Model refinement

Based on the draft reconstruction generated from Model SEED (http://seed-viewer.theseed.org/seedviewer.cgi?page=ModelView) or GEMSiRV, users can curate and refine the reconstruction in GEMSiRV. However, the lack of gene information in imported models may hinder the progress. We, therefore, provide a function to load and update the gene information in GEMSiRV. You can right click on the Gene Index of a model to Load and update gene’s info., and upload the spreadsheet file generated by GBKPaser (http://sb.nhri.org.tw/GEMSiRV/en/GBKParser), e.g. NC_000913.gbk.xls for ECO.

Gene information can be loaded and updated accordingly:

With the aids of simulation and visualization, users can readily identify dead-end metabolites and blocked reactions in the models. Prior to perform simulation, users need to convert the reconstruction into a mathematical model. Therefore, you can right click on a model to Generate simulation tables to generate a model containing a stoichiometric matrix as well as default systems boundaries.

After clicking on Generate simulation tables, three tables including InnerCell Reaction, Exchange Reaction and Metabolite are generated. The prefix “_” used in these three tables for easily distinguishing from the tables required for reconstruction, e.g. Gene Index, Protein Index (optional) and Reaction Index.

_InnerCell Reaction:

_Exchange Reaction:

Please note that the _Exchange Reaction table will be generated only when you have exchanging metabolites (i.e. extracellular metabolites) in the reaction equations.

_Metabolite:

Because growth media for modeled organisms may be similar, an environmental condition can be easily set to a model by right clicking on the model to Define environmental conditions.

Here we use the in silico (computational) minimal media for the model iAF1260 as an example (the text file can be downloaded in http://sb.nhri.org.tw/GEMSiRV/en/Manual). In order to set the system boundaries to the default values, we right click on the model to Generate simulation tables.

The new simulation tables are generated and replace the previous tables. We set a growth medium for modeling the model. We prepare a text file containing the user-defined boundaries and objective, and then right click on the model to Define environmental conditions.

In silico minimal media for the model iAF1260.

A complete medium to simulate all extracellular metabolites can enter/exit the cell freely.

The user-defined system boundaries and the objective are set in the reconstruction model accordingly.

Or you can simply right click on the reaction to update the lower bound (LB), upper bound (UB) or objective coefficient.

You can freely export or save a metabolic model in SBML format or in spreadsheet format by right clicking on a model to Export model (.xml) or to Download model (.xls). Such models generated by GEMSiRV are fully compatible to GEMSiRV for later importing and simulation.

In addiction to the metabolic models saved in SBML format, metabolic reconstructions can be stored in spreadsheet format. The spreadsheet format can store the two-layer relation for gene-protein and protein-reaction associations in network reconstructions. We provide available reconstruction models (GPR) in http://sb.nhri.org.tw/GEMSiRV/en/Metabolic_Models and demonstrate how we use GEMSiRV to reconstruct metabolic networks with GPR relationships.

From reconstruction to model

After clicking on Reconstruction in the menu bar, right click on the Model databases to Import spreadsheets (.xls) for importing the reconstruction file of iAF1260_GPR.xls (download from http://sb.nhri.org.tw/GEMSiRV/en/Metabolic_Models). This reconstruction contains three indices: Gene, Protein and Reaction Index.

Then right click on Reference database to Import database (.xls) for importing the reference database file Ref_BiGG_GPR.xls which is provided in http://sb.nhri.org.tw/GEMSiRV/en/Reference_Databases.

A biomass for E. coli, Biomass_Ecoli_core_N (w/ GAM)-Nmet2, is available in the reference database, you can add the reaction to the reconstruction by right clicking on the main window of Reaction Index to Insert. After submitting the abbreviation of reaction “Biomass_Ecoli_core_N (w/ GAM)-Nmet2”, the related information including reaction name and equation will be conveyed to the reconstruction from the reference database.

Likewise, you can add a new reaction into the reference.

The reaction of Ec_biomass_iAF1260_core_59p81M can be added into the reference database.

