Selection of thermodynamic models for combinatorial control of multiple transcription factors in early differentiation of embryonic stem cells

摘要

Transcription factors (TFs) can have multiple combinatorial ways to regulate the transcription of a target gene. For example, one TF can help another TF to stabilize onto regulatory DNA sequence and the other TF may attract RNA polymerase (RNAP) to start transcription; alternatively, two TFs may both interact with both the DNA sequence and RNAP. The different ways of TF-TF interaction would have different effects on the probability of RNAP's binding onto the promoter sequence and therefore confer different transcriptional efficiencies. In this paper we propose a new approach to select for the thermodynamic model that best describes the TF-TF and TF-RNAP interactions for each target gene. In this approach, time-course microarray data are used to obtain the steady state concentration of the transcripts of the target gene, as well as relative changes of the active concentration for each TF. These inferred concentrations and changes of concentrations are fed into an inference scheme to identify the most compatible thermodynamic model. Such a model represents a particular way of combinatorial control from multiple TFs to a target gene. Applying this approach to time-course microarray data in embryonic stem cells, we have inferred five interaction patterns among three stem cell regulators on ten target genes.