Recurrent neural network models enable prediction and design of health-relevant metabolite dynamics in synthetic human gut communities.

In baboon gut microbiota, most pairwise correlations in bacterial abundances are weak and negative, and bacterial correlation patterns are largely shared across hosts, rather than personalized to each hosts.

A novel mathematical modeling framework for antibiotic-resistant bacteria combining within-host microbiome-pathogen interactions with population-level pathogen epidemiology, demonstrating how antibiotic consumption and ecological competition come together to drive the spread of resistance.

Human gut bacteria alter their metabolism in response to each other's presence, which causes their community dynamics to deviate from predictions that are based on mono-culture data.

Genome-scale metabolic models can be used to predict the engraftment of probiotics.

Longstanding eco-evolutionary theories are tested using shotgun metagenomic data from the human gut microbiome, showing links between community diversity and the evolutionary trajectory of a focal species within the community.

A data-driven within-host model reveals that different antibiotics are associated with divergent effects on antibiotic resistance carriage and abundance in hospitalised patients, with important implications for antibiotic stewardship.

Interactions between fungal and bacterial microbiome members alter Drosophila melanogaster's chemical environment, affecting host behavior and survival.

Killing their neighbors allows bacteria to steal genes, including antibiotic resistance genes, which we observed under a microscope, quantified, modeled, and predicted potentially guiding strategies to combat it.

Analysis of the global genetic requirements and gene expression changes in E. coli in the presence of a simple microbiome revealed pairwise and higher-order interactions, and underlying molecular mechanisms.