Segment particle swarm optimization adoption for large-scale kinetic parameter identification of escherichia coli metabolic network model

• Main Authors: Azrag, Mohammed Adam Kunna, Tuty Asmawaty, Abdul Kadir, Jaber, Aqeel S.
• Format: Article
• Language: English
• Published: Institute of Electrical and Electronics Engineers Inc. 2018
• Subjects: QA75 Electronic computers. Computer science; QA76 Computer software; QH Natural history; TA Engineering (General). Civil engineering (General)
• Online Access: http://umpir.ump.edu.my/id/eprint/33504/1/Segment%20particle%20swarm%20optimization%20adoption%20for%20large-scale%20kinetic%20parameter%20l.pdf; http://umpir.ump.edu.my/id/eprint/33504/; https://doi.org/10.1109/ACCESS.2018.2885118; https://doi.org/10.1109/ACCESS.2018.2885118
• Institution: Universiti Malaysia Pahang

Kinetic parameter identification in the dynamic metabolic model of Escherichia coli (E. coli) has become important and is needed to obtain appropriate metabolite and enzyme data that are valid under in vivo conditions. The dynamic metabolic model under study represents five metabolic pathways with more than 170 kinetic parameters at steady state with a 0.1 dilution rate. In this paper, identification is declared in two steps. The first step is to identify which kinetic parameters have a higher impact on the model response using local sensitivity analysis results upon increasing each kinetic parameter up to 2.0 by steps of 0.5, while the second step uses highly sensitive kinetic results to be identified and minimized the model simulation metabolite errors using real experimental data by adopting. However, this paper focuses on adopting segment particle swarm optimization (PSO) and PSO algorithms for large-scale kinetic parameters identification. Among the 170 kinetic parameters investigated, seven kinetic parameters were found to be the most effective kinetic parameters in the model response after finalizing the sensitivity. The seven sensitive kinetic parameters were used in both the algorithms to minimize the model response errors. The validation results proved the effectiveness of both the proposed methods, which identified the kinetics and minimized the model response errors perfectly.