Data selection schemes

Data selection schemes define which data are considered in the analysis. You can select, for example, which data points or dilution steps you want to include. The default scheme simply selects all data points.

Note: Each assay element must have a data selection scheme assigned. And each Quantitative response assay document must contain at least one data selection scheme.

Outlier detection

Use outlier detection to detect outlying response values. PLA 3.0 supports replicate-based outlier detection (Dixon, Grubbs, and Standard deviation tests) and model-based outlier detection (Studentized residuals test). How PLA 3.0 handles statistical outliers depends on the outlier processing setting. For example, you can exclude them from further analysis or mark them as outliers.

Tip: A range of suitability tests is available to ensure results are valid.

With outlier detection turned on, you can define how PLA 3.0 should process outliers. You can, for example, exclude outlying response values from regression, or only mark them as outliers in plots and reports.

Tip: You can execute outlier tests iteratively and adjust their sensitivity.

Configuration optimization

Use configuration optimization to automatically detect the linear portion of dose-response curves. With optimization turned on, PLA 3.0 calculates all possible configurations of dilution steps as defined by the optimization parameters. If you use a partial dose range for potency calculation, configuration optimization is applied to the partial range only.

Note: Configuration optimization is only performed with the linear parallel-line model.

Two allocation ranges are available for optimization, that is, maximum range and best range. Use maximum range to find the maximum range where all suitability tests are passed, and best range to find the best range for regression using R squared.

Important: Make sure to set up all required parameters. Missing or wrong parameters can result in a large number of configurations that PLA 3.0 cannot calculate in a timely manner.

If you set up configuration optimization such that no configurations remain for testing, PLA 3.0 calculates the assay with deactivated optimization and adds a warning message to the calculation dialog and the report.

Assignment to assay elements

You set up data selection schemes once and assign them to as many elements as you require.

Figure 1. Assignment of data selection schemes to assay elements