Universal Reference Approach to the Creation and Mining of a General Purpose Mouse Gene Expression Database

Bruce Aronow
CHRF 2048
3333 Burnet Ave
Cincinnati, OH 45229
telephone: 513 636-4865
fax:
email: bruce.aronow@chmcc.org
prestype: Platform
presenter: Bruce Aronow

Bruce J Aronow, Sarah Williams, Cathy Ebert, and a consortium of UC/CHMC investigators.
Divisions of Molecular Developmental Biology, Pediatric Informatics, and University of Cincinnati Genome Informatics Core. University of Cincinnati and Children’s Hospital Medical Center, Cincinnati, OH, USA

We have analyzed mRNA expression profiles of 81 normal, developing and disease mouse tissues using Incyte MouseGEM1 microarrays and a single common reference mRNA with a strong emphasis on adult and developing lung, cardiac, CNS, GI, urogenital, immunologic, and endocrine tissues. Duplicate Cy3/Cy5 hybridizations with Agilent Bioanalyzer-graded mRNAs
and day 1 whole mouse mRNA reference demonstrated excellent reproducibility (even with respect to genes expressed at very low level in the reference mRNA), equivalency to dye reversal, and agreement with direct sample comparisons. Use of multiple independent normalization strategies greatly improved quality assurance, optimal replicate correlations, as well as extraction of tissue, organ, and gene ontology-specific expression pattern relationships. Excluding the most over-expressed genes reduced ability to classify tissue specificity, but less so organ origin. Tissues from CNS, immunologic, and GI systems exhibited impressive expression diversity and repertoire specificity suggesting both subtle and intense commitment to tissue-specific gene expression programming. Probing for correlated genes with known biologic relationships within multiple gene ontologies and other known biologic relationships demonstrated great potential of the database to implicate functional associations and potential pathway relationships for unknown genes. These results support the hypothesis that systematic database mining by cross-comparative analysis of diverse biologic systems will greatly augment gene discovery, annotation, and pathway knowledge.
(supported by multiple NIH grants of consortium members and the Howard Hughes Medical Institute)