Stochastic Generalized Lotka-Volterra Model with An Application to Learning Microbial Community Structures
Inferring microbial community structure based on temporal metagenomics data is an important goal in microbiome studies. The deterministic generalized Lotka-Volterra differential (GLV) equations have been used to model the dynamics of microbial data. However, these approaches fail to take random environmental fluctuations into account, which may negatively impact the estimates. We propose a new stochastic GLV (SGLV) differential equation model, where the random perturbations of Brownian motion in the model can naturally account for the external environmental effects on the microbial community. We establish new conditions and show various mathematical properties of the solutions including general existence and uniqueness, stationary distribution, and ergodicity. We further develop approximate maximum likelihood estimators based on discrete observations and systematically investigate the consistency and asymptotic normality of the proposed estimators. Our method is demonstrated through simulation studies and an application to the well-known "moving picture" temporal microbial dataset.
READ FULL TEXT