Sysmetab, 13C metabolic flux analysis with Scilab

3-TestCases

Table of Content

Test Cases

3 Test cases

3.1 Stationary data

As a test case we will use the file e_coli.ftbl file coming from the distribution of influx_s. This file is in the folder SYSMETAB-5.0/EXAMPLES/e_coli

3.1.1 Conversion to the FluxML format

In order to be processed by the sysmetab script, the file has to be converted to the 13CFlux2 FluxML format. This is done as follows

ftbltofml e_coli.ftbl > e_coli.fml

The e_coli.fml file can be edited in any text editor but an XML aware (at least for syntax highlighting) has to be preferred. If you are just curious about the structure of the file you can open it in Google Chrome and fold/unfold elements :

For example, inside the <reactionnetwork> element, here is the <reaction> element associated to the Citrate synthase

3.1.2 Identification of fluxes

The e_coli.fml file is simply passed as first argument to the sysmetab script as follows (the extension of the filename has to be removed)

sysmetab e_coli

Sysmetab then gives (by default) a lot of information about the successive steps of file processing (network analysis, code generation, …). The last output lines should be the following:

Optimization terminated with the following message : Normal termination.

The following free fluxes are poorly determined (stddev>abs(value)): 

eno_xch, fum_a_xch.

You should compute Monte-Carlo statistics (--stats=mc:n) in order to confirm or infirm this.

Solvable free fluxes seem to be : 

Glucupt_1_net, ald_xch, pyk_net, zwf_net, gnd_net, ta_xch, tk1_xch, ppc_xch, out_Ac_net.

Sysmetab here complains about the high values of standard deviation of some fluxes which are likely non-identifiable.

3.1.3 Visualisation of results

All the results, namely optimal flux values and associated standard deviations, individual and group measurement residuals, values of intermediate cumomers, … are contained in the e_coli.fwd file. Like the e_coli.fml file, the fwd file can be opened in your navigator (here in Google Chrome, with folded elements)

For example, opening the <measurement> element (by clicking on the +) allows you to see the contribution of Succinate measurements to overall residual in value of the norm attribute, as well as the scaling factor.

3.1.4 Conversion to spreadsheeet format

The fwd file can be converted to the CSV spreadsheet format like this

fwdtocsv e_coli.fwd > e_coli.csv

3.2 Non-stationary data

As a test case we will use the file e_coli_ns.fml, containing synthetic noisy data which has been generated by a forward non-stationary simulation using the flux estimated values from e_coli. As you can see it below

the measurements are made at times 1,…,10 seconds. The <data> section is 10 times bigger than the same section in e_coli.fml.

3.2.1 Identification of fluxes

The e_coli_ns.fml file is simply passed as first argument to the sysmetab script as follows

sysmetab e_coli_ns

Sysmetab then gives (by default) a lot of information about the successive steps of file processing (network analysis, code generation, …). The last output lines should be the following:

Optimization terminated with the following message : Normal termination.

The following free pool sizes are poorly determined (stddev>abs(value)): 

AKG, Mal, OAA, Pyr.

The following free fluxes are poorly determined (stddev>abs(value)): 

eno_xch, tk2_xch, fum_a_xch.

You should compute Monte-Carlo statistics (--stats=mc:n) in order to confirm or infirm this.

Solvable free fluxes seem to be : 

Glucupt_1_net, ald_xch, pyk_net, zwf_net, gnd_net, ta_xch, tk1_xch, ppc_xch, out_Ac_net.

Solvable pool sizes seem to be : 

Ery4P, Fru6P, FruBP, GA3P, Glc6P, Gnt6P, ICit, PEP, PGA, Rib5P, Sed7P, Suc.

Compared to the stationary case, some information on pool sizes is given

3.2.2 Visualisation of results

All the results, namely optimal flux values and associated standard deviations, individual and group measurement residuals, values of intermediate cumomers, … are contained in the ecoli_ns.fwd file. Like the e_coli_ns.fml file, the fwd file can be opened in your navigator (here in Google Chrome, with folded elements)

Opening the <measurement> element (by clicking on the +) allows you to see the contribution of Succinate measurements to overall residual in value of the norm attribute, as well as the scaling factor.

The time course of simulated measurement at intermediate times is available but only at the Scilab level. In order to be able to access it, you have to prevent sysmetab to launch Scilab . To this purpose, type the following command:

sysmetab --nocalc e_coli_ns

then launch Scilab, open the file e_coli_ns_direct.sce, and execute it. When the optimization terminates, open the file e_coli_ns_plot.sce (this file will be automatically generated in the next release). The following figure should open:

These graphs show the reconstruction of mass isotopomer fractions of measured metabolites using the free fluxes and pool size values obtained after optimization. Comparing these simulated data with the different time measurements marked with circles demonstrates that synthetic data are well recovered. The legend is the following

Created: 3 years 4 hours ago
by Stéphane Mottelet

Updated: 2 years 4 months ago
by Stéphane Mottelet

Old Revisions