bio-curve-fit 0.2.0

Curve fitting algorithms for bio-assays with scikit-learn api

bio-curve-fit

A Python package for fitting common dose-response and standard curve models. Designed to follow the scikit-learn api.

Quickstart

Installation

pip install bio-curve-fit

We recommend using python virtual environments to manage your python packages in an isolated environment. Example:

python -m venv venvname
source venvname/bin/activate

Example usage:

from bio_curve_fit.logistic import FourPLLogistic

# Instantiate model
model = FourPLLogistic()

# create some example data
standard_concentrations = [1, 2, 3, 4, 5]
standard_responses = [0.5, 0.55, 0.9, 1.25, 1.55]


# fit the model
model.fit(
	standard_concentrations, 
	standard_responses, 
)

# interpolate the response for new concentrations
model.predict([1.5, 2.5])

# interpolate the concentration for new responses
model.predict_inverse([0.1, 1.0])

Calculate and plot the curve and limits of detection:

plot_standard_curve(standard_concentrations, standard_responses, model, show_plot=True)

You can also customize the plot arbitrarily using matplotlib. For example, adding labels to the points:

from adjustText import adjust_text

fig, ax = plot_standard_curve_figure(standard_concentrations, standard_responses, model)
texts = []
for x, y in zip(standard_concentrations, standard_responses):
	texts.append(ax.text(x, y, f"x={x:.2f}, y={y:.2f})", fontsize=13, ha="right"))
# Adjust text labels to avoid overlap
adjust_text(texts, ax=ax)

Examples

See the example notebooks for more detailed usage.

Contributing

Contributions are welcome! We built this package to be useful for our own work, but we know there is more to add. Please see CONTRIBUTING.md for more information.