Purchase this article with an account.
D.J. Zack, N. Esumi, Y. Chen, Q. Wang, I. Chowers, J. Qian; Computational Identification of Regulatory Targets of Tissue–Specific Transcription Factors: Application to Retina–Specific Gene Regulation . Invest. Ophthalmol. Vis. Sci. 2005;46(13):2391.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Purpose: The identification of tissue specific gene regulatory networks can yield insights into the molecular basis of a tissue's development, function, and pathology. Our goal was to develop a computational approach for identification of the regulatory target genes of retina–specific transcription factors. Methods: The approach is based on the assumption that genes that are related to retina in terms of expression, disease, and function are more likely to be the target genes of retina–specific transcription factors. A variety of computational techniques for transcriptional target prediction, including phylogenetic footprinting and binding site location constraint, were applied to this set of genes to identify the target genes of a retina specific transcription factor, using CRX as a model system. Mobility shift (EMSA), transient transfection, and chromatin immunoprecipitation (ChIP) assays were performed by standard methods. Results: A list of retina–enriched genes was obtained by integrating Expressed Sequence Tags (EST), Serial Analysis of Gene Expression (SAGE) and microarray datasets. This resulted in a more reliable list of retina–enriched genes, as indicated by the finding that the false discovery rate, as compared to that of the individual datasets, was significantly decreased by the integration. With the combination of various motif search methods, we significantly improved the prediction of CRX target genes. In addition, we applied this approach to two other retina–enriched transcription factors, NRL and NR2E3. To assess the bioinformatics predictions, EMSA, transient transfection, and ChIP experiments were performed with five of the novel predicted CRX targets (ARR3, ABCA4, RP1, BBS4, and GUCY2D), and the results were positive (i.e. consistent with CRX regulation) in 5/5, 3/5, and 4/5 cases, respectively. Conclusions: Our results demonstrate that bioinformatics is a useful approach that can complement more traditional lab–based methods for the modeling of retinal transcriptional networks.
This PDF is available to Subscribers Only