willfinnigan / kinetics

Python package for modelling enzyme reactions using Monte carlo sampling from parameter distributions

kinetics

kinetics is a package for modelling reactions using ordinary differential equations. It's primarily aimed at modelling enzyme reactions, although can be used for other purposes.

See the Documentation for more information.

kinetics uses scipy.integrate.odeint to solve ordinary differential equations, but extends upon this to allow the use of parameter distributions rather than single parameter values. This allows error to be incorporated into the modelling.

kinetics uses scipy's probability distributions, with a large number of distributions to choose from. Typically uniform , normal, log-uniform or log-normal distributions are used.

Documentation: ReadTheDocs

Github: kinetics

Requirements: NumPy, SciPy, matplotlib, tqdm, pandas, SALib, seaborn, and deap.

Installation: pip install kinetics

Citation: Finnigan, W., Cutlan, R., Snajdrova, R., Adams, J., Littlechild, J. and Harmer, N. (2019), Engineering a seven enzyme biotransformation using mathematical modelling and characterized enzyme parts. ChemCatChem.

Features

• Construct systems of ODEs simply by selecting suitable rate equations and naming parameters and species.
• Use either simple parameter values or probability distributions.
• Run sensitivity analysis using SALib
• Easily plot model runs using predefined plotting functions
• Optimisation using genetic algorithm using DEAP (coming soon)

kinetics's Issues

Add more reaction types

Make Sensitivity analysis one command to run

Get setup to use readthedocs, using sphinx

Dealing with non-uniform distributions, and implement other SA methods.

Currently only the Sobol Sensitivity Analysis is implemented through SALib. Implement other methods?

SALib only takes uniform distributions, so using other distributions is a bit of a problem?

If we have a normal distribution can we just approximate this as a uniform distribution for use in SALib by taking the 5th and 95th percentile as the lower and upper bounds?

Method to allow reaction rates to reference other reaction rates

So I'm trying to use this library to represent a comprehensive E.Coli metabolism model but the rate equations I have for numerous reactions are given as being dependent on the rate of another reaction, how would I implement this in the current system, and is there any plans to make this more integrated into the model somehow? Any help would be appreciated. Thanks.

Add tests and exceptions

Add multivariate normal distribution