Amir Pandi
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Amir Pandi has a total of 115 co-authors. Here are the top co-authors by number of shared publications:
- Jean-Loup Faulon
Micalis Institute
10 shared publications - Manish Kushwaha
Université Paris-Saclay, INRAE, AgroParisTech, MIcalis Institute
6 shared publications - Mathilde Koch
Micalis Institute, INRA, AgroParisTech, Université Paris-Saclay, Jouy-en-Josas, France
6 shared publications - Tobias Jürgen Erb
Max Planck Institute for Terrestrial Microbiology
5 shared publications - Ariel B. Lindner
Sorbonne Université, CNRS, ERL U1338 Inserm, Department of Computational, Quantitative and Synthetic Biology, Paris, France
4 shared publications David Adam
Department of Biochemistry and Synthetic Metabolism, Max Planck Institute for Terrestrial Microbiology, Marburg, Germany
4 shared publications
Latest Publications
- 3 April 2026
- Published by Agence Bibliographique de l'Enseignement Supérieur
Synthetic Metabolic Circuits for Bioproduction, Biosensing and Biocomputation
Circuits métaboliques synthétiques pour la bioproduction, la biodétection et le biocalcul: La biologie synthétique est le domaine de la bioingénierie permettant de concevoir, de construire et de tester de nouveaux systèmes biologiques en réécrire le code génétique. Les circuits biologiques synthétiques sont des outils sophistiqués permettant de construire des réseaux biologiques pour des applications médicales, industrielles et environnementales. Cette thèse de doctorat porte sur le développement de voies métaboliques synthétiques conçues à l'aide d'outils informatiques. Ces voies métaboliques sont intégrées à la couche de régulation transcriptionnelle pour développer des biocircuits pour la bioproduction, la biodétection et la biocalcul dans des systèmes cellulaires et acellulaires. Les résultats obtenus durant cette thèse de doctorat révèlent le nouveau potentiel des voies métaboliques dans l'établissement de biocircuits synthétiques. Le volet bioproduction-biodétection de la thèse vise à développer un nouveau biocapteur pour un sucre rare utilisé pour améliorer l'activité catalytique d’enzyme dans la cellule (in vivo). Ce biocapteur a ensuite été implémenté dans un système acellulaire (in vitro) pour découvrir et optimiser le comportement de biocapteurs à base de répresseurs. Une fois optimisé en système acellulaire, notre biocapteur a été utilisé pour surveiller la production enzymatique de sucre rare. Le développement de biocapteurs procaryotes acellulaires, qui reposent principalement sur des répresseurs, permet d'accélérer et de rendre plus efficace le cycle “design-build-test” dans le prototypage des voies métaboliques dans les systèmes acellulaires. L'application de la biodétection des circuits métaboliques pour le diagnostic est la mise en œuvre et l'optimisation des transducteurs métaboliques dans le système acellulaire. Les transducteurs sont des voies métaboliques composées d'au moins une enzyme catalysant un métabolite indétectable en un inducteur transcriptionnel, augmentant ainsi le nombre de petites molécules biologiquement détectables. En tant que nouvelle approche pour effectuer des biocalculs, des circuits métaboliques ont été appliqués pour construire des additionneurs métaboliques et des perceptrons métaboliques. Dans la cellule, trois transducteurs métaboliques et un additionneur métabolique ont été construits et caractérisés. Les systèmes acellulaires permettent d’accélérer la caractérisation de circuits biologiques, de finement régler le niveau d’expression d’un ou plusieurs gènes et facilite l’expression de plusieurs plasmides simultanément. Ceci a permis de construire de multiples transducteurs pondérés et des additionneurs métaboliques. Le modèle basé sur des données expérimentales a permis de concevoir un perceptron métabolique pour construire des classificateurs binaires à quatre entrées. Les additionneurs, perceptrons et classificateurs peuvent être utilisés dans des applications avancées telles que la détection de précision et dans le développement de souches pour le génie métabolique ou la thérapeutique intelligente.
