Boolean Network Model Optimization Based on Neural Network and Genetic Programming

doi:10.16182/j.issn1004731x.joss.25-0334

Abstract

Abstract:

To address the issues of complex node relationships and low accuracy in large-scale Boolean network inference, a new optimization algorithm integrated with long short-term memory (LSTM) networks and genetic programming was proposed. An enhanced LSTM network combined with a self-attention mechanism was designed to extract potential regulatory nodes from time-series data. These nodes were utilized as terminals of the syntax tree for the design of the genetic programming algorithm, and new operators were introduced to optimize Boolean function search. Experimental results have demonstrated that the proposed method significantly outperforms the most advanced existing algorithms in inference accuracy. The Boolean network optimization algorithm provides an effective tool for modeling and simulating dynamic behaviors in complex networks.

Key words: Boolean network, time-series inference, LSTM, self-attention mechanism, genetic programming

CLC Number:

TP301.6

Tang Jinlin, Wang Yan, Liu Xiang, Wang Tuanjie, Ji Zhicheng. Boolean Network Model Optimization Based on Neural Network and Genetic Programming[J]. Journal of System Simulation, 2025, 37(11): 2812-2825.

Figures/Tables 10

Fig. 1

Fig. 2

Fig. 3

Fig. 4

Fig. 5

Fig. 6

Table 1

Table 2

Fig. 7

Fig. 8

References 40

[1]	Kauffman S A. Metabolic Stability and Epigenesis in Randomly Constructed Genetic nets[J]. Journal of Theoretical Biology, 1969, 22(3): 437-467.
[2]	李志强, 宋金利, 杨剑. 概率布尔网络重构[J]. 中国科学(信息科学), 2018, 48(9): 1214-1226.
	Li Zhiqiang, Song Jinli, Yang Jian. Reconstruction of Probabilistic Boolean Networks[J]. Scientia Sinica(Informationis), 2018, 48(9): 1214-1226.
[3]	蒋梦涵, 李海涛. 基于牵制控制的切换布尔网络的分布式集合镇定[J]. 控制理论与应用, 2021, 38(11): 1754-1760.
	Jiang Menghan, Li Haitao. Distributed Set Stabilization of Switched Boolean Networks Via Pinning Control[J]. Control Theory & Applications, 2021, 38(11): 1754-1760.
[4]	刘洋, 刘泽娇, 卢剑权. 状态翻转控制下布尔控制网络的可镇定性和Q学习算法[J]. 控制理论与应用, 2021, 38(11): 1743-1753.
	Liu Yang, Liu Zejiao, Lu Jianquan. State-flipped Control and Q-learning Algorithm for the Stabilization of Boolean Control Networks[J]. Control Theory & Applications, 2021, 38(11): 1743-1753.
[5]	Borriello E, Daniels B C. The Basis of Easy Controllability in Boolean Networks[J]. Nature Communications, 2021, 12(1): 5227.
[6]	Li Lingyu, Sun Liangjie, Chen Guangyi, et al. LogBTF: Gene Regulatory Network Inference Using Boolean Threshold Network Model from Single-cell Gene Expression Data[J]. Bioinformatics, 2023, 39(5): btad256.
[7]	Barman Shohag, Kwon Yung-Keun. A Novel Mutual Information-based Boolean Network Inference Method from Time-series Gene Expression Data[J]. PLoS One, 2017, 12(2): e0171097.
[8]	Lähdesmäki Harri, Shmulevich I, Yli-Harja Olli. On Learning Gene Regulatory Networks Under the Boolean Network Model[J]. Machine Learning, 2003, 52(1): 147-167.
[9]	Shi Ning, Zhu Zexuan, Tang Ke, et al. ATEN: And/Or Tree Ensemble for Inferring Accurate Boolean Network Topology and Dynamics[J]. Bioinformatics, 2020, 36(2): 578-585.
[10]	Zhang Dezhen, Gao Shuhua, Liu Zhiping, et al. LogicGep: Boolean Networks Inference Using Symbolic Regression from Time-series Transcriptomic Profiling Data[J]. Briefings in Bioinformatics, 2024, 25(4): bbae286.
[11]	Nguyen H, Tran D, Tran B, et al. A Comprehensive Survey of Regulatory Network Inference Methods Using Single Cell RNA Sequencing Data[J]. Briefings in Bioinformatics, 2021, 22(3): bbaa190.
[12]	Pušnik Žiga, Mraz Miha, Zimic Nikolaj, et al. Review and Assessment of Boolean Approaches for Inference of Gene Regulatory Networks[J]. Heliyon, 2022, 8(8): e10222.
[13]	沈嘉玮, 才大业, 杨国青, 等. 大规模脉冲神经网络动态加载仿真方法[J]. 系统仿真学报, 2025, 37(2): 541-550.
	Shen Jiawei, Cai Daye, Yang Guoqing, et al. Dynamic Loading Simulation Method for Large-scale Spiking Neural Network[J]. Journal of System Simulation, 2025, 37(2): 541-550.
[14]	李静, 张涛涛, 金凯, 等. 基于时变RBF神经网络的时变非线性系统控制器设计[J]. 系统仿真学报, 2023, 35(10): 2223-2236.
	Li Jing, Zhang Taotao, Jin Kai, et al. Time-varying RBF Neural Network-based Controller Design for a Class of Time-varying Nonlinear Systems[J]. Journal of System Simulation, 2023, 35(10): 2223-2236.
[15]	张少强, 潘镜伊. 基于单细胞RNA测序数据的基因调控网络推断算法综述[J]. 天津师范大学学报(自然科学版), 2024, 44(1): 1-12.
	Zhang Shaoqiang, Pan Jingyi. Review of Gene Regulatory Network Inference Algorithms Based on Single-cell RNA Sequencing Data[J]. Journal of Tianjin Normal University(Natural Science Edition), 2024, 44(1): 1-12.
