Persephone Pipeline Automates Personalization of Human Metabolic and Microbiome Models

Researchers have presented Persephone, a personalization and evaluation pipeline designed to standardize and streamline working with human whole-body metabolic models and community microbiome models. While human metabolism and gut microbiome co-metabolism are difficult to investigate in vivo due to the complexity of the biological systems involved, genome-scale metabolic modeling with whole-body metabolic models offers a computational approach to address this challenge. Because personalizing and analyzing these models typically requires comprehensive metabolic modeling expertise that can be daunting for beginners, Persephone was developed to lower the computational barrier and standardize the workflow.

Persephone provides automated functionalities to construct community microbiome models directly from raw sequencing data or taxonomy-assigned read tables. The pipeline also integrates physiological and metabolomic data to generate personalized whole-body metabolic models. Researchers can further combine these community microbiome models with whole-body models or personalized whole-body models to build integrated community microbiome whole-body metabolic models, capturing the structural interconnections between biofluids and organs.

The pipeline allows users to optimize whole-body metabolic models based on user-defined metabolic reactions and perform statistical analyses to characterize the resulting data. Significant outcomes can be visualized through generated figures, including box plots of flux values for example reactions, outputs from tools such as surfNet, and coverage histograms saved by MARS for whole-genome shotgun sequence datasets. Persephone preserves intermediate outputs and logs all modifications, data adjustments, and warnings encountered throughout the process.

Documented applications of the Persephone pipeline include the investigation of human-microbiome co-metabolism and evaluating how physiological parameters, such as body weight, affect human metabolism.

Persephone is designed for researchers with basic research-computing experience and access to modest computer hardware. Computational execution times vary based on the number of metabolic reactions optimized and whether taxonomic assignment is carried out; running the pipeline on a single computer with a dataset of 100 metagenomic samples can take several days to a week.

The protocol establishes a standardized computational workflow for researchers seeking to evaluate complex biological interactions without requiring advanced modeling expertise.

Subject of Research: Computational modeling of human whole-body metabolism and microbiome interactions

Article Title: Personalization and evaluation of human whole-body metabolic models and community microbiome models using Persephone

Article References: Nap, B., Hensen, T., Barton, W., Sheehy, A., Widder, J., Chenna, S., & Thiele, I. (2026). Personalization and evaluation of human whole-body metabolic models and community microbiome models using Persephone. Nature Protocols. https://doi.org/10.1038/s41596-026-01437-8

Image Credits: AI Generated

DOI: 10.1038/s41596-026-01437-8

Keywords: Persephone, metabolic modeling, whole-body metabolic models, gut microbiome, computational biology, metagenomics, metabolomics, systems biology, Personalization, evaluation, human, whole-body