Massively parallel characterization of engineered transcript isoforms using direct RNA sequencing

AND Logic Based on Suppressor tRNAs Enables Stringent Control of Sliding Base Editors in Pseudomonas putida

Base editors, e.g., cytosine deaminases, are powerful tools for precise DNA editing in vivo, enabling both targeted nucleotide conversions and segment-specific diversification of bacterial genomes. Yet, regulation of their spatiotemporal activity is crucial to avoid off-target effects and enabling controlled evolution of specific genes and pathways. This work reports a strategy for tight control of base-editing devices through subjecting their expression to a genetic AND logic gate in which two chemical inducer inputs are strictly required for cognate activity. The case study involves an archetypal genetic device consisting of a cytosine deaminase (pmCDA1) fused to a T7 RNA polymerase (RNAPT7), which cause intensive diversification of DNA portions bordered by a T7 promoter and a T7 terminator─but whose activity in vivo has been shown unattainable to govern with standard conditional expression systems. By encoding up to three UAG stop codons into the DNA sequence of the pmCDA1-RNAPT7 fusion, which is transcribed by the 3-methylbenzoate inducible promoter Pm, we first broke the structure of the hybrid protein. Then, to overcome the interruptions caused by UAG codons, we placed transcription of a supF tRNA under the control of a cyclohexanone-dependent system. When tested in the soil bacterium and metabolic engineering chassis Pseudomonas putida KT2440, these modifications changed the performance of the sliding base editor from a flawed YES logic to a precise AND logic. We also showed that such a 2-layer control brings about a minimal background activity as compared to a single-input digitalizer circuit. These results show the ability of suppressor tRNA-based logic gates for achieving stringent expression of otherwise difficult to control devices.

CryptKeeper: a negative design tool for reducing unintentional gene expression in bacteria

Foundational techniques in molecular biology-such as cloning genes, tagging biomolecules for purification or identification, and overexpressing recombinant proteins-rely on introducing non-native or synthetic DNA sequences into organisms. These sequences may be recognized by the transcription and translation machinery in their new context in unintended ways. The cryptic gene expression that sometimes results has been shown to produce genetic instability and mask experimental signals. Computational tools have been developed to predict individual types of gene expression elements, but it can be difficult for researchers to contextualize their collective output. Here, we introduce CryptKeeper, a software pipeline that visualizes predictions of Escherichia coli gene expression signals and estimates the translational burden possible from a DNA sequence. We investigate several published examples where cryptic gene expression in E. coli interfered with experiments. CryptKeeper accurately postdicts unwanted gene expression from both eukaryotic virus infectious clones and individual proteins that led to genetic instability. It also identifies off-target gene expression elements that resulted in truncations that confounded protein purification. Incorporating negative design using CryptKeeper into reverse genetics and synthetic biology workflows can help to mitigate cloning challenges and avoid unexplained failures and complications that arise from unintentional gene expression.

CryptKeeper: a negative design tool for reducing unintentional gene expression in bacteria

Foundational techniques in molecular biology-such as cloning genes, tagging biomolecules for purification or identification, and overexpressing recombinant proteins-rely on introducing non-native or synthetic DNA sequences into organisms. These sequences may be recognized by the transcription and translation machinery in their new context in unintended ways. The cryptic gene expression that sometimes results has been shown to produce genetic instability and mask experimental signals. Computational tools have been developed to predict individual types of gene expression elements, but it can be difficult for researchers to contextualize their collective output. Here, we introduce CryptKeeper, a software pipeline that visualizes predictions of bacterial gene expression signals and estimates the translational burden possible from a DNA sequence. We investigate several published examples where cryptic gene expression in E. coli interfered with experiments. CryptKeeper accurately postdicts unwanted gene expression from both eukaryotic virus infectious clones and individual proteins that led to genetic instability. It also identifies off-target gene expression elements that resulted in truncations that confounded protein purification. Incorporating negative design using CryptKeeper into reverse genetics and synthetic biology workflows can help to mitigate cloning challenges and avoid unexplained failures and complications that arise from unintentional gene expression.