- 3 April 2026
- Published by Agence Bibliographique de l'Enseignement Supérieur
Synthetic Metabolic Circuits for Bioproduction, Biosensing and Biocomputation
Circuits métaboliques synthétiques pour la bioproduction, la biodétection et le biocalcul: La biologie synthétique est le domaine de la bioingénierie permettant de concevoir, de construire et de tester de nouveaux systèmes biologiques en réécrire le code génétique. Les circuits biologiques synthétiques sont des outils sophistiqués permettant de construire des réseaux biologiques pour des applications médicales, industrielles et environnementales. Cette thèse de doctorat porte sur le développement de voies métaboliques synthétiques conçues à l'aide d'outils informatiques. Ces voies métaboliques sont intégrées à la couche de régulation transcriptionnelle pour développer des biocircuits pour la bioproduction, la biodétection et la biocalcul dans des systèmes cellulaires et acellulaires. Les résultats obtenus durant cette thèse de doctorat révèlent le nouveau potentiel des voies métaboliques dans l'établissement de biocircuits synthétiques. Le volet bioproduction-biodétection de la thèse vise à développer un nouveau biocapteur pour un sucre rare utilisé pour améliorer l'activité catalytique d’enzyme dans la cellule (in vivo). Ce biocapteur a ensuite été implémenté dans un système acellulaire (in vitro) pour découvrir et optimiser le comportement de biocapteurs à base de répresseurs. Une fois optimisé en système acellulaire, notre biocapteur a été utilisé pour surveiller la production enzymatique de sucre rare. Le développement de biocapteurs procaryotes acellulaires, qui reposent principalement sur des répresseurs, permet d'accélérer et de rendre plus efficace le cycle “design-build-test” dans le prototypage des voies métaboliques dans les systèmes acellulaires. L'application de la biodétection des circuits métaboliques pour le diagnostic est la mise en œuvre et l'optimisation des transducteurs métaboliques dans le système acellulaire. Les transducteurs sont des voies métaboliques composées d'au moins une enzyme catalysant un métabolite indétectable en un inducteur transcriptionnel, augmentant ainsi le nombre de petites molécules biologiquement détectables. En tant que nouvelle approche pour effectuer des biocalculs, des circuits métaboliques ont été appliqués pour construire des additionneurs métaboliques et des perceptrons métaboliques. Dans la cellule, trois transducteurs métaboliques et un additionneur métabolique ont été construits et caractérisés. Les systèmes acellulaires permettent d’accélérer la caractérisation de circuits biologiques, de finement régler le niveau d’expression d’un ou plusieurs gènes et facilite l’expression de plusieurs plasmides simultanément. Ceci a permis de construire de multiples transducteurs pondérés et des additionneurs métaboliques. Le modèle basé sur des données expérimentales a permis de concevoir un perceptron métabolique pour construire des classificateurs binaires à quatre entrées. Les additionneurs, perceptrons et classificateurs peuvent être utilisés dans des applications avancées telles que la détection de précision et dans le développement de souches pour le génie métabolique ou la thérapeutique intelligente.
- Open Access
Serotype-independent inhibition of S. pneumoniae by SMiTE, a commensal-derived bacteriocin
Streptococcus pneumoniae remains a leading cause of disease despite widespread vaccination, highlighting the need for serotype-independent strategies. We recently identified commensal streptococci that produce bacteriocins with anti-pneumococcal activity. Here, we evaluate these bacteriocins as candidates for pneumococcal control. Using cell-free protein synthesis, we screened 58 bacteriocins, the majority of which absent in available pneumococcal genomes, and identified SMiTE as the most potent. Purified SMiTE disrupted pneumococcal membrane integrity, as shown by confocal, transmission, and scanning electron microscopy, induced ATP leakage, and triggered transcriptional responses consistent with envelope stress and metabolic remodeling. In a mouse nasopharyngeal colonization model, intranasal SMiTE treatments reduced pneumococcal loads by 65-fold with no significant weight loss. SMiTE inhibited a broad range of serotypes, with strongest activity against serotype 3, which is poorly controlled by current vaccines, while sparing most oral and upper respiratory tract commensals. Repeated sub-inhibitory exposure did not select resistant mutants in line with the membrane-targeting mechanism. These findings establish SMiTE as a commensal-derived strategy for serotype-independent, microbiota-sparing pneumococcal decolonization.
