Boolean network explained

A Boolean network consists of a discrete set of boolean variables each of which has a Boolean function (possibly different for each variable) assigned to it which takes inputs from a subset of those variables and output that determines the state of the variable it is assigned to. This set of functions in effect determines a topology (connectivity) on the set of variables, which then become nodes in a network. Usually, the dynamics of the system is taken as a discrete time series where the state of the entire network at time t+1 is determined by evaluating each variable's function on the state of the network at time t. This may be done synchronously or asynchronously.^[1]

Boolean networks have been used in biology to model regulatory networks. Although Boolean networks are a crude simplification of genetic reality where genes are not simple binary switches, there are several cases where they correctly convey the correct pattern of expressed and suppressed genes.^{[2] [3]} The seemingly mathematical easy (synchronous) model was only fully understood in the mid 2000s.^[4]

Classical model

A Boolean network is a particular kind of sequential dynamical system, where time and states are discrete, i.e. both the set of variables and the set of states in the time series each have a bijection onto an integer series.

A random Boolean network (RBN) is one that is randomly selected from the set of all possible boolean networks of a particular size, N. One then can study statistically, how the expected properties of such networks depend on various statistical properties of the ensemble of all possible networks. For example, one may study how the RBN behavior changes as the average connectivity is changed.

The first Boolean networks were proposed by Stuart A. Kauffman in 1969, as random models of genetic regulatory networks^[5] but their mathematical understanding only started in the 2000s.^{[6] [7]}

Attractors

Since a Boolean network has only 2^N possible states, a trajectory will sooner or later reach a previously visited state, and thus, since the dynamics are deterministic, the trajectory will fall into a steady state or cycle called an attractor (though in the broader field of dynamical systems a cycle is only an attractor if perturbations from it lead back to it). If the attractor has only a single state it is called a point attractor, and if the attractor consists of more than one state it is called a cycle attractor. The set of states that lead to an attractor is called the basin of the attractor. States which occur only at the beginning of trajectories (no trajectories lead to them), are called garden-of-Eden states^[8] and the dynamics of the network flow from these states towards attractors. The time it takes to reach an attractor is called transient time.

With growing computer power and increasing understanding of the seemingly simple model, different authors gave different estimates for the mean number and length of the attractors, here a brief summary of key publications.^[9]

Author	Year	Mean attractor length	Mean attractor number	comment
Kauffman	1969	\langle A\rangle\sim \sqrt{N}	\langle\nu\rangle\sim \sqrt{N}
Bastolla/ Parisi^[10]	access-date=26 November 2014-->	1998	faster than a power law, \langle A\rangle > N ^{x} \forall x	faster than a power law, \langle\nu\rangle > N ^{x} \forall x	first numerical evidences
Bilke/ Sjunnesson^[11]	access-date=26 November 2014-->	2002		linear with system size, \langle\nu\rangle \sim N
Socolar/Kauffman^[12]	2003		faster than linear, \langle\nu\rangle > N ^{x} with x > 1
Samuelsson/Troein^[13]	access-date=26 November 2014-->	2003		superpolynomial growth, \langle\nu\rangle > N ^{x} \forall x	mathematical proof
Mihaljev/Drossel^[14]	access-date=26 November 2014-->	2005	faster than a power law, \langle A\rangle > N ^{x} \forall x	faster than a power law, \langle\nu\rangle > N ^{x} \forall x

Stability

In dynamical systems theory, the structure and length of the attractors of a network corresponds to the dynamic phase of the network. The stability of Boolean networks depends on the connections of their nodes. A Boolean network can exhibit stable, critical or chaotic behavior. This phenomenon is governed by a critical value of the average number of connections of nodes (

K _{c} ), and can be characterized by the Hamming distance as distance measure. In the unstable regime, the distance between two initially close states on average grows exponentially in time, while in the stable regime it decreases exponentially. In this, with "initially close states" one means that the Hamming distance is small compared with the number of nodes (N ) in the network.

