Importing and exporting¶

SBML¶

SBML is the community exchange format for systems-biology models, and it is how ProcessDB talks to the rest of the field.

Importing¶

File → Import SBML reads a model from an SBML file. ProcessDB maps SBML's vocabulary onto its own:

SBML	ProcessDB
Compartment	Place
Species	State (with its molecule)
Reaction	Process
Reactant, Product	Reactant and Product roles, with stoichiometry
Modifier (by its SBO term)	Enzyme, Catalyst, Activator or Inhibitor; any other modifier becomes a Modifier

SBML distinguishes kinds of inhibitor and activator — competitive, non-competitive, essential, and so on — by SBO term. Those all arrive as plain Inhibitor and Activator roles, because in ProcessDB the mechanism belongs to the rate law, not to the diagram, and is chosen when the law is built.

A progress dialog reports as it works. As with everything else, the imported model is local until you commit — so you can look it over and reload if it is not what you wanted.

Exporting¶

File → Export SBML writes the displayed model out. The save dialog carries the options:

Dialect — SBML L3 V2 or L2 V4;
CellDesigner compatible — for opening the file in CellDesigner;
SBML Qualitative ("qual") — export using the qualitative-models package, for tools that reason about network structure rather than kinetics.

Modifier roles are exported as SBO terms — inhibitor, stimulator, catalyst, enzymatic catalyst, and "modifier of unknown activity" for the Regulator, Mediator and Modifier roles — so a receiving tool knows which way each regulator pushes. Transporters and channels carry no SBO term.

Export carries structure, not kinetics

ProcessDB's SBML export writes the model: compartments, species, reactions, and modifier roles. It does not write rate laws, because those belong to a Model Realization rather than to the model, and a model may have several. Export the structure, and let the receiving tool apply its own kinetics — or keep the quantitative work in ProcessDB and use the model bundle to publish it.

Images¶

File → Export as Image saves the current diagram as a picture — for slides, manuscripts, and lab-meeting handouts. The image shows the diagram as it appears now, including your layout, colours, and whether the state mask is on. It is available while a model's diagram tab is showing.

The model bundle¶

File → Export Model Bundle… writes everything a paper, poster or report needs about a model into one folder, and a zip of the same folder beside it. Choose a name and a location; ProcessDB does the rest on its own, and reports how many files it wrote when it finishes.

The bundle holds:

the model diagram, as SVG and as a PNG at twice screen resolution;
every open graph window, as SVG;
for each MOE, the generated equations three ways — typeset as LaTeX (a fragment to \input, needing amsmath), as MathML, and in the solver's own syntax — with a legend from symbol to state or process;
the parameter values of each parameter set, as CSV, with a column saying whether each value was tuned in that set, inherited from set 0, or left at the generated default;
the last simulation results and all the experimental data, as CSV;
report.html, a self-contained page tying it together.

The report opens in any browser and prints to PDF from there; its tables can be selected and pasted into a word processor, and its equations are real MathML, so they paste into Word as editable mathematics rather than as pictures. It is a starting point, not a finished manuscript — each figure and table also exists as its own file under assets/, which is the copy to take into a paper.

Two things are taken from what is on screen, so open them first:

the diagram is exported only if the model is open — from a MOE or MR tab alone the bundle notes its absence;
a graph is exported only while its window is open. Open the graphs you want in the Parameter Explorer before exporting.

Equations are included for a MOE once they have been generated, and simulation results once it has been simulated in this session. Anything missing is listed at the top of the report rather than silently left out.

Equations typeset with each state's and process's user symbol where one has been set, and its name otherwise — so the symbol table is where to go if the LaTeX should read [ATP] rather than S123.

The command is available from a model tab, a Model Realization tab, or a MOE tab.

JSON for AI tools¶

File → Export JSON (for LLMs) writes the current diagram as a structured file you can attach to a ChatGPT-style conversation, when you want to reason about the model in a tool of your own.

Nothing is sent anywhere by this command — it produces a file, and what happens to it is up to you. To work with an AI assistant inside ProcessDB, with your approval on every change, see The Biology Agent.

Experimental data from Excel¶

Data comes in through an experiment rather than through the File menu: select a data set on an experiment's Experimental Measurements and Data tab and press Import. See Measurements and data.

Exporting unsaved changes¶

File → Export Unsaved Changes… writes everything you have not committed to a local file. It exists for one situation: a save that will not go through and work you cannot afford to lose. See Saving your work.

For colleagues who have ProcessDB, a link beats a file. The share button in the diagram toolbar copies a processdb:// URL that opens the model directly. See Notes, history, and sharing.