Hepatocellular carcinoma (HCC) is one of the major causes of human cancer deaths worldwide. To identify alterations of the genetic program associated with human HCC, we designed a new protocol based on the high-density replica method to analyze protein kinase gene expression in normal liver, HCC, and HCC-derived cell lines. RNA was prepared for reverse transcription and cDNA was used for PCR amplification of the conserved catalytic domain of protein kinase genes. Initially, from a pair of HCC and the adjacent noncancerous tissues, we sequenced 228 samples and identified 26 genes that represent different tyrosine kinase subfamilies. High-density grid filters were then prepared to assist the identification, by hybridization, of genes that are differentially expressed in normal vs HCC samples. Eleven tyrosine kinase genes were tested, and positive signals were reliably scored by doubly offset duplicates and by two independent gene-specific probes. Of the 11 genes tested, PDGF receptor-beta, MEKK-3, axl, and FGFR-4 are preferentially expressed in tumor samples. Additionally, we analyzed protein kinase gene expression in five HCC cell lines and identified distinct kinase gene expression patterns in different cell lines. Our results suggest that multiple kinases are activated in different tumors and confirm that there is molecular heterogeneity in the mechanisms sustaining autonomous cell growth in liver tumor formation.