Probing gene expression: Sequence-specific hybridization on microarrays

Background: DNA microarrays are routinely used to monitor the transcript levels of thousands of genes simultaneously. However, the array design, hybridization conditions, and oligodeoxyribonucleotide probe sequence influence the performance of the DNA microarray platform and must be considered by data analysis. Results: We analyzed the signal intensities of GeneChip microarrays in terms of a microscopic binding model. It considers specific and non-specific transcripts, which both compete for duplex formation with perfect match (PM) and mismatch (MM) oligonucleotide probes. Intensity simulations enable us to judge the accuracy and precision of gene expression measures. The accuracy of the estimated fold changes ranks according to PM-MM > PM > MM, whereas the precision decreases with PM >= MM > PM - MM, where PM-MM denotes the respective intensity difference. Conclusions: MM probes possess the potency to correct the intensity of the respective PM probe for the non-specific background. The middle base related bias of the MM intensity must, however, be considered by improved algorithms of data analysis. Moreover, the knowledge of base pair interactions suggests substituting the complementary mismatches on GeneChips by alternative rules of MM design.