Reserch On Disease Genes Prediction Algorithm In Protein-Protein Interaction Network

Predicting disease genes is important to the prevention,diagnosis and treatment of diseases.Understanding the relationship between causal genes and associated diseases is an important topic in systems biology.Gene expression is implemented by protein synthesis which is controled by DNA.With the increasing of protein-protein interaction data,predicting disease genes based on protein-protein interaction(PPI)networks have became a hot topic in bioinformatics.Based on the PPI networks,this paper analyzes the topology similarity of the entire PPI network,studies the phenotype similarity network and known gene-disease associations,and then summarizes the existing research about disease gene predictions.In view of some existing shortages,new methods are proposed for predicting disease genes in PPI networks.In the existing field of gene prediction,several computational approaches have been proposed to prioritize potential candidate genes relying on PPI networks.In this paper,we propose a new gene prediction method,RWRHN algorithm.RWRHN algorithm calculates the topology similarity of PPI network and reconstructs PPI network by connecting genes of high similarity,then fuse the phenotype similarity network to prioritize candidate genes.The experimental results show that,the RWRHN algorithm has better stability and accuracy than other algorithms in prioritizing potential candidate genes.It is meaningful to combine the topology similarity of PPI network to prioritize potential candidate genes.PPI networks usually have high false positives and false negatives.In order to further improve the accuracy of predicting disease genes,we proposes a gene prediction algorithm,MDTS algorithm.MDTS algorithm constructs a reliable heterogeneous network by fusing multiple networks,a PPI network reconstructed by integrating a variety of biological information and topological similarity,a phenotype similarity network and known gene-disease associations,then devise a random walk-based algorithm on the reliable heterogeneous network to prioritize potential candidate genes.The experimental results show that,the MDTS algorithm has more successful predictions than other algorithms in the cross validation,and greatly increases the prediction ability of disease genes.