Data management and bioinformatics in the analysis of developing tissues

 

Chris Bristow
Department of Chemical Engineering
Princeton University

 

We are developing data management and bioinformatics tools for the analysis of gene expression in developing tissues. Gene expression in development is tightly controlled both in space and in time. Hence, the spatiotemporal component of gene expression is a crucial aspect of data analysis, interpretation, and mining. For example, to describe the results of the genome-wide transcriptional profiling assays, it is not enough to represent a transcriptional change by a scalar: one must determine the spatial pattern of expression of any given target. Also, the design of transcriptional profiling experiments for gene expression in tissues is fundamentally different from that of cell culture experiments. We will discuss these issues in light of our experimental and computational analysis of pattern formation in Drosophila egg development (oogenesis). We are using genome-wide transcriptional profiling assays in combination with genetic perturbations to identify the transcriptional responses to cell communication pathways in Drosophila oogenesis. Our goals are threefold: 1) to identify all the genes responding to two cell communication pathways that have been identified as the key regulators in patterning of the Drosophila egg, 2) to uncover the spatiotemporal patterns of expression of these genes, and 3) to determine their role in establishing the three dimensional eggshell morphology. Because of the scale of this experimental effort and the amount of generated data, we are developing database systems to store, track, integrate, and visualize the emerging heterogeneous datasets. In addition, we are working towards the elucidation of the sequence-specific regulatory patterns (in the first approximation, clusters of known transcription factors) that underlie the observed dynamics of gene expression.

 
Back to Computation Seminars