Machine learning of designed translational control allows predictive pathway optimization in Escherichia coli

Adrian J. Jervis, Pablo Carbonell, Maria Vinaixa, Mark S. Dunstan, Katherine A. Hollywood, Christopher J. Robinson, Nicholas J.W. Rattray, Cunyu Yan, Neil Swainston, Andrew Currin, Rehana Sung, Helen Toogood, Sandra Taylor, Jean Loup Faulon, Rainer Breitling, Eriko Takano, Nigel S. Scrutton

Research output: Contribution to journalArticle

4 Citations (Scopus)

Abstract

The field of synthetic biology aims to make the design of biological systems predictable, shrinking the huge design space to practical numbers for testing. When designing microbial cell factories, most optimization efforts have focused on enzyme and strain selection/engineering, pathway regulation, and process development. In silico tools for the predictive design of bacterial ribosome binding sites (RBSs) and RBS libraries now allow translational tuning of biochemical pathways; however, methods for predicting optimal RBS combinations in multigene pathways are desirable. Here we present the implementation of machine learning algorithms to model the RBS sequence-phenotype relationship from representative subsets of large combinatorial RBS libraries allowing the accurate prediction of optimal high-producers. Applied to a recombinant monoterpenoid production pathway in Escherichia coli, our approach was able to boost production titers by over 60% when screening under 3% of a library. To facilitate library screening, a multiwell plate fermentation procedure was developed, allowing increased screening throughput with sufficient resolution to discriminate between high and low producers. High producers from one library did not translate during scale-up, but the reduced screening requirements allowed rapid rescreening at the larger scale. This methodology is potentially compatible with any biochemical pathway and provides a powerful tool toward predictive design of bacterial production chassis.

LanguageEnglish
Pages127-136
Number of pages10
JournalACS Synthetic Biology
Volume8
Issue number1
Early online date18 Dec 2018
DOIs
Publication statusPublished - 18 Jan 2019

Fingerprint

Binding sites
Ribosomes
Escherichia coli
Learning systems
Binding Sites
Libraries
Screening
Synthetic Biology
Monoterpenes
Chassis
Biological systems
Computer Simulation
Fermentation
Learning algorithms
Industrial plants
Enzymes
Tuning
Throughput
Machine Learning
Phenotype

Keywords

  • machine learning
  • pathway engineering
  • ribosome binding site
  • synthetic biology
  • terpenoids
  • translational tuning

Cite this

Jervis, A. J., Carbonell, P., Vinaixa, M., Dunstan, M. S., Hollywood, K. A., Robinson, C. J., ... Scrutton, N. S. (2019). Machine learning of designed translational control allows predictive pathway optimization in Escherichia coli. ACS Synthetic Biology, 8(1), 127-136. https://doi.org/10.1021/acssynbio.8b00398
Jervis, Adrian J. ; Carbonell, Pablo ; Vinaixa, Maria ; Dunstan, Mark S. ; Hollywood, Katherine A. ; Robinson, Christopher J. ; Rattray, Nicholas J.W. ; Yan, Cunyu ; Swainston, Neil ; Currin, Andrew ; Sung, Rehana ; Toogood, Helen ; Taylor, Sandra ; Faulon, Jean Loup ; Breitling, Rainer ; Takano, Eriko ; Scrutton, Nigel S. / Machine learning of designed translational control allows predictive pathway optimization in Escherichia coli. In: ACS Synthetic Biology. 2019 ; Vol. 8, No. 1. pp. 127-136.
@article{4e40c11389384258b5aa99d16bfd50e1,
title = "Machine learning of designed translational control allows predictive pathway optimization in Escherichia coli",
abstract = "The field of synthetic biology aims to make the design of biological systems predictable, shrinking the huge design space to practical numbers for testing. When designing microbial cell factories, most optimization efforts have focused on enzyme and strain selection/engineering, pathway regulation, and process development. In silico tools for the predictive design of bacterial ribosome binding sites (RBSs) and RBS libraries now allow translational tuning of biochemical pathways; however, methods for predicting optimal RBS combinations in multigene pathways are desirable. Here we present the implementation of machine learning algorithms to model the RBS sequence-phenotype relationship from representative subsets of large combinatorial RBS libraries allowing the accurate prediction of optimal high-producers. Applied to a recombinant monoterpenoid production pathway in Escherichia coli, our approach was able to boost production titers by over 60{\%} when screening under 3{\%} of a library. To facilitate library screening, a multiwell plate fermentation procedure was developed, allowing increased screening throughput with sufficient resolution to discriminate between high and low producers. High producers from one library did not translate during scale-up, but the reduced screening requirements allowed rapid rescreening at the larger scale. This methodology is potentially compatible with any biochemical pathway and provides a powerful tool toward predictive design of bacterial production chassis.",
keywords = "machine learning, pathway engineering, ribosome binding site, synthetic biology, terpenoids, translational tuning",
author = "Jervis, {Adrian J.} and Pablo Carbonell and Maria Vinaixa and Dunstan, {Mark S.} and Hollywood, {Katherine A.} and Robinson, {Christopher J.} and Rattray, {Nicholas J.W.} and Cunyu Yan and Neil Swainston and Andrew Currin and Rehana Sung and Helen Toogood and Sandra Taylor and Faulon, {Jean Loup} and Rainer Breitling and Eriko Takano and Scrutton, {Nigel S.}",
year = "2019",
month = "1",
day = "18",
doi = "10.1021/acssynbio.8b00398",
language = "English",
volume = "8",
pages = "127--136",
journal = "ACS Synthetic Biology",
issn = "2161-5063",
publisher = "American Chemical Society",
number = "1",

