Boolean Networks for Gene Regulatory Modeling: A Comprehensive Review

Introduction

Gene regulatory networks (GRNs) govern the expression dynamics of genes through interactions among transcription factors, signaling molecules, and other regulatory elements. Understanding these networks is fundamental to deciphering cellular differentiation, response to stimuli, and disease progression. Boolean networks (BNs) provide a discrete, logical abstraction of GRNs that captures the essential regulatory logic while remaining computationally tractable for large-scale analysis [1]. In a BN, each gene is represented as a binary node (ON/OFF), and its state is updated according to a Boolean function that depends on the states of its regulators. This formalism has been applied extensively to model biological processes ranging from cell cycle control to immune responses, and increasingly in veterinary and comparative biology contexts [84].

The utility of BNs lies in their ability to reveal emergent dynamical properties such as attractors (stable states or cycles) that correspond to cellular phenotypes, including quiescence, proliferation, or differentiation [2]. Recent advances in inference algorithms, control theory, and integration with omics data have expanded the applicability of BNs to complex, partially observed systems [3, 4]. This review provides an exhaustive technical overview of Boolean network modeling for GRNs, with emphasis on formal definitions, dynamical analysis, inference methods, control strategies, and applications relevant to veterinary medicine and diagnostics.

Formal Definition of Boolean Networks

A Boolean network is defined as a tuple (V, F) where V = {x1, x2, ..., xn} is a set of n binary variables (nodes), and F = {f1, f2, ..., fn} is a set of Boolean functions [5]. Each node xi at time t takes a value xi(t) in {0, 1}. The update rule for node i is given by:

xi(t+1) = fi(xj1(t), xj2(t), ..., xjk(t))

where the arguments are the regulators of node i. The network topology is defined by the wiring diagram, a directed graph where edges represent regulatory interactions (activation or inhibition). The Boolean functions can be expressed in disjunctive normal form, as truth tables, or using logical operators (AND, OR, NOT, XOR, etc.) [6].

A key property of Boolean functions is canalization, where a single input value can determine the output regardless of other inputs. Canalization has been shown to stabilize network dynamics and reduce nonlinearity, making BNs more robust to perturbations [6, 7]. Kadelka et al. [81] performed a meta-analysis of published Boolean models and identified design principles such as the prevalence of canalizing functions and modularity.

Threshold Boolean networks (TBNs) are a subclass where each node's update is determined by a weighted sum of inputs compared to a threshold. TBNs have been used to model the tryptophan operon in Escherichia coli [8] and have been critically assessed for their ability to describe GRN dynamics [9, 10]. The generalization power of TBNs has been studied by Ruz and Cho [11], showing that they can capture complex regulatory logic with fewer parameters.

Dynamics and Attractors

The dynamics of a BN are governed by the update scheme. In synchronous updating, all nodes are updated simultaneously at each time step. In asynchronous updating, nodes are updated one at a time in a deterministic or stochastic order. Asynchronous schemes are often more biologically realistic because molecular processes occur at different timescales [12, 13]. Bensussen et al. [12] provided an analytical comparison of synchronous and asynchronous update schemes for biological BNs.

The state space of a BN consists of 2^n possible configurations. From any initial state, the system evolves deterministically (or stochastically under asynchronous updates) and eventually reaches an attractor: either a fixed point (steady state) or a limit cycle (periodic sequence of states). Attractors correspond to distinct cellular phenotypes, such as cell types or disease states [2]. The set of states that lead to a given attractor is its basin of attraction.

Attractor analysis is a central task in BN modeling. Tools such as biobalm [14] and GRiNS [15] enable efficient attractor detection. Pastva et al. [13] investigated why motif-avoidant attractors are rare in asynchronous BNs, revealing structural constraints. Bavisetty et al. [7] showed that attractors are less stable than their basins, a phenomenon termed the coherence gap, which is influenced by canalization.

Mutually inhibiting teams of nodes have been identified as a predictive framework for structure-dynamics relationships [16]. The concept of trap spaces (subsets of states that are closed under dynamics) is used for node and edge control identification [17]. Gedeon [58, 65] explored how network topology and interaction logic determine the set of states a network can support, using lattice structures.

