Advances and applications of machine learning and intelligent optimization algorithms in genome-scale metabolic network models

Abstract

Due to the increasing demand for microbially manufactured products in various industries, it has become important to find optimal designs for microbial cell factories by changing the direction of metabolic flow and its flux size by means of metabolic engineering such as knocking out competing pathways and introducing exogenous pathways to increase the yield of desired products. Recently, with the gradual cross-fertilization between computer science and bioinformatics fields, machine learning and intelligent optimization-based approaches have received much attention in Genome-scale metabolic network models (GSMMs) based on constrained optimization methods, and many high-quality related works have been published. Therefore, this paper focuses on the advances and applications of machine learning and intelligent optimization algorithms in metabolic engineering, with special emphasis on GSMMs. Specifically, the development history of GSMMs is first reviewed. Then, the analysis methods of GSMMs based on constraint optimization are presented. Next, this paper mainly reviews the development and application of machine learning and intelligent optimization algorithms in genome-scale metabolic models. In addition, the research gaps and future research potential in machine learning and intelligent optimization methods applied in GSMMs are discussed.

Keywords

Genome-scale metabolic models / Machine learning / Intelligent optimization / Metabolic engineering

References

[1] Notebaart RA, van Enckevort FH, Francke C, et al.. Accelerating the reconstruction of genome -scale metabolic networks. BMC Bioinform, 2006, 7: 296.

[2] Francke C, Siezen RJ, Teusink B. Reconstructing the metabolic network of a bacterium from its genome[J]. Trends Microbiol, 2005, 13(11): 550-558.

[3] Arakawa K, Yamada Y, Shinoda K, et al.. GEM system: automatic prototyping of cell-wide metabolic pathway models from genomes[J]. BMC Bioinform, 2006, 7: 168-178.

[4] Lianzhong AL, Chengjie HOU. Research progress of lactic acid bacteria genome scale metabolic models. J Food Sci Technol, 2021, 39(3): 1-10

[5] Lin YP, Wang QH. Regulation and adaptive evolution of industrial microorganisms towards genetic and environmental disturbances. Chin J Biotech, 2019, 35(10): 1925-1941. (in Chinese).

[6] Bastanlar Y, Ozuysal M. Introduction to machine learning. Methods Mol Biol, 2014.

[7] Bin J, Fan W, Zhou JH, et al.. Application of intelligent optimization algorithms to wavelength selection of near-infrared spectroscopy. Spectrosc Spectr Anal, 2017, 37: 95-102

[8] Feist AM, Palsson BO. The growing scope of applications of genome-scale metabolic reconstructions: the case of E.coli. Nat Biotechnol, 2008, 26(6): 659-667.

[9] Edwards JS, Palsson BO. Systems properties of the haemophilus influenzae Rd metabolic genotype[J]. J Biol Chem, 1999, 274(25): 17410-17416.

[10] Zhang CY, Wu YK, Xu XH, et al.. Current status and future perspectives of metabolic network models of industrial microorganisms. Chin J Biotech, 2021, 37(3): 860-873

[11] Edwards JS, Palsson BØ. The Escherichia coli MG1655 in silico metabolic genotype: its definition, characteristics, and capabilities. Proc Natl Acad Sci USA, 2000, 97(10): 5528-5533.

[12] Reed JL, Vo TDTT, Schilling CH, et al.. An expanded genome-scale model of Escherichia coli K-12 (iJR904 GSM/GPR). Genome Biol, 2003, 4(9): R54.

[13] Feist AM, Henry CS, Reed JL, et al.. A genome-scale metabolic reconstruction for Escherichia coli K-12 MG1655 that accounts for 1260 ORFs and thermodynamic information. Mol Syst Biol, 2007, 3: 121.

[14] O’Brien EJ, Lerman JA, Chang RL, et al.. Genome-scale models of metabolism and gene expression extend and refine growth phenotype prediction. Mol Syst Biol, 2013, 9(1): 693

[15] Carrera J, Estrela R, Luo J, et al.. An integrative, multi-scale, genome-wide model reveals the phenotypic landscape of Escherichia coli. Mol Syst Biol, 2014, 10(7): 735.

[16] Khodayari A, Maranas CD. A genome-scale Escherichia coli kinetic metabolic model k-ecoli457 satisfying flux data for multiple mutant strains. Nat Commun, 2016, 7: 13806.

[17] Navid A, Almaas E. Genome-level transcription data of Yersinia pestis analyzed with a new metabolic constraint-based approach[J]. BMC Syst Biol, 2012, 6(1): 150-167.

