Optimizing Subnetwork Markers for the Classification of Breast Cancer Metastases: A Simulated Annealing Strategy

摘要：

Although the combined use of gene expression profiles and protein-protein interaction networks has demonstrated remarkable successes in the classification of breast cancer metastases, the development of computational methods for identifying predictive subnetwork markers, as the primary step of these network-based methods, remains a great challenge. Existing methods typically identify subnetwork markers using greedy search algorithms, which may sacrifice the optimality of the identified subnetwork markers and thus impair the prediction power of the successive learning machines. In this paper, we adopt a global optimization method called simulated annealing to further optimize subnetworks identified by the existing greedy search method. Experimental results show that the proposed simulated annealing algorithm can effectively improve the optimality of the subnetwork markers and thus benefit the successive learning machines for the classification of breast cancer metastases.

关键词： subnetwork optimization classification simulated annealing

作者: MingYin Peibei Shi Wangshu Zhang Rui Jiang

作者单位: MOE Key Laboratory of Bioinformatics and Bioinformatics Division TNLIST/Department of Automation, Tsinghua University, Beijing 100084 China

会议类型: 国际会议

会议名称: 2011 International Conference on Medical Information and Bioengineering(2011年医药信息与生物工程国际会议 ICMIB 2011)

会议地点: 成都

会议语种:英文

页码: 18-22

在线出版日期: 2011-01-14（万方平台首次上网日期，不代表论文的发表时间）