Parameter sets and MOE functions¶

Parameter sets¶

A parameter set is a named collection of parameter values for one MOE. Every MOE has at least set 0, which holds the MOE-wide tuned values; further sets let you keep alternatives side by side — a fit to one data series, a published set, a hypothetical knockdown.

On the Parameter Sets tab:

• the Select Set dropdown chooses which set the tab shows, and which the Simulate button will open the Parameter Explorer with;
• Notes describes what the set is for, and is edited right there, up to 500 characters. It is the only thing that distinguishes "Set 3" from "Set 4" when a MOE is reopened months later, and it is saved with the set by the ordinary Commit;
• the table lists every parameter with its Name — a state's own name, for an initial condition — its Value, and where that value came from;
• + creates a new set, copying its starting values from set 0. − deletes the selected set — set 0 cannot be deleted.

Double-clicking a parameter name jumps to the rate law that introduced it, which is the quickest way to answer "what is Km_Glcse_P447?".

How a value is resolved¶

A parameter set does not store a row for every parameter. It stores only what you changed, and everything else falls through to a default. The From column tells you which layer answered:

This is why adding a parameter to the model never requires you to go back and fill it into existing sets: a set covers every parameter automatically, and carries rows only for the ones you have deliberately tuned.

It also means editing a parameter that a set had merely inherited creates a row for it in that set — which is the point: the set is now saying something specific about that parameter.

Sets and the Parameter Explorer¶

The Parameter Explorer works on one set at a time and has its own controls for them:

• New creates a set from the values currently loaded;
• Open switches to another set;
• Open Ref loads a second set as the reference set, so its values appear in a Ref Value column beside the working ones;
• Make Ref makes the current set the reference.

A reference set is how you compare two fits numerically rather than by eye — and how you keep a known-good set visible while you experiment.

Values edited in the Parameter Explorer are saved with its Save button. See Saving your work.

MOE functions¶

A MOE function is a named algebraic definition attached to one MOE and one of its experiments: a name, a formula, and the constants the formula uses. They are written verbatim into the end of the equation listing, so what you type is what the simulator sees.

They are the right home for things that are about how one experiment was read out rather than about the biology:

• an instrument response or a scaling between model units and assay units;
• a derived quantity you measured but the model does not have as a state — a total across compartments, a ratio, a normalised fluorescence;
• a labelled fraction expressed the way your mass spectrometer reports it.

Edit them on the MOE Functions tab. The upper table lists the functions, with a search box that reports how many of them match; press + to add one, and it is created for the experiment selected in the MOE's experiment list. Its experiment can be changed by double-clicking the Experiment column until the function has been saved; after that, make a new function for the other experiment instead.

Selecting a function shows its parameters below, in Parameters of the selected function. A new parameter starts at 1, and its Value takes numbers only — the cell turns red as soon as the text stops being a number, and a value cannot be left blank — so a function never reaches the equations with an undefined constant. Define a function once and it can be plotted, compared to data, and fitted like any other model variable.

The broader sibling is the MR function, which belongs to the realization and appears in every MOE built on it. Both are editable from this tab, on their respective sub-tabs.