Simbiology Build Model logo

Simbiology Build Model

OrganizationPopular
matlab
simbiology-build-model

Build, modify, and diagram SimBiology models — API reference, helper functions, and layout patterns. Use when constructing or editing models programmatically or visually.

Overview

Publishermatlab
Repositorymatlab-agentic-toolkit
Skill namesimbiology-build-model
Stars
1.2K
Forks
138
Bundled files
22
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 22 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by matlab on GitHub. Read the source before you install it.

Installation

Install the Simbiology Build Model AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/matlab/matlab-agentic-toolkit/tree/main/skills-catalog/computational-biology/simbiology-build-model
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/matlab/matlab-agentic-toolkit.git /tmp/matlab-agentic-toolkit
mkdir -p .claude/skills
cp -r /tmp/matlab-agentic-toolkit/skills-catalog/computational-biology/simbiology-build-model .claude/skills/simbiology-build-model
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Simbiology Build Model in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Simbiology Build Model on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Simbiology Build Model is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Build SimBiology Models

API reference, helper functions, and patterns for building, modifying, and diagramming SimBiology models. Works in all MATLAB environments (desktop, headless, batch, remote). Diagram/layout features require the Model Builder app and are activated only when the user requests visual output.

When to Use

  • Building or modifying SimBiology models (compartments, species, reactions, parameters, rules, events, doses, observables, variants)
  • Opening, saving, or loading models in the Model Builder / Analyzer apps
  • Designing or adjusting diagram layouts
  • Keywords: "build", "create", "modify", "add compartment/species/reaction", "diagram", "layout"

When NOT to Use

  • Simulation and analysis (use simbiology-simulate-model)
  • Parameter fitting, population modeling, NCA (use simbiology-fit-model)

Must-Follow Rules

1. Add helper scripts to the MATLAB path first

Run at the start of every session:

matlab
addpath(fullfile('<WORKSPACE_ROOT>', '.claude', 'skills', 'simbiology-build-model', 'scripts'));
disp('Helper scripts added to path.')

2. Only open the Builder when needed

Do NOT open the Model Builder by default. Open it when:

  • The user explicitly requests a diagram, layout, or visual
  • The input is an .sbproj file — use loadViaBuilder to preserve diagram layout/styling (sbioloadproject loses this data)

A model is fully functional without a diagram — it can be simulated, fitted, and analyzed using only the model object on sbioroot.

3. Model construction uses standard SimBiology API

Build models using addcompartment, addspecies, addreaction, etc. directly. This works in all environments: desktop, headless, batch, remote.

matlab
model = sbiomodel('MyModel'); disp(model.uuid)
comp = addcompartment(model, 'Central', 1);
addspecies(comp, 'Drug', 100);
addparameter(model, 'ke', 0.1);
rx = addreaction(model, 'Central.Drug -> null');
kl = addkineticlaw(rx, 'MassAction');
kl.ParameterVariableNames = {'ke'};

For standard PK models (1- or 2-compartment with standard dosing and elimination), use references/pk-library-guidance.md as the reference for correct parameterization, naming, and rules. When no diagram is needed, call PKModelDesign directly. When the user requests a diagram, construct the model manually following the same PK library conventions but use addAndPositionCompartment for layout control (see Rule 8b).

4. Write reactions in the biological forward direction

The diagram renders arrows on products and plain lines on reactants (based on the forward direction of the reaction string). Writing a reaction backwards produces incorrect arrows even if the kinetics are equivalent.

matlab
% CORRECT — L and R get plain lines, C gets an arrow
addreaction(model, 'cell.L + cell.R <-> cell.C');

% WRONG — same kinetics but L and R get arrows (they're "products" now)
addreaction(model, 'cell.C <-> cell.L + cell.R');

Guidelines:

  • Binding: write A + B -> C (substrates on left, complex on right)
  • Degradation/elimination: write Drug -> null (not null -> Drug)
  • Synthesis: write null -> mRNA (not mRNA -> null)
  • Transport: write Source.Drug -> Dest.Drug (source on left)

5. Always use qualified names for species and reaction-scoped parameters

Always reference species and reaction-scoped parameters by their qualified name. If any of the names are not valid MATLAB variable names, surround them with square brackets before building the qualified name.

