Free IC50 & EC50 Calculator
Paste a concentration–response table or a whole plate, fit a four-parameter logistic curve, and read IC50 or EC50 with a 95% confidence interval, Hill slope, fitted asymptotes and Z′-factor. Several compounds at once, entirely in your browser.
IC50 with 95% CI
Weighted 4PL nonlinear regression reporting the interval, not just a bare number.
Fit diagnostics
Residual structure, extrapolated-asymptote warnings, Hill slope and Z′-factor for screening quality.
Data stays local
No account and no upload — all fitting happens in the browser session.
How to calculate IC50 step by step
- 1
Import the data: paste a plate export or a two-column concentration + response table, or upload .csv / .xls / .xlsx.
- 2
Define the dose series: enter the top concentration and the dilution factor; replicates are grouped automatically.
- 3
Choose the direction and normalisation: inhibition (IC50) or stimulation (EC50), with optional 0–100% control normalisation.
- 4
Fit the 4PL model: constrain Top and Bottom when your controls define them, otherwise let all four parameters float.
- 5
Check the fit before reporting: residual scatter, CI width, whether IC50 sits inside the tested range, and Z′-factor.
- 6
Export: CSV of fitted parameters per compound, or a PDF report with the curve and parameter table.
The 4PL model behind the number
Response = Bottom + (Top − Bottom) / (1 + (C / IC50)^HillSlope)
Bottom and Top are the fitted asymptotes — the response at zero and at saturating concentration.
HillSlope describes the steepness; values far from 1 suggest cooperativity or a dose-range problem.
IC50 (or EC50) is the inflection point: the concentration at half the fitted response window.
Fitting on log concentration keeps the parameter estimates stable and the CI symmetric on the log scale.
IC50 & EC50 FAQ
How is IC50 calculated?
IC50 is not read off the raw data — it is a fitted parameter. Response is regressed against log concentration with a four-parameter logistic model, Response = Bottom + (Top − Bottom) / (1 + (C / IC50)^HillSlope), and the inflection parameter is the IC50. Labreador fits the curve by weighted nonlinear least squares and reports the 95% confidence interval alongside the point estimate.
What is the difference between IC50 and EC50?
They are the same mathematical parameter applied to opposite directions of effect: IC50 is the concentration giving 50% inhibition (a decreasing curve), EC50 the concentration giving 50% of the maximal effect (an increasing curve). Both come from the same 4PL fit; only the sign of the Hill slope differs.
Do I need to normalise my data first?
Only if you want a relative IC50 that is comparable between plates. Set the vehicle/untreated control to 0% and the maximal-inhibition control to 100%, then fit. An absolute IC50 fitted on raw signal is still valid within one experiment.
Why is my IC50 unreliable?
Three classic warning signs: the fitted IC50 falls outside the tested concentration range, the 95% CI spans more than an order of magnitude, or an asymptote is fitted far beyond any observed response. All three mean the dose range, not the model, needs fixing. Labreador flags each of them automatically.
Is this IC50 calculator free?
Yes — no account, no trial, no upload. Concentration–response data is fitted locally in your browser, and the results can be exported as CSV or a PDF report.
Related reading: ·