Quantitative Bioinformatics from Imaging Data

Invited Talk presented on July 9, 2008 by Ruben Schilling at MPI for Developmental Biology, Department Evolutionary Biology, Tübingen, Germany.

Abstract: Mutant phenotypes of many popular model organisms are extensively used to provide evidence of gene function. Due to a revolution in high-throughput microscopy, digital images provide high content, quantitative information to assist us in answering questions about development. In this talk we outline two new studies: We show results from our work with Christoph Dieterich on the QTL analysis of P. pacificus, where we quantify phenotypic features from image sequences. Our goal is to generate a candidate list of quantitative traits for P. pacificus lineages and to find clusters in such features. In the second part we show results from our effort towards elucidating regulatory networks in the development of mid gestation mouse embryos. Traditionally the analysis of genetic transcription rates was recorded in textual annotations from photographs of in-situ patterns or physical slices. Our approach is the quantitative analysis of expression patterns directly on 3D image data, which fits many categories of patterns. A recent study on fused in-situ and microarray data in Drosophila showed benefits for finding temporally co-expressed genes.