Skip to content

Model Realizations

A Model Realization (MR) is a quantitative version of a model. The model says what can happen; the realization says how fast, and from what starting point.

Keeping the two apart is what lets you ask sharp questions. One model with three realizations — mass action, Michaelis–Menten, and a modular rate law — is three competing quantitative hypotheses about the same mechanism, and they can be compared against the same data without any risk that the diagrams have drifted apart.

Creating one

Right-click a model — in the Models browser or on its tab — and choose Create Model Realization.

ProcessDB does not leave you with a blank slate. A new realization is seeded with a working starting point:

  • a canonical mass-action rate law for every process, with its parameters;
  • an initial condition of 1.0 for every state;
  • boundary status for every state that only ever plays a modifier role — an enzyme that is never made or consumed anywhere in the model is held constant by default, with a boundary formula of 1.0.

That is enough to generate equations and run a simulation immediately. It is not enough to be right — the numbers are placeholders — but starting from something that runs is much easier than starting from nothing.

The Model Realization tab

The MR tab has two regions, and a panel that belongs to it.

The tree, on the left, shows the realization, its Models of Experiments, and the experiments each MOE includes, with a search box above. Selecting an item — by clicking, or with the arrow keys — shows it on the right; double-clicking a MOE or experiment opens it as its own tab. An experiment dragged from the Browser onto a MOE here is added to that MOE.

The detail panel, on the right, shows the model diagram when the realization itself is selected. This diagram is in MR context: right-clicking a process here offers Edit Rate Law…, and the selection here drives the Parameter Explorer's Selected display mode, so clicking a process there shows exactly that process's parameters. With a MOE or an experiment selected in the tree, the detail panel shows that instead.

Action Items is a panel of the main window rather than a part of the tab — see Arranging the panels. It follows the realization behind the current tab, a MOE tab included, and opens itself along the bottom edge the moment a realization has something that still needs filling in.

Action Items

The Action Items table has one row for each thing that must have a value before the equations mean anything:

Kind Target Formula
Rate Law a process its active rate law
Initial Condition a state its starting value
Boundary Formula a boundary state the formula that fixes it

Rows with nothing filled in are highlighted and sorted to the top, processes before states, so the panel answers "what is left to do?" at a glance. While a realization's rate laws are still loading, the panel says so rather than showing every row blank.

The panel opens by itself when a realization needs attention and keeps out of the way otherwise. Hide it — with the ✕ in its header or its button in the panel bar — and it stays hidden while those items remain; it asks again only when something new comes up. Its button is always in the bar.

You can edit directly in the table: type into the Formula column and the value is applied. A rate law typed for a process that has none is created on the spot, and a new name in a formula becomes a parameter. Selecting a row selects the corresponding process or state in the diagram, so you can see what you are editing, and right-clicking a row offers Show in Diagram, Properties, and, for a process, Edit Rate Law.... Ctrl/⌘-click a symbol in the Target or Formula column to jump to it.

Two buttons act on everything at once:

  • Fill Defaults builds a rate law for every process that has none, asking first which kind to build. It never touches a rate law that already exists.
  • Change All Rate Laws… switches every process to one kind, replacing what it is using now. The previous rate law for each process is kept as an inactive alternative, so switching back is not destructive.

Both use each regulator's role default for how activators and inhibitors enter a modular law; the per-regulator choice is made in the Edit Rate Law dialog.

A Display Names toggle swaps the S…/P… symbols for state and process names throughout the table. Formulas are read-only while it is on — the editor works in symbols — so turn it off before editing.

A rate law that names a deleted state

If you remove a state from the model after writing a rate law that mentions it, the Action Items row is highlighted and its tooltip names the missing state: This rate law refers to S25420 (SDH_FAD), which is not a state of this model — the equations would use a constant in its place. Fix it: re-derive the law with Default… or Fill Defaults, or add the state back.

What comes next

A realization on its own cannot be simulated — it has no experiments and so no time span, no perturbations, and no data. Right-click it and choose Create MOE to build a Model of Experiment.

In this section