Fig. 2

In this toy problem, ASVs are phylogenetically related by the given tree. Black circles correspond to ASVs found in the affected cohort and white circles correspond to ASVs in the control cohort. Although the there is a disease-specific clade enriched in the affected cohort, we would not necessarily detect this differential abundance using traditional analysis. On the ASV level, each ASV is only found in one individual, creating too sparse of a person-vs-ASV matrix to detect differential enrichment. Aggregating to the genus level, the affected and unaffected cohorts have similar abundance of genus 1 (blue) as well as genus 2 (red), and there is no genus-level differential enrichment