Expression Microarray Cross Hybridization

    Cross-hybridization can be seen in expression profiling microarrays, arising from small regions of sequence similarity found among repetitive elements as well as between members of gene families. Some of the signals as measured on microarrays may be purely artifactual for this reason and lead to misinterpretation of expression data. Individual yeast genes selected by computer-based similarity analysis were amplified, labeled and hybridized to a whole genome microarray to better quantify this effect. We found the intensity of signal overlap for other sequence-similar yeast ORFs approached that of the perfect match. This was true for both gene families and repetitive elements. In addition, it is unlikely human CoT-1 DNA will block all the repetitive elements on human microarrays equally because many repetitive sequences that occur in genomic DNA are rare in coding and vice versa.  Cross-hybridization of labeled products can be seen even among relatively short stretches of DNA (>20bp) and is an important consideration when designing microarrays.

 

    BLAST alignment of a 1200 bp segment of YBR112C showing the regions of sequence similarity responsible for cross-hybridization to other Yeast ORFs (image captured from NCBI’s BLAST output).  Dashed lines represent gaps in the aligned sequence. Descending from top to bottom, the high-homology hits were YBR112C, YOL051W, YKL054C, YBR289W, YMR043W, YMR164C, YBL081W, YKL032C, YNL161W, YPL089C, YDL005C