ISBIS4 Abstract

Contact Author's Name: Nam-Ky Nguyen
Title of Abstract: Experimental Designs for 2-colour cDNA Microarray Experiments
Author(s): Nam-Ky Nguyen & Emlyn Williams
Affiliation: University of New England & CSIRO Forestry

Experimental Designs for 2-colour cDNA Microarray Experiments

In microarrays the DNA probe sequences are immobilized at pre-determined positions. Then two compared target samples are labeled with different dyes, added and allowed to hybridize. After washing away the non-specifically bound targets and dyes, the signal ntensity of the remaining bound target sequences is proportional to the amount of target sequence. Different dyes are used to label the compared DNA samples. The most commonly used dyes are Cy3 and Cy5 (green and red). To measure the intensities, the array is scanned at wavelengths specific for these dyes, and the relative abundance of the samples on this array can be calculated.

Kerr & Churchill [Biostatistics (2001) 183-201] showed how the samples can be paired onto arrays with a catalogue of A-optimal incomplete block designs (IBDs). This catalogue, obtained by the graph theoretic approach, provides comprehensive listing of IBDs up to 10 varieties (type of tissues, drug treatments, etc.) with equal as well as unequal replications.

In this talk we will give quick overview on commonly used designs for microarray experiments and on IBD alternatives and describe an efficient algorithm to extend the IBD catalogue of Kerr & Churchill . We aim at IBDs with up to 100 varieties; in the case of equal replications, some comparisons are made with the design generation package CycDesigN by Whitaker et al. (http://www.ffp.csiro.au/tigr/software/cycdesign/). We also describe method of converting these IBDs into efficient row-column designs if Red/Green dye effects are assumed.