Robust Statistical Methods for Longitudinal Microbiome Studies

PROJECT SUMMARY Large-scale longitudinal microbiome studies are increasingly common as they allow investigators to study temporal patterns of the microbiome, elucidating the forces that shape the microbiome and enabling the development of microbiota-based interventions. Unfortunately, despite the potential of longitudinal microbiome studies, few methods exist for analyzing these studies. Also, the few extant tailored methods are limited, either failing to accommodate the characteristics of microbiome data or failing to properly accommodate the longitudinal structure, leading to potentially spurious findings. This proposal aims to fill the critical gaps in methodology by addressing four major areas. Specifically, we aim to develop a comprehensive and coherent suite of statistical tools for (1) addressing batch effects in longitudinal microbiome data – a pressing problem as studies are getting bigger or integrated; (2) improved identification of individual taxa associated with crucial biomedical exposures or outcomes over time; (3) identifying microbiome interaction network dynamics between taxa within longitudinal microbiome data; (4) visualizing the longitudinal microbiome data. These approaches are all based on rigorous prior data emphasizing the importance of the problems as well as the limitations or absence of existing strategies. Our work is motivated by and will directly enable analyses within three important longitudinal microbiome profiling studies: few studies have the combination of large cohorts of patients with numerous follow-ups and rich covariates. Consequently, our methods have the potential to accelerate understanding of the roles of the microbiome in diseases and therapeutics, especially in the motivating studies that include populations of infants with diabetes and immunocompromised individuals (patients undergoing kidney or bone marrow transplant). Ultimately, the deployment of our methods could guide the design and development of novel microbiota-based preventive and therapeutic interventions.