Skip to content

Home

RMA kinetics is a library providing models for simulating synthetic serum marker dynamics in various context, including constitutive and drug induced expression.

Currently, there are three main models:

  1. ConstitutiveRMA
  2. TetRMA
  3. ChemogeneticRMA

A basic web application is available for simple testing. Please see the web app guide for more information.

Installation

A python package is available on PyPI and can be installed with uv, pip, and possibly other pip compatible alternatives.

To get started with uv, create a new project with uv init in a clean working directory, or use uv add directly. For example,

uv

uv add rma-kinetics
# or uv pip install rma-kinetics

The same can be done with standard pip,

pip

pip install rma-kinetics

Source

To install the package from source, make sure git is installed on your system and follow the workflow detailed below.

  1. Clone the repository
git clone https://github.com/szablowskilab/rma-kinetics.git
cd rma-kinetics
  1. Create a new virtual environment. Below is an example using UV to sync dependencies (Linux/Macos).
uv sync
source .venv/bin/activate # optional if using `uv add` later
  1. Install the rma-kinetics package (uv pip install -e .).

You can then import the package in scripts or make modifications to source for testing.

Quick Start

from rma_kinetics.models import ConstitutiveRMA
import matplotlib.pyplot as plt

# initialize constitutive RMA model
model = ConstitutiveRMA(
    rma_prod_rate=7e-3,
    rma_rt_rate=1,
    rma_deg_rate=7e-3
)

# simulate model and plot plasma RMA
solution = model.simulate(t0=0, t1=72, y0=(0,0))

print(f"Plasma RMA at final timepoint: {solution.plasma_rma[-1]}")
print(f"Brain RMA at final timepoint: {solution.brain_rma[-1]}")

solution.plot_plasma_rma()
plt.gcf()

All RMA models can be run by calling the simulate method detailed below.

simulate

Simulates model within the given time interval.

Wraps diffrax.diffeqsolve with specific defaults for RMA model simulation.

Parameters:

Name Type Description Default
t0 float

Start time of integration.

 required
t1 float

Stop time of integration.

 required
y0 PyTree[float]

Tuple of initial conditions.

 required
dt0 float | None`

Initial step size if using adaptive step sizes, or size of all steps if using constant stepsize.

 None
sampling_rate float

Sampling rate for saving solution.

 1
stepsize_controller AbstractStepSizeController`

Determines how to change step size during integration.

 PIDController(rtol=1e-05, atol=1e-05)
max_steps int

Max number of steps before stopping.

 4096
solver AbstractSolver

Differential equation solver.

 Kvaerno3()
adjoint AbstractAdjoint

How to differentiate.

 RecursiveCheckpointAdjoint()
throw bool

Raise an exception if integration fails.

 True

Returns:

Name Type Description
solution Solution

A solution object (parent of diffrax.Solution) with added plotting methods.

Simulations return a Solution object which can be used to inspect the results.

Please see the API reference or examples for more details.

Citation

Will be updated shortly.