Bayesian inference based model and design for gene regulation networks

Abstract

As we are beginning to understand the mechanisms governing biological systems, we are starting to identify potential ways of guiding or controlling the behavior of cellular and molecular systems. Rationally reengineering organisms for biomedical or biotechnological purposes has become the central aim of the fledgling discipline of system biology. A key challenge in the post genome era is to identify genome-wide transcriptional regulatory networks, which specify the interactions between transcription factors and their target genes. Numerous methods have been developed for constructing gene regulatory networks from microarray data. However, most of them are based on coarse-grained qualitative models. There often are many possible alternative biological circuits that are capable of executing a particular biological function. Some topologies may be more favorable because of robustness. Distinguishing models and finding the most suitable ones is an important challenge in systems biology, as such model ranking, by experimental evidence, will help to judge the support of the working hypotheses forming each model. The objectives of this thesis include 1) to develop a binding energy based gene regulation model to quantify the transcriptional regulatory network at the steady state. 2) to propose a binding energy based gene regulation model to quantify the dynamics of the transcriptional regulatory network. 3) to introduce a new design framework to search for biological circuits that can maintain signaling sensitivity while minimizing noise propagation. 4) to propose a BRN model to analyze the dynamics of biochemical reaction system using the random graph theory. The significance and contributions of the thesis can be highlighted as follows A regulatory model is proposed to quantify the transcriptional regulatory network. Multiple quantities, including binding affinity, regulatory efficiency and the activity level of transcription factor (TF) are incorporated into a general learning model. The sequence features of the promoter are exploited to derive the binding energy. Comparing with the previous models that only employ microarray data, the proposed transcriptional regulatory model can bridge the gap between the relative background frequency of the observed nucleotide and the gene's transcription rate. Moreover, the kinetic parameters introduced in the proposed model can reveal more biological sense than some previous models can do. A Bayesian inference based regulatory model is presented to quantify the transcriptional dynamics. The model relies on a continuous time, differential equation description of transcriptional dynamics where TFs are treated as latent on/off variables and are modeled using a switching stochastic process. Multiple quantities, including binding energy, binding affinity and the activity level of transcription factor (TF) are incorporated into a general learning model. The sequence features of the promoter and the occupancy of nucleosomes are exploited to derive the binding energy. By introducing a switching stochastic process, the proposed model can not only incorporate both activation and repression, but allows any non-trivial interaction between TFs, including 'AND' and 'OR' gates. A new design framework is introduced to select the biological circuits qualifying for buffering noise without a reduction in sensitivity, focusing on cases where the noise is characterized by rapid fluctuations. By systematically analyzing three-component circuits, we rank these biological circuits and identify three basic biological motifs allowing for the buffering noise while maintaining sensitivity to long-term changes in input signals. We discuss in detail a particular implementation in the control of nutrient homeostasis in yeast. The principal component analysis of the posterior provides insight into the nature of the reaction between nodes. The biochemical reaction network model is proposed based on the complex network theory and the dynamics of the network is analyzed on the molecular-scale. Given the initial state and the evolution rules of the biochemical network, we demonstrated how the biochemical reaction network achieving homeostasis via simulation. The dynamics of the biochemical reaction network is studied in perspective of average degree and edges. The network features of biochemical reaction system were analyzed in both its initial state and equilibrium state. This work may give one direction to study biochemical reaction system using complex network theory. This thesis mainly focuses on two kinds of problems. The first one is the reverse engineering problem - to quantify the transcriptional regulatory network based on the observed gene expression data. The second one is the design problem - to find the most suitable biological circuit determining some biological function. The proposed methods and design framework have been applied to real biological datasets and demonstrated the effectiveness.

Date of Award	3 Oct 2012
Original language	English
Awarding Institution	City University of Hong Kong
Supervisor	Hanxiong LI (Supervisor)