Back

State of health of the gut microbiome assessed through a quantitative metabolomic-based marker panel

DeLeon, O.; Frith, M.; Sidebottom, A. M.; Koval, J.; Lolans, K.; Ceccato, H. D.; Cham, C. M.; Fei, N.; Wall, E.; Kupfer, S.; Sims, A.; Pierre, J. F.; Chang, E. B.

2025-02-18 gastroenterology
10.1101/2025.02.14.24319815 medRxiv
Show abstract

Defining the "health" of the gut microbiome or microbial organ has been elusive and controversial. This in turn makes defining gut dysbiosis (imbalances in gut microbiota) an enigma particularly in determining what, how, and when to intervene to restore eubiosis. Taxonomical profiles of gut microbiota have typically been used but are unreliable because of highly variable interindividual taxa and arbitrary functional assignments generalized to microbial groups. We therefore developed a functional metabolomic marker panel, selected through a supervised, machine-learning approach, using a healthy adult discovery cohort. The panels performance and utility were evaluated individually and together using a composite Z score through; (1) two validation cohorts of self-described healthy adults, 2) a large, external inflammatory bowel disease cohort, and 3) a patient with functional bowel disease to direct dietary intervention. The findings of this study provide proof of concept that quantitative functional profiles of the gut microbiome can be useful in assessing the health of the gut microbiome and for clinical application.

Matching journals

The top 7 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.