Identifying hot cancer and novel cell-cycle genes from literature mining and protein interaction data

Loic Royer

Dresden University, BIOTEC, Dresden, Germany

A recent genome-wide high-throughput RNAi screen on HeLa cells identified over 850 novel cell-cycle genes not previously annotated as such ( Kittler et al. Nature 2007) .

Our analysis shows that 24% of these novel cell cycle genes can be confirmed. We show that guilt-by-association in protein interaction networks leads to statistically significant prediction, whereas protein sequence homology and literature co-occurrence with cell-cycle terminology do not.

Back