Signal Treatment and Analysis

Sharing Meaningful and Relevant Assay Data

Make the most out of HTRF® assay data

Learn how to ensure the maximum significance of result interpretation by addressing the following points:

  • How to perform a ratiometric data analysis step that will clear results from background or compound interference, medium effects or pipetting variations
  • How to perform 4PL 1/y^2 curve fitting for cytokine assays in order to accurately measure samples across wide ranges of concentrations.

hand dude
hand fam
hand meclunette
black-hnd black

Ratiometric data analysis: a straightforward way to eliminate compound interference or normalize data between assays

What it is and what it adds to data analysis

The ratiometric analysis of data is a unique feature of HTRF assays which result into significant improvement of data quality. It relies on measuring fluorescence at 2 different wavelengths from the donor and the acceptor (See table 1) and processing the resulting signals into a single value that compensates for the following risks:

  • Well-to-well variations that may arise from pipetting error or imprecision.
  • Compounds and/or media components added in the plate that may change the photophysical properties in a given well, and the degree to which this occurs can vary from sample to sample.

Correction by the ratiometric analyses will provide more robust datasets between replicates (intra-assay) or between assay runs (inter-assay).

XL665 d2 Green dye
Eu3+ cryptate 620 nm
665 nm
620 nm
665 nm
Tb3+ cryptate 620 nm
665 nm
620 nm
665 nm
620 nm
520 nm

Table 1: recommended wavelengths to measure for the ratiometric reduction of data

How it works: the ratiometric analysis in action

Data analysis: case of standard curves in a competitive assay

Ratio and delta ratio

Five standards were plated and incubated with HTRF reagents. Their emissions at 665 nm (Acceptor) and 620 nm (Donor) were measured after incubation (Table 2.a).

The ratio must be calculated for each well individually. The mean and standard deviation can then be worked out from replicates. A 10^4 multiplying factor is introduced for easier data processing (Table 2.b)

HTRF ratio formula

The delta ratio (ΔR) reflecting the “specific signal” is obtained by simply subtracting the background from the signal of each positive point (Table 2.c).

HTRF delta ratio formula

A 665 nm
B 620 nm
Background 2040 40765
Std 0 45999 41442
Std 1 40615 41000
Std 2 29212 41732
Std 3 15249 40124
Std 4 6258 39124

a. Standard emissions at 665 nm and 620 nm

Arrow rightArrow down


b. HTRF ratio

Arrow rightArrow down

Delta ratio

c. Delta ratio

Table 2: Ratiometric reduction of a standard curve

Assay window

The window is obtained by dividing the maximum signal ratio value by the minimum signal ratio value (Table 3).

HTRF assay window formula
Background 500
Std 0 11100 (Max)
Std 1 9906
Std 2 7000
Std 3 3800
Std 4 1600 (Min)

Arrow rightArrow down

Assay window
Table 3: Determination of the assay window

Data normalization for comparing two assays

Delta F for inter-assay comparisons

Delta F is used for the comparison of day-to-day runs of the same assay or assays run by different users. It reflects the signal to background of the assay. The negative control plays the role of an internal assay control.

HTRF delta F formula

The following table only exemplifies the normalization of one of the assays (#1), but both assays compared should be treated this way.

Assay #1 Channel A 665 nm Channel B 620 nm Ratio
Background 2140 42765 500
std 0 75241 43242 17400
Std 1 69319 45100 15370
Std 2 49115 44732 10980
Std 3 25098 43124 5820
Std 4 9991 43924 2275

Arrow rightArrow down

Table 4: Determination of Delta F

Delta F / Delta F max enables the comparison of two experiments

This calculation is used for normalizing the signal in competitive assays. This is done for both assays.

HTRF delta F ratio formula
Assay #1 ΔF
std 0 3377%
Std 1 2971%
Std 2 2094%
Std 3 1063%
Std 4 355%

HTRF assays before adjustment with delta F ratio

Arrow rightArrow down

Assay #1 ΔF/ΔF max
std 0 100%
Std 1 88%
Std 2 62%</td >
Std 3 31%
Std 4 10%

HTRF assays adjusted with delta F ratio

Table 5: Determination of ΔF/ΔF max

Determination of the negative control

Sandwich and direct binding assays

Sandwich assays’ negative control should involve both antibody-coupled HTRF reagents to test for their respective specificity and ensure they do not generate FRET signal in the absence of their target proteins.

When performing a direct binding assay (immunocompetitive assays), we recommend you perform a cryptate blank negative control with all assay components but the acceptor conjugate.

Negative control for two assay formats
Figure 1: Negative control for sandwich assays and direct binding assays

4PL 1/y2 fitting for Cisbio cytokine assays

The 4 Parameter Logistic (4PL) curve is the most commonly recommended curve for fitting an ELISA standard curve (Fig. 1 for example).

4PL regression enables the accurate measurement of an unknown sample across a wider range of concentrations than linear analysis, making it ideally suited to the analysis of biological systems like cytokine releases. This is especially true in the low-end concentrations of the standard curve, where data points would be “lost” in a linear regression.

No need for a degree in Statistics to use this equation and analyze data. Software programs like Prism or Excel allow you to run a 4PL analysis without getting into the math, and there are free online software able to run this analysis.

4PL 1/y2 curve of a cytokine assay
Figure 1: examplified 4PL 1/y2 curve of a cytokine assay. Note that standard concentration is NOT in a logarithmic scale

Even though linear regression is easy to use and can be run with a very low number of standard points, it is not considered the best fitting method for biological phenomena like cytokine release, especially in an immunoassay. The main drawback is that it is only applicable for samples that fall within the linear range of the assay, thus reducing analysis flexibility (dilutions …).

The 4PL equation includes 4 variable parameters related to the curve:

  • Estimated response at concentration zero
  • Estimated response at maximal signal
  • Slope factor
  • Mid-range concentration (or “point of inflexion”)

To get the most out of your data, we add a 1/y2 weighting to the equation, thus making it a 4PL 1/y2 fitting. The 1/y2 correction basically considers the changes of variance occurring with an increase in signal and provides.

  • A broader range of concentrations for analysis
  • Accuracy in the low/high ends of the standard curve

How to run 4PL 1/y2 analysis

Graphpad Prism V7 of higher

If you are using GraphPad Prism 7: choose the “four parameters” equation available in the software. Among the 2 equations suggested, we recommend the use of “four parameters” equation, the other “4PL” equation is NOT SUITED for analyzing cytokine assays.

Graphpad Prism Version lower than 7

If you are using an older version than Graphpad Prism 7, we have detailed the step-by-step procedure to run the analysis in this video:

Since Graphpad Prism Version 6, the 4PL equation has been already built into the software, but if you have any doubts, you can download the following Prism file with the HTRF 4PL already built in.

Free online software

If you do not have any version of Graphpad Prism, we have tested several free online software programs and we highly recommend This free online analysis software will give you access to easy data analysis, analysis customization and export capabilities.