Inference and Reconstruction of Boolean Networks

Inferring a BN from experimental data is a challenging inverse problem. Approaches can be categorized as data-driven, prior-guided, or hybrid. Data-driven methods use time-series transcriptomic data to learn Boolean functions via symbolic regression or feature selection [18]. Zhang et al. [19] developed LogicSR, a prior-guided symbolic regression method for single-cell transcriptomics. LogicGep [62] uses symbolic regression from time-series profiling data.

Bayesian methods have been employed for topology inference under partial observability. Alali and Imani [3] proposed a Bayesian lookahead perturbation policy for inference of regulatory networks [69], and later extended it to reinforcement learning for causal inference [96]. Bayesian optimization has also been applied to state and parameter estimation of dynamic networks with binary space [55].

Knowledge graphs can augment logical GRN inference. Li et al. [20] introduced KGBN, which uses knowledge graphs to constrain and optimize logical models. The integration of regulatory networks with genome-scale metabolic models is addressed by gMISpy [21], enabling combined regulatory-metabolic analysis.

Single-cell RNA sequencing (scRNA-seq) provides high-resolution data for BN inference. scBoolSeq links scRNA-seq statistics with Boolean dynamics [61]. The Augusta pipeline [79] converts RNA-seq data into GRNs and Boolean models. Chevalier et al. [4] demonstrated data-driven inference of BNs from transcriptomes to predict cellular differentiation and reprogramming.

Evolutionary computation has been applied to reconstruct threshold networks, as demonstrated for the tryptophan operon [8]. Integer linear programming has been used for contrasting state interventions in BNs [22]. The RBI algorithm [23] combines regulatory and metabolic network models for optimal mutant strain design.

Control and Intervention in Boolean Networks

Controlling the dynamics of BNs to achieve desired phenotypes (e.g., reversing a disease state) is a major goal. Control strategies can be classified as node control (perturbing individual genes) or edge control (modifying interactions). Murrugarra et al. [24] developed modular control approaches for BN models. Plaugher and Murrugarra [91] reviewed phenotype control techniques.

Trap spaces provide a basis for identifying control targets [17]. Integer linear programming has been used to find optimal interventions [22]. Pareto-optimal interventions using signal temporal logic were proposed by Hosseini et al. [25]. Deep reinforcement learning has been applied for intervention in partially observable regulatory networks [26]. An optimal Bayesian intervention policy in response to unknown dynamic cell stimuli was developed by Hosseini and Imani [27].

Set stabilization of logical control networks using minimum node control was addressed by Liu et al. [74]. Global stabilization of BNs with applications to biomolecular network control was studied by Rafimanzelat [28]. The concept of equilibrium interventions for regulatory networks was introduced by Hosseini and Imani [73].

Quantum algorithms have emerged for efficient attractor search [29] and logical network analysis [94]. These approaches promise exponential speedup for certain combinatorial problems in BN control.

Applications in Veterinary and Comparative Biology

Boolean network modeling has been applied to a range of biological systems relevant to veterinary medicine. In poultry, BNs have been used to model host-pathogen interactions, such as the response to Salmonella infection (see Salmonella in Chickens). The GRN of Pseudomonas aeruginosa CCBH4851 was modeled using Boolean formalism [83], providing insights into antibiotic resistance mechanisms relevant to veterinary pathogens.

In livestock, BNs have been applied to study mammary gland development and mastitis (see Bovine Mastitis Caused by Staphylococcus aureus). The tryptophan operon model in E. coli [8] serves as a paradigm for bacterial gene regulation in enteric pathogens of veterinary importance.

Comparative oncology benefits from BN models of cancer progression. Kittaneh et al. [2] used ensemble threshold Boolean modeling to identify robust attractors and regulatory drivers in pediatric leukemia, with implications for veterinary oncology. Bhattacharjee et al. [30] identified optimal drug combinations for prostate cancer using BNs, a methodology transferable to canine and feline prostate cancer.

