No abstract

No abstract

Microarray methods have become a basic tool in studies of global gene expression and changes in transcript levels. Affymetrix microarrays from the HGU133 series contain multiple probe-sets complementary to the same gene (4742 genes are represented by more than one probe-set in a microarray HGU133A). Individual probe-sets annotated to the same gene often show different hybridization signals and even opposite trends, which may result from some of them matching transcripts of more than one gene and from the existence of different splice-variant transcripts. Existing methods that redefine probe-sets and develop custom probe-set definitions use mathematical tools such as Matlab or the R statistical environment with the Bioconductor package (Gentleman et al., 2004, Genome Biol. 5: 280) and thus are directed to researchers with a good knowledge of bioinformatics. We propose here a new approach based on the principle that a probe-set which hybridizes to more than one transcript can be recognized because it produces a signal significantly different from others assigned to the particular gene, allowing it to be detected as an outlier in the group and eliminated from subsequent analyses. A simple freeware application has been developed (available at www.bioinformatics.aei.polsl.pl) that detects and removes outlying probe-sets and calculates average signal values for individual genes using the latest annotation database provided by Affymetrix. We illustrate this procedure using microarray data from our experiments aiming to study changes of transcription profile induced by ionizing radiation in human cells.

No abstract

No abstract