Statistical aspects of design and validation of microtitre-plate-based linear and non-linear parallel in vitro bioassays |
| |
Authors: | Hanne Zimmermann Daniel Gerhard Theo Dingermann Professor Ludwig A Hothorn |
| |
Institution: | 1. Roche Diagnostics GmbH, Mannheim, Germany;2. Leibniz University Hanover, Institute of Biostatistics, Hanover, Germany |
| |
Abstract: | Assay validation was performed using four consecutive experiments with the related statistical evaluation. A cell-based assay on microtitre plates measured repeatedly within 1 day and on consecutive days was chosen as the model. The following problems were addressed: (i) choosing an appropriate design on a plate to avoid heterogeneities, (ii) quantification of all sources of variability and (iii) selecting between linear and non-linear parallel line assays. A mixed model was used with the random factors: rows, columns and plates and fixed effect factors with either linear or non-linear parallel line models. |
| |
Keywords: | Assay validation Linear parallel-line assay Mixed model approach Non-linear parallel-line assay Variance component analysis |
|