Integrating microbiome data visualization into FAIRDatabase using edge functions

Microbiome research continues to grow, so does the volume of data it produces. Yet privacy constraints on human-associated samples and the compositional nature of sequencing outputs make quick exploratory analysis difficult. This study extends the FAIRDatabase, an open-source, privacy-compliant infrastructure for microbiome data, with a visualization module designed to tackle both challenges. The module performs composition-aware beta diversity analysis using centered log-ratio transformation and the Aitchison distance metric. All computations are run within Supabase edge functions, which makes sure that sensitive data never leave the secure environment. To guide the design, requirements were derived from prior work and literature, covering compositional data analysis, beta diversity visualization, and principles for clear data interpretation. The resulting tool supports interactive heatmaps and principal coordinates analysis (PCoA) plots, with options for metadata-based coloring, variance explained labels, and color palettes chosen for accessibility and interpretability. In order to evaluate the module, ten participants have performed tasks on the module and filled in the system usability scale (SUS) questionnaire, which resulted in a mean SUS score of 86.8. Three of whom have been interviewed. They valued being able to quickly explore data without downloading files or facing contractual obstacles. Overall, this work shows that edge functions can support composition-aware microbiome analysis without compromising data security. It offers a starting point for building privacy-preserving visualization tools in research areas where data sensitivity is a significant concern.