RPE cells form a single cell layer underneath the photoreceptor outer segments (POS) and are responsible for regeneration of photopigments and phagocytosis in growing POS. Impairment of RPE cells is the primary cause of age-related macular degeneration (AMD), the leading cause of blindness among the elderly in industrialized countries. The cause of RPE dysfunction in AMD is not known; however, since AMD is a late onset and multifactorial disease, chronic disruption of the transcriptional machinery of RPE is likely one major mechanism underlying various RPE abnormalities. Thus better understanding RPE-specific transcription will help uncover the etiology of AMD. The overall goal of the project is to develop computational methods and algorithms that can systematically identify putative DNA motifs responsible for RPE specificity. The hypothesis is that, among other factors, RPE-specific expression is regulated by transcription factors binding to specific DNA sequences and the accessibility of these sites in the neighborhood of RPE-specific genes. Such mechanisms should be evolutionarily conserved in mammals. To identify RPE-specific genes, a data library will be created through thorough analysis of gene expression experiments. With this data library, we will then develop computational algorithms based on mathematical laws to find a combination of DNA k-mer patterns enriched in RPE-specific genes. Moreover, regulatory signals entering the nucleus first encounter chromatin, not DNA, and histones binding to DNA were also found in a sequence-specific manner. We thus characterize only those sites compatible with relatively """"""""free"""""""" chromosomal regions and consistent with genes in the test dataset as potential candidate sites of cis-regulation. Computational predictions will be verified experimentally in the laboratories of collaborators. Taken together, these approaches will allow us to identify key regulatory elements leading to RPE specificity. It will allow eye researchers to address questions that are currently not easily answered. It is also anticipated that the resulting computational approaches would be applicable much more broadly. ? ?
