A new method for gene functional prediction based on homologous expression profile

It is a project with significant challenge to predict functions of genes in the post-genomics era. Most function annotation systems is available to predict functions of part genes, but the rate of annotation is very low and a large number of genes can not be annotated, what's more, the measure of credibility isn't quite sure. Aiming at the problem, we address the new concept of functional expression profile using knowledge system existed, and consider the method of gene cluster mapping associated with hub genes to predict functions of genes which are not still annotated based on the association between gene expression and gene function. At last we applied the method to colon data set. The results implied the prediction efficiency and credibility have been improved significantly by the method.