Open AccessSerotype-independent inhibition of S. pneumoniae by SMiTE, a commensal-derived bacteriocin
Streptococcus pneumoniae remains a leading cause of disease despite widespread vaccination, highlighting the need for serotype-independent strategies. We recently identified commensal streptococci that produce bacteriocins with anti-pneumococcal activity. Here, we evaluate these bacteriocins as candidates for pneumococcal control. Using cell-free protein synthesis, we screened 58 bacteriocins, the majority of which absent in available pneumococcal genomes, and identified SMiTE as the most potent. Purified SMiTE disrupted pneumococcal membrane integrity, as shown by confocal, transmission, and scanning electron microscopy, induced ATP leakage, and triggered transcriptional responses consistent with envelope stress and metabolic remodeling. In a mouse nasopharyngeal colonization model, intranasal SMiTE treatments reduced pneumococcal loads by 65-fold with no significant weight loss. SMiTE inhibited a broad range of serotypes, with strongest activity against serotype 3, which is poorly controlled by current vaccines, while sparing most oral and upper respiratory tract commensals. Repeated sub-inhibitory exposure did not select resistant mutants in line with the membrane-targeting mechanism. These findings establish SMiTE as a commensal-derived strategy for serotype-independent, microbiota-sparing pneumococcal decolonization.
- Top Cited PapersOpen Access
- 3 April 2025
- Published by Springer Nature in Nature Communications
- Vol. 16 (1) , 1-12
CodonTransformer: a multispecies codon optimizer using context-aware neural networks
Degeneracy in the genetic code allows many possible DNA sequences to encode the same protein. Optimizing codon usage within a sequence to meet organism-specific preferences faces combinatorial explosion. Nevertheless, natural sequences optimized through evolution provide a rich source of data for machine learning algorithms to explore the underlying rules. Here, we introduce CodonTransformer, a multispecies deep learning model trained on over 1 million DNA-protein pairs from 164 organisms spanning all domains of life. The model demonstrates context-awareness thanks to its Transformers architecture and to our sequence representation strategy that combines organism, amino acid, and codon encodings. CodonTransformer generates host-specific DNA sequences with natural-like codon distribution profiles and with minimum negative cis-regulatory elements. This work introduces the strategy of Shared Token Representation and Encoding with Aligned Multi-masking (STREAM) and provides a codon optimization framework with a customizable open-access model and a user-friendly Google Colab interface.
- 3 April 2025
- Published by Springer Nature in Nature Communications
- Vol. 16 (1) , 1-12
Top Cited PapersOpen AccessCodonTransformer: a multispecies codon optimizer using context-aware neural networks
Degeneracy in the genetic code allows many possible DNA sequences to encode the same protein. Optimizing codon usage within a sequence to meet organism-specific preferences faces combinatorial explosion. Nevertheless, natural sequences optimized through evolution provide a rich source of data for machine learning algorithms to explore the underlying rules. Here, we introduce CodonTransformer, a multispecies deep learning model trained on over 1 million DNA-protein pairs from 164 organisms spanning all domains of life. The model demonstrates context-awareness thanks to its Transformers architecture and to our sequence representation strategy that combines organism, amino acid, and codon encodings. CodonTransformer generates host-specific DNA sequences with natural-like codon distribution profiles and with minimum negative cis-regulatory elements. This work introduces the strategy of Shared Token Representation and Encoding with Aligned Multi-masking (STREAM) and provides a codon optimization framework with a customizable open-access model and a user-friendly Google Colab interface.