Measuring the burden of hundreds of BioBricks defines an evolutionary limit on constructability in synthetic biology

Engineered DNA will slow the growth of a host cell if it redirects limiting resources or otherwise interferes with homeostasis. Escape mutants that alleviate this burden can rapidly evolve and take over cell populations, making genetic engineering less reliable and predictable. Synthetic biologists often use genetic parts encoded on plasmids, but their burden is rarely characterized. We measured how 301 BioBrick plasmids affected Escherichia coli growth and found that 59 (19.6%) were burdensome, primarily because they depleted the limited gene expression resources of host cells. Overall, no BioBricks reduced the growth rate of E. coli by >45%, which agreed with a population genetic model that predicts such plasmids should be unclonable. We made this model available online for education ( https://barricklab.org/burden-model ) and added our burden measurements to the iGEM Registry. Our results establish a fundamental limit on what DNA constructs and genetic modifications can be successfully engineered into cells.

Transfer learning for cross-context prediction of protein expression from 5'UTR sequence

Model-guided DNA sequence design can accelerate the reprogramming of living cells. It allows us to engineer more complex biological systems by removing the need to physically assemble and test each potential design. While mechanistic models of gene expression have seen some success in supporting this goal, data-centric, deep learning-based approaches often provide more accurate predictions. This accuracy, however, comes at a cost - a lack of generalization across genetic and experimental contexts that has limited their wider use outside the context in which they were trained. Here, we address this issue by demonstrating how a simple transfer learning procedure can effectively tune a pre-trained deep learning model to predict protein translation rate from 5' untranslated region (5'UTR) sequence for diverse contexts in Escherichia coli using a small number of new measurements. This allows for important model features learnt from expensive massively parallel reporter assays to be easily transferred to new settings. By releasing our trained deep learning model and complementary calibration procedure, this study acts as a starting point for continually refined model-based sequence design that builds on previous knowledge and future experimental efforts.

Data hazards in synthetic biology

Data science is playing an increasingly important role in the design and analysis of engineered biology. This has been fueled by the development of high-throughput methods like massively parallel reporter assays, data-rich microscopy techniques, computational protein structure prediction and design, and the development of whole-cell models able to generate huge volumes of data. Although the ability to apply data-centric analyses in these contexts is appealing and increasingly simple to do, it comes with potential risks. For example, how might biases in the underlying data affect the validity of a result and what might the environmental impact of large-scale data analyses be? Here, we present a community-developed framework for assessing data hazards to help address these concerns and demonstrate its application to two synthetic biology case studies. We show the diversity of considerations that arise in common types of bioengineering projects and provide some guidelines and mitigating steps. Understanding potential issues and dangers when working with data and proactively addressing them will be essential for ensuring the appropriate use of emerging data-intensive AI methods and help increase the trustworthiness of their applications in synthetic biology.

Engineering is evolution: a perspective on design processes to engineer biology

Careful consideration of how we approach design is crucial to all areas of biotechnology. However, choosing or developing an effective design methodology is not always easy as biology, unlike most areas of engineering, is able to adapt and evolve. Here, we put forward that design and evolution follow a similar cyclic process and therefore all design methods, including traditional design, directed evolution, and even random trial and error, exist within an evolutionary design spectrum. This contrasts with conventional views that often place these methods at odds and provides a valuable framework for unifying engineering approaches for challenging biological design problems.

Nanoporous Amorphous Carbon Monolayer Derived from Fullerene Film

Carbon materials derived from fullerene are reported recently with unique structures and properties. However, only micrometer size samples can be obtained that limits the studying and exploration for membrane applications. Here, the preparation of a centimeter-size nanoporous amorphous carbon monolayer is reported by rapid pyrolyzing a Langmuir-Blodgett film of fullerene. The sample is fully characterized and the results indicate that the amorphous carbon monolayer derived from fullerene is metastable and insulating. The ionic transmembrane transport study demonstrates that the membrane is porous and cation selective with a selectivity of 26%. This work provides new insights into the controlled synthesis of large-size metastable amorphous carbon monolayer.

Design Approaches to Expand the Toolkit for Building Cotranscriptionally Encoded RNA Strand Displacement Circuits

Cotranscriptionally encoded RNA strand displacement (ctRSD) circuits are an emerging tool for programmable molecular computation, with potential applications spanning in vitro diagnostics to continuous computation inside living cells. In ctRSD circuits, RNA strand displacement components are continuously produced together via transcription. These RNA components can be rationally programmed through base pairing interactions to execute logic and signaling cascades. However, the small number of ctRSD components characterized to date limits circuit size and capabilities. Here, we characterize over 200 ctRSD gate sequences, exploring different input, output, and toehold sequences and changes to other design parameters, including domain lengths, ribozyme sequences, and the order in which gate strands are transcribed. This characterization provides a library of sequence domains for engineering ctRSD components, i.e., a toolkit, enabling circuits with up to 4-fold more inputs than previously possible. We also identify specific failure modes and systematically develop design approaches that reduce the likelihood of failure across different gate sequences. Lastly, we show the ctRSD gate design is robust to changes in transcriptional encoding, opening a broad design space for applications in more complex environments. Together, these results deliver an expanded toolkit and design approaches for building ctRSD circuits that will dramatically extend capabilities and potential applications.

