Potency Assays

STATLIA MATRIX performs the three potency methods referenced in the USP 1032, 1033, 1034 and EP 5.3 guidelines.

See your data in STATLIA MATRIX.

Relative Potency and Parallelism Testing

pot8The relative potency of a test sample is the amount of biological activity it produces compared to an equal amount of a reference standard under the same conditions.  Relative potency is measured between a series of dilution doses from both materials. The responses from both dilution curves are constrained to fit regression curves having an identical curve shape that provides the best fit for both curves. The distance between the two curves on the log dose axis is the log of the relative potency, and its antilog is the relative potency. This relative potency is equivalent to the ratio of the IC50s between the two curves. Since the two curves are constrained to the same shape, all ICnn ratios will be the same as the IC50 ratio (e.g. black bars between constrained curves in graph on right). Determining the relative potency from unconstrained curves is not accurate since the two curves are never precisely the same shape and ICnn ratios will vary throughout the length of the curves​. An estimation of the confidence limits of the relative potency determination between two dilution curves is required to measure its reliability. These confidence limits are de­termined using the Monte Carlo method (see the Tech Note: Monte Carlo Method), which is more stable and precise than linear approximation or profile method estimates. The log of the upper/lower confidence limits of the relative potency estimate, the RP CL Ratio, provides a stable metric of reliability across a wide range of relative potency values.

Testing for parallelism (similarity) between the two dilution curves is a prerequisite for determining the relative potency of two bioactive substances in biological systems. When the two substances are not parallel (not similar), there is no meaningful relative potency between the reference standard and the test sample. Parallelism testing is also used for matrix effects (linearity), cross-reactivity, interfering substances, concentration estimation, and inhibition studies.

The dilution series can span a full dose response nonlinear curve or a more limited linear set of doses.  The advantage of a full dose response curve is that some differences between substances only appear at high or low doses, and relative potency is more accurately determined.  But the more limited number of doses needed for a linear comparison is an advantage for animal studies and requires simpler math computations.

The three main approaches to assessing the parallelism between substances are the Equivalence Method (empirical test), the Chi-Square Method (direct measure of parallelism), and the F Test (hypothesis test).  The first method compares the confidence intervals of the regression parameters, and the other two approaches utilize the residual method from regression statistics (also called the extra sum of squares method).   All of these methods are cited in the USP 1032, 1033, 1034 and EP 5.3 guidelines and are available in STATLIA MATRIX.

Each Potency Method Uses the Most Accurate Computations In the Industry

Residual Parallelism

Residual Parallelism includes the Chi-Square Method and the F Test and is a measure of the similarity between the weighted residuals2 of the individual dilutions of the two 5PL or 4PL curves or the linear regression of two lines.

With the Chi-Square Method, STATLIA MATRIX uses the weighted residuals2 between the observed points and their respective curves to provide an actual measure of the amount of parallelism between the two curves or lines to determine parallelism.

Confidence Interval Parallelism

Confidence Interval Parallelism (also called the Equivalence Method), is a measure of the similarity between the asymptotic end points plus the slope of the inflection point of the two curves. For linear regression, the similarity of the slopes is measured.

STATLIA MATRIX uses Monte Carlo determinations to compute the confidence interval limits of each curve or line’s coefficients and then compares them to determine parallelism.

Even Your Most Difficult Potency Assays Are Easily Handled in STATLIA MATRIX

Curves Without Plateaus at One or Both Ends

Curves without plateaus

Some assay curves don’t reach saturation plateaus within a usable dose range. Potency assays with curves that do not have plateaus at one or both ends are not an issue with STATLIA MATRIX.


Hooks are often a problem when developing and running assays. And they can appear at different dilution doses from assay to assay. With STATLIA MATRIX, hooks and other nonmonotonic conditions are readily identified and can easily be masked.


Asymmetric Curves

Many bioassay curves are too asymmetric to be fit accurately with the symmetric 4PL, like the bioassay above. The 5PL curve and dilution squared residuals (left two graphs) and the 4PL curve and dilution squared residuals (right two graphs) were computed using the same data. The dilution squared residuals graphs show the squared residual errors (red bars) between the observed responses and predicted responses off the curve divided by their estimated variance for each dilution of that curve. See Tech Note: 5PL and 4PL Curve Fitting for more information.

These bad fits at the upper and lower ends of the 4PL curves show why bad fits affect parallelism and potency results. STATLIA MATRIX’s 5PL is the gold standard in analysis software and will easily fit your assays’ curves whatever their shapes. The manuscript: The Five Parameter Logistic: A Characterization And Comparison With The Four Parameter Logistic, has been requested over 5.000 times and been cited more than 300 times. STATLIA MATRIX is the industry software with the best weighting to meet the FDA guidelines for “appropriate weighting”. See the Tech Note: Curve Weighting for more information.

Ill-Behaved Assays

Ill-Behaved Assays

Developing a reliable cell-based assay can be a real challenge. STATLIA MATRIX has a wealth of tools to help you refine these tests and manage those bad boys.

Relative Potencies from The Same Unknown Are Pulled from The Database and Combined

STATLIA MATRIX automatically collects all relative potencies that have been computed for identical unknowns in any previous assay. These relative potencies are listed in each assay report along with their logarithmic and arithmetic (not shown) metrics. Failed test sample’s results can be automatically suppressed, if desired.

Control dilution curves can also be assayed. Control relative potencies are compared to reference statistics from your pooled assays or tolerance ranges set by your laboratory for Pass/Fail determinations.

Your Previously Run Assays Show If Your Potency Test is Stable and Reliable

Standard Curve RSSEs

Unknown Curve RSSEs

Rel Pot Confidence Limit Distribution

Use STATLIA MATRIX graphs and metrics of the data from a pool of your previously run assays to determine the stability and reliability of your potency test.  The standard and unknown Fit RSSE distribution’s show the suitability of the test for reliable potency determinations.  The red line shows the acceptance thresholds, either computed automatically by the software or set by the laboratory. The cumulative percentage of curves at each level of Fit RSSE is plotted on the right axis in green.

The relative potencies (X-axis) versus their confidence limit ratios (Y-axis) are plotted for all unknowns (blue dots) in the pooled previously run assays. The log of the upper/lower confidence limits of the relative potency estimate (RP CL Ratio), determined using Monte Carlo estimations (see the Tech Note: Monte Carlo Method), show the reliability and stability of the relative potency results across the range of values tested.  A rise in the RP CL Ratio at either end shows the limits of reliability for relative potency determinations.