Thumbnails:
List:
Year:
Category:
Session:
Poster:
Getting poster data...
Jun Seita, Debashis Sahoo, David L. Dill, Irving L. Weissman (Stanford University, Stanford, CA, USA)
Gene expression profiling using microarray has been limited to profile difference of gene expression at comparison setting since probesets for different genes have different sensitivities. Conventional methods obtain “fold-change” between 2 or more samples, and if a change is not statistically significant, it is scored as “not significant” whether it is expressed or not. Thus scientists could obtain relative difference for limited number of genes and the result is comparison-pair specific. We tackled this limitation with the hypothesis that if we accumulate a very large number of microarray datasets, meta-analysis could be applied to it to compute dynamic range of each probeset. Further more, meta analysis of data distribution could reveal “actively expressed” range for each probeset with statistical significance. Then mapping individual sample data against those meta-analysis results enables to profile, not relative, but absolute gene expression. The strategy is implemented in web-based interface named “Gene Expression Commons” (https://gexc.stanford.edu/ ).