Skip to Main Content
The cancer classification is a major and important study field in the medical research, and DNA microarrays have been proved to provide useful and great information for cancer classification at molecular level, compared with traditional clinical or histopathological information. Bio-markers from microarray gene expression analysis have developed to a new approach for cancer classification but still face the problems such as: different genetic signatures for same cancer under different methods; disregards of small but consistent changes in expression; and lack of biological systematic opinion. Here, this paper proposes biological pathway conducting cancer classification based on gene expression data with pathway information in KEGG. There are experiments on four different data-sets for breast, colon, gloimas and lymphoma cancer: the accuracy of these classifications are all promoted about 10%, and even achieve 100% in LOOCV on gloimas data;the pathway conducting classifiers show more significant biological functional features than genetic bio-markers.