Information on the carcinogenic potential of chemicals is primarily available for High Production Volume (HPV) products. Because of the limited knowledge gain from routine cancer bioassays and the fact that HPV chemicals are tested only, there is the need for more cost-effective and informative testing strategies. Here we report the application of advanced genomics to a cellular transformation assay to identify toxicity pathways and gene signatures predictive for carcinogenicity. Specifically, genome-wide gene expression analysis and quantitative real time polymerase chain reaction (qRT-PCR) were applied to untransformed and transformed mouse fibroblast Balb/c 3T3 cells that were exposed to either 2, 4-diaminotoluene, benzo(a)pyrene, 2-acetylaminoflourene, or 3-methycholanthrene at IC20 conditions for 24 and 120 h, respectively. Then, bioinformatics was applied to define toxicity pathways and a gene signature predictive of the carcinogenic risk of these chemicals. Although bioinformatics revealed distinct differences for individual chemicals at the gene-level pathway, analysis identified common perturbation that resulted in an identification of 14 genes whose regulation in cancer tissue had already been established. Strikingly, this gene signature was identified in short-term (24 and 120 h) untransformed and transformed cells (3 weeks), therefore demonstrating robustness for its predictive power. The developed testing strategy thus identified commonly regulated carcinogenic pathways and a gene signature that predicted the risk for carcinogenicity for three well-known carcinogens. Overall, the testing strategy warrants in-depth validation for the prediction of carcinogenic risk of industrial chemicals in in vitro carcinogenicity assay.