opencobra / mass-toolbox Goto Github PK

If MathSBML is not installed, the following error occurs:

Instead of this error though, it should say that MathSBML is not installed and where it can be obtained.

When loading a given model (BIOMD0000000064 in biomodels database) I get the following error:

After the error is raised Mathematica continues running the command for at least 1min (after that I aborted the evaluation).

model = ExampleData[{"Toolbox", "Glycolysis"}];
adjustUnits[{k["vh"] -> 1}, model] (*-> Millimole Hour^-1 Liter^-1*)

vs.

adjustUnits[{k["vh"] -> 1}, {r["vh", {m["h", "c"]}, {}, {1}, True]}] (*-> Hour^-1*)

the Ignore option allows on to exclude certain species from the generated mass-action rate-laws. They're not excluded though if they are the only species in a reaction. adjustUnits is not aware of that.

Km["ATP", "HX1"]

In the toolbox versions after 1.0.6, importing this file with equilibrator data takes longer than with v 1.0.6, and then when using it to calculate deltaG, e.g.:

rxn = str2mass["ENO: 2pg_c <=> h2o_c + pep_c"]
calcDeltaG[{rxn}, bigg2equilibrator, is -> 0.25 Mole Liter^-1, pH -> 7]

there are a number of errors related to unknown units.

Add an annotation attribute to MASSmodel that can host MIRIAM URNs and code to automatically resolve them.

Reproduces the error:

plotSimulation[simulate[glycolysis, {t, -10, 10}][[1]], 
 PlotFunction -> Plot, Tooltipped -> True, Legend -> True]

and how meaningful are they in mass-action sense, for example:

In[48]:= model=Import["/Users/niko/Arbejder/Models/MASS-Models/models/Yeast_Glycolysis_Smallbone_2013/Yeast_Glycolysis_Smallbone_2013.m.gz"];
modelKOtdh2=deleteReaction[model,"TDH_TDH2"];
simulate[modelKOtdh2,{t,0,1*^5}]

results in

During evaluation of In[48]:= simulate::missingIC: Missing initial conditions encountered for {Underscript[energy_charge, Global],Underscript[fit_conc, Global],Underscript[sum_PXG, Global]}. >>
Out[50]= $Aborted

The results of simulating a yeast glycolysis model with MASS v1.1.7 do not match the results obtained from Copasi and MASS v1.0.6

Results from MASS v.1.1.7:

Results from MASS v1.0.6:

When having the following enzyme mechanism:

catalyticBranch = {"E_PFK1[c] + f6p[c] <=> E_PFK1[c]&f6p",
"E_PFK1[c]&f6p + atp[c] <=> E_PFK1[c]&f6p&atp",
"E_PFK1[c]&f6p&atp <=> E_PFK1[c]&fdp&adp",
"E_PFK1[c]&fdp&adp <=> E_PFK1[c]&fdp + adp[c]",
"E_PFK1[c]&fdp <=> E_PFK1[c] + fdp[c]"};

and constructing the enzyme module with one inhibitor, as follows:

enzymeModel =
constructEnzymeModule[Mechanism -> catalyticBranch,
Activators -> {}, ActivationSites -> 0,
Inhibitors -> {m["pep", "c"]}, InhibitionSites -> 4];

All the inhibition reactions have the same name, which doesn't seem right...

Not sure where the LessEqual comparison would be attempted by Mathematica. This weird output is very annoying and distracts from the actual error message.

Either adjust units could be modified to take the transport issue into account or Solve (inside of calcPERC) should solve for the PERCs using the units of the Keqs, species etc.

test

Allow models to store things in minutes, hours, liters etc. rather than just mL and sec.

With the model MODEL1303260018, when doing the following:

newODE = model["CustomODE"] /.Toolboxspecies["ATP", "cell"][t] -> Toolboxspecies["ATP", "cell"];
updateCustomODE[model, newODE];

The ODEs that have changed in newODE are simply duplicated in the model, instead of being substituted.

module["ReverseRateConstants"] returns empty

The set function works if there is no volume on the left side of the equation, but when there is a volume there, the getODE function does not recognize that the ODE's are for the same variables

Take care of the most urgent ones

simulate[model, {t, 0, 0}]

currently returns incomprehensible gibberish like

Part::partw: Part 1 of {} does not exist. >>

Unit conversions and getJacobian hang in M9. After an abort (Cmd + .) they return their results immediately. No glitch in M8

Reproduce problem:

model = ExampleData[{"Toolbox", "Glycolysis"}];
getJacobian[model]

or

Convert[51. Second, Hour]

... probably shouldn't do that

constructModel[{reaction["b1", {}, {metabolite["x1", None]}, {1}, False]}, Parameters -> {rateconst["b1", True] -> 0.1 Hour^-1}]

does not raise an error although the correct unit should be

rateconst["b1", True] -> 0.1 * Hour^-1 * Liter]

in maps

CustomRateLaws might include rate laws that have ignored (Ignore attribute) species variables in them. Rate constants and equilibrium constants will get wrong units in these cases.

editModel and editModelInPlace already exist so one could add editParameters, editInitialCondtions, etc. to the get, set and update functions.

Todo:

• So far, editModel can only deal with list of rules

After installing the new MASS Toolbox version (v1 .1.0) with WebInstall, I got the following errors when executing this command: "model = biomodel2model["MODEL1303260018"];":

StringJoin::string: String expected at position 2 in Milli <> ToLowerCase[1/Moles].
Quantity::compat: Millimoles/Liters and Millimoles¹/Liters¹ are incompatible units

When #issue38 was reported there were no problems loading the above model.

While trying to reproduce an old problem, I noticed that if I try to build a very simple model and convert it to SBML, it crashes due to some units problem.

The notebook is also attached.
build_simple_model.zip

Ideas for 'How to ...' notebooks:

• How to build a toy model from scratch
• How to import a model from SBML and analyze it
• How to QC/QA a model
• How to do analyze the consequences of a single gene knockout
• How to visualize data on a pathway map
• How to effectively utilize ODE Events for perturbations

...

In Copasi, I used the attached model (in the zip file: Smallbone.xml) to add the following reaction:
acaldt: acetaldehyde{extracellular} -> acetaldehyde{cell}

then exported the model in SBML (level 2 version 4) (attached as Smallbone_acaldt.xml, in the zip file), and tried using it in the MASS toolbox.

I can import it, however it crashes when I simulate it. The code+output is the following:

The original model (Smallbone.xml) works fine with the MASS toolbox, and both the original and the edited models work fine with Copasi.

models.zip

Integrate SBML test suite into the unit testing suite.