Epithelial-mesenchymal plasticity (EMP) has been modeled using multilevel Boolean formalisms [31, 32]. These models are relevant to understanding metastasis in companion animals. The hybrid E/M phenotypes in EMP were captured by a multilevel formalism [31], and perturbation-induced EMT transitions were studied computationally [32].

Developmental biology applications include modeling of preimplantation development [33], cardiac differentiation [34], and hematopoietic stem cell differentiation [35]. These models provide insights into cell fate decisions that are conserved across mammals, including veterinary species.

Software and Tooling for Boolean Network Modeling

A growing ecosystem of software tools supports BN modeling. The following table summarizes key tools and their features:

The Mermaid diagram below illustrates a typical workflow for Boolean network modeling from data to intervention:

flowchart TD
    A["Experimental Data<br>(RNA-seq, scRNA-seq,<br>time-series")] --> B[Data Preprocessing<br>Discretization, Normalization]
    B --> C[Network Inference<br>Symbolic Regression, Bayesian,<br>Knowledge-guided]
    C --> D[Boolean Network Model<br>Nodes, Functions, Topology]
    D --> E[Dynamical Analysis<br>Attractor Detection, Basin Analysis,<br>Stability Assessment]
    E --> F[Model Validation<br>Comparison with Perturbation Data,<br>Mutant Phenotypes]
    F --> G[Control Strategy Design<br>Node/Edge Interventions,<br>Optimal Drug Targets]
    G --> H[In Silico Perturbation<br>Simulation of Interventions]
    H --> I[Experimental Validation<br>Follow-up wet-lab experiments]
    I --> J[Refined Model]

Challenges and Future Directions

Despite their utility, Boolean networks face several challenges. The discretization of continuous gene expression into binary values can lose information, although multilevel generalizations exist [31, 36]. The choice of update scheme (synchronous vs. asynchronous) significantly affects dynamics, and there is no universal best practice [12, 13]. Model inference [3] from noisy, sparse data remains difficult, particularly for large networks [69].

Scalability is a concern: the state space grows exponentially with the number of nodes. Quantum algorithms offer potential solutions for attractor search [29]. The integration of BNs with other modeling frameworks, such as metabolic models [21] and Petri nets [36], is an active area of research.

In veterinary medicine, the application of BNs to non-model species is limited by the availability of high-quality regulatory data. However, the increasing accessibility of transcriptomic and epigenomic data for livestock, poultry, and companion animals is expected to drive future modeling efforts. Cross-species comparisons using BNs can reveal conserved regulatory principles and host-specific adaptations relevant to zoonotic diseases and veterinary therapeutics.

Conclusion

Boolean networks provide a powerful and interpretable framework for modeling gene regulatory networks. Their ability to capture discrete logical interactions, predict attractor landscapes, and identify control targets makes them invaluable for systems biology. Recent advances in inference algorithms, control theory, and software tools have expanded their applicability to complex biological systems, including those of veterinary importance. Continued development of scalable, data-driven methods and integration with multi-omics data will further enhance the utility of Boolean networks in veterinary diagnostics and therapeutics.

References

[1] Hotstegs CLO, Llano JP, Kestler HA. A Dataset of Benchmark Boolean Models for Gene Regulatory Networks. Sci Data. 2026. URL: https://pubmed.ncbi.nlm.nih.gov/42251065/

[2] Kittaneh H, Castiglione F, Liso A et al. Ensemble threshold Boolean modeling reveals robust attractors and regulatory drivers in pediatric leukemia. Comput Biol Med. 2026. URL: https://pubmed.ncbi.nlm.nih.gov/42090947/

[3] Alali M, Imani M. Bayesian Topology Inference of Regulatory Networks under Partial Observability. Results Control Optim. 2025. URL: https://pubmed.ncbi.nlm.nih.gov/41867493/

[4] Chevalier S, Becker J, Gui Y et al. Data-driven inference of Boolean networks from transcriptomes to predict cellular differentiation and reprogramming. NPJ Syst Biol Appl. 2025. URL: https://pubmed.ncbi.nlm.nih.gov/41006334/

[5] Dague P. A semantics for Boolean networks consistent with regulatory threshold constraints. J Theor Biol. 2026. URL: https://pubmed.ncbi.nlm.nih.gov/41587654/