}

Jervis, AJ, Carbonell, P, Vinaixa, M, Dunstan, MS, Hollywood, KA, Robinson, CJ, Rattray, NJW, Yan, C, Swainston, N, Currin, A, Sung, R, Toogood, H, Taylor, S, Faulon, JL, Breitling, R, Takano, E & Scrutton, NS 2019, 'Machine learning of designed translational control allows predictive pathway optimization in Escherichia coli' ACS Synthetic Biology, vol. 8, no. 1, pp. 127-136. https://doi.org/10.1021/acssynbio.8b00398

Machine learning of designed translational control allows predictive pathway optimization in Escherichia coli. / Jervis, Adrian J.; Carbonell, Pablo; Vinaixa, Maria; Dunstan, Mark S.; Hollywood, Katherine A.; Robinson, Christopher J.; Rattray, Nicholas J.W.; Yan, Cunyu; Swainston, Neil; Currin, Andrew; Sung, Rehana; Toogood, Helen; Taylor, Sandra; Faulon, Jean Loup; Breitling, Rainer; Takano, Eriko; Scrutton, Nigel S.

In: ACS Synthetic Biology, Vol. 8, No. 1, 18.01.2019, p. 127-136.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Machine learning of designed translational control allows predictive pathway optimization in Escherichia coli

AU - Jervis, Adrian J.

AU - Carbonell, Pablo

AU - Vinaixa, Maria

AU - Dunstan, Mark S.

AU - Hollywood, Katherine A.

AU - Robinson, Christopher J.

AU - Rattray, Nicholas J.W.

AU - Yan, Cunyu

AU - Swainston, Neil

AU - Currin, Andrew

AU - Sung, Rehana

AU - Toogood, Helen

AU - Taylor, Sandra

AU - Faulon, Jean Loup

AU - Breitling, Rainer

AU - Takano, Eriko

AU - Scrutton, Nigel S.

PY - 2019/1/18

Y1 - 2019/1/18

N2 - The field of synthetic biology aims to make the design of biological systems predictable, shrinking the huge design space to practical numbers for testing. When designing microbial cell factories, most optimization efforts have focused on enzyme and strain selection/engineering, pathway regulation, and process development. In silico tools for the predictive design of bacterial ribosome binding sites (RBSs) and RBS libraries now allow translational tuning of biochemical pathways; however, methods for predicting optimal RBS combinations in multigene pathways are desirable. Here we present the implementation of machine learning algorithms to model the RBS sequence-phenotype relationship from representative subsets of large combinatorial RBS libraries allowing the accurate prediction of optimal high-producers. Applied to a recombinant monoterpenoid production pathway in Escherichia coli, our approach was able to boost production titers by over 60% when screening under 3% of a library. To facilitate library screening, a multiwell plate fermentation procedure was developed, allowing increased screening throughput with sufficient resolution to discriminate between high and low producers. High producers from one library did not translate during scale-up, but the reduced screening requirements allowed rapid rescreening at the larger scale. This methodology is potentially compatible with any biochemical pathway and provides a powerful tool toward predictive design of bacterial production chassis.

AB - The field of synthetic biology aims to make the design of biological systems predictable, shrinking the huge design space to practical numbers for testing. When designing microbial cell factories, most optimization efforts have focused on enzyme and strain selection/engineering, pathway regulation, and process development. In silico tools for the predictive design of bacterial ribosome binding sites (RBSs) and RBS libraries now allow translational tuning of biochemical pathways; however, methods for predicting optimal RBS combinations in multigene pathways are desirable. Here we present the implementation of machine learning algorithms to model the RBS sequence-phenotype relationship from representative subsets of large combinatorial RBS libraries allowing the accurate prediction of optimal high-producers. Applied to a recombinant monoterpenoid production pathway in Escherichia coli, our approach was able to boost production titers by over 60% when screening under 3% of a library. To facilitate library screening, a multiwell plate fermentation procedure was developed, allowing increased screening throughput with sufficient resolution to discriminate between high and low producers. High producers from one library did not translate during scale-up, but the reduced screening requirements allowed rapid rescreening at the larger scale. This methodology is potentially compatible with any biochemical pathway and provides a powerful tool toward predictive design of bacterial production chassis.

KW - machine learning

KW - pathway engineering

KW - ribosome binding site

KW - synthetic biology

KW - terpenoids

KW - translational tuning

UR - http://www.scopus.com/inward/record.url?scp=85059811943&partnerID=8YFLogxK

UR - https://pubs.acs.org/journal/asbcd6

U2 - 10.1021/acssynbio.8b00398

DO - 10.1021/acssynbio.8b00398

M3 - Article

VL - 8

SP - 127

EP - 136

JO - ACS Synthetic Biology

T2 - ACS Synthetic Biology

JF - ACS Synthetic Biology

SN - 2161-5063

IS - 1

ER -