Probabilistic Model-based Transcription Regulatory Network Construction
Probabilistic models are models that are built based on observed data and can be used to infer unknown relationships. In biology, probabilistic models are used to construct transcription regulatory networks based on biological experiment dataset, such as gene expression data, promoter sequences, and knowledge of transcription factor (TF) binding sites. Probabilistic models use the stochastic variables to account for measurement noise, variability in the biological system, and those not captured by general models. The probabilistic nature of the model can determine the regulation relationships that are significant in a given experimental condition. The popular probabilistic models used to construct transcription regulatory networks include Gaussian Graphical model (GGM), Probabilistic Bayesian Networks (PBN), and Probabilistic Boolean Networks (PBN) (Friedman 2004; Huang et al. 2009).
Variables and Topology
- Huang Y, Tien-Isabel IM, Wang Y (2009) A survey of statistical models for reverse engineering gene regulatory network. IEEE Signal Proc Mag 26(1):76–97Google Scholar