- Open Access
Generalizable prediction of liquid-liquid phase separation from protein sequence
Liquid-liquid phase separation (LLPS) is a physicochemical process through which a homogeneous liquid solution spontaneously separates into distinct liquid phases with different compositions and properties. Driven by weak multivalent interactions, LLPS in living systems enables dynamic compartmentalization of biomolecules to promote or regulate various cellular processes. Despite recent advances, predicting phase-separating proteins and their key LLPS-driving regions remains limited by the versatility of models. Here, we developed Phaseek, a generalizable LLPS predicting model that combines contextual sequence encoding with statistical protein graph representations. Phaseek accurately predicts LLPS-prone proteins in diverse biological contexts and identifies key regions and effects of point mutations. Proteome-wide analysis across 18 species suggests evolutionary conservation of LLPS among orthologs and associated biological processes. Additionally, predictions by Phaseek highlight the key physicochemical and structural properties associated with LLPS. Provided as an open-access model with a user-friendly implementation, Phaseek serves as a multipurpose LLPS predictor for advancing fundamental and applied research.
Open AccessGeneralizable prediction of liquid-liquid phase separation from protein sequence
Liquid-liquid phase separation (LLPS) is a physicochemical process through which a homogeneous liquid solution spontaneously separates into distinct liquid phases with different compositions and properties. Driven by weak multivalent interactions, LLPS in living systems enables dynamic compartmentalization of biomolecules to promote or regulate various cellular processes. Despite recent advances, predicting phase-separating proteins and their key LLPS-driving regions remains limited by the versatility of models. Here, we developed Phaseek, a generalizable LLPS predicting model that combines contextual sequence encoding with statistical protein graph representations. Phaseek accurately predicts LLPS-prone proteins in diverse biological contexts and identifies key regions and effects of point mutations. Proteome-wide analysis across 18 species suggests evolutionary conservation of LLPS among orthologs and associated biological processes. Additionally, predictions by Phaseek highlight the key physicochemical and structural properties associated with LLPS. Provided as an open-access model with a user-friendly implementation, Phaseek serves as a multipurpose LLPS predictor for advancing fundamental and applied research.
- Open Access
- 27 December 2024
- Published by Oxford University Press (OUP) in Nucleic Acids Research
- Vol. 53 (3)
Phage-mediated intercellular CRISPRi for biocomputation in bacterial consortia
Coordinated actions of cells in microbial communities and multicellular organisms enable them to perform complex tasks otherwise difficult for single cells. This has inspired biological engineers to build cellular consortia for larger circuits with improved functionalities while implementing communication systems for coordination among cells. Here, we investigate the signalling dynamics of a phage-mediated synthetic DNA messaging system and couple it with CRISPR interference to build distributed circuits that perform logic gate operations in multicellular bacterial consortia. We find that growth phases of both sender and receiver cells, as well as resource competition between them, shape communication outcomes. Leveraging the easy programmability of DNA messages, we build eight orthogonal signals and demonstrate that intercellular CRISPRi (i-CRISPRi) regulates gene expression across cells. Finally, we multiplex the i-CRISPRi system to implement several multicellular logic gates that involve up to seven cells and take up to three inputs simultaneously, with single- and dual-rail encoding: NOT, YES, AND and AND-AND-NOT. The communication system developed here lays the groundwork for implementing complex biological circuits in engineered bacterial communities, using phage signals for communication.
- 27 December 2024
- Published by Oxford University Press (OUP) in Nucleic Acids Research
- Vol. 53 (3)
Open AccessPhage-mediated intercellular CRISPRi for biocomputation in bacterial consortia
Coordinated actions of cells in microbial communities and multicellular organisms enable them to perform complex tasks otherwise difficult for single cells. This has inspired biological engineers to build cellular consortia for larger circuits with improved functionalities while implementing communication systems for coordination among cells. Here, we investigate the signalling dynamics of a phage-mediated synthetic DNA messaging system and couple it with CRISPR interference to build distributed circuits that perform logic gate operations in multicellular bacterial consortia. We find that growth phases of both sender and receiver cells, as well as resource competition between them, shape communication outcomes. Leveraging the easy programmability of DNA messages, we build eight orthogonal signals and demonstrate that intercellular CRISPRi (i-CRISPRi) regulates gene expression across cells. Finally, we multiplex the i-CRISPRi system to implement several multicellular logic gates that involve up to seven cells and take up to three inputs simultaneously, with single- and dual-rail encoding: NOT, YES, AND and AND-AND-NOT. The communication system developed here lays the groundwork for implementing complex biological circuits in engineered bacterial communities, using phage signals for communication.