For N-K-model^[15] the network is stable if

K<K _{c} , critical if K = K _{c} , and unstable if K>K _{c} .

The state of a given node

n _{i} is updated according to its truth table, whose outputs are randomly populated. p _{i} denotes the probability of assigning an off output to a given series of input signals.

If

p _{i} = p = const. for every node, the transition between the stable and chaotic range depends on p . According to Bernard Derrida and Yves Pomeau^[16], the critical value of the average number of connections is K _{c} =1 /[ 2 p( 1- p)] .

If

K is not constant, and there is no correlation between the in-degrees and out-degrees, the conditions of stability is determined by \langle K ^{in} \rangle ^{[17] [18] [19]} The network is stable if \langle K ^{in} \rangle <K _{c} , critical if \langle K ^{in} \rangle = K _{c} , and unstable if \langle K ^{in} \rangle >K _{c} .

The conditions of stability are the same in the case of networks with scale-free topology where the in-and out-degree distribution is a power-law distribution:

P(K) \propto K ^{-\gamma} , and \langle K ^{in} \rangle = \langle K ^{out} \rangle , since every out-link from a node is an in-link to another.^[20]

Sensitivity shows the probability that the output of the Boolean function of a given node changes if its input changes. For random Boolean networks,

q _{i} =2 p _{i} ( 1- p _{i} ) . In the general case, stability of the network is governed by the largest eigenvalue λ _{Q} of matrix Q , where Q _{ij} = q _{i} A _{ij} , and A is the adjacency matrix of the network.^[21] The network is stable if λ _{Q} < 1 , critical if λ _{Q} =1 , unstable if λ _{Q} > 1 .

Variations of the model

Other topologies

One theme is to study different underlying graph topologies.

• The homogeneous case simply refers to a grid which is simply the reduction to the famous Ising model.
• Scale-free topologies may be chosen for Boolean networks.^[22] One can distinguish the case where only in-degree distribution in power-law distributed,^[23] or only the out-degree-distribution or both.

Other updating schemes

Classical Boolean networks (sometimes called CRBN, i.e. Classic Random Boolean Network) are synchronously updated. Motivated by the fact that genes don't usually change their state simultaneously,^[24] different alternatives have been introduced. A common classification^[25] is the following:

• Deterministic asynchronous updated Boolean networks (DRBNs) are not synchronously updated but a deterministic solution still exists. A node i will be updated when t ≡ Q_i (mod P_i) where t is the time step.^[26]
• The most general case is full stochastic updating (GARBN, general asynchronous random boolean networks). Here, one (or more) node(s) are selected at each computational step to be updated.
• The Partially-Observed Boolean Dynamical System (POBDS)^{[27] [28] [29] [30]} signal model differs from all previous deterministic and stochastic Boolean network models by removing the assumption of direct observability of the Boolean state vector and allowing uncertainty in the observation process, addressing the scenario encountered in practice.
• Autonomous Boolean networks (ABNs) are updated in continuous time (t is a real number, not an integer), which leads to race conditions and complex dynamical behavior such as deterministic chaos.^{[31] [32]}

Application of Boolean Networks

Classification

• The Scalable Optimal Bayesian Classification^[33] developed an optimal classification of trajectories accounting for potential model uncertainty and also proposed a particle-based trajectory classification that is highly scalable for large networks with much lower complexity than the optimal solution.

References

• Dubrova, E., Teslenko, M., Martinelli, A., (2005). *Kauffman Networks: Analysis and Applications, in "Proceedings of International Conference on Computer-Aided Design", pages 479-484.

