Review articleFrom experiment-driven database analyses to database-driven experiments in Arabidopsis thaliana transcription factor research

- Transcription factors are particularly amenable for database-driven analyses.
- 47 of the 60 A. thaliana TF families in PlnTFDB 3.0 are represented in binding site studies.
- 984 plus 12 unique A. thaliana TFs are identified for which binding sites have been determined.
- It is proposed that binding sites in combinatorial elements for interacting TFs may differ from binding sites of the single TFs.
- Approaches to determine binding site specificity for interacting TFs are discussed.

Experiment-driven database analysis is employed in forward genetics to predict the function of genes assocíated with a mutant phenotype. These analyses subsequently lead to database-driven experiments involving reverse genetics to verify functional predictions based on bioinformatic analyses. Genomic transcription factors (TFs) are key regulators of gene expression by binding to short regulatory sequences and by interacting with other TFs. Currently more than 2400 TFs are predicted for A. thaliana. As DNA-binding proteins they are particularly amenable to database-driven experiments, especially when their binding site specificities are known. Databases are available for predicting binding sites for specific TFs in regulatory sequences. Since most of these bioinformatically identified binding sites may not be functional, additional experiments for identifying the actual in vivo binding sites for TFs are required. Recently, large scale approaches were employed to determine binding sites for many A. thaliana TFs. With these approaches binding sites for 984 unique TFs were determined experimentally. An area deserving further research is proposed for interacting TFs. Most of the A. thaliana genes are under combinatorial control, and in vivo interacting TFs, similar to mammalian TFs, may bind to combinatorial elements in which the binding sites vary from those detected with the single TFs.