A Novel Data-Driven Boolean Model for Genetic Regulatory Networks

Abstract

A Boolean model is a simple, discrete and dynamic model without the need to consider the effects at the intermediate levels. However, little effort has been made into constructing activation, inhibition, and protein decay networks, which could indicate the direct roles of a gene (or its synthesized protein) as an activator or inhibitor of a target gene. Therefore, we propose to focus on the general Boolean functions at the subfunction level taking into account the effectiveness of protein decay, and further split the subfunctions into the activation and inhibition domains. As a consequence, we developed a novel data-driven Boolean model; namely, the Fundamental Boolean Model (FBM), to draw insights into gene activation, inhibition, and protein decay. This novel Boolean model provides an intuitive definition of activation and inhibition pathways and includes mechanisms to handle protein decay issues. To prove the concept of the novel model, we implemented a platform using R language, called FBNNet. Our experimental results show that the proposed FBM could explicitly display the internal connections of the mammalian cell cycle between genes separated into the connection types of activation, inhibition and protein decay. Moreover, the method we proposed to infer the gene regulatory networks for the novel Boolean model can be run in parallel and; hence, the computation cost is affordable. Finally, the novel Boolean model and related Fundamental Boolean Networks (FBNs) could show significant trajectories in genes to reveal how genes regulated each other over a given period. This new feature could facilitate further research on drug interventions to detect the side effects of a newly-proposed drug.

Excitation and inhibition imbalance affects dynamical complexity through symmetries

“…The state of biological systems generally-and gene regulatory systems specifically-can be simply represented by Boolean states (0,1) where 0 indicates low concentration or activity and 1 indicates high concentration or activity. Following the typical construction [14], our model consists of a set of n ordered nodes with a binary state represented by a Boolean vector x(t) = (x 1 (t), x 2 (t), . .…”

“…In these models, the state of each component-a gene, protein, or RNA-is described by a binary value, and the interactions between components-binding, chemical reaction, and so on-are described by Boolean functions. Prior work has extensively studied the interaction functions [10][11][12] to model probabilistic [13] and multi-level [14] interactions, or to stabilize existing sequences of reactions [15]. Other work has focused on the intensive study of specific network topologies [2,3,[16][17][18][19] and local structures that are typically referred to as motifs [20].…”

Challenges and opportunities in network-based solutions for biological questions

“…A host of network-based biological models can capture dynamic relationships, particularly the relationship between genes, proteins and other cellular entities in gene regulatory networks (GRNs; [ 33 ]). GRNs are flexible and enable temporal representation of node states that incorporate uncertainty in stochastic (as opposed to deterministic) models, thus making them amenable to Boolean [ 34 , 35 ] and Bayesian network approaches [ 36 ].…”

Codependency and mutual exclusivity for gene community detection from sparse single-cell transcriptome data

“…Many mathematical methods have been proposed to infer GRNs from bulk transcriptome data, including the use of Boolean networks [1, 2, 3, 4], Bayesian networks [5, 6, 7, 8], mutual information [9, 10, 11] and linear regression [12, 13]. Although most of these methods enable us to capture codependency and regulatory interactions from a dataset with a limited sample size, they are not suitable for inferring regulatory relationships on the basis of temporal information or sparse expression.…”