External links

• Analysis of Dynamic Algebraic Models (ADAM) v1.1
• bioasp/bonesis: Synthesis of Most Permissive Boolean Networks from network architecture and dynamical properties
• CoLoMoTo (Consortium for Logical Models and Tools)
• DDLab
• NetBuilder Boolean Networks Simulator
• Open Source Boolean Network Simulator
• JavaScript Kauffman Network
• Probabilistic Boolean Networks (PBN)
• RBNLab
• A SAT-based tool for computing attractors in Boolean Networks

Notes and References

1. Naldi. A.. Monteiro. P. T.. Mussel. C.. Kestler. H. A.. Thieffry. D.. Xenarios. I.. Saez-Rodriguez. J.. Helikar. T.. Chaouiya. C.. Cooperative development of logical modelling standards and tools with CoLoMoTo. Bioinformatics. 25 January 2015. 31. 7. 1154–1159. 10.1093/bioinformatics/btv013. 25619997. free.
2. Albert. Réka. Othmer. Hans G. The topology of the regulatory interactions predicts the expression pattern of the segment polarity genes in Drosophila melanogaster. Journal of Theoretical Biology. July 2003. 223. 1. 1–18. 10.1016/S0022-5193(03)00035-3. 12782112. 6388622. q-bio/0311019. 2003JThBi.223....1A. 10.1.1.13.3370.
3. Li. J.. Bench. A. J.. Vassiliou. G. S.. Fourouclas. N.. Ferguson-Smith. A. C.. Green. A. R.. Imprinting of the human L3MBTL gene, a polycomb family member located in a region of chromosome 20 deleted in human myeloid malignancies . Proceedings of the National Academy of Sciences. 30 April 2004 . 101. 19 . 7341–7346 . 10.1073/pnas.0308195101. 15123827 . 409920. 2004PNAS..101.7341L . free.
4. Book: Drossel. Barbara. Schuster. Heinz Georg. Chapter 3. Random Boolean Networks. December 2009. 10.1002/9783527626359.ch3. 0706.3351. Reviews of Nonlinear Dynamics and Complexity. Wiley. 69–110. 9783527626359. Random Boolean Networks. 119300231.
5. Kauffman. Stuart. Homeostasis and Differentiation in Random Genetic Control Networks. Nature. 11 October 1969. 224. 5215. 177–178. 10.1038/224177a0. 5343519. 1969Natur.224..177K . 4179318.
6. Book: Aldana. Maximo. Coppersmith. Susan. Susan Coppersmith . Kadanoff. Leo P.. Perspectives and Problems in Nonlinear Science: A Celebratory Volume in Honor of Lawrence Sirovich . Boolean Dynamics with Random Couplings. 2003. 23–89. 10.1007/978-0-387-21789-5_2. nlin/0204062. 978-1-4684-9566-9. 15024306.
7. nlin.AO/0408006. Gershenson. Carlos. Introduction to Random Boolean Networks. In Bedau, M., P. Husbands, T. Hutton, S. Kumar, and H. Suzuki (Eds.) Workshop and Tutorial Proceedings, Ninth International Conference on the Simulation and Synthesis of Living Systems (ALife IX). Pp. 2004. 160–173. 2004. 2004nlin......8006G.
8. Book: Wuensche. Andrew. Exploring discrete dynamics : [the DDLab manual : tools for researching cellular automata, random Boolean and multivalue neworks [sic] and beyond]]. 2011. Luniver Press. Frome, England. 9781905986316. 16. 12 January 2016. 4 February 2023. https://web.archive.org/web/20230204160659/https://books.google.com/books?id=qsktzY_Vg8QC&pg=PA16. live.
9. Greil. Florian. Boolean Networks as Modeling Framework. Frontiers in Plant Science. 2012. 3. 178. 10.3389/fpls.2012.00178. 22912642. 3419389. free.
