Learning causal networks using inducible transcription factors and transcriptome-wide time series

A small molecule transcriptional induction system is used to induce transcription factors (TFs) in Saccharomyces cerevisiae. Subsequent gene expression changes are used to train parametric models and make predictions of regulator-gene connections. The majority of experiments were sampled at t = 0, 5, 10, 15, 20, 30, 45, and 90 minutes following beta-estradiol addition under phosphate limitation. Induction of a target TF is detectable in <5 minutes and reaches saturation within ~10 minutes following β-estradiol addition at a median level 53-fold higher than at t = 0 min. We estimate that at 90 min. following beta-estradiol addition, the median TF is 28.4-fold more highly expressed than the native-promoter driven TF allele. Gene expression patterns are captured over time using Agilent 2-color microarrays.