Prediction of Disease-Causing Genes in Breast Cancer by Graph Mining in Biological Networks

Traditional studies in breast cancer, do not locate exactly the casual breast cancer genes in the genome andthey often detect a region containing many candidates’ genes. Gene prioritization problem tries to rank thecandidate genes from most to least promising. It can lead to faster discovery of novel casual genes and canlead to better diagnostic accuracy and treatment in breast cancer.Despite many advances in medical science and biology, many people die each year from breast cancer. Thisshows that science still has a long way to go to cure this cancer. Breast cancer is a very complex and deadlydisease. Many of the genes associated with the disease are not yet known, for example, known genes in breastcancer, such as BRCA۱ and BRCA۲, account for only ۵% of breast cancer cases. In this study, proteinnetwork data source and network-based algorithms were evaluated and then Network propagation algorithmapplied to prioritize candidate genes for breast cancer.The results showed that protein networks have a significant impact on the quality of the gene prioritizationapproach. This approach outperformed previously published algorithms (e.g., DIR, and ENDEAVOUR) inevaluation metrices such as AUC, average rank, and TOP ۵% metrics.

کلیدواژه ها:

Candidate Gene Prioritization ، Biological Complex Network ، Graph mining ، Breast Cancer ، Network Propagation