ELISA data analyzer
Paste standards and sample ODs. Fits a 4PL standard curve and reads sample concentrations from it.
Paste standards and sample ODs. Fits a 4PL standard curve and reads sample concentrations from it.
An ELISA gives an optical density (OD) for every well, but what you want is a concentration. The standards, wells with known amounts of analyte, define a standard curve that maps OD to concentration. This analyzer fits that curve with a four-parameter logistic model, reports how well it fits, and reads every sample OD back through the curve to give its concentration, corrected for any dilution you made before loading the plate.
The parameters are found by non-linear least squares: a log-logit regression gives starting values, then a Levenberg–Marquardt refinement minimises the squared OD residuals of all standards.R² is 1 minus the residual sum of squares over the total sum of squares of the standard ODs. Recovery in the standards table is the back-calculated concentration as a percentage of the nominal one; values within 80–120% are the usual acceptance range.
This is the 4PL equation solved for concentration. It only works between the two plateaus, so samples with an OD at or below A are reported as below curve and those at or above D asabove curve. Samples that interpolate outside the range of your standards are marked as extrapolated, because the curve is unverified there. The concentration in the well is then multiplied by the sample's dilution factor to give the concentration in the original sample.
Standards: one per line, concentration first then OD, separated by spaces, tabs or commas. Replicate ODs can follow as extra columns and are averaged. Include the zero standard or blank if you have it. Units are whatever your standards are in (ng/mL, pg/mL, U/mL) and the results come back in the same units.
Samples: one per line, a name, the OD and an optional dilution factor. If the name is missing the samples are numbered. If you subtracted a blank from your standards, enter the same blank OD in the field under the curve and it is subtracted from every OD before fitting.
The same 4PL model gives potency values in the IC50 calculator, and total protein from a linear standard curve is covered by theBradford assay calculator.
At least four points, and ideally six to eight two-fold dilutions plus a blank, spanning both plateaus. The four parameters (bottom, top, midpoint and slope) cannot be estimated from fewer points, and the fit is only reliable inside the range of the standards.
ELISA signal is sigmoidal on a log concentration scale: it flattens at low concentrations (background) and at high concentrations (saturation). A four-parameter logistic follows that shape over the whole range, whereas a straight line only fits a narrow middle section.
How much of the variation in the standards' OD the curve explains. Values above 0.99 are typical for a good ELISA curve. A lower value usually points to a pipetting error in one standard, a saturated top standard or a wrong blank.
The sample OD is outside the range covered by the standards, so its concentration cannot be interpolated. Dilute the sample further if it is above the curve, or concentrate it or use a more sensitive assay if it is below.
The concentration read from the curve is the concentration in the well. Multiplying by the dilution factor gives the concentration in the original sample. A sample diluted 1:10 before loading has a dilution factor of 10.