[1]
[2]
iMM904, and e_coli_core. GDLS had the minimum elapsed
time, but its success ratio was low. As described in Section 3.2, many GDLS strategies resulted in target metabolite
production without growth or growth-coupled production
when the best PR was evaluated. However, when the worst
PR at GR maximization was evaluated, growth-coupled
production was rarely achieved. As optGene returns the
best solution at that point when the specified computation
time runs out, designating a longer computation time may
increase the success ratio.
TrimGdel, optGene, and GDLS are in descending order
of the average number of deleted genes, which is consistent with the descending order of the success ratio. One
reason for the high success ratio of TrimGdel is the success
in searching for larger gene deletion strategies. The elapsed
time of TrimGdel, which was less than 20 m for every case,
even for genome-scale models, is acceptable for actual use.
The average numbers of deleted genes for TrimGdel were
8.90%, 6.02%, and 9.04% of whole genes, but they were
less than 5% for more than a half of the gene deletion
strategies.
For TrimGdel, the number of gene deletions was between
1% and 5% of whole genes for 66.5%, 50.5%, and 56.8% of
the successful gene deletion strategies for iML1515,
iMM904, and e_coli_core, respectively. On the other hand,
for GDLS, the number of gene deletions was 1% or less of
whole genes for 100%, 100%, 33.3% of the successful gene
deletion strategies for iML1515, iMM904, and e_coli_core,
respectively. These numbers were 100%, 80.1%, and 4% for
optGene, respectively. It was seen that GDLS and optGene
could derive gene deletion strategies whose sizes are 1% or
more in a small model. However, we can conclude that
TrimGdel is much more suitable for deriving gene deletion
strategies whose sizes are between 1% and 5% of whole
genes for genome-scale models.
Biotin Production.
In all cases described in Table 7, the values of biotin synthase and MALCOAMT are identical, which may imply
that malonyl CoA should be produced in abundance in biotin production. The sum of the absolute reaction rates of all
reactions (total flux) is correlated with the number of gene
deletions and the growth rate but not with the biotin production rate. Because more gene deletions lead to more
repressed reactions, it results in a lower total flux and GR.
However, there seem to be other factors that determine
whether growth-coupled production of biotin occurs.
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
W. B. Copeland et al., “Computational tools for metabolic engineering,” Metabolic Eng., vol. 14, no. 3, pp. 270–280, 2012.
A. P. Burgard, P. Pharkya, and C. D. Maranas, “OptKnock:
A bilevel programming framework for identifying gene knockout
strategies for microbial strain optimization,” Biotechnol. Bioeng.,
vol. 84, no. 6, pp. 647–657, 2003.
P. Pharkya, A. P. Burgard, and C. D. Maranas, “OptStrain:
A computational framework for redesign of microbial production
systems,” Genome Res., vol. 14, no. 11, pp. 2367–2376, 2004.
P. Pharkya and C. D. Maranas, “An optimization framework for
identifying reaction activation/inhibition or elimination candidates for overproduction in microbial systems,” Metabolic Eng.,
vol. 8, no. 1, pp. 1–13, 2006.
K. R. Patil, I. Rocha, J. F€
orster, and J. Nielsen, “Evolutionary programming as a platform for in silico metabolic engineering,” BMC
Bioinf., vol. 6, no. 1, 2005, Art. no. 308.
S. Ranganathan, P. F. Suthers, and C. D. Maranas, “OptForce: An
optimization procedure for identifying all genetic manipulations
leading to targeted overproductions,” PLoS One Comput. Biol,
vol. 6, no. 4, 2010, Art. no. e1000744.
I. Rocha et al., “OptFlux: An open-source software platform for in silico
metabolic engineering,” BMC Syst. Biol., vol. 4, no. 1, pp. 1–12, 2010.
Y. Toya and H. Shimizu, “Flux analysis and metabolomics for systematic metabolic engineering of microorganisms,” Biotechnol.
Adv., vol. 31, no. 6, pp. 818–826, 2013.
J. D. Orth, I. Thiele, and B. Palsson, “What is flux balance analysis?,” Nature Biotechnol., vol. 28, no. 3, pp. 245–248, 2010.
D. S. Lun et al., “Large-scale identification of genetic design strategies using local search,” Mol. Syst. Biol., vol. 5, no. 1, 2009,
Art. no. 296.
G. Rockwell, N. J. Guido, and G. M. Church, “Redirector: Designing cell factories by reconstructing the metabolic objective,” PLoS
One Comput. Biol.. vol. 9, no. 1, 2013, Art. no. e1002882.
L. Yang, W. R. Cluett, and R. Mahadevan, “EMILiO: A fast algorithm for genome-scale strain design,” Metabolic Eng., vol. 13,
no. 3, pp. 272–281, 2011.
D. Egen and D. S. Lun, “Truncated branch and bound achieves
efficient constraint-based genetic design,” Bioinformatics, vol. 28,
no. 12, pp. 1619–1623, 2012.
N. E. Lewis et al., “Omic data from evolved E coli are consistent
with computed optimal growth from genome-scale models,” Mol.
Syst. Biol., vol. 6, no. 1, 2010, Art. no. 390.