Design and Analysis of Massively Parallel Reporter Assays Using FORECAST

Machine learning is revolutionizing molecular biology and bioengineering by providing powerful insights and predictions. Massively parallel reporter assays (MPRAs) have emerged as a particularly valuable class of high-throughput technique to support such algorithms. MPRAs enable the simultaneous characterization of thousands or even millions of genetic constructs and provide the large amounts of data needed to train models. However, while the scale of this approach is impressive, the design of effective MPRA experiments is challenging due to the many factors that can be varied and the difficulty in predicting how these will impact the quality and quantity of data obtained. Here, we present a computational tool called FORECAST, which can simulate MPRA experiments based on fluorescence-activated cell sorting and subsequent sequencing (commonly referred to as Flow-seq or Sort-seq experiments), as well as carry out rigorous statistical estimation of construct performance from this type of experimental data. FORECAST can be used to develop workflows to aid the design of MPRA experiments and reanalyze existing MPRA data sets.

Accuracy and data efficiency in deep learning models of protein expression

Synthetic biology often involves engineering microbial strains to express high-value proteins. Thanks to progress in rapid DNA synthesis and sequencing, deep learning has emerged as a promising approach to build sequence-to-expression models for strain optimization. But such models need large and costly training data that create steep entry barriers for many laboratories. Here we study the relation between accuracy and data efficiency in an atlas of machine learning models trained on datasets of varied size and sequence diversity. We show that deep learning can achieve good prediction accuracy with much smaller datasets than previously thought. We demonstrate that controlled sequence diversity leads to substantial gains in data efficiency and employed Explainable AI to show that convolutional neural networks can finely discriminate between input DNA sequences. Our results provide guidelines for designing genotype-phenotype screens that balance cost and quality of training data, thus helping promote the wider adoption of deep learning in the biotechnology sector.

Hypermutation of specific genomic loci of Pseudomonas putida for continuous evolution of target genes

The ability of T7 RNA polymerase (RNAPT7 ) fusions to cytosine deaminases (CdA) for entering C➔T changes in any DNA segment downstream of a T7 promoter was exploited for hyperdiversification of defined genomic portions of Pseudomonas putida KT2440. To this end, test strains were constructed in which the chromosomally encoded pyrF gene (the prokaryotic homologue of yeast URA3) was flanked by T7 transcription initiation and termination signals and also carried plasmids expressing constitutively either high-activity (lamprey's) or low-activity (rat's) CdA-RNAPT7 fusions. The DNA segment-specific mutagenic action of these fusions was then tested in strains lacking or not uracil-DNA glycosylase (UDG), that is ∆ung/ung+ variants. The resulting diversification was measured by counting single nucleotide changes in clones resistant to 5-fluoroorotic acid (5FOA), which otherwise is transformed by wild-type PyrF into a toxic compound. Although the absence of UDG dramatically increased mutagenic rates with both CdA-RNAPT7 fusions, the most active variant - pmCDA1 - caused extensive appearance of 5FOA-resistant colonies in the wild-type strain not limited to C➔T but including also a range of other changes. Furthermore, the presence/absence of UDG activity swapped cytosine deamination preference between DNA strands. These qualities provided the basis of a robust system for continuous evolution of preset genomic portions of P. putida and beyond.

The context matrix: Navigating biological complexity for advanced biodesign

Synthetic biology offers many solutions in healthcare, production, sensing and agriculture. However, the ability to rationally engineer synthetic biosystems with predictable and robust functionality remains a challenge. A major reason is the complex interplay between the synthetic genetic construct, its host, and the environment. Each of these contexts contains a number of input factors which together can create unpredictable behaviours in the engineered biosystem. It has become apparent that for the accurate assessment of these contextual effects a more holistic approach to design and characterisation is required. In this perspective article, we present the context matrix, a conceptual framework to categorise and explore these contexts and their net effect on the designed synthetic biosystem. We propose the use and community-development of the context matrix as an aid for experimental design that simplifies navigation through the complex design space in synthetic biology.