Abstract
Transcription factors (TFs) are known to co-occur with cooperativity to regulate genes. Genome-wide location data can help us understand how an individual TF regulates its target gene, but how TFs cooperate to regulate their target genes still needs further study. In this study, to explore dynamic property of gene expression profiles, an approach based on the stochastic system model and the maximum likelihood estimation method is proposed to integrate genome-wide location data and expression profiles to reveal how TFs cooperate to regulate their target genes from the systems biological perspective in the yeast cell cycle. Based on a stochastic dynamic model, a new measurement of cooperativity among TFs is developed according to the regulatory abilities of cooperative TF pairs and the number of their occurrences. Our method is applied to the yeast cell cycle to successfully reveal many cooperative TF pairs confirmed by previous experiments, and other TF pairs mentioned potentially with cooperativity in our results can provide a direction for future experiments. Furthermore, our method also provides quantitative regulatory ability of the individual TF and the cooperativity of TFs to their target genes. Finally, a cooperative TF network of cell cycle is constructed based on significant cooperative TF pairs.