Combining mechanistic and machine learning models for predictive engineering and optimization of tryptophan metabolism

Jie Zhang; Søren D. Petersen; Tijana Radivojevic; Andrés Ramirez; Andrés Pérez-Manríquez; Eduardo Abeliuk; Benjamín J. Sánchez; Zak Costello; Yu Chen; Michael J. Fero; Hector Garcia Martin; Jens Nielsen; Jay D. Keasling; Michael K. Jensen

doi:10.1038/s41467-020-17910-1

Combining mechanistic and machine learning models for predictive engineering and optimization of tryptophan metabolism

Jie Zhang, Søren D. Petersen, Tijana Radivojevic, Andrés Ramirez, Andrés Pérez-Manríquez, Eduardo Abeliuk, Benjamín J. Sánchez, Zak Costello, Yu Chen, Michael J. Fero, Hector Garcia Martin, Jens Nielsen, Jay D. Keasling, Michael K. Jensen^*

^*Corresponding author for this work

Research output: Contribution to journal › Journal article › Research › peer-review

166 Downloads (Pure)

Abstract

Through advanced mechanistic modeling and the generation of large high-quality datasets, machine learning is becoming an integral part of understanding and engineering living systems. Here we show that mechanistic and machine learning models can be combined to enable accurate genotype-to-phenotype predictions. We use a genome-scale model to pinpoint engineering targets, efficient library construction of metabolic pathway designs, and high-throughput biosensor-enabled screening for training diverse machine learning algorithms. From a single data-generation cycle, this enables successful forward engineering of complex aromatic amino acid metabolism in yeast, with the best machine learning-guided design recommendations improving tryptophan titer and productivity by up to 74 and 43%, respectively, compared to the best designs used for algorithm training. Thus, this study highlights the power of combining mechanistic and machine learning models to effectively direct metabolic engineering efforts.

Original language	English
Article number	4880
Journal	Nature Communications
Volume	11
Issue number	1
Number of pages	13
ISSN	2041-1723
DOIs	https://doi.org/10.1038/s41467-020-17910-1
Publication status	Published - 1 Dec 2020

Bibliographical note

Funding Information:
This work was supported by the Novo Nordisk Foundation and the European Commission Horizon 2020 programme (grant agreement no. 722287 and no. 686070). This work was also part of the DOE Agile BioFoundry (http://agilebiofoundry.org), supported by the U.S. Department of Energy, Energy Efficiency and Renewable Energy, Bioenergy Technologies Office, and the DOE Joint BioEnergy Institute (http://www.jbei.org), supported by the Office of Science, Office of Biological and Environmental Research, through contract DE-AC02-05CH11231 between Lawrence Berkeley National Laboratory and the U.S. Department of Energy. The Department of Energy will provide public access to these results of federally sponsored research in accordance with the DOE Public Access Plan (http://energy.gov/downloads/doe-public-access-plan). H.G.M. was also supported by the Basque Government through the BERC 2014-2017 program, and by Spanish Ministry of Economy and Competitiveness MINECO: BCAM Severo Ochoa excellence accreditation SEV-2013-0323. This work was also supported by the Chilean economic development agency, Corfo, through grant 17IEAT-73382.

Publisher Copyright:
© 2020, The Author(s).

Access to Document

10.1038/s41467-020-17910-1

FulltextFinal published version, 1.11 MB

OpenUrl availability

Full text

Cite this

Zhang, J., Petersen, S. D., Radivojevic, T., Ramirez, A., Pérez-Manríquez, A., Abeliuk, E., Sánchez, B. J., Costello, Z., Chen, Y., Fero, M. J., Martin, H. G., Nielsen, J., Keasling, J. D., & Jensen, M. K. (2020). Combining mechanistic and machine learning models for predictive engineering and optimization of tryptophan metabolism. Nature Communications, 11(1), Article 4880. https://doi.org/10.1038/s41467-020-17910-1

@article{d3f43470124242828114549377408fd6,

title = "Combining mechanistic and machine learning models for predictive engineering and optimization of tryptophan metabolism",

abstract = "Through advanced mechanistic modeling and the generation of large high-quality datasets, machine learning is becoming an integral part of understanding and engineering living systems. Here we show that mechanistic and machine learning models can be combined to enable accurate genotype-to-phenotype predictions. We use a genome-scale model to pinpoint engineering targets, efficient library construction of metabolic pathway designs, and high-throughput biosensor-enabled screening for training diverse machine learning algorithms. From a single data-generation cycle, this enables successful forward engineering of complex aromatic amino acid metabolism in yeast, with the best machine learning-guided design recommendations improving tryptophan titer and productivity by up to 74 and 43%, respectively, compared to the best designs used for algorithm training. Thus, this study highlights the power of combining mechanistic and machine learning models to effectively direct metabolic engineering efforts.",

author = "Jie Zhang and Petersen, {S{\o}ren D.} and Tijana Radivojevic and Andr{\'e}s Ramirez and Andr{\'e}s P{\'e}rez-Manr{\'i}quez and Eduardo Abeliuk and S{\'a}nchez, {Benjam{\'i}n J.} and Zak Costello and Yu Chen and Fero, {Michael J.} and Martin, {Hector Garcia} and Jens Nielsen and Keasling, {Jay D.} and Jensen, {Michael K.}",

