GitHub

HetaSimulator is an open-source simulation and parameters estimation (fitting) platform for the Heta modeling language. The main purpose of the package is to establish the linkage between emerging QSP frameworks and fast computational methods (parallel simulations, automatic differentiation, etc.).

Introduction

Heta language is a domain-specific modeling language (DSL) for dynamic models used in quantitative systems pharmacology (QSP) and systems biology (SB). Heta models can be translated into variety of formats like Simbiology, Matlab, mrgsolve, DBSolve and many others.

This package provides the Julia-based simulation engine for Heta-based models and modeling platforms. Users can simulate QSP models in heta format as well as ODE systems in general form using HetaSimulator without additional tools.

The main features of the package are

  • rich capabilities for estimating parameters' values based on experimental datasets
  • working with multiple models including multi-scenarios fitting
  • parallel simulations
  • storing data and scenarios in the unified formats
  • datasets and scenarios import from CSV/Excel tables
  • full Heta standard support
  • storing models, scenarios and data in the Platform object for easy management of platform components
  • utilizing the features of open-source projects like Julia and SciML ecosystem.

Installation

It is assumed that you have Julia installed. The latest Julia release can be downloaded from julialang.org

To install or update HetaSimulator run the code below in Julia environment:

julia> ]
(@v1.10) pkg> add HetaSimulator

Internally HetaSimulator installs Heta compiler as an artifact.

Basic usage

Create a model in Heta format or use your Heta-based platform. Here we will use an example model with two species and one reaction.

// index.heta file in directory "my_project"
comp1 @Compartment .= 1.5;

s1 @Species {compartment: comp1, output: true} .= 12;
s2 @Species {compartment: comp1, output: true} .= 0;

r1 @Reaction {actors: s1 => s2, output: true} := k1 * s1 * comp1;

k1 @Const = 1e-3;

To read more about Heta Heta specifications

using HetaSimulator, Plots

# set the absolute or relative path to the project directory
platform = load_platform("./my_project")
# wait for the platform compilation...

# get the default Heta model
model = platform.models[:nameless]

# single simulation and plot
results = Scenario(model, (0., 1200.)) |> sim
plot(results)

Plot

# transform results to data frame
df = DataFrame(results)
...
9×4 DataFrame
 Row │ t             s1        s2           scope  
     │ Float64       Float64   Float64      Symbol 
─────┼─────────────────────────────────────────────
   1 │    0.0555525  11.9993   0.000666611  ode_
   2 │    0.611077   11.9927   0.00733069   ode_

Architecture

The user of HetaSimulator typically deals with the following three types:

  • Model - an ODE model, containing rhs, rules, initial parameters and vector of events.
  • Scenario - scenario representing a specific model's setup for simulations or fitting. This setup can include initial parameters and events, output variables etc. In case of fitting Scenario should also include experimental data. A common usage of Scenario can be model's simulation with different drugs (parameters and events setup). Different Scenario's can be united to run multi-scenarios simulations and fitting.
  • Platform - container for different Models and Scenarios.

The user can perform the following three operations with both Model, Scenario and Platform

  • sim - run a single or multi-scenarios simulations.
  • fit - fit a model to experimental data.
  • mc - run Monte-Carlo or virtual patients simulations.

See documentation for detailed overview of HetaSimulator types and functions' arguments.

Getting help

License

This package is distributed under the terms of the MIT License.

Copyright 2020-2024, InSysBio LLC

Authors and history

  • Ivan Borisov
  • Evgeny Metelkin