[18] Lee N, Lee CH, Lee D, et al.. Genome-scale metabolic network reconstruction and in silico analysis of hexanoic acid producing Megasphaera elsdenii[J]. Micro-organisms, 2020, 8(4): 539-550

[19] Chung WY, Zhu Y, Mahamad MM, et al.. Novel antimicrobial development using genome-scale metabolic model of gram-negative pathogens: a review [J]. J Antibiot, 2021, 74(2): 95-104.

[20] GI Guzmán, Utrilla J, Nurk S, et al. Model-driven discovery of underground metabolic functions in Escherichia coli. Proc Natl Acad Sci USA 2015;112(3):929.

[21] Liu LM, Liu T, Zou W. Constraint-based algorithms for genome scale metabolic model-a review[J]. Chinese J Bioprocess Eng, 2012, 010(006): 70-77

[22] Orth JD, Thiele I, Palsson BØ. What is flux balance analysis?. Nat Biotechnol, 2010, 28: 245-248.

[23] Schilling CH, Edwards JS, Palsson BØ. Toward metabolic phenomics: analysis of genomic data using flux balances. Biotechnol Prog, 1999, 15: 288-295.

[24] Schellenberger J, Que R, Fleming RMT, Thiele I, Orth JD, Feist AM, Zielinski DC, Bordbar A, Lewis NE, Rahmanian S, et al.. Quantitative prediction of cellular metabolism with constraint-based models: The COBRA Toolbox v2.0. Nat Protoc, 2011, 6: 1290-1307.

[25] Schuetz R, Kuepfer L, Sauer U. Systematic evaluation of objective functions for predicting intracellular fluxes in Escherichia coli. Mol Syst Biol, 2007, 3(1): 119.

[26] Fan SC, Zhang ZY, Zou W, et al.. Development of a minimal chemically defined medium for Ketogulonicigenium vulgare WSH001 based on its genome-scale metabolic model. J Biotechnol, 2014, 169: 15-22.

[27] Ye C, Xu N, Chen HQ, et al.. Reconstruction and analysis of a genome-scale metabolic model of the oleaginous fungus Mortierella alpina. BMC Syst Biol, 2015, 9: 1.

[28] Segre D, Vitkup D, Church GM. Analysis of optimality in natural and perturbed metabolic networks. Proc Natl Acad Sci USA, 2002, 99: 15112-15117.

[29] Alper H, Jin YS, Moxley JF, et al.. Identifying gene targets for the metabolic engineering of lycopene biosynthesis in Escherichia coli. Metab Eng, 2005, 7: 155-164.

[30] Ren S, Bo Z, Qian X. Adaptive bi-level programming for optimal gene knockouts for targeted overproduction under phenotypic constraints[J]. BMC Bioinform, 2013, 14(Suppl 2): S17.

[31] Arif MA, Mohamad MS, Abd Latif MS, et al.. A hybrid of Cuckoo Search and Minimization of Metabolic Adjustment to optimize metabolites production in genome-scale models. Comput Biol Med, 2018, 102: 112-119.

[32] Cambridge UC, Cohen PC. Applied multiple regression/ correlation analysis for the behavioral sciences. J Royal Statist Soc Series D (The Statistician), 2003, 52(4): 691

[33] Carbonell P, Jervis AJ, Robinson CJ, et al.. An automated design-build-test-learn pipeline for enhanced microbial production of fine chemicals. Commun Biol, 2018, 1: 66.

[34] Giuseppe M, Guido Z, Claudio A. Multimodal regularized linear models with flux balance analysis for mechanistic integration of omics data[J]. Bioinformatics, 2021, 20: 20

[35] Menard S. Logistic Regression[J]. Am Stat, 2004, 58(4): 364

[36] Liao JG, Chin KV. Logistic regression for disease classification using microarray data: model selection in a large p and small n case[J]. Bioinformatics, 2007, 23(15): 1945-1951.

[37] Balázs, Szappanos, Károly, et al. An integrated approach to characterize genetic interaction networks in yeast metabolism.[J]. Nat Genet 2011:43:656-662

[38] Saunders C, Stitson MO, Weston J, et al.. Support Vector Machine[J]. Computer Science, 2002, 1(4): 1-28

[39] Shaked I, Oberhardt M, Atias N, et al.. Metabolic network prediction of drug side effects[J]. Cell Syst, 2016, 2(3): 209-213.

[40] Nandi S, Subramanian A, Sarkar RR. An integrative machine learning strategy for improved prediction of essential genes in Escherichia coli metabolism using flux-coupled features[J]. Mol BioSyst, 2017, 13: 1584-1596.

[41] Li G, Rabe K S, Nielsen J, et al. Machine Learning Applied to Predicting Microorganism Growth Temperatures and Enzyme Catalytic Optima[J]. ACS Synth Biol 2019; 8(6).

