4PL IC50 & Dose-Response Calculator
Free IC50 and EC50 calculator. Four-parameter logistic (4PL) curve fitting with Hill slope, confidence intervals and ROUT outlier removal.
How it works
Dose-response data is fitted to a four-parameter logistic model: y = Bottom + (Top − Bottom) / (1 + (x/IC50)^Hill). All four parameters are estimated — pinning the asymptotes to 0 and 100 biases IC50 whenever real data does not reach both plateaus. Fitting runs as a log-space grid search followed by coordinate descent. The 95% confidence interval is derived from the fit covariance, se = √(diag((JᵀJ)⁻¹)·σ²), and is withheld entirely when the dose range cannot identify the midpoint.
Frequently asked questions
What is IC50 and how is it calculated?
IC50 is the concentration of a compound that produces half of its maximal inhibitory effect. It is obtained by fitting a sigmoidal curve to dose-response data and reading the concentration at the curve’s midpoint. A four-parameter logistic fit estimates the upper plateau, lower plateau, IC50 and Hill slope together; fixing the plateaus at 0 and 100 percent introduces systematic error when the assay does not fully saturate at both ends.
Why is my IC50 different from GraphPad Prism?
The most common cause is that the two fits are not estimating the same parameters. If one fit holds the top and bottom plateaus fixed while the other estimates them, the midpoints will differ — often substantially. Other causes are different weighting schemes, whether outliers were excluded before fitting, and whether the concentration axis was fitted in linear or logarithmic space. Fitting log(IC50) is standard because its sampling distribution is closer to normal.
Method reference
Motulsky HJ & Brown RE (2006) BMC Bioinformatics 7:123.