The computational approach to the study of transcriptional regulation networks has become more attractive and feasible with the rapid accumulation of complete genome sequences and the advance of high-throughput expression profiling technology. In this review, current computational approaches for understanding the transcriptional regulatory network, including promoter prediction, transcription factor binding site identification, combinatorial regulatory elements prediction, and transcription factor target gene identification, are discussed. The role of comparative genomics in transcription regulatory region analysis is also reviewed. (C) 2003 Elsevier Inc. All rights reserved.