BIOLOGY-DRIVEN CLUSTERING OF MICRO ARRAY DATA: Applications to the NCI60 Data Set

摘要

In this note, we describe an approach to the analysis of microarray data that uses significant biological information early in the process. By combining information about the biological function and chromosomal location of genes with microarray expression data, we are able to get a more comprehensive picture of the heterogeneity of different kinds of cancer. We also get information about the importance of different chromosomes and biological processes for distinguishing cancers. In general, methods that use existing biological knowledge are likely to provide more meaningful and more interpretable results than completely unsupervised methods.