- Open Access
qByte: Open-source isothermal fluorimeter for democratizing analysis of nucleic acids, proteins and cells
Access to affordable and reliable scientific instrumentation remains a significant barrier to the democratization of healthcare and scientific research. In the field of biotechnology, in particular, the complexity, cost, and infrastructure requirements of many instruments continue to limit their accessibility, especially in resource-limited environments. Despite the recent increase in the development of open-source tools, driven by advances in digital fabrication and electronic prototyping, few of these projects have reached large-scale implementation or validation in real-world settings. Here, we present qByte, an open-source, 8-tube isothermal fluorimeter designed to overcome these barriers by offering a cost-effective ($60) yet production-ready solution. qByte leverages standard digital manufacturing and Printed Circuit Board (PCB) assembly techniques and is designed to be portable, making it ideal for both laboratory and field use. The device has been benchmarked against commercial real-time thermocyclers and spectrophotometers, showing comparable results across four key applications: nucleic acid amplification and detection, analysis of protein activity and stability, genetic construct characterization, and bacterial viability tests. Further testing including the on-site diagnosis of human parasites in Kumasi, Ghana, and field work in Patagonia, Chile, validated qByte’s reliability in real-world conditions. Taken together, our results proved qByte as flexible and reliable equipment for a variety of biological tests and applications, while its affordability and open-source design simplify further development and allow adaptation to the needs of future users.
Open AccessqByte: Open-source isothermal fluorimeter for democratizing analysis of nucleic acids, proteins and cells
Access to affordable and reliable scientific instrumentation remains a significant barrier to the democratization of healthcare and scientific research. In the field of biotechnology, in particular, the complexity, cost, and infrastructure requirements of many instruments continue to limit their accessibility, especially in resource-limited environments. Despite the recent increase in the development of open-source tools, driven by advances in digital fabrication and electronic prototyping, few of these projects have reached large-scale implementation or validation in real-world settings. Here, we present qByte, an open-source, 8-tube isothermal fluorimeter designed to overcome these barriers by offering a cost-effective ($60) yet production-ready solution. qByte leverages standard digital manufacturing and Printed Circuit Board (PCB) assembly techniques and is designed to be portable, making it ideal for both laboratory and field use. The device has been benchmarked against commercial real-time thermocyclers and spectrophotometers, showing comparable results across four key applications: nucleic acid amplification and detection, analysis of protein activity and stability, genetic construct characterization, and bacterial viability tests. Further testing including the on-site diagnosis of human parasites in Kumasi, Ghana, and field work in Patagonia, Chile, validated qByte’s reliability in real-world conditions. Taken together, our results proved qByte as flexible and reliable equipment for a variety of biological tests and applications, while its affordability and open-source design simplify further development and allow adaptation to the needs of future users.
- Open Access
CodonTransformer: a multispecies codon optimizer using context-aware neural networks
The genetic code is degenerate allowing a multitude of possible DNA sequences to encode the same protein. This degeneracy impacts the efficiency of heterologous protein production due to the codon usage preferences of each organism. The process of tailoring organism-specific synonymous codons, known as codon optimization, must respect local sequence patterns that go beyond global codon preferences. As a result, the search space faces a combinatorial explosion that makes exhaustive exploration impossible. Nevertheless, throughout the diverse life on Earth, natural selection has already optimized the sequences, thereby providing a rich source of data allowing machine learning algorithms to explore the underlying rules. Here, we introduce CodonTransformer, a multispecies deep learning model trained on over 1 million DNA-protein pairs from 164 organisms spanning all kingdoms of life. The model demonstrates context-awareness thanks to the attention mechanism and bidirectionality of the Transformers we used, and to a novel sequence representation that combines organism, amino acid, and codon encodings. CodonTransformer generates host-specific DNA sequences with natural-like codon distribution profiles and with negative cis-regulatory elements. This work introduces a novel strategy of Shared Token Representation and Encoding with Aligned Multi-masking (STREAM) and provides a state-of-the-art codon optimization framework with a customizable open-access model and a user-friendly interface.
