Improving flux predictions by integrating data from multiple strains

Abstract

Supplementary data are available at Bioinformatics online.

Essentiality of local topology and regulation in kinetic metabolic modeling

“…These genes encoded 28 of the 51 flexible and highly active reactions in PPP, TCA cycle, glyoxylate cycle, polysaccharide synthesis, mitochondrial transporters, and byproduct formation (Appendix file 1 Material S2 and Appendix file 5 Table S4) (Blank et al, 2005). Decrem and four other methods—pFBA, FBA, RELATCH and REPPS (Long & Reed, 2016) were applied to flux prediction in these mutants, with the specified extracellular exchange fluxes and actual nutrient concentrations constraining the feasible region (the upper/lower bounds of corresponding reactions).…”

“…Thirty-eight single-gene-deletion mutants with experimental 13 C-MFA fluxes, growth rates, nutrient uptake properties and several extracellular exchange fluxes 44 were used (Supplementary Information: Material S2; Supplementary Table 4). Decrem and four other methods: pFBA, FBA, RELATCH and REPPS 45 were used for flux prediction in these mutants. The predictions by Decrem showed the highest correlations with the experimentally measured fluxes in almost all mutant strains for both GSMs (Fig.…”

“…So far, the largest metabolic kinetic model, k-ecoli457 of E. coli , only contains 457 reactions, 337 metabolites and 295 substrate-level regulatory interactions according to the computationally predicted kinetic parameters (Khodayari & Maranas, 2016). Alternatively, a type of indirect kinetic method is approximating the possible flux boundary using its transcriptomics (proteomics) data to describe coarse-grained feasible region of reactions though stoichiometric matrix (Long & Reed, 2016; Sánchez et al, 2017; Tian & Reed, 2018), which has shown limited improvement (Machado & Herrgård, 2014). In addition, a recent study found experimental fluxes cannot be accurately predicted from Michaelis-Menten kinetics without considering metabolic regulators (Hackett et al, 2016).…”

“…For example, the flux balance analysis (FBA) predicts metabolic flux distribution by maximizing the fluxes toward the biomass production based on experimentally measured nutrient uptake (Fell & Small, 1986; Orth et al, 2010), while the minimization of metabolic adjustments (MOMA) and regulatory on/off minimization (ROOM) minimize the divergence between the reference fluxes and the prediction of perturbed fluxes for mutant strains (Segrè et al, 2002; Shlomi et al, 2005). Furthermore, multi-omics experimental data such as transcriptomes and proteomes are now integrated into CBMs to reinforce the algorithm’s performance, and such condition-specific cell state information helps to shrink the feasible regions of CBMs and identify the activated reactions (Kim & Reed, 2012; Long & Reed, 2016; Sánchez et al, 2017; Tian & Reed, 2018; Xie et al, 2017).…”

Genome-scale modeling of yeast: chronology, applications and critical perspectives

“… 2016 ). So, if we take this into account while performing computational simulations, we will likely improve the understanding of genotype–environment–phenotype relationships and, consequently, the rational design of cell factories (Long and Reed 2017 ). Lastly, in nature, yeast species present several interactions with other microorganisms and the compounds they secret can influence their co-habitants (Jouhten et al.…”