[16]	鲍艳春, 石彩霞, 张传强, 等. 深度学习在基因组学中的研究进展[J]. 遗传, 2024, 46(9): 701-715.
	Bao Yanchun, Shi Caixia, Zhang Chuanqiang, et al. Progress on Deep Learning in Genomics[J]. Hereditas, 2024, 46(9): 701-715.
[17]	Hassan Ismail Fawaz, Forestier Germain, Weber Jonathan, et al. Deep Learning for Time Series Classification: A Review[J]. Data Mining and Knowledge Discovery, 2019, 33(4): 917-963.
[18]	Wen Qingsong, Zhou Tian, Zhang Chaoli, et al. Transformers in Time Series: A Survey[EB/OL]. (2022-02-15) [2025-03-09]. .
[19]	Zheng Yi, Liu Qi, Chen Enhong, et al. Time Series Classification Using Multi-channels Deep Convolutional Neural Networks[C]//Web-Age Information Management. Cham: Springer International Publishing, 2014: 298-310.
[20]	Lecun Y, Bottou L, Bengio Y, et al. Gradient-based Learning Applied to Document Recognition[J]. Proceedings of the IEEE, 1998, 86(11): 2278-2324.
[21]	Yuan Ye, Bar-Joseph Z. Deep Learning for Inferring Gene Relationships from Single-cell Expression Data[J]. Proceedings of the National Academy of Sciences of the United States of America, 2019, 116(52): 27151-27158.
[22]	Zampieri G, Vijayakumar S, Yaneske E, et al. Machine and Deep Learning Meet Genome-scale Metabolic Modeling[J]. PLoS Computational Biology, 2019, 15(7): e1007084.
[23]	Mahendran Nivedhitha, Durai Raj Vincent P M. A Deep Learning Framework with an Embedded-based Feature Selection Approach for the Early Detection of the Alzheimer's Disease[J]. Computers in Biology and Medicine, 2022, 141: 105056.
[24]	Hochreiter Sepp, Schmidhuber Jürgen. Long Short-term Memory[J]. Neural Computation, 1997, 9(8): 1735-1780.
[25]	Gyamfi K S, Brusey J, Gaura E. Differential Radial Basis Function Network for Sequence Modelling[J]. Expert Systems With Applications, 2022, 189: 115982.
[26]	Duan Haoshan, Meng Xi, Tang Jian, et al. Time-series Prediction Using a Regularized Self-organizing Long Short-term Memory Neural Network[J]. Applied Soft Computing, 2023, 145: 110553.
[27]	Al-Ajeli Ahmed, Alshamery Eman S. A Multi-output Deep Learning Model for Fault Diagnosis Based on Time-series Data[J]. International Journal of Prognostics and Health Management, 2024, 15(1): 1-10.
[28]	Huang Zhixin, Lin Jiaxiang, Lin Lizheng, et al. A Sensitive LSTM Model for High Accuracy Zero-inflated Time-series Prediction[J]. IEEE Access, 2024, 12: 171527-171539.
[29]	Wang Yongjian, Qian Cheng, Qin Si. Attention-mechanism Based DiPLS-LSTM and Its Application in Industrial Process Time Series Big Data Prediction[J]. Computers & Chemical Engineering, 2023, 176: 108296.
[30]	Wen Xianyun, Li Weibang. Time Series Prediction Based on LSTM-attention-LSTM Model[J]. IEEE Access, 2023, 11: 48322-48331.
[31]	Barman Shohag, Kwon Yung-Keun. A Boolean Network Inference from Time-series Gene Expression Data Using a Genetic Algorithm[J]. Bioinformatics, 2018, 34(17): i927-i933.
[32]	Barman Shohag, Kwon Yung-Keun. A Neuro-evolution Approach to Infer a Boolean Network from Time-series Gene Expressions[J]. Bioinformatics, 2020, 36(S2): i762-i769.
[33]	Liu Xiang, Wang Yan, Liu Shan, et al. MIFuGP: Boolean Network Inference from Multivariate Time Series Using Fuzzy Genetic Programming[J]. Information Sciences, 2024, 680: 121129.
[34]	Mei Yi, Chen Qi, Lensen Andrew, et al. Explainable Artificial Intelligence by Genetic Programming: A Survey[J]. IEEE Transactions on Evolutionary Computation, 2023, 27(3): 621-641.
[35]	邱俊杰, 郑红, 程云辉. 基于多尺度LSTM预测模型研究[J]. 系统仿真学报, 2022, 34(7): 1593-1604.
	Qiu Junjie, Zheng Hong, Cheng Yunhui. Research on Prediction of Model Based on Multi-scale LSTM[J]. Journal of System Simulation, 2022, 34(7): 1593-1604.
[36]	Bahdanau Dzmitry, Cho Kyunghyun, Bengio Yoshua. Neural Machine Translation by Jointly Learning to Align and Translate[EB/OL]. (2014-09-01) [2025-03-12]. .
[37]	Zhang Yang, Liu Caiqi, Liu Mu jiein, et al. Attention Is All You Need: Utilizing Attention in AI-enabled Drug Discovery[J]. Briefings in Bioinformatics, 2024, 25(1): bbad467.
[38]	马仑, 杨跃, 王迨贺, 等. 联合自注意力机制与权值共享的人体行为识别模型[J]. 系统仿真学报, 2025, 37(9): 2409-2419.
	Ma Lun, Yang Yue, Wang Daihe, et al. A Model Combining Self-attention and Weight Sharing for Human Activity Recognition[J]. Journal of System Simulation, 2025, 37(9): 2409-2419.
[39]	Xie Huayang, Zhang Mengjie. Impacts of Sampling Strategies in Tournament Selection for Genetic Programming[J]. Soft Computing, 2012, 16(4): 615-633.
[40]	Müssel Christoph, Hopfensitz Martin, Kestler Hans A. BoolNet-an R Package for Generation, Reconstruction and Analysis of Boolean Networks[J]. Bioinformatics, 2010, 26(10): 1378-1380.