[6] Kadelka C. Canalization as a stabilizing principle of gene regulatory networks: a discrete dynamical systems perspective. NPJ Syst Biol Appl. 2026. URL: https://pubmed.ncbi.nlm.nih.gov/41580437/

[7] Bavisetty VSN, Wheeler M, Kadelka C. Attractors are less stable than their basins: Canalization creates a coherence gap in gene regulatory networks. bioRxiv. 2025. URL: https://pubmed.ncbi.nlm.nih.gov/41279382/

[8] Ruz GA, Ledger T, Macauley M. Evolutionary computation for reconstructing threshold networks of the tryptophan operon in Escherichia coli. Biosystems. 2026. URL: https://pubmed.ncbi.nlm.nih.gov/41421779/

[9] Correction to: Critical assessment of the ability of Boolean threshold models to describe gene regulatory network dynamics. PNAS Nexus. 2025. URL: https://pubmed.ncbi.nlm.nih.gov/40978086/

[10] Kadelka C, Hari K. Critical assessment of the ability of Boolean threshold models to describe gene regulatory network dynamics. PNAS Nexus. 2025. URL: https://pubmed.ncbi.nlm.nih.gov/40809087/

[11] Ruz GA, Cho AD. Generalization power of threshold Boolean networks. Biosystems. 2025. URL: https://pubmed.ncbi.nlm.nih.gov/40885514/

[12] Bensussen A, Arciniega-González JA, Álvarez-Buylla ER et al. Analytical approach of synchronous and asynchronous update schemes applied to solving biological Boolean networks. PLoS One. 2025. URL: https://pubmed.ncbi.nlm.nih.gov/40938909/

[13] Pastva S, Park KH, Huvar O et al. An open problem: Why are motif-avoidant attractors so rare in asynchronous Boolean networks? J Math Biol. 2025. URL: https://pubmed.ncbi.nlm.nih.gov/40504255/

[14] Trinh VG, Park KH, Pastva S et al. Mapping the attractor landscape of Boolean networks with biobalm. Bioinformatics. 2025. URL: https://pubmed.ncbi.nlm.nih.gov/40327535/

[15] Harlapur P, Bv H, Jolly MK. GRiNS: a python library for simulating gene regulatory network dynamics. BMC Bioinformatics. 2025. URL: https://pubmed.ncbi.nlm.nih.gov/41107706/

[16] Shyam S, S NN, Anand V et al. Mutually inhibiting teams of nodes: A predictive framework for structure-dynamics relationships in gene regulatory networks. Phys Biol. 2025. URL: https://pubmed.ncbi.nlm.nih.gov/41038227/

[17] Cifuentes-Fontanals L, Tonello E, Siebert H. Node and edge control strategy identification via trap spaces in Boolean networks. BMC Bioinformatics. 2025. URL: https://pubmed.ncbi.nlm.nih.gov/41057758/

[18] Xu X, Li Y, Zhong D et al. An approach to inferring gene regulatory networks via boolean modeling and feature selection. Neural Netw. 2026. URL: https://pubmed.ncbi.nlm.nih.gov/41205357/

[19] Zhang D, Liu ZP, Gao R. LogicSR: prior-guided symbolic regression for gene regulatory network inference from single-cell transcriptomics data. Brief Bioinform. 2025. URL: https://pubmed.ncbi.nlm.nih.gov/41269283/

[20] Li LX, Zhang Y, Aguilar B et al. KGBN: Augmenting and optimizing logical gene regulatory networks using knowledge graphs. bioRxiv. 2026. URL: https://pubmed.ncbi.nlm.nih.gov/41659556/

[21] Rodriguez-Flores CJ, Barrena N, Paulevé L et al. gMISpy: integration of complex regulatory networks and genome scale metabolic models. Bioinformatics. 2026. URL: https://pubmed.ncbi.nlm.nih.gov/41965071/

[22] Bampos C, Megalooikonomou V. Integer linear programming for contrasting state interventions in Boolean networks. PeerJ. 2026. URL: https://pubmed.ncbi.nlm.nih.gov/41809699/