  • Species: CompartmentName.SpeciesName (e.g., Central.Drug, Peripheral.[Drug-bound])
  • Reaction-scoped parameters: ReactionName.ParameterName (e.g., Elimination.ke)

Qualification is always exactly one level deep — the immediate parent compartment only. Multi-level paths like Body.Central.Drug are invalid in reaction strings. This is never ambiguous because compartment names must be globally unique across the entire model (SimBiology enforces this regardless of nesting depth). So Central.Drug is always sufficient.

Compartment naming rules:

  • Names must be unique across the entire model — no two compartments can share a name even at different nesting levels
  • If you need hierarchical naming, use underscores: Body_Central (not nested compartments both named Central)
  • Species names must be unique within a compartment but can repeat across different compartments (disambiguated by Compartment.Species)

6. Use modern property names (Value, Units, Constant)

SimBiology objects (species, compartments, parameters) share a unified property interface. Always use the modern names:

ModernDeprecated (do NOT use)Applies to
ValueInitialAmount, Capacityspecies, compartments, parameters
UnitsInitialAmountUnits, CapacityUnits, ValueUnitsspecies, compartments, parameters
ConstantConstantAmount, ConstantCapacity, ConstantValuespecies, compartments, parameters
matlab
sp.Value = 100;       % NOT sp.InitialAmount
sp.Units = 'milligram';  % NOT sp.InitialAmountUnits
sp.Constant = false;  % NOT sp.ConstantAmount

comp.Value = 1;       % NOT comp.Capacity
comp.Units = 'liter'; % NOT comp.CapacityUnits
comp.Constant = true; % NOT comp.ConstantCapacity

p.Value = 0.1;        % NOT redundant, but never use p.ValueUnits or p.ConstantValue
p.Units = '1/hour';
p.Constant = true;

7. Close Builder and Analyzer before sbioreset

sbioreset does NOT close these apps, leaving orphaned windows:

matlab
try mb = SimBiology.web.desktophandler.getModelBuilder();
    if ~isempty(mb) && isfield(mb,'webWindow') && isvalid(mb.webWindow), mb.webWindow.close(); end
catch, end
try ma = SimBiology.web.desktophandler.getModelAnalyzer();
    if ~isempty(ma) && isfield(ma,'webWindow') && isvalid(ma.webWindow), ma.webWindow.close(); end
catch, end
pause(1); sbioreset;

8. Diagram rules (only when user requests a diagram)

The following rules apply ONLY when the user asks for a diagram or layout. Skip all of these for pure model construction.

Model size limit (precondition): Layout helpers bail out above 400 total blocks (species + reactions). For large models, skip automated layout — use simple grid positioning instead (reactions at midpoints of connected species).

a. Use addAndPositionCompartment for diagram layout

When building a diagram, use addAndPositionCompartment instead of raw addcompartment + setBlock — it atomically creates, positions, and validates each compartment.

matlab
% speciesInfo: cell array of structs with .Name, .Value, .Position
speciesInfo = {
    struct('Name', 'Drug', 'Value', 100, 'Position', [40, 30, 50, 16]);
    struct('Name', 'DrugBound', 'Value', 0, 'Position', [140, 30, 100, 16])
};
[comp, sp] = addAndPositionCompartment(model, 'Central', 1, [20, 20, 280, 80], speciesInfo);
% sp is a SimBiology Species ARRAY — index with sp(1), sp(2), NOT sp{1}

b. Diagram build order

  1. Open the Builder first — the diagram does not exist until the Builder creates it. All simbio.diagram.* calls and addAndPositionCompartment will fail without this step.
  2. Plan [x y w h] positions for ALL compartments up front (leave 80 px gaps minimum)
  3. Build ONE compartment at a time with addAndPositionCompartment
  4. Add ALL parameters (including rule LHS targets), reactions, rules, doses, events
  5. repositionAllReactions(model) then checkDiagramLayout(model) — fix until zero violations
  6. positionAncillaryBlocks(model) — positions rule/parameter blocks in a grid to the right

c. Leave 80 px gaps between connected compartments

Inter-compartment reaction nodes (15×15) are placed in these gaps by repositionAllReactions. Without adequate gaps, reaction lines cross through compartment blocks. For compartments with many shared reactions (3+), increase to 120 px.

