Georg Gerber, Ph.D.
, assistant professor, Harvard University
The human microbiome is highly dynamic on multiple timescales, changing dramatically during development of the gut in childhood, with diet, or due to medical interventions. Understanding and being able to manipulate these dynamics is essential for the rational design of microbiome-based diagnostics and therapeutics. However, analysis of longitudinal microbiome data is hampered by a paucity of tailored and principled computational methods that address inherent challenges of these data including temporally irregular and sparse sampling, experimental noise, and complex dependency structures. Gerber will present several novel Bayesian machine learning methods that they have developed to overcome these challenges. The first, MC-TIMME (Microbial Counts Trajectories Infinite Mixture Model Engine), is a non-parametric Bayesian model for clustering microbiome time-series data that they have applied to gain insights into the temporal response of human and animal microbiota to antibiotics, infectious, and dietary perturbations. The second, MDSINE (Microbial Dynamical Systems INference Engine), is a method for efficiently inferring dynamical systems models from microbiome time-series data and predicting temporal behaviors of the microbiota, which they have applied to developing bacteriotherapies for C. difficile
infection and inflammatory bowel disease. The third, Microbiome Interpretable Temporal Rule Engine (MITRE), is a method for predicting host status from microbiome time-series data, which achieves high accuracy while maintaining interpretability by learning predictive rules over automatically inferred time-periods and phylogenetically related microbes. Gerber's team is currently using MITRE to analyze data from human cohorts, specifically for developing a microbiome-based diagnostic for C. difficile