[23] Ananda R, Daud KM, Zainudin S. RBI: a novel algorithm for regulatory-metabolic network model in designing the optimal mutant strain. PeerJ Comput Sci. 2025. URL: https://pubmed.ncbi.nlm.nih.gov/40575010/

[24] Murrugarra D, Veliz-Cuba A, Dimitrova E et al. Modular Control of Boolean Network Models. Bull Math Biol. 2025. URL: https://pubmed.ncbi.nlm.nih.gov/40461704/

[25] Hosseini SH, Aksaray D, Imani M. Pareto-Optimal Interventions in Gene Regulatory Networks using Signal Temporal Logic. Proc Am Control Conf. 2025. URL: https://pubmed.ncbi.nlm.nih.gov/41773284/

[26] Hosseini SH, Imani M. Deep Reinforcement Learning for Intervention of Partially Observable Regulatory Networks. Proc Am Control Conf. 2025. URL: https://pubmed.ncbi.nlm.nih.gov/41726383/

[27] Hosseini SH, Imani M. An optimal Bayesian intervention policy in response to unknown dynamic cell stimuli. Inf Sci (N Y). 2024. URL: https://pubmed.ncbi.nlm.nih.gov/39464381/

[28] Rafimanzelat MR. Global stabilization of Boolean networks with applications to biomolecular network control. Sci Rep. 2025. URL: https://pubmed.ncbi.nlm.nih.gov/40307350/

[29] Rossini M, Weidner FM, Ankerhold J et al. A novel quantum algorithm for efficient attractor search in gene regulatory networks. Patterns (N Y). 2025. URL: https://pubmed.ncbi.nlm.nih.gov/41040963/

[30] Bhattacharjee P, Kumar AP, Datta A. Boolean network-based identification of optimal drug combinations for prostate cancer. Comput Biol Chem. 2026. URL: https://pubmed.ncbi.nlm.nih.gov/41548507/

[31] Hari K, Tripathi S, Anand V et al. A multilevel formalism to model the hybrid E/M phenotypes in epithelial-mesenchymal plasticity. Biophys J. 2025. URL: https://pubmed.ncbi.nlm.nih.gov/41272973/

[32] Ramirez D, Kessler DA, Lu M et al. A computational approach for perturbation-induced EMT transitions. NPJ Syst Biol Appl. 2025. URL: https://pubmed.ncbi.nlm.nih.gov/41233338/

[33] Bolteau M, Messaoudi C, David L et al. Deep exploration of logical models of cell differentiation in human preimplantation embryos. NPJ Syst Biol Appl. 2025. URL: https://pubmed.ncbi.nlm.nih.gov/40425601/

[34] Zeegelaar R, Argyris G, Post JN. Molecular insights into heart field-specific cardiomyocyte differentiation - A computational study. PLoS One. 2026. URL: https://pubmed.ncbi.nlm.nih.gov/41490449/

[35] Herrera J, Bensussen A, García-Gómez ML et al. A system-level model reveals that transcriptional stochasticity is required for hematopoietic stem cell differentiation. NPJ Syst Biol Appl. 2024. URL: https://pubmed.ncbi.nlm.nih.gov/39639033/

[36] Andreas E, Quist E, Gedeon T. Petri net on expanded networks computes the dynamics of finite multilevel models. Interface Focus. 2025. URL: https://pubmed.ncbi.nlm.nih.gov/40862234/

[37] Li W, Provata A, MacCarthy T. An extensive study of two-node McCulloch-pitts networks. J Theor Biol. 2026. URL: https://pubmed.ncbi.nlm.nih.gov/41740882/

[38] Demongeot J, Diallo AK, Hazgui H et al. Boolean Networks with Classic and New Updating Modes Applied to Genetic Regulation in Some Familial Diseases. Int J Mol Sci. 2025. URL: https://pubmed.ncbi.nlm.nih.gov/41465403/

[39] Jog A, Sverdlov R, Schuster S et al. A reversible genetic NOR gate in plants using translational repression. Trends Biotechnol. 2026. URL: https://pubmed.ncbi.nlm.nih.gov/41058342/

