Identification of proteins that can be therapeutically targeted is an important problem in molecular biology. Transcriptomics approaches such as coexpression network analysis have been previously proposed as tools facilitating drug targets discovery. To assess whether coexpression network analysis is applicable to prediction of novel anticancer drug targets, we compared known targets of 103 antineoplastic drugs with those of 776 drugs irrelevant to cancer in terms of their position in the coexpression network of glioblastoma ~ one of the most malignant human cancer types. Affymetrix GeneChip expression data for 93 glioblastoma surgery samples were analyzed. We were able to identify coexpression modules associated with such processes as proliferation, immune response, neurotransmission, ATP synthesis, extracellular matrix formation and others. Anticancer drug targets were four-fold over-represented in the coexpression module associated with cell proliferation and mitosis relative to the other modules. Network connectivity of drug targets within the mitotic module was found to be highly correlated with the number of anticancer drugs acting upon them. Our results support the hypothesis that hubs in the mitotic module represent potential anticancer drug targets, and confirm applicability of coexpression network analysis to anticancer drug targets identification.