DBD provides transcription factor predictions for more than 150 completely sequenced genomes available for browsing and download. Predictions are based on presence of sequence specific DNA binding domain assignments using hidden Markov models from the SUPERFAMILY and PFAM databases. Evaluation shows that our predictions are 97% accurate and give at least 65% coverage. Focusing on individual genomes show that DBD identifies many novel factors for example, corresponding to a 90% increase in the known mouse transcription factor repertoire. Users can browse factors by genome or domain type, search for particular factors or submit a sequence for prediction.