Proper analysis is crucial for transforming raw optical density (OD) readings into meaningful quantitative results. The process relies on generating a standard curve from known concentrations to interpolate the unknown sample values.
Collect OD Values: Export the absorbance readings from your microplate reader for the appropriate wavelength (e.g., 450 nm for TMB with acid stop).
Average Replicates: Calculate the average OD value for each standard, control, and sample duplicate (or triplicate). Also, calculate the average for your blank well(s).
Blank Correction: Subtract the average OD of the blank wells (which contain only substrate and stop solution) from the average OD of every other well on the plate. This yields the corrected absorbance.
Corrected OD = Average Sample OD - Average Blank OD
Note: Many modern plate reader software and ELISA analysis packages can perform this step automatically.
The standard curve is the heart of ELISA quantification. It represents the relationship between the known concentrations of the standards and their corrected OD values.
Plot the Data: On a graph, plot the known concentration of each standard on the x-axis (usually logarithmic) against its corrected average OD on the y-axis.
Select the Curve Fit: The relationship is rarely linear across the entire range. The most appropriate curve fit must be selected. The best fit is typically determined by the software to achieve the highest R² value (closest to 1.0).
Four-Parameter Logistic (4PL) or Five-Parameter Logistic (5PL) Curve Fit: This is the gold standard for immunoassays like ELISA. It models the sigmoidal (S-shaped) curve, accurately capturing the non-linear relationship at the upper and lower asymptotes (saturation and background levels).
Log-Log Linear Fit: Sometimes used for a simpler approximation, but it is less accurate at the extremes of the curve compared to 4PL.
You must never use a simple linear regression on non-linear data, as it will produce highly inaccurate results.
Assess Curve Quality: A good standard curve must meet two key criteria:
High R-squared (R²) Value: This value indicates how well the curve fits your data points. An R² value of >0.99 is generally excellent and expected for a reliable assay. An R² value below 0.98 may indicate a problem with the standards or the assay itself.
Accurate Point Replication: The replicates for each standard point should be very close together (low %CV). High variation suggests pipetting errors or uneven washing.
Once a robust standard curve is generated, the software uses its equation to calculate the concentration of your unknown samples.
Interpolation, Not Extrapolation: For each sample's corrected OD value, the software finds the corresponding concentration on the standard curve. This is called interpolation.
Critical Rule: You can only reliably calculate concentrations for samples whose OD values fall within the range of your standard curve (i.e., between the lowest and highest standard points).
If a sample's OD is above the highest standard, the sample must be diluted and re-run. Report the result as ">Upper Limit of Quantification (ULOQ)".
If a sample's OD is below the lowest standard (but above the blank), the concentration is too low to quantify accurately. It may be reported as "<Lower Limit of Quantification (LLOQ)" or "Below the range of detection". Do not extrapolate the curve, as this is highly unreliable.
Apply the Dilution Factor: If you diluted your samples prior to adding them to the plate, you must multiply the concentration obtained from the curve by the dilution factor.
*Example: If you diluted a serum sample 1:10, and the result from the curve is 45 pg/mL, the actual concentration is 45 x 10 = 450 pg/mL.*
| Standard Point | Concentration (pg/mL) | Avg. Corrected OD |
|---|---|---|
| S1 | 0 | 0.015 |
| S2 | 31.25 | 0.095 |
| S3 | 62.5 | 0.250 |
| S4 | 125 | 0.600 |
| S5 | 250 | 1.200 |
| S6 | 500 | 1.950 |
| S7 | 1000 | 2.300 |
Unknown Sample: Corrected OD = 0.850
Curve Fit: 4-Parameter Logistic (4PL)
Software Calculation: The analysis software plugs OD=0.850 into the 4PL equation of the standard curve and returns a value of ~180 pg/mL.
Dilution Factor: If the sample was diluted 1:5, the final concentration is 180 pg/mL * 5 = 900 pg/mL.
Most modern plate readers come with dedicated software (e.g., SoftMax Pro, Gen5) that automates all these steps: blank subtraction, replication averaging, curve fitting, and sample concentration calculation. You can also perform this analysis in graph-making software like GraphPad Prism (highly recommended), Excel (with add-ins for 4PL), or other statistical packages.
By rigorously following these steps, you ensure your ELISA results are both accurate and reproducible.