Open AccessCodonTransformer: a multispecies codon optimizer using context-aware neural networks
The genetic code is degenerate allowing a multitude of possible DNA sequences to encode the same protein. This degeneracy impacts the efficiency of heterologous protein production due to the codon usage preferences of each organism. The process of tailoring organism-specific synonymous codons, known as codon optimization, must respect local sequence patterns that go beyond global codon preferences. As a result, the search space faces a combinatorial explosion that makes exhaustive exploration impossible. Nevertheless, throughout the diverse life on Earth, natural selection has already optimized the sequences, thereby providing a rich source of data allowing machine learning algorithms to explore the underlying rules. Here, we introduce CodonTransformer, a multispecies deep learning model trained on over 1 million DNA-protein pairs from 164 organisms spanning all kingdoms of life. The model demonstrates context-awareness thanks to the attention mechanism and bidirectionality of the Transformers we used, and to a novel sequence representation that combines organism, amino acid, and codon encodings. CodonTransformer generates host-specific DNA sequences with natural-like codon distribution profiles and with negative cis-regulatory elements. This work introduces a novel strategy of Shared Token Representation and Encoding with Aligned Multi-masking (STREAM) and provides a state-of-the-art codon optimization framework with a customizable open-access model and a user-friendly interface.
- Top Cited PapersOpen Access
- 8 November 2023
- Published by Springer Nature in Nature Communications
- Vol. 14 (1) , 1-14
Cell-free biosynthesis combined with deep learning accelerates de novo-development of antimicrobial peptides
Bioactive peptides are key molecules in health and medicine. Deep learning holds a big promise for the discovery and design of bioactive peptides. Yet, suitable experimental approaches are required to validate candidates in high throughput and at low cost. Here, we established a cell-free protein synthesis (CFPS) pipeline for the rapid and inexpensive production of antimicrobial peptides (AMPs) directly from DNA templates. To validate our platform, we used deep learning to design thousands of AMPs de novo. Using computational methods, we prioritized 500 candidates that we produced and screened with our CFPS pipeline. We identified 30 functional AMPs, which we characterized further through molecular dynamics simulations, antimicrobial activity and toxicity. Notably, six de novo-AMPs feature broad-spectrum activity against multidrug-resistant pathogens and do not develop bacterial resistance. Our work demonstrates the potential of CFPS for high throughput and low-cost production and testing of bioactive peptides within less than 24 h.
- 8 November 2023
- Published by Springer Nature in Nature Communications
- Vol. 14 (1) , 1-14
Top Cited PapersOpen AccessCell-free biosynthesis combined with deep learning accelerates de novo-development of antimicrobial peptides
Bioactive peptides are key molecules in health and medicine. Deep learning holds a big promise for the discovery and design of bioactive peptides. Yet, suitable experimental approaches are required to validate candidates in high throughput and at low cost. Here, we established a cell-free protein synthesis (CFPS) pipeline for the rapid and inexpensive production of antimicrobial peptides (AMPs) directly from DNA templates. To validate our platform, we used deep learning to design thousands of AMPs de novo. Using computational methods, we prioritized 500 candidates that we produced and screened with our CFPS pipeline. We identified 30 functional AMPs, which we characterized further through molecular dynamics simulations, antimicrobial activity and toxicity. Notably, six de novo-AMPs feature broad-spectrum activity against multidrug-resistant pathogens and do not develop bacterial resistance. Our work demonstrates the potential of CFPS for high throughput and low-cost production and testing of bioactive peptides within less than 24 h.
