Critical assessment of genome-scale metabolic Models of Arabidopsis thaliana

A Genome-Scale Metabolic Model (GSMM) include all known reactions of an organism, together with the known Gene-Protein-Reaction (GPR) relations. GSMMs are considered as important tools in Systems Biology and have enabled researchers to perform systems level study of living organisms in silico. Flux balance analysis (FBA) as a constraint-based technique enables computation of reaction fluxes through GSMMs in steady state and prediction of cellular behavior in specific conditions. The quality of the GSMMs is important, as a defective model can lead to erroneous and unreliable predictions. Consequently, quality control of GSMMs before any computation is necessary. Five GSMMs for A. thaliana have been reconstructed to date, including Arameta (Poolman et al., ۲۰۰۹), AraGEM (Dal’Molin et al., ۲۰۱۰), Aramodel۲۰۱۲ (Mintz-Oron et al., ۲۰۱۲), Aracell (Cheung et al., ۲۰۱۳) and Aramodel۲۰۱۵ (Seaver et al., ۲۰۱۵). Here, we have evaluated the quality of these models. For evaluation of the previously-published models, we analyzed their different computational properties, including the number of blocked reactions, mass balance of reactions and existence of erroneous energy generating cycles (EGC). Our study suggests that all of the five models include mass-unbalanced reactions. Furthermore, all models except Arecell included EGC. However, Aracell had the maximum number of blocked reactions. These results obviously show that simulation using these models may result in wrong predictions and manual curations before any computational study is necessary step.

کلیدواژه ها:

Genome-scale metabolic network ، FBA ، manual model curation ، Quality control