Prediction Of Pathogenicity Of Retinal Pigment Based On Human Gene Linkage

The achievement of human genome maps marks that human being has entered post-genomic era. People begin to study the structure, expression, restore and function of genes, which provides conditions for excavating the pathogenesis of genetic diseases and for diagnosing patients’ disease genes on a molecular level.In the wake of developments in complex network, studying human diseases through biology and complex network becomes a hot research direction. We can analyze the pathogeneses of genetic diseases on a molecular level by building biomolecular networks which have biological meaning, such as protein-protein interaction network and metabolic network. Then we can predict disease genes of genetic diseases.However, the diagnostic methods of disease genes at present still have many shortcomings. This paper improves several shortcomings and predicts Retinitis Pigmentosa’s disease genes based on vertex similarity theory. This paper’s major work is as follows:(1) We proposed one new calculation methods of similarity based on shortest paths. We analyzed and compared the local calculation methods of similarity not only based on common neighbor but also based on shortest path and we also introduced global calculation methods of similarity.(2) We proposed the prediction scheme of disease gene. We analyzed and compared each gene’s topological position in three metabolic pathways of Retinitis Pigmentosa and finally we confirmd that RHO was the core gene. According to the genes and relationships neatened from STRING, we constructed a weighted human gene connectom(HGC). Finally we designed experiment to compare the prediction effect of each method.