This repository contains the scripts for the paper Ducci G. et al. paper, including:
To focus on the different metabolic rearrangements characterizing FGFR3-wt and FGFR3 altered groups, we used a computational pipeline that integrates exometabolomics and transcriptomics data, i.e. based on the ENGRO2 core model as a metabolic network
-
model.xml: this file containts the stoichiometric metabolic model (i.e. ENGRO2) used both the RAS computation and flux sampling simulation.
-
ras_adata: this file containts the Reaction Activity Scores computed for the five cancer cell lines (RT112, 5637, UMUC3, HT1376 and J82), using ENGRO2 core model as a metabolic network. Transcript levels were converted into RASs, namely enzymatic activities according to the expression of their associated genes and the relationship among them encoded within GPR (Gene-Protein-Reaction) associations.
-
data: folder containing mean and std of exchanges rates of 19 metabolites from exometabolomics (i.e. alanine, aspartate, cystine, glucose, glutamate, glutamine, glycine, histidine, isoleucine, lactate, leucine, lysine, methionine, phenylalanine, pyruvate, serine, threonine, tyrosine, valine). These values are used as constraints for the model, in terms of consumed and excreted metabolites to fuel cell metabolism, and used the upper and lower values of exchange rates as upper and lower bounds of the corresponding exchange reactions.
-
map.svg: svg file containing the graphical representation of the ENGRO2 metabolic model. This map is used to report both RAS and fluxes differences between FGFR3-altered and FGFR3-wild type.
-
utils.py: various python objects (functions, variables) used for the pipeline
-
libraries.py python libraries used for the computational pipeline
-
1_transcriptomics_statistics.py: script to generate statistical analysis for the RASs values
-
2_create_metabolic_network.py: script to generate the metabolic network of the five cancer cell line using transcriptomics constraints (i.e. RAS values) and exo-metabolic constraints. After the creation of metabolic network, flux sampling is used to generate to corresponsing fluxes with the Optgp algorithm (10 batches, thinning =100, 1000 samples for each batch)
-
3_analysis_fluxes_statistiche.py: script used to generate statistical analysis for the flux values
-
4_mapping.py: script used to generate the maps for
- FGFR3-altered vs FGFR3-wild type: RAS differences and Flux differences
- CCLE and TCGA data: RAS differences
All the outputs are generated in the output directory.