Adaptive Statistical Methods for Human Microbiome Studies

Hyunwook Koh, AMS SUNY Korea

Abstract: The human microbiome is the totality of all microbes inhabiting in different organs (e.g., gut, mouth, skin) of the human body. The roles of the microbiome on human health or disease have been increasingly studied by the recent advance in high-throughput sequencing technologies. For example, the microbiome perturbation has been associated with a variety of health or disease status (e.g., obesity, diabetes, cancer, brain disorder), medical interventions (e.g., antibiotic or probiotic use) and behavioral or environmental factors (e.g., diet, residence, smoking, birth mode). However, the high complexity of the microbiome data (e.g., high-dimensionality, compositionality, zero-inflation, phylogenetic correlations) makes the downstream data analysis challenging, and thus there is a strong need for more sophisticated statistical methods. In this seminar, we study the human microbiome from the very beginnings of the subject to recent statistical methods (e.g., a- and b-diversity analysis). Then, we further discuss the potential and promise in statistical method development.