[40] Taheri-Ledari M, Marashi SA, Ghahremani MH et al. Calibrating a parameterized stochastic Boolean network model of gene regulation using a single steady-state gene expression profile. Math Biosci. 2025. URL: https://pubmed.ncbi.nlm.nih.gov/40935080/

[41] Jothiprakasam T, Jhunjhunwala S. Phenomenological modeling of gene transcription by approximating cooperativity of transcription factors improves prediction and reduces complexity in gene regulatory network models. J Theor Biol. 2026. URL: https://pubmed.ncbi.nlm.nih.gov/40925464/

[42] Cadena M, Barreto GE. Exploring the interplay between circadian rhythms and obesity: A Boolean network approach to understanding metabolic dysregulation. PLoS One. 2025. URL: https://pubmed.ncbi.nlm.nih.gov/40924721/

[43] Bv H, Billakurthi HS, Adigwe S et al. Emergent dynamics of cellular decision making in multi-node mutually repressive regulatory networks. J R Soc Interface. 2025. URL: https://pubmed.ncbi.nlm.nih.gov/40829640/

[44] Huang X, Hullur AR, Jafari E et al. Leveraging transcription factor physical proximity for enhancing gene regulation inference. Bioinformatics. 2025. URL: https://pubmed.ncbi.nlm.nih.gov/40662784/

[45] Martinez-Padilla RC, Mendoza L, Miranda-Rios J. Impacts of a high-glucose diet or starvation on microRNA-transcription factor networks in Caenorhabditis elegans through Boolean mathematical modeling. Biosystems. 2025. URL: https://pubmed.ncbi.nlm.nih.gov/40425126/

[46] Hemedan AA, Satagopam V, Schneider R et al. Protocol to identify regulatory modules in Parkinson's disease progression using miRNA data and Boolean modeling. STAR Protoc. 2025. URL: https://pubmed.ncbi.nlm.nih.gov/40347476/

[47] Fareed MM, Shityakov S. Quantifying pleiotropy through directed signaling networks: A synchronous Boolean network approach and in-silico pleiotropic scoring. Biosystems. 2025. URL: https://pubmed.ncbi.nlm.nih.gov/39988275/

[48] Gong JR, Lee CK, Kim HM et al. Control of Cellular Differentiation Trajectories for Cancer Reversion. Adv Sci (Weinh). 2025. URL: https://pubmed.ncbi.nlm.nih.gov/39661721/

[49] Peng D, Cahan P. OneSC: a computational platform for recapitulating cell state transitions. Bioinformatics. 2024. URL: https://pubmed.ncbi.nlm.nih.gov/39570626/

[50] Singh A, Dwivedi A. Network dynamics investigation of omics-data-driven circadian-hypoxia crosstalk logical model in gallbladder cancer reveals key therapeutic target combinations. Integr Biol (Camb). 2024. URL: https://pubmed.ncbi.nlm.nih.gov/39499101/

[51] Moya-Cuevas J, Ortiz-Gutiérrez E, López-Sánchez P et al. A Model for the Gene Regulatory Network Along the Arabidopsis Fruit Medio-Lateral Axis: Rewiring the Pod Shatter Process. Plants (Basel). 2024. URL: https://pubmed.ncbi.nlm.nih.gov/39458874/

[52] Ciccone MD, Messina CD. Translating weighted probabilistic bits to synthetic genetic circuits. Plant Genome. 2025. URL: https://pubmed.ncbi.nlm.nih.gov/39425499/

[53] Kubaczka E, Gehri M, Marlhens JJM et al. Energy Aware Technology Mapping of Genetic Logic Circuits. ACS Synth Biol. 2024. URL: https://pubmed.ncbi.nlm.nih.gov/39378113/

[54] Vivanco E, Goles E, Montalva-Medel M et al. Dynamical robustness of a Boolean model for the human gonadal sex determination. Comput Biol *** Disclaimer: This article is for educational and informational purposes only. It is not intended to substitute for professional veterinary advice, diagnosis, treatment, or regulatory guidance. Always consult a licensed veterinarian or qualified specialist regarding animal health, disease diagnosis, and therapeutic decisions.