The microarray technology has been developed and widely used for high-throughput genome-wide measurements of gene expression. Arabidopsis thaliana has been utilized as a plant model for gene expression research since its complete genome sequence was identified in December 2000. At present, commercially produced Arabidopsis microarrays (Affymetrix and Agilent Technologies), representing more than 21,000 genes, can be adopted for analysis.
By using rice FOX lines for microarray analysis, it is possible to isolate important Arabidopsis genes producing the specific phenotypes of rice FOX lines. Therefore, microarray experiments were conducted using Arabidopsis overexpressors of useful genes identified in the rice FOX project. If gene expression profiles are obtained using overexpressors showing similar useful phenotypes, it would be possible to re-extract useful Arabidopsis genes from among genes commonly exhibiting increased or decreased expression from these data sets (Figure).
For this project, we used the Agilent Arabidopsis 2 Oligo Microarray kit (Agilent Technologies, Japan). Each RNA sample for Cy5- and Cy3-labeled cDNA probes was isolated from the aerial organ of the respective Arabidopsis overexpressors. All experiments were repeated more than twice. The hybridized, washed material on each glass slide was scanned using the Agilent DNA microarray scanner (model G2565BA; Agilent Technologies). We used the Feature Extraction and Image Analysis Software (Agilent Technologies) for establishing the location and delineation of every spot in the array. For integration of each intensity, filtration and normalization was adopted, and to calculate the expression ratio and p-value of each spot we used the default parameters. Identification of genes with reliably altered levels was achieved using a false discovery rate procedure from our experimental data set (1). The calculation was performed using the statistical analysis software R that includes a module for performing the q-value calculation. All default parameters in the q-value module were used. The genes were suggested to have altered expression levels if the q-values for differential expression were below 0.005 in 2 experiments. GeneSpring GX (Agilent Technologies) was used for the gene-clustering analysis.
1. Storey JD, Tibshirani R. Statistical methods for identifying differentially expressed genes in DNA microarrays. Methods Mol Biol 224: 149-157 (2003).