Repository navigation
Tutorials
Eduard Kerkhoven edited this page Sep 6, 2026
·
8 revisions
The RAVEN Toolbox contains several tutorials which highlight the basic RAVEN functionality. While several scripts only demonstrate the RAVEN functions, the other ones have questions, which require manual code changes to solve them.
NOTE: The tutorial scripts are located in RAVENdir/tutorial.
The following tutorials are available:
| Tutorial | Script Name | Brief Description |
|---|---|---|
| Tutorial 1 | tutorial1 |
Use the Penicillium chrysogenum GEM iAL1006 to show some functionality of RAVEN, with a focus on GEM import, modification, simulations and the ways to interpret the simulation results |
| Tutorial 2 | tutorial2 |
Deal with the RAVEN model format and the most basic aspects of GEM modelling: build a simple model from scratch, set parameters and perform simple simulations |
| Tutorial 3 | tutorial3 |
Run FBA simulations, gene deletion analysis, and use GEMs as a scaffold for interpreting microarray data. Use a simplified model of yeast metabolism as an example. |
| Tutorial 4 | tutorial4 |
While having a supplied version of the small yeast model with errors inserted, find and fix these errors to ensure that the model could predict growth |
| Tutorial 5 | tutorial5 |
Show how to de novo generate a draft GEM from KEGG, based on protein sequences in a FASTA file, and doing some functionality checks on the model. |
| Tutorial 6 | tutorial6 |
Show how to de novo generate a combined draft GEM from KEGG and MetaCyc. Shows the case study for bacterium Streptomyces coelicolor together with additional curation steps |
| Tutorial 7 | tutorial7 |
Use the biomass/ functions to calibrate a model's biomass pseudoreaction against measured condition-specific composition data, with a template data file (biomassConditions_template.tsv) to adapt for a new model |
In addition, a book chapter is available from bioRxiv, which details a protocol where protein homology is used to generate a draft GEM using a functional GEM from another species as a template. It comes with its own GitHub repository.
- Introduction
- Installation
- External Databases
- Getting Started
- Model Reconstruction from KEGG
- Option 1: Based on KEGG Organism Code
- Option 2: Based on Homology Search Against KEGG Orthology Specific HMMs
- Option 2-a: Use Pre-Trained HMMs
- Option 2-b: de novo Generate HMMs
- Development Policy
- Known Issues
- Developer Protocols