Functional modules based organization for association rules of gene expression profiles in cancer cell lines / 第三军医大学学报

ABSTRACT

Objective To apply apriori algorithm to mine the gene expression profiles will produce a huge number of association rules.A new approach to organize the mined association rules on the modularized gene functional category level is proposed.Methods The association rules containing the genes frequently appearing in the head parts of all the association rules were selected and grouped together.The gene functional categories,which were non-randomly enriched with the genes in the body parts of each grouped association rules,were found with statistical test and were used to form the body parts of the functional modular level association rules.Results By applying this new method to NCI60 data set,some cancer associated genes and their related functional category level expression association rules were discovered.Conclusion Gene expression profiles of diseases can be efficiently mined by association rules on the modularized gene functional category level.