Microarrays of surface-linked oligonucleotide probes are an emerging technology for understanding the genome-wide relationships among gene sequences and cellular functions. Microarrays can potentially measure the inter-relationships among the expression levels of all an organism's genes in parallel, enhancing our ability to use sequences to understand disease processes. Chemically synthesized oligonucleotide probes can be designed to differentiate between closely related homologues and alternatively spliced mRNAs; arrays based on probes derived from cDNA libraries cannot perform such measurements. Estimates of the degree of cross-hybridization (specificity) of that probe to other biological targets represented in the sample are essential requirements for the extraction of meaningful data and results. To provide such estimates, we have characterized the performance of in situ-synthesized oligonucleotide microarrays fabricated by a novel, highly parallel printing process. The microarrays contained 25-mer oligonucleotide probes for yeast b-actin (ACT1) and 4 homologous actin-related genes (ARP1, ARP2, ARP3 and ARP5). Cy3-labelled actin-specific targets were hybridized in the presence of Cy5-labelled targets derived from the actin-related genes. For the arrays and protocols used here, most actin-specific 25-mer probes did not hybridize to the actin-related targets. Those probes that showed detectable cross-hybridization were readily predicted using BLAST analysis. Cross-hybridizing probes could also be predicted with reasonable accuracy by including probes complementary to homologous targets on the experimental array. Low ratios of cross-hybridization control probe signal-to-target-specific probe signal were indicative of lower levels of cross-hybridization. This approach may offer a general method for experimentally verifying probe specificity.
This is a preview of subscription content, access via your institution