Highly efficient factorial designs for cDNA microarray experiments: use of approximate theory together with a step-up...

Area of Science:

• Bioinformatics
• Statistical Genomics
• Experimental Design

Background:

• cDNA microarray experiments require efficient factorial designs for accurate analysis.
• Existing methods for small, highly efficient designs have limitations in exactness and optimization.
• Unequal importance of main effects and interactions needs consideration in design.

Purpose of the Study:

• To develop a general method for obtaining highly efficient, small factorial designs for cDNA microarray experiments.
• To address limitations of naive discretization in achieving exact designs and improving efficiency.
• To create designs robust to dye-color effects and heteroscedasticity.

Main Methods:

• Utilized approximate theory to derive an optimal design measure.
• Developed a step-up/down procedure for discretizing the design measure.
• Focused on baseline and all-to-next parametrizations, adaptable to hybrids and others.

Main Results:

• The proposed step-up/down procedure effectively yields exact designs of stipulated sizes.
• The resulting designs demonstrate significant improvements in efficiency compared to naive methods.
• Designs exhibit robustness against potential dye-color effects and heteroscedasticity.

Conclusions:

• The developed method provides a robust and efficient approach for factorial design in cDNA microarrays.
• The step-up/down procedure overcomes discretization challenges, ensuring design exactness and efficiency.
• This approach is versatile and applicable across various parametrizations, enhancing experimental reliability.