Skip to content

usage-guide.md

Latest commit

 

History

History
77 lines (50 loc) · 1.85 KB

usage-guide.md

File metadata and controls

77 lines (50 loc) · 1.85 KB

Model Curation - Usage Guide

Overview

Validate, gap-fill, and curate genome-scale metabolic models using memote for quality assessment and COBRApy for corrections.

Prerequisites

pip install memote cobra

Quick Start

Tell your AI agent what you want to do:

  • "Validate my metabolic model with memote"
  • "Gap-fill my model to grow on M9 media"
  • "Find dead-end metabolites in my model"

Example Prompts

Quality Assessment

"Run memote on my model and show the quality score"

"Check which SBML compliance tests my model fails"

Gap-Filling

"My model can't grow - find reactions to add"

"Gap-fill my model using the BiGG universal model"

Debugging

"Find all unbalanced reactions in my model"

"Which metabolites are dead-ends?"

Standards Compliance

"Add SBO annotations to my model"

"Fix GPR rules to standard format"

What the Agent Will Do

  1. Load the metabolic model
  2. Run memote quality tests
  3. Identify issues (dead-ends, unbalanced reactions, missing annotations)
  4. Suggest or implement fixes
  5. Re-test to confirm improvements
  6. Export curated model

Tips

  • Memote score - Aim for >70% for publication-ready models
  • Dead-ends - Often indicate missing transport or exchange reactions
  • Mass balance - Check proton accounting (common source of imbalance)
  • GPR format - Use 'and' for complexes, 'or' for isozymes
  • Iterative curation - Run tests after each major change