布尔网络	算法	噪声1%				噪声5%
布尔网络	算法	召回率	精度	假阳性率	F1值	召回率	精度	假阳性率	F1值
Protein	ATEN	0.853±0.010	0.851±0.011	0.063±0.005	0.852±0.009	0.888±0.013	0.782±0.012	0.092±0.006	0.832±0.012
	MIBNI	1.000	0.444	0.375	0.615	1.000	0.444	0.375	0.615
	Best-Fit	0.875	0.875	0.037	0.875	1.000	0.727	0.107	0.842
	NNBNI	1.000	0.800	0.071	0.889	1.000	0.615	0.179	0.762
	NNGSP	1.000	1.000	0	1.000	0.750±0	1.000	0	0.857±0.073
cAMP	ATEN	0.909±0.012	0.714±0.003	0.075±0.001	0.800±0.004	0.727±0.021	0.571±0.016	0.113±0.001	0.640±0.011
	MIBNI	0.909	0.417	0.264	0.571	0.818	0.375	0.283	0.514
	Best-Fit	0.728	0.571	0.113	0.640	0.909	0.476	0.208	0.625
	NNBNI	0.909	0.667	0.200	0.769	0.909	0.476	0.207	0.624
	NNGSP	0.873±0.036	0.960±0.060	0.008±0.002	0.914±0.038	1.000±0	0.890±0.078	0.026±0.012	0.941±0.024
Th-Cell	ATEN	0.955±0.009	0.750±0.008	0.057±0.002	0.840±0.010	0.864±0.009	0.655±0.012	0.082±0.003	0.754±0.011
	MIBNI	0.636	0.389	0.180	0.483	0.591	0.361	0.189	0.448
	Best-Fit	1.000	0.786	0.049	0.880	0.909	0.606	0.106	0.727
	NNBNI	0.909	0.476	0.207	0.626	0.818	0.419	0.205	0.554
	NNGSP	0.950±0.040	0.976±0.024	0.004±0.004	0.963±0.033	0.855±0.045	0.990±0.040	0.002±0.002	0.917±0.035

