In silico experiment

• Yeast S. cerevisiae, 6221 genes, 80 expression conditions for each (from P. Brown’s lab)

• Cluster the profiles

• Take 600 bp upstream sequences for each gene

• Look for overrepresented sequence patterns in all clusters of upstream sequences

• Estimate the statistical significance of the patterns (repeat the pattern discovery on randomized data) - Group the similar patterns together and generalize to profiles each group

• Search the discovered patterns in TF databases and literature