Virtual embryos as tools for 3D gene expression analyses.
1) Life Sciences and Genomics Division, Lawrence Berkeley National Laboratory, Berkeley, CA.
2) Computer Science Division, University of California, Berkeley, CA.
3) Institute for Data Analysis and Visualization, University of California, Davis, CA.
The Berkeley Drosophila Transcription Network Project (BDTNP) is a multidisciplinary collaboration for studying the developmental regulatory network of Drosophila blastoderm embryos. One component of this project maps the blastoderm expression patterns of 37 principal developmental regulatory genes and hundreds of their targets at cellular resolution, and uses these data to model potential regulatory interactions. We have now generated such 3-D data for 24 of the principal regulators and over 80 putative target genes, the latter selected using BDTNP ChIP-chip binding data and BDGP expression data. Gene expression data in regulatory factor mutant embryos and other Drosophila species is also being collected. Because each imaged embryo contains expression information of only two genes, expression data from hundreds of embryos is mapped onto a virtual embryo to allow many genes' expression to be compared and modeled within each cohort. These virtual embryos contain nuclei placed to match the average density pattern and embryo shape for each cohort. This allows temporal comparison within each nucleus between earlier expression of regulators in one cohort to the later expression of target gene patterns in another cohort, as well as better estimates of the developmental increase in complexity. Gene expression in such virtual embryos can be viewed with our tool called PointCloudXplore, which provides realistic interactive exploration of the 3D expression data as well as abstract views for analyzing the correlation between expression patterns within the N-dimensional gene expression space.
Last modified January 8, 2007.