Microarray experiments

Microarray Platform

To measure the gene expression level at mRNA level, we took on the Affymetrix GeneChip® system. We used the Mouse Genome 430 2.0 array and Human Genome U133 Plus 2.0 array platforms for transcriptome profiling of mouse and human samples, respectively.

Probe annotation

The probe sets were annotated as follows:
The eleven PM probe sequences of a set were aligned to sequences on RefSeq , GenBank and UniGene using BLAST program. When more than 9 of the 11 probes in a set perfectly matched with a given transcript, we judged this set to be able to specifically recognize the transcript. If they matched with reverse strand of a given transcript, we judged this set as the antisense targeting probe set. From the matching between probe sets and transcripts, we assigned the gene identifier (GeneID) that associated with the transcript to the probe set. If a probe set assigned more than one GeneID, we judged the probe set possibly cross-hybridize transcripts derived from different gene loci. We categorized all probe sets into six classes as shown in the below table.

A single gene, matched RefSeq entry
B single gene, matched GenBank entry but not RefSeq entry
C single gene, matched only UniGene entry
D single gene, anti-sense (possibly designed to the complementary strand of the intended reference gene)
E multiple genes
X nonqualified

You can retrieve the information regarding gene annotation of each probe set from Probe annotation in RefDIC Tools. The annotation is going to be updated once a couple of months.

Data processing algorithms

We used two types of algorithm: MAS5 (the standard algorhithm provided by Affymetrix, see here for more detail) and gcRMA, for estimation RNA expression levels from signal intensities of each probe on GeneChip. For the calculation we used affy package implemented in Bioconductor.