Abbreviation: Ec_biomass_iAF1260_core_59p81M

Equation:

(0.000223) 10fthf[c] + (0.000223) 2ohph[c] + (0.5137) ala-L[c] + (0.000223) amet[c] + (0.2958) arg-L[c] + (0.2411) asn-L[c] + (0.2411) asp-L[c] + (59.984) atp[c] + (0.004737) ca2[c] + (0.004737) cl[c] + (0.000576) coa[c] + (0.003158) cobalt2[c] + (0.1335) ctp[c] + (0.003158) cu2[c] + (0.09158) cys-L[c] + (0.02617) datp[c] + (0.02702) dctp[c] + (0.02702) dgtp[c] + (0.02617) dttp[c] + (0.000223) fad[c] + (0.007106) fe2[c] + (0.007106) fe3[c] + (0.2632) gln-L[c] + (0.2632) glu-L[c] + (0.6126) gly[c] + (0.2151) gtp[c] + (54.462) h2o[c] + (0.09474) his-L[c] + (0.2905) ile-L[c] + (0.1776) k[c] + (0.01945) kdo2lipid4[e] + (0.4505) leu-L[c] + (0.3432) lys-L[c] + (0.1537) met-L[c] + (0.007895) mg2[c] + (0.000223) mlthf[c] + (0.003158) mn2[c] + (0.003158) mobd[c] + (0.01389) murein5px4p[p] + (0.001831) nad[c] + (0.000447) nadp[c] + (0.011843) nh4[c] + (0.04148) pe160[p] + (0.02233) pe160[c] + (0.02632) pe161[c] + (0.04889) pe161[p] + (0.1759) phe-L[c] + (0.000223) pheme[c] + (0.2211) pro-L[c] + (0.000223) pydx5p[c] + (0.000223) ribflv[c] + (0.2158) ser-L[c] + (0.000223) sheme[c] + (0.003948) so4[c] + (0.000223) thf[c] + (0.000223) thmpp[c] + (0.2537) thr-L[c] + (0.05684) trp-L[c] + (0.1379) tyr-L[c] + (0.000055) udcpdp[c] + (0.1441) utp[c] + (0.4232) val-L[c] + (0.003158) zn2[c] --> (59.81) adp[c] + (59.81) h[c] + (59.806) pi[c] + (0.7739) ppi[c]

Right click on a reconstruction to Generate simulation tables can convert the reconstruction to a model. Then you can set the system boundaries for simulation.

Draft reconstruction and network refinement

As described previously, we can draft a reconstruction for a genetically related species (e.g. Salmonella) with the existing E. coli model in GEMSiRV. Therefore, we import the file NC_003197.gbk-blast.xls and draft a reconstruction with reference to iAF1260_GPR.

Then we can refine the draft reconstruction by adding metabolic reactions with gene-protein-reaction associations, some existing reactions in the reference database can be conveyed to the reconstruction. For example, Salmonella is reported to be able to synthesize cobalamin due to its metabolic genes (operon) STM2016-STM2035. Therefore, we can manually add those associated reactions and proteins to the draft reconstruction.

Reaction CPC2MT

Name precorrin-2 C20-methyltransferase

Equation [c] : amet + copre2 --> ahcys + copre3 + h

Locus

Gene

Protein

Reaction

CPC2MT

STM2024

cbiL

CbiL

After clicking into the Protein Index of NC_003197.gbk-blast.xls, right click on the main window of protein index to insert the protein abbreviation CbiL, the associated gene STM2024 and a note Added to synthesize cobalamin.

The gene-protein association will be automatically brought into the Gene Index table.

After clicking into the Reaction Index of NC_003197.gbk-blast.xls, right click on the main window of reaction index to insert the reaction abbreviation CPC2MT, the associated protein CbiL, a note Added to synthesize cobalamin and the confidence score 3 for genetic evidence.

The reaction information including name and equation will be automatically brought into the Reaction Index table.

Likewise, the protein-reaction association will be automatically brought into the Reaction Index table.