Developed by Esraa Gabal, PhD candidate at Baloni lab, HSCI department, Purdue University
Integration of transcriptomics with GEM iMM1865:
in R:
step 1 : Acquire the Entrez gene ID and gene length for the transcriptomics data genes using biomaRt package in R
step 2: Map the transcriptomics data against the gene list from iMM1865 model using entrez_id as the shared column ... note to maintain the gene order as that in model using the function left_join() in R
step 3: Replace the NA in the datasets (model genes which are not present in the transcriptomics dataste and hence, no gene count for them) by zero through data[is.na(data)] <- 0 and then export the data as csv file
in Matlab : (needs GUROBI solver and Cobratoolbox installation)
step 4: Either separate the TPM_normalized csv of samples as individual csv files and run the script for iMAT integration OR keep them in one file, and make loop to read through each coulumn and construct an individual GEM of it
step 5: Export the indiviudal GEM models
metabolic flux analysis
in Python: (needs GUROBI solver and CobraPy installation)
step 6: Read the iMAT GEMs models and process with FBA analysis to define the flux state while optimizing the objective function ("BIOMASS_reaction")
step 7 : Read the iMAT GEMs models and process with FVA analysis to define the flux distribution of metabolic reactions
step 8 : Read the iMAT GEMs models and process with flux sampling using the OptGPSampler algorithm with 1000 iterations to investigate the flux changes in metabolic reactions
metabolic flux visualization
in R:
process the outputs of FBA, FVA, and flux sampling which are exported either as .xlsx or .csv format and visulize using tidyplot or ggplot2
for FBA output, it was normalized using SIGDA in R as identified in the methods part within the paper