Schematic description of the Remove-Renormalize-Relate (3R) approach adopted for identifying the core-associated taxa for each study. Briefly, to measure the community association of a given taxon (X) in a study cohort A, we first removed the abundances of X from all microbiomes of this cohort and recomputed the abundances by renormalizing the abundances of the remaining taxa. Community variations of the resulting microbiome profiles were then computed using principal coordinate analysis (PCoA). Subsequently, these community variations were related with the abundance of X using the envfit approach. In scenarios where the abundance of X is strongly associated with the compositions of all other members in a gut microbial community, community-wide variations as measured using specific principal coordinates (or PCoA axes) will show a strong correlation (computed using the envfit approach) with the abundance of X (as shown in this representative example). The 3R analysis was applied to all 201 taxa across a given study and taxa were ranked based on their R2values. Taxa with R2in the top 70% were selected. In parallel, we also identified the prevalent taxa in the dataset (65% prevalence) (bottom panel). Taxa that satisfied both the properties were then selected as the core-associated taxa of the study cohort.
