Back

Pan-microbiome analysis along the human respiratory axis reveals an ecological continuum in health and collapse in disease

Shinde, T.; Kumpitsch, C.; Zhang, Y.; Franzosa, E. A.; Mohammadzadeh, R.; Weinberger, V.; Eagan, T. M.; Huttenhower, C.; Foris, V.; Moissl-Eichinger, C.

2025-09-04 bioinformatics
10.1101/2025.09.04.674177 bioRxiv
Show abstract

The human respiratory tract (RT) harbors complex microbial communities whose functions are critical to health and disease. Yet, current insights remain fragmented across anatomical sites, populations, and clinical states, limiting the fields ability to define common patterns in health and disease. Here, we present the first global respiratory pan-microbiome atlas, a resource integrating over 4,000 metagenomes across upper, intermediate, and lower RT from diverse cohorts encompassing health, pneumonia, COVID-19, and cystic fibrosis. Standardized taxonomic profiling reveals marked biogeographic structure: in health, lower RT communities largely represent filtered subsets of upper RT microbiota. Respiratory disease disrupts this organization, with reproducible depletion of core taxa at specific locations such as Rothia mucilaginosa and Fusobacterium pseudoperiodonticum, the latter being present in 88% of healthy sputum samples. Source-tracking analyses further support the collapse of inter-compartmental connectivity in disease and show the emergence of invasive taxa of unclear origin. Finally, prevalence-based models outperform abundance-based models in detecting disease-associated disruptions, providing greater sensitivity to shifts in community stability. Altogether, this atlas defines the healthy RT microbiome as a spatially structured ecosystem and provides a foundational reference for advancing personalized care and systems-level models of respiratory disease.

Matching journals

The top 3 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.