Skip to content

Your first model

This walkthrough builds a two-step enzymatic pathway from nothing, gives it kinetics, attaches a time course from the bench, and fits it. It touches every major part of ProcessDB in about half an hour. Substitute your own biology freely — the sequence of steps is what matters.

The example: substrate S in the cytosol is converted by enzyme E1 to intermediate M, which enzyme E2 converts to product P. You have measured M and P over time.

1. Create the model

In the Models browser, click New Model. A model called "New Model" appears and opens in a tab. Right-click its tab (or the entry in the browser) and choose Properties to give it a real name and a description.

2. Give it a compartment

Every state lives in a place. In the Places browser, click New Place and name it Cytosol — or, if a suitable place already exists, skip this and use it.

Drag the place from the browser into the diagram. It appears as a large rounded rectangle: a container the states will sit inside.

3. Add the species

In the Molecules and Complexes browser, click New Molecule for each of S, M, P, E1, and E2 — again, reuse existing molecules where you can, since that is what makes models comparable across the database.

Now drag each molecule from the browser onto the Cytosol place in the diagram. Each drop creates a state: "S in Cytosol", "M in Cytosol", and so on.

4. Draw the processes

Hold Alt and drag from the S node to the M node. Watch for the green dot that marks a valid target. Releasing creates a process with S as reactant and M as product, and opens its Properties dialog — name it S to M and close the dialog.

Repeat from M to P for the second step.

The status bar tells you what your keys will do

The Gesture Suggestions bar at the bottom of the diagram updates as you press and release modifier keys. Leave it on while you are learning.

5. Add the enzymes

Hold Alt and drag from E1 to the process node of S to M. The Add Reactant or Modifier dialog opens. Choose the role Enzyme and press Add. Do the same for E2 and M to P.

An enzyme is a modifier: it changes the rate but is not consumed, so the Stoichiometry field stays disabled. Had you chosen Reactant, it would have asked how many molecules the process consumes.

6. Tidy up and save

Use the Layout button in the diagram toolbar to arrange things, and Fit to window to see it all. Then press the Save button in the main toolbar (or File → Commit). Until you do, everything you have drawn exists only on your computer.

7. Make it quantitative

Right-click the model in the Models browser and choose Create Model Realization. ProcessDB creates a realization and seeds a starting point: a mass-action rate law for every process, and an initial condition of 1.0 for every state.

The realization opens in its own tab, and the Action Items panel opens along the bottom of the window whenever a realization has something still to fill in: it lists every rate law, initial condition, and boundary formula, with anything blank highlighted and sorted to the top.

Two things to do here:

  • Give the enzymatic steps proper rate laws. Right-click the S to M process in the diagram and choose Edit Rate Law…, then press Default… and pick Mass action / Michaelis–Menten — because the process has an enzyme, that produces a Michaelis–Menten form with a kcat and a Km. See Rate laws for the other choices.
  • Set the initial conditions. In the Action Items table, type the starting concentration for each state directly into the Formula column. Mark anything held constant — a buffered substrate, a large external pool — as a boundary state in that state's Properties.

8. Describe the experiment

In the Experiments browser, click New Experiment and name it. Open it and use the Experimental Measurements and Data tab:

  1. Add a measurement for each thing you measured — "M concentration", "P concentration" — and set its Variable to the model state it corresponds to.
  2. Add a data set under each measurement (one per replicate or condition) and give it a weight if some data sets are more reliable than others.
  3. Select a single data set and press Import to read the time course from an Excel file. You choose the sheet and drag out the region holding the times and values.

If the experiment involved a perturbation — adding substrate at t = 0, infusing something, clamping a concentration — record it on the State Protocols tab. See Protocols.

9. Join them: create the MOE

Right-click the Model Realization and choose Create MOE. Name it after the question you are asking.

In the MOE tab, press + under Experiments in MOE — or simply drag your experiment from the Experiments browser onto the list — to include the experiment you just described.

Set the Start / End Time to cover your data, then look at the Equations tab. This is the complete mathematical model: differential equations for every state, one copy per experiment, plus rate laws, protocol equations, and parameter values. If a banner says the equations predate your latest edits, press Regenerate.

10. Simulate

Press ▶ Simulate. The Parameter Explorer opens in its own window, compiles the equations, and runs the solver.

In the Parameter Explorer:

  • The Graphs panel on the left plots simulation output. Press Add to create a graph, then drop in the variables you care about — your measured data appears on the same axes as the corresponding simulated variables.
  • The Parameters panel on the right lists every parameter. Switch the Display filter to Adjustable to hide the clutter, then edit a value and press Simulate again to see what it does.

11. Fit

Once the shape is roughly right, let the optimizer do the rest:

  1. Tick the Adj box for the parameters you want estimated, and give each a Lower and Upper bound.
  2. Press Optimize, choose Particle Swarm, and press Start.
  3. Watch the objective function fall. When it settles, press Keep Best Parameters to move the fitted values into the parameter set.

See Fitting to data for weights, residuals, and what to do when a fit will not converge.

12. Save everything

Two saves are involved, and they are separate:

  • In the Parameter Explorer, press its own Save button to keep edited parameter values and your graph layout.
  • In the main window, press Save (File → Commit) to write the model, realization, experiment, and MOE to the database.

That is the whole loop. From here you can add a second Model Realization with different kinetics, add more experiments to the MOE, hand the model to the Biology Agent and ask it what the literature says you are missing, or write it up — File → Export Model Bundle… collects the diagram, equations, parameters, plots, and data into files for a paper.