10. Bastolla. U.. Parisi. G.. The modular structure of Kauffman networks. Physica D: Nonlinear Phenomena. May 1998. 115. 3–4. 219–233. 10.1016/S0167-2789(97)00242-X. cond-mat/9708214 . 1998PhyD..115..219B . 1585753.
11. Bilke. Sven. Sjunnesson. Fredrik. Stability of the Kauffman model. Physical Review E. December 2001. 65. 1. 016129. 10.1103/PhysRevE.65.016129. 11800758. cond-mat/0107035 . 2001PhRvE..65a6129B . 2470586.
12. Socolar. J.. Kauffman. S.. Scaling in Ordered and Critical Random Boolean Networks. Physical Review Letters. February 2003. 90. 6. 068702. 10.1103/PhysRevLett.90.068702. 12633339. 2003PhRvL..90f8702S. cond-mat/0212306 . 14392074.
13. Samuelsson. Björn. Troein. Carl. Superpolynomial Growth in the Number of Attractors in Kauffman Networks. Physical Review Letters. March 2003. 90. 9. 10.1103/PhysRevLett.90.098701. 2003PhRvL..90i8701S. 12689263. 098701.
14. Mihaljev. Tamara. Drossel. Barbara. Scaling in a general class of critical random Boolean networks. Physical Review E. October 2006. 74. 4. 046101. 10.1103/PhysRevE.74.046101. 17155127. cond-mat/0606612 . 2006PhRvE..74d6101M . 17739744.
15. Kauffman . S. A. . 1969 . Metabolic stability and epigenesis in randomly constructed genetic nets . Journal of Theoretical Biology . 22 . 3 . 437–467 . 10.1016/0022-5193(69)90015-0. 5803332 . 1969JThBi..22..437K .
16. Random Networks of Automata: A Simple Annealed Approximation. Europhysics Letters (EPL). 1986-01-15. 45–49. 1. 2. 10.1209/0295-5075/1/2/001. B. Derrida. Y. Pomeau. 1986EL......1...45D. 160018158. 2016-01-12. 2020-05-17. https://web.archive.org/web/20200517203049/http://stacks.iop.org/0295-5075/1/i=2/a=001?key=crossref.8cf81041d96c6144f3a397d9cf72cf09%2F. live.
17. Phase transitions and antichaos in generalized Kauffman networks. Physics Letters A. 1995-01-02. 331–334. 196. 5–6. 10.1016/0375-9601(94)00876-Q. Ricard V.. Solé. Bartolo. Luque. 1995PhLA..196..331S.
18. Phase transitions in random networks: Simple analytic determination of critical points. Physical Review E. 1997-01-01. 257–260. 55. 1. 10.1103/PhysRevE.55.257. Bartolo. Luque. Ricard V.. Solé. 1997PhRvE..55..257L.
19. From topology to dynamics in biochemical networks . Chaos: An Interdisciplinary Journal of Nonlinear Science. 2001-12-01. 1054-1500. 809–815. 11. 4. 10.1063/1.1414882. 12779520 . Jeffrey J.. Fox. Colin C.. Hill. 2001Chaos..11..809F. free.
20. A natural class of robust networks. Proceedings of the National Academy of Sciences. 2003-07-22. 0027-8424. 166377. 12853565. 8710–8714. 100. 15. 10.1073/pnas.1536783100. Maximino. Aldana. Philippe. Cluzel. 2003PNAS..100.8710A. free.
21. The effect of network topology on the stability of discrete state models of genetic control. Proceedings of the National Academy of Sciences. 2009-05-19. 0027-8424. 2688895. 19416903. 8209–8214. 106. 20. 10.1073/pnas.0900142106. Andrew. Pomerance. Edward. Ott. Michelle. Girvan. Michelle Girvan . Wolfgang. Losert. 0901.4362. 2009PNAS..106.8209P. free.
22. Aldana. Maximino. Boolean dynamics of networks with scale-free topology. Physica D: Nonlinear Phenomena. October 2003. 185. 1. 45–66. 10.1016/s0167-2789(03)00174-x. cond-mat/0209571. 2003PhyD..185...45A.
