. Our approach enables the inference of microbiota dynamics from a single time point, paving the way for tailored dietary interventions, which we refer to as Optibiomics.