The forecasting power of the microbiome

Abstract

Microorganisms are informative biological integrators of past and present environmental abiotic and biotic conditions. At the same time, they are directly involved in ecosystem processes. Unfortunately, the complexity of microbial communities has so far resulted in most studies being descriptive. Here, we suggest that signals in the microbiome data can be used to forecast future ecosystem processes. The combination of omics with various statistical learning approaches, selected based on accuracy-interpretability and bias-variance trade-offs, will be key to attain this goal, as exemplified by recent studies. The time is ripe for microbial ecologists to fully exploit the forecasting power of microbiomes.

Keywords: microbiome; omics data; predictive models; statistical learning.