Code availability for the manuscript titled "Reconciled rat and human metabolic networks for comparative toxicogenomics analyses and biomarker predictions" by Blais et al.
The article is now published and available at http://www.nature.com/articles/ncomms14250
This MATLAB script imports iRno and iHsa into the RAVEN toolbox and performs the checkTasks function for 327 metabolic tasks described in Supplementary Table 4. Most tasks capture metabolic functions that are common to both rats and humans. Some tasks are meant to fail in both organisms while others are meant to fail in only one organism, consistent with known species-specific differences between rat and human metabolism.
This script was used to generate Fig. 2a and Supplementary Fig. 3. For Supplementary Fig. 3, this R script inputs orthology annotations from multiple databases by reading in data from Supplementary Table 2 to generate rat gene-protein-reaction (GPR) relationship rules based on human GPR rules from HMR2 (after replacing Ensembl Gene Identifiers with Entrez gene identifiers). For Fig. 2a, this R script compares the complexity of gene-protein-reaction (GPR) directly between rat and human metabolic models by reading in data from Supplementary Table 3. This script inputs gene-protein-reaction (GPR) relationship information from a superset of reactions included in iRno and iHsa to compare the global distribution of GPR sizes between the rat and human models. The goal of this script was to demonstrate how GPR sizes varied between species when filtering orthology annotations for the automated conversion of iHsa to iRno as well as after manual curation.
Inputs raw gene expression microarray data
Outputs gene expression changes
Helper functions: ncomm_helper.R
Inputs gene expression changes, rat and human GPR rules
Outputs TIMBR reaction weights
Helper functions: ncomm_helper.R
Inputs TIMBR reaction weights, rat and human metabolic networks
Outputs raw TIMBR predictions
Helper functions: timbr.m and ncomm_blais_model2irrev.m
Inputs raw TIMBR predictions
Outputs normalized TIMBR production scores, manuscript figures
Helper functions: ncomm_helper.R
R source file that defines several helper functions used in the R scripts above.
MATLAB implementation of the TIMBR (transcriptionally-inferred metabolic biomarker response) algorithm
Modified version of xls2model from the COBRA toolbox (www.github.com/opencobra/cobratoolbox) that fixes a minor bug associated with importing Excel-formatted versions of rat and human metabolic networks in the ratcon database
Modified version of convertToIrreversible from the COBRA toolbox (www.github.com/opencobra/cobratoolbox) that facilitates the mapping of TIMBR reaction weights to irreversible reactions