d. Post-placement validation is mandatory

After placing all blocks:

matlab
repositionAllReactions(model);
results = checkDiagramLayout(model);
if results.nTotal > 0
    for i = 1:numel(model.Reactions)
        pos = computeSafeReactionPosition(model, model.Reactions(i));
        simbio.diagram.setBlock(model.Reactions(i), 'Position', pos);
    end
    results = checkDiagramLayout(model);
end
positionAncillaryBlocks(model);  % must run LAST, after all objects exist

e. Always use the safe-open pattern for the Builder

Never call simBiologyModelBuilder(model) without first checking isAppOpen('builder'). If open, close it, wait 2s, then reopen.

matlab
if isAppOpen('builder')
    try
        mb = SimBiology.web.desktophandler.getModelBuilder();
        if ~isempty(mb) && isfield(mb, 'webWindow') && isvalid(mb.webWindow)
            mb.webWindow.close();
        end
    catch, end
    pause(2);
end
% If Analyzer is open, it already has a model loaded — open Builder
% without an argument so it picks up the Analyzer's active model.
% Passing a model argument when Analyzer is open can cause conflicts.
if isAppOpen('analyzer')
    simBiologyModelBuilder();
else
    simBiologyModelBuilder(model);
end

f. Never close the Builder to make modifications

The model handle is on sbioroot — all code works on the live model and updates the diagram in real time. Only close when the user explicitly asks.

Helper Functions (scripts/)

Model construction (always available)

FunctionSignaturePurpose
getModelByUUIDmodel = getModelByUUID(uuid)Look up model by UUID

Diagram & Builder (only when user requests diagram/layout)

FunctionSignaturePurpose
addAndPositionCompartment[comp,sp] = addAndPositionCompartment(model,name,cap,compPos,speciesInfo,Name=Value)Create compartment + species and position atomically. sp is a Species array — index with sp(1), NOT sp{1}. Options: FontWeight ("bold"), TextLocation ("center"), Padding (20), AutoExpand (true), AutoFixPositions (true)
checkDiagramLayoutresults = checkDiagramLayout(model)Containment + line-through-block + overlap checks
computeSafeReactionPositionpos = computeSafeReactionPosition(model,rxn)Crossing-free reaction node position
repositionAllReactionsnFixed = repositionAllReactions(model)Batch-reposition all reactions (up to 3 passes)
positionAncillaryBlocksn = positionAncillaryBlocks(model)Grid-position rule/parameter blocks to the right of compartments
openLiveBuilderopenLiveBuilder(model)Open Builder with safe-open pattern
isAppOpentf = isAppOpen(appName)Check if Builder/Analyzer is open
loadViaBuildermodel = loadViaBuilder(filePath)Load .sbproj preserving diagram
saveViaBuildersaveViaBuilder(filePath)Save from Builder preserving diagram
lineIntersectsRecthit = lineIntersectsRect(x1,y1,x2,y2,rect)Shared geometry helper (used internally by layout scripts)

checkDiagramLayout output

matlab
results.nTotal        % total violations (must be 0 before presenting)
results.nContainment  % species outside parent compartment
results.nLineThrough  % connection lines through unrelated blocks
results.nOverlap      % blocks <10px apart

API Quick Reference

Model

  • sbiomodel(name) — create model; model.uuid — unique ID
  • sbioloadproject('file.sbproj') — returns a struct with the model name as field; extract dynamically:
    matlab
    proj = sbioloadproject('file.sbproj');
    fn = fieldnames(proj);
    model = proj.(fn{1});
  • copyobj(model) — deep clone (does NOT copy diagram layout); verify(model) — check consistency
  • sbioreset — clear all models (close apps first!)

Compartments

  • addcompartment(model, name, value)
  • comp.Value, comp.Constant, comp.Units, model.Compartments
  • Units consistency: If any component has units, ALL components must have units (compartments, species, parameters). Missing units cause 'Dimensional analysis failed' errors at simulation/fit time.