[42] Jain AK, Mao J, Mohiuddin KM. Artificial neural networks: a tutorial[J]. Computer, 2015, 29(3): 31-44.

[43] Hailin, Meng, Jianfeng, et al. quantitative design of regulatory elements based on high-precision strength prediction using artificial neural network[J]. PLoS One, 2013; 8(4):e60288.

[44] Zhou, YK, Li, et al. MiYA, an efficient machine-learning workflow in conjunction with the YeastFab assembly strategy for combinatorial optimization of heterologous metabolic pathways in Saccharomyces cerevisiae[J]. Metab Eng 2018.

[45] Jervis A J, Carbonell P, Vinaixa M, et al. Machine Learning of Designed Translational Control Allows Predictive Pathway Optimization in Escherichia coli[J]. ACS Synth Biol, 2018, 8(1).

[46] Awad M, Khanna R. Deep Neural Networks[C]// Apress. Apress, 2015;127–147.

[47] Guo W, You X, Feng X. DeepMetabolism: A deep learning system to predict phenotype from genome sequencing[J]. 2017.

[48] Yousoff S, ‘Amirah Baharin, Abdullah A. Differential Search Algorithm in Deep Neural Network for the Predictive Analysis of Xylitol Production in Escherichia Coli[J]. Springer, Singapore, 2017.

[49] Jae, Yong, Ryu, et al. Deep learning enables high-quality and high-throughput prediction of enzyme commission numbers[J]. Proc Natl Acad Sci USA, 2019; 116(28):13996–14001.

[50] Kotopka BJ, Smolke CD. Model-driven generation of artificial yeast promoters[J]. Nat Commun 2020; 11(1).

[51] Wang Y, Wang H, Wei L, et al.. Synthetic promoter design in Escherichia coli based on a deep generative network[J]. Nucleic Acids Res, 2020, 12: 12

[52] Hastie T, Tibshirani R, Friedman J. Ensemble Learning[M], 2009. New York, Springer.

[53] Heckmann D, Lloyd C J, Mih N, et al. Machine learning applied to enzyme turnover numbers reveals protein structural correlates and improves metabolic models[J]. Nat Commun 2018; 9(1).

[54] Oyetunde T, Liu D, Martin HG, et al. Machine learning framework for assessment of microbial factory performance[J]. PLoS One, 2019; 14(1).

[55] Giordano PC, Beccaria AJ, Goicoechea HC, et al. Optimization of the hydrolysis of lignocellulosic residues by using radial basis functions modeling and particle swarm optimization[J]. Bioche Eng J 2013; 80(Complete): 1–9.

[56] Viswanadham S, Meyer AG, Piyush R, et al. Predicting growth conditions from internal metabolic fluxes in an in-silico model of E. coli[J]. PLoS One, 2014; 9(12): e114608.

[57] Zampieri G, Coggins M, Valle G, et al. A poly-omics machine-learning method to predict metabolite production in CHO cells[C]// Int Electron Confer Metabolomics. 2017.

[58] Whitley D. A genetic algorithm tutorial. Stat Comput, 1994, 4(2): 65-85.

[59] Patil K R, Rocha I, Förster J, et al. Evolutionary programming as a platform for in silico metabolic engineering[J]. BMC Bioinform 2005; 6.

[60] Ismail MA, Deris S, Mohamad MS, et al.. A hybrid of Newton method and genetic algorithm for constrained optimization method of the production of metabolic pathway[J]. Life Sci J, 2014, 11(9): 409-414

[61] Arfian IM, Safaai D, Saberi MM, et al.. A newton cooperative genetic algorithm method for In Silico optimization of metabolic pathway production[J]. PLoS ONE, 2015, 10(5. ArticleID: e0126199

[62] Nair G, Jungreuthmayer C, Hanscho M, et al.. Designing minimal microbial strains of desired functionality using a genetic algorithm[J]. Algorithms Mol Biol AMB, 2015, 10(1): 29-29.

[63] Shabestary K, Hudson EP. Computational metabolic engineering strategies for growth-coupled biofuel production by Synechocystis[J]. Metab Eng Commun, 2016, 3: 216-226.

[64] Patane A, Santoro A, Costanza J, et al.. Pareto optimal design for synthetic biology[J]. IEEE Trans Biomed Circuits Syst, 2015, 9(4): 555-571.

[65] Alter TB, Blank LM, Ebert BE. Genetic optimization algorithm for metabolic engineering revisited. 2018.

[66] Patane A, Jansen G, Conca P, et al.. Multi-objective optimization of genome-scale metabolic models: the case of ethanol production[J]. Ann Oper Res, 2019, 276(1–2): 211-227.

