Beta Diversity and Distance-Based Analysis of Microbiome Data

Distance-based analysis of microbiome beta diversity can be a powerful tool for discovering novel associations between microbial composition and a wide variety of phenotypes. Key advantages to distance-based analysis are the flexible form of association between microbiome and outcome the potential for increased statistical power and the ability to account for biological structure in the data (e.g., phylogenetic information). In this chapter, we begin by outlining common beta diversity metrics (distance or dissimilarity metrics). We then describe methods for data visualization, including principal coordinate analysis, and for formal hypothesis testing, including regression-based kernel association tests and the sum of powered score tests. The chapter concludes with a discussion of the strengths of distance-based analysis, its limitations, and areas for future investigation.