Species

  • addspecies(comp, name, value)
  • sp.Value, sp.Units, sp.BoundaryCondition, sp.Constant
  • sp.Parent.Name — parent compartment; model.Species

Parameters

  • addparameter(model, name, value) — model-scoped
  • addparameter(kineticLaw, name, value) — reaction-scoped
  • p.Value, p.Units, p.Constant, model.Parameters

Reactions

  • addreaction(model, 'A -> B') — forward; 'A <-> B' — reversible
  • kl = addkineticlaw(rx, 'MassAction'); kl.ParameterVariableNames = {'k1'} (or {'kf','kr'} for reversible)
  • Multi-compartment: 'Central.Drug -> Peripheral.Drug'

Rules, Events, Doses

  • addrule(model, 'x = expr', ruleType) — 'initialAssignment', 'repeatedAssignment', 'rate'
  • Rule/reaction exclusivity: A species that is the LHS of a rate rule or repeated assignment rule cannot also appear as a reactant or product in a reaction. Use one or the other to govern its dynamics.
  • Rule LHS requirement: The LHS must be an existing species, parameter, or compartment with Constant = false. Create the parameter before the rule (not after as a fix — this ensures diagram blocks exist for layout):
    matlab
    p = addparameter(model, 'RO', 0); p.Constant = false;
    addrule(model, 'RO = Complex / (Target + Complex)', 'repeatedAssignment');
  • addevent(model, 'trigger', {'action1', 'action2'})
  • Event action requirement: Any parameter modified in an event action must have Constant = false (parameters default to true):
    matlab
    p = sbioselect(model, 'Type', 'parameter', 'Name', 'kgrow');
    p.Constant = false;
    addevent(model, 'Tumor.Cancer < 1e6', {'kgrow = kgrow * 0.5'});
  • sbiodose(name, 'schedule') / sbiodose(name, 'repeat')
  • Schedule: .TargetName, .Amount, .Time, .Rate
  • Repeat: .TargetName, .Amount, .StartTime, .Interval, .RepeatCount
  • adddose(model, d) / removedose(model, d)

Observables & Variants

  • addobservable(model, name, expression) — use ./ and .* for element-wise ops
  • StatesToLog caveat: If the observable references a constant parameter (e.g., 'Drug ./ Vd'), add that parameter to StatesToLog — otherwise it logs as NaN. Observables themselves are auto-logged when their dependencies are present (do NOT add observables to StatesToLog — it only accepts species, parameters, and compartments).
  • v = addvariant(model, name) then addcontent(v, {'type','name','prop',val})
  • v.Content, getvariant(model, name)

Simulation Config

  • cs = getconfigset(model, 'active')
  • cs.StopTime, cs.SolverType ('ode15s', 'ode45', 'sundials')
  • cs.RuntimeOptions.StatesToLog — 'all' or handle array

Selection

  • sbioselect(model, 'Type', type, 'Name', name)
  • For reactions: use 'Reaction' property (not 'Name')

Core Patterns

Model creation (standard — no diagram)

matlab
model = sbiomodel('MyModel');
disp(model.uuid)
comp = addcompartment(model, 'Central', 1);
addspecies(comp, 'Drug', 100);
addparameter(model, 'ke', 0.1);

Reactions

matlab
% MassAction (always use qualified species names)
rx = addreaction(model, 'Central.Drug -> null');
kl = addkineticlaw(rx, 'MassAction');
kl.ParameterVariableNames = {'ke'};

% Custom rate
rx = addreaction(model, 'Central.E + Central.S <-> Central.ES');
rx.ReactionRate = 'kf*Central.E*Central.S - kr*Central.ES';

% Multi-compartment transfer
rx = addreaction(model, 'Central.Drug -> Peripheral.Drug');

% Species with invalid MATLAB variable names
rx = addreaction(model, 'Central.[Drug-bound] -> Central.[Drug-free]');

Removing components

matlab
delete(sbioselect(model, 'Type', 'species', 'Name', 'Drug'));
delete(sbioselect(model, 'Type', 'reaction', 'Reaction', 'Drug -> null'));
removedose(model, model.Doses(1));  % doses use removedose, not delete

Doses

matlab
% Bolus
d = sbiodose('Dose', 'schedule');
d.TargetName = 'Drug'; d.Amount = 100; d.Time = 0;
adddose(model, d);

% Repeat dose
d = sbiodose('RepeatDose', 'repeat');
d.TargetName = 'Drug'; d.Amount = 50;
d.StartTime = 0; d.Interval = 8; d.RepeatCount = 3;
adddose(model, d);

Events, variants, observables

matlab
% Event modifying a parameter — mark non-constant first
p = sbioselect(model, 'Type', 'parameter', 'Name', 'ke');
p.Constant = false;
ev = addevent(model, 'time >= 10', {'ke = ke * 2'}); ev.Name = 'EnzymeInduction';

