Power and sample-size estimation in human microbiome research

Human microbiome research has become pivotal in advancing our understanding of complex diseases such as diabetes, inflammatory bowel disease, and cancer. Much of this work relies on comparing microbial communities across health and disease states, or case-control cohorts, using high-throughput metagenomic sequencing. Yet the very nature of sequencing-derived microbiome data makes robust cohort design and power-based sample-size estimation unusually difficult. Unlike other omics, microbiome profiles are compositional, sparse, and often zero inflated, properties that complicate statistical modeling and inflate sample-size requirements. These challenges are further compounded by the diversity of analytical frameworks—ranging from diversity indices to causal inference—each built on different statistical assumptions and optimized for a distinct research hypothesis. This review synthesizes current approaches around the study design and sample-size estimation in microbiome research, aiming to provide clinicians and researchers with practical guidance for navigating the statistical complexities unique to this field.