Strip-plot designs, mixed models, and comparisons between linear and non-linear models for microtitre plate...

Area of Science:

• Experimental design
• Biostatistics
• Cell biology

Background:

• Cell culture bioassays are vital tools in biological research.
• Ensuring statistical legitimacy and practicality in bioassay design is essential for reliable results.
• Existing bioassay designs may not adequately address all sources of variability and error.

Purpose of the Study:

• To develop statistically legitimate and practical experimental designs for cell culture bioassays.
• To demonstrate the utility of strip-plot layouts in bioassay development.
• To provide a framework for improving the continuous improvement of biological assays through statistical rigor.

Main Methods:

• Application of strip-plot layouts for cell culture bioassays.
• Incorporation of randomization of experimental units.
• Utilizing linear or non-linear models, and fixed or mixed models.
• Extension of non-linear mixed models to address serial dilution error.

Main Results:

• Strip-plot designs offer statistical legitimacy and practicality for cell culture bioassays.
• These designs accommodate various statistical models (linear/non-linear, fixed/mixed).
• The approach provides protection from location effects and can handle serial dilution error.

Conclusions:

• Statistically sound experimental design, including randomization, is a powerful tool for bioassay improvement.
• Attention to statistical detail is fundamental for identifying subtle effects in biological assays.
• The proposed designs enhance the reliability and accuracy of cell culture bioassays.