D. Gu, C. Zhang, S. Zhou, L. Wei, and Q. Hua, “IdealKnock:
A framework for efficiently identifying knockout strategies leading to targeted overproduction,” Comput. Biol. Chem., vol. 61,
pp. 229–237, 2016.
S. Ohno, H. Shimizu, and C. Furusawa, “FastPros: Screening of
reaction knockout strategies for metabolic engineering,” Bioinformatics, vol. 30, no. 7, pp. 981–987, 2014.
T. Tamura, “Grid-based computational methods for the design of
constraint-based parsimonious chemical reaction networks to simulate metabolite production: GridProd,” BMC Bioinf., vol. 19,
no. 1, 2018, Art. no. 325.
A. von Kamp and S. Klamt, “Enumeration of smallest intervention
strategies in genome-scale metabolic networks,” PLoS One Comput. Biol., vol. 10, no. 1, 2014, Art. no. e1003378.
O. H€adicke and S. Klamt, “Computing complex metabolic intervention strategies using constrained minimal cut sets,” Metabolic
Eng., vol. 13, no. 2, pp. 204–213, 2011.
B. J. Harder, K. Bettenbrock, and S. Klamt, “Model-based metabolic engineering enables high yield itaconic acid production by
escherichia coli,” Metabolic Eng., vol. 38, pp. 29–37, 2016.
TAMURA: TRIMMING GENE DELETION STRATEGIES FOR GROWTH-COUPLED PRODUCTION IN CONSTRAINT-BASED METABOLIC
[21] P. Schneider, A. von Kamp, and S. Klamt, “An extended and generalized framework for the calculation of metabolic intervention
strategies based on minimal cut sets,” PLoS One Comput. Biol.,
vol. 16, no. 7, 2020, Art. no. e1008110.
[22] D. Banerjee et al., “Genome-scale metabolic rewiring improves
titers rates and yields of the non-native product indigoidine at
scale,” Nature Commun., vol. 1, no. 1, pp. 1–11. 2020.
[23] D. Machado and M. J. Herrgard, and I. Rocha, “Stoichiometric
representation of gene–protein–reaction associations leverages
constraint-based analysis from reaction to gene-level phenotype
prediction,” PLoS One Comput. Biol., vol. 12, no. 10, 2016,
Art. no. e1005140.
[24] Z. Razaghi-Moghadam and Z. Nikoloski, “GeneReg: A constraintbased approach for design of feasible metabolic engineering strategies
at the gene level,” Bioinformatics, vol. 37, no. 12, pp. 1717–1723, 2021.
[25] T. Tamura, A. Muto-Fujita, Y. Tohsato, and T. Kosaka, “Gene deletion
algorithms for minimum reaction network design by mixed-integer
linear programming for metabolite production in constraint-based
models: gDel_minRN,” 2021, doi: 10.21203/rs.3.rs-611371/v1.
[26] I. Rocha, P. Maia, M. Rocha, and F. C. Ferreira, “OptGene:
A framework for in silico metabolic engineering,” in Proc. 10th
Int. Chem. Biol. Eng. Conf., 2008, pp. 218–219.
[27] J. M. Monk et al., “iML1515, a knowledgebase that computes Escherichia coli traits,” Nature Biotechnol., vol. 35, no. 10,
pp. 904–908, 2017.
[28] M. L. Mo, B. Palsson, and M. J. Herrgard, “Connecting extracellular metabolomic measurements to intracellular flux states in
yeast,” BMC Syst. Biol., vol. 3, no. 1, 2009, Art. no. 37.
[29] J. D. Orth, R. M. Fleming, and B. O. Palsson, “Reconstruction and
use of microbial metabolic networks: The core Escherichia coli
metabolic model as an educational guide,” EcoSal Plus, vol. 4,
pp. 1–48, 2010.
1549
[30] L. Heirendt et al., “Creation and analysis of biochemical constraint-based models using the COBRA Toolbox v. 3.0,” Nature
Protoc., vol. 14, no. 3, pp. 639–702, 2019.
[31] S. Liu, P. Kang, Z. Cui, Z. Wang, and T. Chen, “Increased riboflavin production by knockout of 6-phosphofructokinase I and
blocking the entner–doudoroff pathway in escherichia coli,”
Biotechnol. Lett., vol. 38, no. 8, pp. 1307–1314, 2016.
[32] A. Schrijver, Theory of Linear and Integer Programming. Hoboken,
NJ, USA: Wiley, 1998.
[33] E. Brunk et al., “Recon3D enables a three-dimensional view of
gene variation in human metabolism,” Nature Biotechnol., vol. 36,
no. 3, pp. 272–281, 2018.
Takeyuki Tamura (Member, IEEE) received the
BE, ME, and the PhD degrees in informatics from
Kyoto University, Japan, in 2001, 2003, and 2006,
respectively. He joined Bioinformatics Center,
Institute for Chemical Research, Kyoto University
as a postdoctoral fellow, in 2006. He worked as
an assistant professor from Dec. 2007 to Sep.
2017, and works as an associate professor from
Oct. 2017. His research interests include bioinformatics and the theory of combinatorial optimization for graphs and networks.
" For more information on this or any other computing topic,
please visit our Digital Library at www.computer.org/csdl.
...