The rapid expansion of microbiome research across human, animal, and environmental systems has generated complex, high-dimensional datasets that challenge conventional analytical frameworks. Microbiome measurements typically originate from zero-inflated count data that are subsequently normalized to enable cross-sample comparisons, leading to compositional structures. Such representations induce inherent dependencies among features and violate assumptions underlying many classical statistical methods based on correlations or covariances. A wide range of normalization strategies—such as centered, additive, or isometric log-ratio transformations and their robust variants—are currently used, often without clear consensus or full understanding of their analytical consequences.

Microbial ecosystems are also inherently dynamic. The microbiome can vary substantially within the same individual over time, making longitudinal designs essential for capturing temporal patterns and causal mechanisms. However, these studies introduce additional methodological challenges, including intra-individual correlation, temporal dependence, irregular sampling, and increased model complexity.

This Special Session aims to provide a dedicated forum for advances in bioinformatics, biostatistics, and computational intelligence that address sparsity, compositional constraints, temporal structure, high dimensionality, and the intricate interaction networks linking bacteria, fungi, and other microorganisms across ecosystems.

By fostering interdisciplinary collaboration within a One Health perspective, the session seeks to promote robust, interpretable, and scalable analytical strategies that advance microbiome science and strengthen its impact on human, animal, and environmental health.