期刊名称：International Journal of Hybrid Information Technology
印刷版ISSN：1738-9968
出版年度：2016
卷号：9
期号：4
页码：291-304
DOI：10.14257/ijhit.2016.9.4.25
出版社：SERSC
摘要：Cancer is one of complex diseases that are a big threat to mankind's health and life so far, and it is well known that the somatic mutation is an important factor leading to cancer development. Finding these important somatic mutation or driver mutation is of great benefit to the gene therapy of cancer patients. However, it is difficult to distinguish driver mutations from a great number of passenger mutations because of mutational heterogeneity, which is the key factor to deal with the problem of cancer treatment. In this study, we present an efficient way Multi-population Genetic Algorithm (MPGA) integrated with chaos algorithm to find important mutated cancer genes, which can be transformed into the maximum weight submatrix problem. The experiments on the simulated and several real mutation datasets indicate that the presented methods performs more efficiently and can find more driver genes. Comparing with other relevant methods, MPGA method is proved the most robust one among these approaches. Analyzing the experimental results obtained indicates that these important pathways rediscovered play a key role in cancer development.
