Abstract
<title>Abstract</title> <p>Microbiome studies commonly characterise taxa by their prevalence across samples and their average abundance within samples. However, these descriptors do not capture how abundance is distributed among hosts and therefore cannot distinguish taxa that maintain stable abundance from those that occasionally undergo extreme expansion. Such sporadic dominance events may have disproportionate functional consequences, including transient amplification of antimicrobial resistance gene (ARG) abundance. Here, we introduce bloominess, a metric for quantifying sporadic dominance by measuring the divergence between mean and median abundance across samples in which a taxon is present. Using a genome-resolved chicken caecal microbiome atlas, we applied this framework to thousands of metagenome-assembled genomes (MAGs). Bloominess revealed that sporadic dominance was constrained at both low and high prevalence but varied widely among taxa of intermediate prevalence. Taxonomic groups differed markedly in bloominess distributions, with several clinically relevant taxa exhibiting sporadic dominance. When integrated with antimicrobial resistance (AMR) annotations, bloominess distinguished stable background reservoirs of resistance from taxa capable of transiently amplifying resistance signals. Together, bloominess provides a complementary ecological axis for interpreting microbiome organisation and shows that AMR risk reflects both resistance gene content and the ecological behaviour of the taxa that carry it.</p>