Because the liver is the primary target organ for chemicals and pharmaceuticals, evaluation of these substances' liver toxicity is of critical importance. New evaluation methods without animal testing (i.e., in vitro and/or in silico) are eagerly anticipated, both for animal welfare and for decreasing cost. Also, the importance of mechanistic interpretation of the output derived from non-animal testing has been increasing. Accordingly, we investigated the potential for evaluating liver toxicity by applying the adverse outcome pathway (AOP) concept using gene set enrichment analysis (GSEA) from gene expression (GEx) data. A case study targeting hepatocellular fatty degeneration (HFD) is reported and discussed. We first identified the events detectable in an in vitro system by comparing the GEx data from the rat primary hepatocyte (in vitro) and rat liver (in vivo) treated with a chemical with the ability to induce HFD as one of the phenotypes in a 28-day repeated-dose toxicity test. Then, the scores based on GSEA were calculated after establishing the gene sets for each event leading to HFD. As a result, the mechanistic information leading to HFD was obtained from the score calculated based on the GSEA and the usefulness of the transcriptome-driven evaluation using AOP was demonstrated.