Validation and Routine of Ligand-Binding Assays, a Bayesian Perspective

The way to apply Bayesian modeling is illustrated to validate a ligand-binding assay such as ELISA. Hierarchical non-linear model and the associated predictive distribution of back-calculated responses allow quantifying the total uncertainty of every future measurement with the assays, through the use of precision and risk profiles. It is also shown how the obtained posterior distribution of the parameters can be used as prior for new the calibration curves during routine.