Abstract
Changes in the gut microbiome are often associated with disease. One of the major goals in microbiome research is determining which components of this complex system are responsible for the observed differences in health state. Most studies apply a reductionist approach, wherein individual organisms are evaluated independently of the surrounding context of the microbiome. While such methods have yielded valuable insights into the microbiome, they fail to identify patterns that may be obscured by contextual variation. A recent report by Schubert et al. [A. M. Schubert, H. Sinani, and P. D. Schloss, mBio 6(4):e00974-15, 2015, doi: 10.1128/mBio.00974-15] communicates an alternative approach to the study of the microbiome's association with host health. By coupling a multifactored experimental design with regression modeling, the authors are able to profile context-dependent changes in the microbiome and predict health status. This work underscores the value of incorporating model-based procedures into the investigation of the microbiome and illustrates the potential clinical transformations that may arise through their use.