Comparison And Analysis Of Different Genome Scale Metabolic Models Of Yarrowia Lipolytica

With the development of sequencing technology and the emergence of various omics data,the construction of the genome-scale metabolic models(GEMs)can successfully put different metabolic data together and analyze microorganism metabolic network characteristics.This makes GEM to be an indispensable tool for studying the complex physiological metabolism of microorganisms.In recent years,Yarrowia lipolytica is known for its superior characteristics in the production of oils and fatty acid-derived compounds,and can be used for metabolic production of various organic acids and polyalcohols,which has attracted the attention of researchers.More and more tools made for automating construction of GEMs have led to an increase of published models.The same species often has more than one GEM constructed by different model creators.However,the exact procedure and resources used depend on the model creators'expertise and research interests,such that independently created models for the same organism are often very different,which might result in significant difficulties for model users to choose an appropriate model.It is therefore necessary to understand and analyze the existing models comprehensively.In this study,we have compared and analyzed six published GEMs of Y.lipolytica in order to find out some common or specific problems involved in the construction of GEMs for the same species and provide suggestions to modify to improve the accuracy of GEMs.In this study,different simulation results of the published GEMs of Y.lipolytica were analyzed based on experimental data of Y.lipolytica in order to understand the differences between different models and discover if there are common problems in models.First,specific growth rates(?)were simulated by different models,then we found that the molecular weight of biomass in models have significant influences on the accuracy of simulation results of ?.Besides,after the simulation of intracellular flux distribution,the analysis revealed that there were some wrong descriptions on the reversibility of reactions of currency metabolites in models.Finally,the simulation of specific metabolite formation rates was carried out for different targeting products based on different models,and large discrepancies among different models were observed and the corresponding solutions were obtained.Based on the above findings,we have therefore modified and improved the accuracy of the existing Y.lipolytica GEMs,together with several suggestions for better constructing and refining genome scale metabolic network models.