23. Drossel. Barbara. Greil. Florian. Critical Boolean networks with scale-free in-degree distribution. Physical Review E. 4 August 2009. 80. 2. 026102. 10.1103/PhysRevE.80.026102. 19792195. 0901.0387. 2009PhRvE..80b6102D. 2487442.
24. Book: Harvey. Imman. Bossomaier. Terry. Husbands. Phil. Harvey. Imman. Time out of joint: Attractors in asynchronous random Boolean networks. Proceedings of the Fourth European Conference on Artificial Life (ECAL97). 1997. 67–75. MIT Press. 9780262581578. 2020-09-16. 2023-02-04. https://web.archive.org/web/20230204160659/https://books.google.com/books?id=ccp8fzlyorAC&pg=PA67. live.
25. Book: Gershenson. Carlos. Standish. Russell K. Bedau. Mark A. Classification of Random Boolean Networks. Proceedings of the Eighth International Conference on Artificial Life. 2002. 8. 1–8. 12 January 2016. cs/0208001. Artificial Life. Cambridge, Massachusetts, USA. 9780262692816. 2002cs........8001G. 4 February 2023. https://web.archive.org/web/20230204160659/https://books.google.com/books?id=si_KlRbL1XoC&pg=PA1. live.
26. Book: Gershenson. Carlos. Broekaert. Jan. Aerts. Diederik. Advances in Artificial Life . Contextual Random Boolean Networks . 14 September 2003. 2801. 615–624. 10.1007/978-3-540-39432-7_66. nlin/0303021. Lecture Notes in Computer Science. 7th European Conference, ECAL 2003. Dortmund, Germany. 978-3-540-39432-7. 4309400.
27. Imani. M.. Braga-Neto. U. M.. 2017-01-01. Maximum-Likelihood Adaptive Filter for Partially Observed Boolean Dynamical Systems. IEEE Transactions on Signal Processing. 65. 2. 359–371. 10.1109/TSP.2016.2614798. 1053-587X. 1702.07269. 2017ITSP...65..359I. 178376.
28. Book: 972–976. Imani. M.. Braga-Neto. U. M.. en-US. 10.1109/GlobalSIP.2015.7418342. Optimal state estimation for boolean dynamical systems using a boolean Kalman smoother. 2015. 978-1-4799-7591-4. 2015 IEEE Global Conference on Signal and Information Processing (GlobalSIP). 8672734.
29. Book: Imani. M.. Braga-Neto. U. M.. en-US. 10.1109/ACC.2016.7524920. 2016 American Control Conference (ACC). 227–232. 2016. 978-1-4673-8682-1. 7210088.
30. Book: Imani. M.. Braga-Neto. U.. 2016 IEEE 55th Conference on Decision and Control (CDC) . Point-based value iteration for partially-observed Boolean dynamical systems with finite observation space . 2016-12-01. 4208–4213. 10.1109/CDC.2016.7798908. 978-1-5090-1837-6. 11341805.
31. Zhang. Rui. Cavalcante. Hugo L. D. de S.. Gao. Zheng. Gauthier. Daniel J.. Socolar. Joshua E. S.. Adams. Matthew M.. Lathrop. Daniel P.. Boolean chaos. Physical Review E. 80. 4. 2009. 045202. 1539-3755. 10.1103/PhysRevE.80.045202. 19905381. 0906.4124. 2009PhRvE..80d5202Z. 43022955.
32. Cavalcante. Hugo L. D. de S.. Gauthier. Daniel J.. Socolar. Joshua E. S.. Zhang. Rui. On the origin of chaos in autonomous Boolean networks. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences. 368. 1911. 2010. 495–513. 1364-503X. 10.1098/rsta.2009.0235. 20008414. 0909.2269. 2010RSPTA.368..495C. 426841.
33. Hajiramezanali, E. & Imani, M. & Braga-Neto, U. & Qian, X. & Dougherty, E.. Scalable Optimal Bayesian Classification of Single-Cell Trajectories under Regulatory Model Uncertainty. ACMBCB'18. https://dl.acm.org/citation.cfm?id=3233689