T-LGL	ATEN	0.885±0.007	0.575±0.010	0.057±0.003	0.697±0.009	0.845±0.014	0.564±0.011	0.057±0.007	0.677±0.013
	MIBNI	0.808	0.292	0.171	0.429	0.769	0.278	0.175	0.408
	Best-Fit	0.961	0.510	0.081	0.666	0.883	0.434	0.101	0.582
	NNBNI	0.808	0.488	0.074	0.609	0.730	0.432	0.084	0.543
	NNGSP	0.881±0.042	0.935±0.0183	0.006±0.001	0.907±0.034	0.708±0.0612	0.761±0.079	0.019±0.005	0.733±0.067

布尔网络	算法	召回率	精度	假阳性率	F1值
Protein	ATEN	0.875±0.016	0.636±0.014	0.143±0.008	0.737±0.023
	MIBNI	0.750	0.353	0.393	0.48
	Best-Fit	1.000	0.444	0.357	0.615
	NNBNI	1.000	0.571	0.214	0.727
	NNGSP	0.750±0	1.000	0	0.857+0.073
cAMP	ATEN	0.818±0.024	0.45±0.033	0.208±0.009	0.581±0.018
	MIBNI	0.727	0.348	0.283	0.470
	Best-Fit	0.818	0.391	0.264	0.529
	NNBNI	0.818	0.409	0.213	0.545
	NNGSP	1.000±0	0.733±0.080	0.076±0.016	0.846±0.033
Th-cell	ATEN	0.682±0.018	0.600±0.020	0.082±0.005	0.638±0.019
	MIBNI	0.500	0.344	0.172	0.407
	Best-Fit	0.591	0.394	0.164	0.473
	NNBNI	0.682	0.288	0.303	0.405
	NNGSP	0.682±0.046	0.652±0.061	0.066±0.008	0.667+0.076
T-LGL	ATEN	0.615±0.023	0.340±0.018	0.104±0.012	0.438±0.022
	MIBNI	0.462	0.245	0.124	0.320
	Best-Fit	0.346	0.188	0.131	0.243
	NNBNI	0.538	0.311	0.113	0.393
	NNGSP	0.577±0.077	0.600±0.069	0.034±0.016	0.588+0.106