Tuning the Performance of Synthetic Riboswitches using Machine Learning

Groher, Ann-Christin; Jäger, Sven; Schneider, Christopher; Groher, Florian; Hamacher, Kay; Suess, Beatrix

doi:10.1021/acssynbio.8b00207

Cited by 41 publications

(29 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For example, to date, <1000 total toehold switches have been designed and tested [2][3][4][5][6]9,15,16 . While a recent attempt was made to apply deep learning to a riboswitch dataset with 263 variants 22 , the lack of high-throughput datasets has generally limited the synthetic biology community's ability to analyze this type of response molecule using deep-learning techniques. High-throughput assays that utilize deep sequencing to analyze fluorescencesorted bacteria have previously been used to characterize the translation of Escherichia coli mRNA [23][24][25][26][27] ; in this study, in order to improve our understanding and ability to predict new functional RNA-based response elements, we synthesized and characterized an extensive in vivo library of toehold switches using a high-throughput flow-seq (also known as sort-seq) 23,24 pipeline for subsequent exploration using various machine-learning and deep-learning architectures.…”

Section: Resultsmentioning

confidence: 99%

A deep learning approach to programmable RNA switches

et al. 2020

View full text Add to dashboard Cite

Engineered RNA elements are programmable tools capable of detecting small molecules, proteins, and nucleic acids. Predicting the behavior of these synthetic biology components remains a challenge, a situation that could be addressed through enhanced pattern recognition from deep learning. Here, we investigate Deep Neural Networks (DNN) to predict toehold switch function as a canonical riboswitch model in synthetic biology. To facilitate DNN training, we synthesize and characterize in vivo a dataset of 91,534 toehold switches spanning 23 viral genomes and 906 human transcription factors. DNNs trained on nucleotide sequences outperform (R2 = 0.43–0.70) previous state-of-the-art thermodynamic and kinetic models (R2 = 0.04–0.15) and allow for human-understandable attention-visualizations (VIS4Map) to identify success and failure modes. This work shows that deep learning approaches can be used for functionality predictions and insight generation in RNA synthetic biology.

show abstract

Section: Resultsmentioning

confidence: 99%

A deep learning approach to programmable RNA switches

et al. 2020

View full text Add to dashboard Cite

show abstract

“…The goal of this technique is to explore efficient synthetic routes to produce a target molecule from a host plant species. In addition, convolutional neural networks combined with linear regression models (Groher et al, 2018;Carbonell et al, 2018) can be used to help in optimizing plasmid copy number and selecting promoter region which can be helpful in plant transformation experiments. Recently, an advanced deep learning-based method DeepCRISPR (Chuai et al, 2018) is capable of predicting on-target to be knocked out and off-target sites of single-guide RNAs efficiently.…”

Section: Role Of Artificial Intelligence In Plant Metabolic Pathway Ementioning

confidence: 99%

An outlook on metabolic pathway engineering in crop plants

Sahoo

Mohapatra²,

Mishra

2020

Arch. Agric. Environ. Sci.

View full text Add to dashboard Cite

To produce the essential secondary metabolites, plants are the major and important target source materials for conducting the high-profile metabolic engineering studies. Metabolic pathway engineering of both microorganism targets and plants target contribute towards important drug discovery. In order to efficiently work out in advanced plant metabolic pathway engineering techniques, a detailed knowledge and expertise is essentially needed regarding the plant cell physiology and the mechanics of plant metabolism. Mathematical and statistical models to scale and map the genome for integrative metabolic pathway activity, signal transduction mechanism in the genome, gene regulation and the networks of protein-protein interaction can provide the in-depth knowledge to work efficiently on plant metabolic pathway engineering studies. Incorporation of omics data into these statistical and mathematical models is crucial in the case of drug discovery using the plant system. Recently, artificial intelligence concept and approaches are experimentally applied for efficient and accurate metabolic engineering in plants.

show abstract

“…As recent results suggest synthetic riboswitches are amenable to improvement with machine learning approaches 51 , 52 , we sought to further optimize sequences. To that end, we constructed two optimization pipelines, coined NuSpeak (Fig.…”

Section: Resultsmentioning

confidence: 99%

Sequence-to-function deep learning frameworks for engineered riboregulators

et al. 2020

View full text Add to dashboard Cite

While synthetic biology has revolutionized our approaches to medicine, agriculture, and energy, the design of completely novel biological circuit components beyond naturally-derived templates remains challenging due to poorly understood design rules. Toehold switches, which are programmable nucleic acid sensors, face an analogous design bottleneck; our limited understanding of how sequence impacts functionality often necessitates expensive, time-consuming screens to identify effective switches. Here, we introduce Sequence-based Toehold Optimization and Redesign Model (STORM) and Nucleic-Acid Speech (NuSpeak), two orthogonal and synergistic deep learning architectures to characterize and optimize toeholds. Applying techniques from computer vision and natural language processing, we ‘un-box’ our models using convolutional filters, attention maps, and in silico mutagenesis. Through transfer-learning, we redesign sub-optimal toehold sensors, even with sparse training data, experimentally validating their improved performance. This work provides sequence-to-function deep learning frameworks for toehold selection and design, augmenting our ability to construct potent biological circuit components and precision diagnostics.

show abstract

Tuning the Performance of Synthetic Riboswitches using Machine Learning

Cited by 41 publications

References 36 publications

A deep learning approach to programmable RNA switches

A deep learning approach to programmable RNA switches

An outlook on metabolic pathway engineering in crop plants

Sequence-to-function deep learning frameworks for engineered riboregulators

Contact Info

Product

Resources

About