Binding Curve Viewer

Binding Curve Viewer is a set of tools for visualizing and simulating the equilibrium and kinetics of binding reactions with 1:1 stoichiometry and no cooperativity, for planning and validating protein–ligand binding experiments.

• Online tools: https://www.eplatton.net/binding-curve-viewer/
• Documentation: https://www.eplatton.net/binding-curve-viewer/doc/

Purpose

Binding constants (K_d, k_on, k_off, IC₅₀ and K_i) are estimated by fitting a model to assay data, so their accuracy depends on the assay design, the readout noise, the fitting model and the incubation time. Binding Curve Viewer shows the exact theoretical curves and, in version 2.0, simulates complete experiments, so that a design and an analysis method can be tested before the experiment is run.

Contents

Visualization tools (Python/Bokeh). src/binding_curve_viewer_v2.py generates six interactive web pages: determination of the dissociation constant, kinetics of association and dissociation, iso-affinity plot, competitive binding kinetics (association and dissociation), and simulation of high-throughput screening. Version 2.0 calculates the exact IC₅₀ and apparent IC₅₀ in closed form.

Interactive simulation tools (R Shiny). Four standalone applications in src/shiny-apps/:

Application	Description
kd-sim	Monte Carlo simulation of saturation binding; Kd fitted with the exact quadratic equation and with the hyperbola
kin-sim	Monte Carlo simulation of association and dissociation; kon fitted with the exact second-order and the pseudo-first-order equations
comp-sim	Monte Carlo simulation of competitive binding; apparent IC50, Ki by the Cheng–Prusoff and Wang's group equations, and Ki fitted directly with Wang's exact equation
time-sim	Apparent Kd of a saturation experiment read before equilibrium, and the incubation time needed

The three Monte Carlo applications report the distributions of the fitted parameters (histograms or normal Q–Q plots), fit-outcome counts and parametric bootstrap Kolmogorov–Smirnov tests. Runs are reproducible from the random seed.

Installation

Requirements:

• Python ≥ 3.7.6 with Bokeh ≥ 3.4.0 and NumPy ≥ 1.26.0
• R ≥ 4.3 with shiny ≥ 1.13.0, bslib ≥ 0.4.0, ggplot2 ≥ 3.5.0 and minpack.lm

git clone https://github.com/ydu-sci/Binding_Curve_Viewer.git
cd Binding_Curve_Viewer
pip install -r requirements.txt

install.packages(c("shiny", "bslib", "ggplot2", "minpack.lm"))

Usage

Visualization tools (the six HTML pages are written to the current directory):

python src/binding_curve_viewer_v2.py

Simulation tools, from the R console in the repository folder:

shiny::runApp("src/shiny-apps/kd-sim")

On the web server, a Monte Carlo run is limited to 1000 simulated experiments. In local mode, 5000, 10,000 or 100,000 are available:

Sys.setenv(BINDING_APP_LOCAL = "true")
shiny::runApp("src/shiny-apps/kd-sim")

Documentation:

cd docs
pip install -r requirements.txt
make html

Repository structure

src/binding_curve_viewer_v2.py   visualization tools (Python/Bokeh)
src/shiny-apps/                  interactive simulation tools (R Shiny)
src/scripts/                     scripts for the figures of the 2024 article
docs/                            Sphinx documentation

Version 1.0

The version described in the 2024 article in J. Chem. Inf. Model. is available under the tag v1.0.0.

How to cite

Du, Y. Binding Curve Viewer: Visualizing the Equilibrium and Kinetics of Protein–Ligand Binding and Competitive Binding. J. Chem. Inf. Model. 2024, 64 (10), 4180–4192. https://doi.org/10.1021/acs.jcim.4c00130

License

The code is released under the MIT License (see LICENSE.txt). The web pages are made available under the CC BY 4.0 license.