Culture-independent studies of the human microbiome and microbial communities using multiple high-throughput functional profiling technologies have become a powerful tool for surveying the whole community. This has been highlighted by an increase in longitudinal and population-level microbiome-wide studies that rely on multi-omics profiling to simultaneously characterize community function, dynamics, and biochemical signatures across diverse disease states and environments, providing massive opportunities to decode connections between microbiome and a series of diseases and health conditions. On the other hand, the field has not yet reached the maturity attained in other established molecular epidemiology fields such as cancer biomarker discovery and genome-wide association studies for making the leap from omics profiling to rational microbiome-based therapeutics. One primary limitation to leveraging this large body of microbiome data is computational and statistical. Among these are the technical nature of the data associated: high-dimensionality, count and compositional data structure, sparsity (zero-inflation), over-dispersion, and hierarchical, spatial, and temporal dependence, among others. These topics will be discussed within the framework microbiome association analysis and new statistical methods and computational tools to combat these challenges, while maintaining both statistical rigor and biological relevance, will be proposed.