1. Name the response before the midpoint
EC50 asks which concentration reaches half of a defined effective response. IC50 asks which concentration reaches half of a defined inhibitory response. The label alone is incomplete: state the biological readout, direction, conditions, concentration unit, and response transformation.
Also define “50.” A relative midpoint is halfway between the fitted lower and upper responses. An absolute 50% endpoint is where a curve crosses a fixed normalized response. They need not be the same concentration.
| Term | Response question | Required context |
|---|---|---|
| EC50 | Which concentration reaches half of a defined effect? | Effect definition, model, unit, tested range |
| IC50 | Which concentration reaches half of a defined inhibition? | Activity or inhibition definition, controls, model, unit |
| Relative midpoint | Where is halfway between the fitted top and bottom? | Both fitted asymptotes |
| Absolute 50% | Where does the curve cross a fixed 50% response? | Normalization and crossing definition |
2. Calculate the two worked relative midpoints
Fit response against concentration with an equation whose orientation is stated explicitly. In the increasing example, effect = 5 + (95 − 5) ÷ [1 + (12 ÷ concentration)^1.2]. Substituting concentration = 12 nM makes the denominator 2 and returns 50%, halfway between the fitted 5% and 95% responses: EC50 = 12 nM.
For the decreasing example, remaining activity = 8 + (98 − 8) ÷ [1 + (concentration ÷ 27)^1.1]. At concentration = 27 nM, the fitted activity is 53%, halfway between 98% and 8%: IC50 = 27 nM. These values cannot rank the assays against each other because they answer different response questions.
Relative midpoint response = lower + (upper − lower) ÷ 2- Plot the retained observations and controls before fitting.
- Choose and record the exact increasing or decreasing equation, response transformation, and concentration unit.
- Fit the model without rounding its parameter estimates.
- Substitute the fitted midpoint concentration into the equation and confirm that it returns the response halfway between the fitted asymptotes.
- Check coverage, residuals, uncertainty, controls, and suitability before reporting EC50 or IC50.
EC50 and IC50 answer different response questions
EC50: increasing effect
Halfway through the fitted effect range.
IC50: decreasing activity
Halfway through the fitted activity range.
3. Preserve the controls behind a normalized response
A common inhibition calculation uses positive and negative controls to map raw response onto a percentage scale. If those controls drift, the normalized curve and midpoint can move even when test wells do not.
Retain raw response, control values, formula, and normalized response. Inspect control variability and plate position. Do not clip values above 100% or below 0% unless the approved method defines that transformation.
4. Decide whether the midpoint is reportable
Confirm that the midpoint lies within the tested concentration range and that observations provide information on both sides. Check fitted asymptotes, residual patterns, parameter uncertainty, replicate variability, control behavior, and method-specific suitability.
If the response does not reach a required plateau or the midpoint is outside the range, software may still return a number. Report the predefined qualifier or inability to estimate instead of false precision.
Limits on interpretation and comparison
Neither EC50 nor IC50 is an assay-independent property. Incubation time, biological system, substrate, readout, normalization, concentration range, and model convention can change the estimate.
Log-scale summaries are often useful because concentration ratios can be more meaningful than raw differences, but the statistical treatment must match the comparison objective. IC50 cannot be converted directly to Ki without the required binding model and experimental assumptions.
Frequently asked questions
Is a lower IC50 always a better compound?
A lower IC50 indicates greater apparent potency under that assay’s conditions, but efficacy, selectivity, mechanism, exposure, toxicity, and assay artifacts also matter.
Should EC50 and IC50 be log transformed?
Fitting is often parameterized on log concentration, and between-run summaries often use log values. The correct approach depends on the statistical objective; always report the final concentration and unit clearly.
Can IC50 be converted directly to Ki?
Not universally. Conversion depends on binding model, substrate or ligand concentration, affinity assumptions, and experimental design.