2. MethodologyΒΆ

Starting from the Recon 2.2 consensus human metabolic network [1][2], we reconstructed a network for H1 hESCs to generate a model that recapitulates the observed production of nucleic acids, amino acids, lipids, and other metabolites required for H1 hESC growth. We:

  1. Added new metabolites, transport reactions, and exchange reactions based on mTESR1 stem cell culture media composition;

  2. Added new metabolites to the biomass reaction and set the stoichiometric coefficients of these metabolites to their measured concentrations in H1, mammalian and yeast cells. We also added new reactions from KEGG to ensure the model could simulate the production of these metabolites;

  3. Updated DNA composition based on the genome sequence of H1.

With our hESC reconstruction, we used H1 transcriptomic data to determine the subset of metabolic reactions that are active. We then used kinetic data from SABIO-RK [3] and H1 proteomic data from PaxDB [4][5] to constrain the model. We simulate the model using flux balance analysis (FBA) [6].