note = "Funding Information: This work was supported by the Novo Nordisk Foundation and the European Commission Horizon 2020 programme (grant agreement no. 722287 and no. 686070). This work was also part of the DOE Agile BioFoundry (http://agilebiofoundry.org), supported by the U.S. Department of Energy, Energy Efficiency and Renewable Energy, Bioenergy Technologies Office, and the DOE Joint BioEnergy Institute (http://www.jbei.org), supported by the Office of Science, Office of Biological and Environmental Research, through contract DE-AC02-05CH11231 between Lawrence Berkeley National Laboratory and the U.S. Department of Energy. The Department of Energy will provide public access to these results of federally sponsored research in accordance with the DOE Public Access Plan (http://energy.gov/downloads/doe-public-access-plan). H.G.M. was also supported by the Basque Government through the BERC 2014-2017 program, and by Spanish Ministry of Economy and Competitiveness MINECO: BCAM Severo Ochoa excellence accreditation SEV-2013-0323. This work was also supported by the Chilean economic development agency, Corfo, through grant 17IEAT-73382. Publisher Copyright: {\textcopyright} 2020, The Author(s).",

year = "2020",

month = dec,

day = "1",

doi = "10.1038/s41467-020-17910-1",

language = "English",

volume = "11",

journal = "Nature Communications",

issn = "2041-1723",

publisher = "Nature Publishing Group",

number = "1",

}

Zhang, J, Petersen, SD, Radivojevic, T, Ramirez, A, Pérez-Manríquez, A, Abeliuk, E, Sánchez, BJ, Costello, Z, Chen, Y, Fero, MJ, Martin, HG, Nielsen, J, Keasling, JD & Jensen, MK 2020, 'Combining mechanistic and machine learning models for predictive engineering and optimization of tryptophan metabolism', Nature Communications, vol. 11, no. 1, 4880. https://doi.org/10.1038/s41467-020-17910-1

TY - JOUR

T1 - Combining mechanistic and machine learning models for predictive engineering and optimization of tryptophan metabolism

AU - Zhang, Jie

AU - Petersen, Søren D.

AU - Radivojevic, Tijana

AU - Ramirez, Andrés

AU - Pérez-Manríquez, Andrés

AU - Abeliuk, Eduardo

AU - Sánchez, Benjamín J.

AU - Costello, Zak

AU - Chen, Yu

AU - Fero, Michael J.

AU - Martin, Hector Garcia

AU - Nielsen, Jens

AU - Keasling, Jay D.

AU - Jensen, Michael K.

N1 - Funding Information: This work was supported by the Novo Nordisk Foundation and the European Commission Horizon 2020 programme (grant agreement no. 722287 and no. 686070). This work was also part of the DOE Agile BioFoundry (http://agilebiofoundry.org), supported by the U.S. Department of Energy, Energy Efficiency and Renewable Energy, Bioenergy Technologies Office, and the DOE Joint BioEnergy Institute (http://www.jbei.org), supported by the Office of Science, Office of Biological and Environmental Research, through contract DE-AC02-05CH11231 between Lawrence Berkeley National Laboratory and the U.S. Department of Energy. The Department of Energy will provide public access to these results of federally sponsored research in accordance with the DOE Public Access Plan (http://energy.gov/downloads/doe-public-access-plan). H.G.M. was also supported by the Basque Government through the BERC 2014-2017 program, and by Spanish Ministry of Economy and Competitiveness MINECO: BCAM Severo Ochoa excellence accreditation SEV-2013-0323. This work was also supported by the Chilean economic development agency, Corfo, through grant 17IEAT-73382. Publisher Copyright: © 2020, The Author(s).

PY - 2020/12/1

Y1 - 2020/12/1

N2 - Through advanced mechanistic modeling and the generation of large high-quality datasets, machine learning is becoming an integral part of understanding and engineering living systems. Here we show that mechanistic and machine learning models can be combined to enable accurate genotype-to-phenotype predictions. We use a genome-scale model to pinpoint engineering targets, efficient library construction of metabolic pathway designs, and high-throughput biosensor-enabled screening for training diverse machine learning algorithms. From a single data-generation cycle, this enables successful forward engineering of complex aromatic amino acid metabolism in yeast, with the best machine learning-guided design recommendations improving tryptophan titer and productivity by up to 74 and 43%, respectively, compared to the best designs used for algorithm training. Thus, this study highlights the power of combining mechanistic and machine learning models to effectively direct metabolic engineering efforts.

AB - Through advanced mechanistic modeling and the generation of large high-quality datasets, machine learning is becoming an integral part of understanding and engineering living systems. Here we show that mechanistic and machine learning models can be combined to enable accurate genotype-to-phenotype predictions. We use a genome-scale model to pinpoint engineering targets, efficient library construction of metabolic pathway designs, and high-throughput biosensor-enabled screening for training diverse machine learning algorithms. From a single data-generation cycle, this enables successful forward engineering of complex aromatic amino acid metabolism in yeast, with the best machine learning-guided design recommendations improving tryptophan titer and productivity by up to 74 and 43%, respectively, compared to the best designs used for algorithm training. Thus, this study highlights the power of combining mechanistic and machine learning models to effectively direct metabolic engineering efforts.

U2 - 10.1038/s41467-020-17910-1

DO - 10.1038/s41467-020-17910-1

M3 - Journal article

C2 - 32978375

AN - SCOPUS:85091480477

SN - 2041-1723

VL - 11

JO - Nature Communications

JF - Nature Communications

IS - 1

M1 - 4880

ER -

Combining mechanistic and machine learning models for predictive engineering and optimization of tryptophan metabolism

Abstract

Bibliographical note

Access to Document

OpenUrl availability

Fingerprint

Cite this