Getting started

From a column of numbers to a publication-ready fit — with the uncertainty spelled out — in about a minute.

1 · Format your data

Enter one dose, response pair per line. Values can be separated by a comma, space, or tab. Doses must be positive (a log axis can't show zero), and you need at least four points to fit four parameters. Set the dose unit (µM, nM, mM…) so the axis and the literature comparison read in the right scale.

1, 2 10, 5 100, 38 1000, 93 3000, 98

If you have replicates, enter each measurement on its own line at the same dose — the fit uses them all.

2 · Fit the curve

Press Fit curve. DoseFit returns the four parameters plus a goodness-of-fit:

  • EC50 / IC50 — the dose at the half-maximal response. For an inhibition curve this is your IC50.
  • Hill slope — how steep the transition is. Around 1 is typical; steeper means a sharper switch.
  • Top & bottom — the upper and lower plateaus the curve approaches.
  • R² — how much of the variation the fit explains.

3 · Read the uncertainty, not just the number

Every parameter shows a 95% confidence interval. This is the most important habit the tool encourages: a tight interval means the value is well-determined; a wide one means your data can't pin it down yet. A common cause is a dose range that doesn't bracket the EC50 — if the curve never crosses half-max, the EC50 is an extrapolation and its interval will balloon. The fix is experimental, not numerical: add doses that span the transition.

4 · Handle real-world data

  • Weighting — if your measurement error grows with the signal (common in biology), switch to 1/Y or 1/Y² so big responses don't dominate the fit.
  • Outlier flags — points far from the curve get a highlight ring. They are flagged, never removed — you decide whether a point is a genuine result or a pipetting slip.
  • Log axis — dose-response is almost always read on a log dose axis; toggle it to match how you'll present the figure (it doesn't change the fit).

5 · Compare to the literature

In the Compare to literature card, pick a target (or paste a ChEMBL id) and a measure (IC50/EC50/Ki). DoseFit pulls the published activity distribution and shows where your fitted potency lands — “within the published interquartile range,” “more potent than ~75% of ligands,” and so on. Remember this is the spread across all known ligands for that target, a sanity-check on scale, not a claim about your specific compound.

6 · Export & share

Download the figure as a 2× PNG for slides and print, or as true vector SVG for journals and further editing — the confidence band, EC50 marker, and a stats inset are baked in. Copy results puts the parameters and intervals on your clipboard as a single line for your notebook or methods section.

Open the fitter →   Read the methods