CarveMe

CarveMe automates top-down genome-scale metabolic model reconstruction for microbial species and communities from annotated genome sequences.

Key Features:

• Universal Model-Based Reconstruction: CarveMe uses a manually curated, simulation-ready universal model as a template for building single-species and community models.
• Top-Down Model Carving: The tool employs a top-down approach to extract organism-specific metabolic networks from the universal model based on genome annotation.
• Gap-Filling Support: CarveMe can predict uptake and secretion capabilities without requiring gap-filling, while also supporting optional gap-filling with custom media libraries.
• Microbial Community Modeling: The tool generates community models by merging single-species models.
• Ensemble Modeling: CarveMe supports ensemble modeling to explore how alternative solutions lead to diverse network structures and phenotype predictions.
• Orthology-Based Search: The tool offers orthology-based search using eggNOG-mapper (kegg-bigg branch) with adjustable blasting options.
• Customizable Universe Models: Users can select from provided templates or create customized universe models during the carving process.
• SBML Compatibility: CarveMe generates models compatible with various SBML formats.

Scientific Applications:

• Single-Species Metabolic Modeling: CarveMe reconstructs genome-scale metabolic models for individual microbial species to study cellular metabolism and guide re-engineering efforts.
• Microbial Community Analysis: The tool enables fast generation of community models to unravel cross-feeding interactions among microbial species.
• Phenotype Prediction: CarveMe models reproduce experimental phenotypes including substrate utilization and gene essentiality.
• Large-Scale Model Databases: The tool supports creation of extensive bacterial model databases, demonstrated with 5587 bacterial models and 74 human gut bacteria models.

Methodology:

CarveMe uses a top-down approach starting from a manually curated universal model and annotated genome sequences to build single-species and community models. The method involves model carving based on genome annotation, with optional gap-filling and ensemble generation to explore alternative network structures.