TY - JOUR T1 - Inferring the quasipotential landscape of microbial ecosystems with topological data analysis JF - bioRxiv DO - 10.1101/584201 SP - 584201 AU - William K. Chang AU - Libusha Kelly Y1 - 2019/01/01 UR - http://biorxiv.org/content/early/2019/03/21/584201.abstract N2 - The dynamics of high-dimensional, nonlinear systems drive biology at all scales, from gene regulatory networks to ecosystems. Microbial ecosystems (‘microbiomes’) exemplify such systems due to their richness and the small length- and time-scales of complex ecological and evolutionary dynamics. Microbes inhabit, respond to, and alter environments ranging from the human gut to the ocean. Here, using information theory and topological data analysis [1] (TDA), we model microbiome dynamics as motion on a potential energy-like landscape, called the quasipotential, identifying attractor states and trajectories that characterize ecological processes including disease progression in the human microbiome and geochemical cycling in the oceans. Our approach allows holistic analysis and prediction of large-scale dynamics in generalized complex systems that are difficult to reduce to their underlying interactions. ER -