Multiplicity Quantifies Intra‐Cluster Diversity in Metagenomic Analysis

ABSTRACT

Clustering data into operational units is essential in biodiversity studies, allowing researchers to organize observations into biologically meaningful entities such as species, proteins, or genes. However, this step inevitably leads to information loss by concealing intra‐cluster variability—an important component of functional biodiversity—and the extent of this loss depends on the clustering scheme. Here we introduce multiplicity, a novel index grounded in Hill numbers within the alpha–beta–gamma diversity framework, to quantify the diversity masked through clustering and capture intra‐cluster heterogeneity. We develop two versions of the index for contexts in which distances between subunits are either unknown (inventory multiplicity) or known (distance‐based multiplicity). We demonstrate the distinctiveness of multiplicity relative to established diversity indices through simulations. When applied to real‐world soil metagenomic data, multiplicity revealed hidden biological variability within clusters and informed the choice of clustering schemes. This index complements traditional diversity measures by providing a more complete picture of community functional diversity and guiding methodological decisions that balance clustering granularity with diversity preservation. Both versions of multiplicity are implemented and publicly available in the R package divermeta.