Compositional microbiome-based signatures associate with general health status

M. Pujolassos${}^a$, A. Kurilshikov${}^b$, S. Zhernakova${}^b$ and M.L. Calle${}^c$

${}^a$University of Vic – Central University of Catalonia, Spain, ${}^b$University Medical Center Groningen, Groningen, The Netherlands, ${}^c$Institut de Recerca i Innovació en Ciències de la Vida i de la Salut a la Catalunya Central, Spain

Growing evidence links gut microbiome composition to human health. However, many disease‑associated microbial signatures come mainly from clinical cohorts, leaving their relevance to general‑population health unclear. This gap, together with the challenges of analysing compositional microbiome data, still limits their use in preventive medicine.

To address this challenge, we adopted a two‑stage aproach combining (1) discovery of disease‑associated microbial signatures from public datasets and (2) external validation in a general population cohort. Bacterial signatures were identified using a compositional data analysis method, coda4microbiome [1], which performs variable selection via penalized regression applied to all‑pairs of log‑ratios. Using coda4microbiome in large public microbiome datasets, we identified 16 disease‑specific bacterial signatures, which were then quantified in a general‑population cohort, LifeLines‑DMP [2], to assess their association with overall health status and well-being.

Most signatures showed consistent associations linked to poorer overall health. Results support the idea that shifts in microbial compsition may reflect early physiological alterations preceding clinical diagnosis, highlighting the value of microbiome‑based compositional biomarkers in preventive healthcare.

Keywords: compositional data analysis, microbiome, microbial signatures, general health prediction