[67] Colorni A, Dorigo M, Maniezzo V (1991) Ant system: an autocatalytic optimization process. Technical Report No. 91–016, Politecnico di Milano, Italy

[68] Chong SK, Mohamad MS, Salleh AM, et al.. A hybrid of ant colony optimization and minimization of metabolic adjustment to improve the production of succinic acid in Escherichia coli[J]. Comput Biol Med, 2014, 49C: 74-82.

[69] Shi JL, Salleh A, Mohamad MS, et al.. Identification of gene knockout strategies using a hybrid of an ant colony optimization algorithm and flux balance analysis to optimize microbial strains[J]. Comput Biol Chem, 2014, 53: 175-183.

[70] Salleh A, Mohamad MS, Deris S, et al.. Gene knockout identification for metabolite production improvement using a hybrid of genetic ant colony optimization and flux balance analysis[J]. Biotechnol Bioprocess Eng, 2015, 20(4): 685-693.

[71] Kennedy I, Eberhart R. “Particle swarm optimization”. Pmc. IEEE int. Conf. On Neural Network, 1995; 1942–1948.

[72] Govind N, Christian J, Jăźrgen Z. Optimal knockout strategies in genome-scale metabolic networks using particle swarm optimization. 2017.

[73] Lee MK, Mohamad MS, Choon YW, et al.. A Hybrid of Particle Swarm Optimization and Minimization of Metabolic Adjustment for Ethanol Production of Escherichia Coli[M], 2019. Cham, Springer

[74] Lee MK, Mohamad MS, Choon YW, et al. Comparison of optimization-modelling methods for metabolites production in Escherichia coli[J]. J Integ Bioinform 2020.

[75] Storn R, Price K. Differential evolution–a simple and efficient heuristic for global optimization over continuous spaces. J Global Optim, 1997, 11: 341-359.

[76] Rashid AH, Choon YW, Mohamad MS, et al.. Producing succinic acid in yeast using a hybrid of differential evolution and flux balance analysis[J]. Int J Bio-Sci Bio-Technol, 2013, 5(6): 91-100.

[77] Wang FS, Wu WH. Optimal design of growth-coupled production strains using nested hybrid differential evolution[J]. J Taiwan Inst Chem Eng, 2015, 54: 57-63.

[78] Daud KM, Zakaria Z, Shah ZA, et al.. A hybrid of differential search algorithm and flux balance analysis to: Identify knockout strategies for in silico optimization of metabolites production[J]. Int J Adv Soft Comput Applic, 2018, 10(2): 84-107

[79] KMD A, Msmb C, Zz A, et al. A non-dominated sorting differential search algorithm flux balance analysis (ndsDSAFBA) for in silico multiobjective optimization in identifying reactions knockout[J]. Computers Biol Med 113.

[80] Pham DT, Ghanbarzadeh A, Koc E, Otri S, Rahim S, Zaidi M: The Bees Algorithm-A novel tool for complex optimisation problems. Intellig Product Mach Syst (2006).

[81] Yin LH, Choon YW, Mohamad MS, et al. Prediction of vanillin and glutamate productions in yeast using a hybrid of continuous bees algorithm and flux balance analysis (CBAFBA)[J]. Curr Bioinform 2014.

[82] Lee SS, Choon YW, Chai LE, et al. A hybrid of artificial bee colony and flux balance analysis for identifying optimum knockout strategies for producing high yields of lactate in Escherichia Coli[C]// Springer Berlin Heidelberg. Springer Berlin Heidelberg, 2013.

[83] Koo CL, Mohamad MS, Ornatu S, et al. A gene knockout strategy for succinate production using a hybrid algorithm of bees algorithm and minimization of metabolic adjustment. IEEE, 2014.

[84] Choon YW, Mohamad MS, Deris S, et al.. A hybrid of bees algorithm and flux balance analysis (BAFBA) for the optimisation of microbial strains[J]. Int J Data Min Bioinform, 2014, 10(2): 225-238.

[85] Choon YW, Mohamad MS, Deris S, et al.. Differential bees flux balance analysis with OptKnock for In Silico microbial strains optimization[J]. PLoS ONE, 2014, 9(7. ArticleID: e102744

[86] Tang PW, Choon YW, Mohamad MS, et al.. Optimising the production of succinate and lactate in Escherichia coli using a hybrid of artificial bee colony algorithm and minimisation of metabolic adjustment[J]. J Biosci Bioeng, 2015, 119(3): 363-368.

[87] Hon MK, Mohamad MS, Salleh AM, et al. Identifying a gene knockout strategy using a hybrid of simple constrained artificial bee colony algorithm and flux balance analysis to enhance the production of succinate and lactate in Escherichia Coli[J]. Interdiscip Sci Comput Life Sci 2019.

Funding

National key research and development program of China(2020YFA0908303)

RIGHTS & PERMISSIONS

Jiangnan University