- Open Access
Cell-free biosynthesis combined with deep learning accelerates de novo-development of antimicrobial peptides
Bioactive peptides are key molecules in health and medicine. Deep learning holds a big promise for the discovery and design of bioactive peptides. Yet, suitable experimental approaches are required to validate candidates in high throughput and at low cost. Here, we established a cell- free protein synthesis (CFPS) pipeline for the rapid and inexpensive production of antimicrobial peptides (AMPs) directly from DNA templates. To validate our platform, we used deep learning to design thousands of AMPs de novo. Using computational methods, we prioritized 500 candidates that we produced and screened with our CFPS pipeline. We identified 30 functional AMPs, which we characterized further through molecular dynamics simulations, antimicrobial activity and toxicity. Notably, six de novo-AMPs feature broad-spectrum activity against multidrug-resistant pathogens and do not develop bacterial resistance. Our work demonstrates the potential of CFPS for production and testing of bioactive peptides within less than 24 hours and <10$ per screen.
Open AccessCell-free biosynthesis combined with deep learning accelerates de novo-development of antimicrobial peptides
Bioactive peptides are key molecules in health and medicine. Deep learning holds a big promise for the discovery and design of bioactive peptides. Yet, suitable experimental approaches are required to validate candidates in high throughput and at low cost. Here, we established a cell- free protein synthesis (CFPS) pipeline for the rapid and inexpensive production of antimicrobial peptides (AMPs) directly from DNA templates. To validate our platform, we used deep learning to design thousands of AMPs de novo. Using computational methods, we prioritized 500 candidates that we produced and screened with our CFPS pipeline. We identified 30 functional AMPs, which we characterized further through molecular dynamics simulations, antimicrobial activity and toxicity. Notably, six de novo-AMPs feature broad-spectrum activity against multidrug-resistant pathogens and do not develop bacterial resistance. Our work demonstrates the potential of CFPS for production and testing of bioactive peptides within less than 24 hours and <10$ per screen.
- Top Cited PapersOpen Access
- 5 July 2022
- Published by Springer Nature in Nature Communications
- Vol. 13 (1) , 1-15
A versatile active learning workflow for optimization of genetic and metabolic networks
Optimization of biological networks is often limited by wet lab labor and cost, and the lack of convenient computational tools. Here, we describe METIS, a versatile active machine learning workflow with a simple online interface for the data-driven optimization of biological targets with minimal experiments. We demonstrate our workflow for various applications, including cell-free transcription and translation, genetic circuits, and a 27-variable synthetic CO2-fixation cycle (CETCH cycle), improving these systems between one and two orders of magnitude. For the CETCH cycle, we explore 1025 conditions with only 1,000 experiments to yield the most efficient CO2-fixation cascade described to date. Beyond optimization, our workflow also quantifies the relative importance of individual factors to the performance of a system identifying unknown interactions and bottlenecks. Overall, our workflow opens the way for convenient optimization and prototyping of genetic and metabolic networks with customizable adjustments according to user experience, experimental setup, and laboratory facilities.
- 5 July 2022
- Published by Springer Nature in Nature Communications
- Vol. 13 (1) , 1-15
Top Cited PapersOpen AccessA versatile active learning workflow for optimization of genetic and metabolic networks
Optimization of biological networks is often limited by wet lab labor and cost, and the lack of convenient computational tools. Here, we describe METIS, a versatile active machine learning workflow with a simple online interface for the data-driven optimization of biological targets with minimal experiments. We demonstrate our workflow for various applications, including cell-free transcription and translation, genetic circuits, and a 27-variable synthetic CO2-fixation cycle (CETCH cycle), improving these systems between one and two orders of magnitude. For the CETCH cycle, we explore 1025 conditions with only 1,000 experiments to yield the most efficient CO2-fixation cascade described to date. Beyond optimization, our workflow also quantifies the relative importance of individual factors to the performance of a system identifying unknown interactions and bottlenecks. Overall, our workflow opens the way for convenient optimization and prototyping of genetic and metabolic networks with customizable adjustments according to user experience, experimental setup, and laboratory facilities.