% Event modifying a species (species default Constant=false — no extra step)
ev = addevent(model, 'time >= 10', {'Drug = 50'}); ev.Name = 'RescueDose';

v = addvariant(model, 'HighDose'); addcontent(v, {'parameter','ke','Value',0.5});
obs = addobservable(model, 'DrugConc', 'Drug ./ Central');

Saving Models

  • Standard: save('mymodel.mat', 'model') / loaded = load('mymodel.mat'); model = loaded.model; Do NOT use sbiosaveproject (deprecated, requires base workspace hacks).
  • With diagram (.sbproj): saveViaBuilder('name.sbproj') / loadViaBuilder(path) (requires Builder open)
  • Switching models: ask user to save first → close Builder → pause(2) → simBiologyModelBuilder(newModel)
  • See references/app-lifecycle-guidance.md for full switching/coordination patterns.

Diagram Basics (only when user requests diagram/layout)

Coordinate system

Position = [x y width height] where (x, y) is top-left corner.

Standard block sizes

Block TypeDefault SizeNotes
Species[50, 16]Scale width: <=5 chars → 50, 6-12 → 100, 13+ → 130
Reaction[15, 15]
Rule[20, 20]

Compartment sizing from content

Species CountSizeNotes
1160 x 100Single species, centered
2 (isolated)240 x 170Vertically stacked
2 (in chain)400 x 100Side by side
3-5160+n*50 x 100+n*35Scale to content
6+240+n*50 x 220+n*35Row layout, multiple rows if needed

Internal padding: 30 px minimum on all sides.

Row-based species placement (3+ species)

Distribute species evenly in a horizontal row at y + height/2 - 8, with 40 px margin from compartment edges. Scale species width by name length: <=5 chars → 50, 6-12 → 100, 13+ → 130.

Species ordering by connection direction

When a compartment has 2+ species that connect to different external compartments, order them so each species faces its connections. This prevents connection lines from crossing through sibling species.

  • Horizontal neighbors: place the species connecting LEFT on the left edge, species connecting RIGHT on the right edge.
  • Vertical stacking: place the species connecting UP/LEFT on top, the species connecting DOWN/RIGHT on bottom.
  • Example: Blood has Neutrophil (connects left to Transit3) and Lymphocyte (connects upper-left to Spleen). Put Lymphocyte on top and Neutrophil on bottom so lines don't cross.

Layout rules

Hard requirements:

  1. No block overlap (except species inside compartments)
  2. Labels visible — not overlapping other blocks
  3. Horizontal/vertical alignment for same-tier compartments
  4. Inter-compartment reactions outside both compartments (in the gap)
  5. No connection lines through unrelated blocks
  6. 50 px minimum reaction-to-species distance

Preferred placement: 7. Elimination/degradation reactions inside their parent compartment (communicates the process occurs locally, not across a boundary)

Flow direction by model type

Model TypeFlow Direction
PKPDPD upper-left, PK lower-right
MetabolicTop-to-bottom
PBPKCirculation-based columns
Simple PKLeft-to-right or diagonal

References

Load on demand for detailed guidance:

  • references/layout-strategy-guidance.md — strategy selection, pre-build checklist, 7 recipes
  • references/pbpk-layout-guidance.md — PBPK circulation and ACAT chain layouts
  • references/evacuation-procedure-guidance.md — 5-phase rearrangement for existing models
  • references/api-cheatsheet-guidance.md — full simbio.diagram API (getBlock, setBlock, lines, clones)
  • references/app-lifecycle-guidance.md — switching models, Analyzer coordination
  • references/diagram-styling-guidance.md — colors, fonts, cloning mechanics
  • references/pk-library-guidance.md — PKModelDesign for standard PK models

Copyright 2026 The MathWorks, Inc.


Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Simbiology Build Model AI skill do?

Build, modify, and diagram SimBiology models — API reference, helper functions, and layout patterns. Use when constructing or editing models programmatically or visually.

Why use Simbiology Build Model on TypingMind?

Because you install it once and use it with any model. Simbiology Build Model is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Simbiology Build Model in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/matlab/matlab-agentic-toolkit/tree/main/skills-catalog/computational-biology/simbiology-build-model. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Simbiology Build Model?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Simbiology Build Model?

As many as you like. As long as a model supports skills, you can use Simbiology Build Model with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Simbiology Build Model AI skill free?

It is published on GitHub by matlab. Check the repository for licensing terms. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