INTRODUCTION Cells respond to their environment, make decisions, build structures, and coordinate tasks. Underlying these processes are computational operations performed by networks of regulatory proteins that integrate signals and control the timing of gene expression. Harnessing this capability is critical for biotechnology projects that require decision-making, control, sensing, or spatial organization. It has been shown that cells can be programmed using synthetic genetic circuits composed of regulators organized to generate a desired operation. However, the construction of even simple circuits is time-intensive and unreliable. RATIONALE Electronic design automation (EDA) was developed to aid engineers in the design of semiconductor-based electronics. In an effort to accelerate genetic circuit design, we applied principles from EDA to enable increased circuit complexity and to simplify the incorporation of synthetic gene regulation into genetic engineering projects. We used the hardware description language Verilog to enable a user to describe a circuit function. The user also specifies the sensors, actuators, and “user constraints file” (UCF), which defines the organism, gate technology, and valid operating conditions. Cello (www.cellocad.org) uses this information to automatically design a DNA sequence encoding the desired circuit. This is done via a set of algorithms that parse the Verilog text, create the circuit diagram, assign gates, balance constraints to build the DNA, and simulate performance. RESULTS Cello designs circuits by drawing upon a library of Boolean logic gates. Here, the gate technology consists of NOT/NOR logic based on repressors. Gate connection is simplified by defining the input and output signals as RNA polymerase (RNAP) fluxes. We found that the gates need to be insulated from their genetic context to function reliably in the context of different circuits. Each gate is isolated using strong terminators to block RNAP leakage, and input interchangeability is improved using ribozymes and promoter spacers. These parts are varied for each gate to avoid breakage due to recombination. Measuring the load of each gate and incorporating this into the optimization algorithms further reduces evolutionary pressure. Cello was applied to the design of 60 circuits for Escherichia coli , where the circuit function was specified using Verilog code and transformed to a DNA sequence. The DNA sequences were built as specified with no additional tuning, requiring 880,000 base pairs of DNA assembly. Of these, 45 circuits performed correctly in every output state (up to 10 regulators and 55 parts). Across all circuits, 92% of the 412 output states functioned as predicted. CONCLUSION Our work constitutes a hardware description language for programming living cells. This required the co-development of design algorithms with gates that are sufficiently simple and robust to be connected by automated algorithms. We demonstrate that engineering principles can be applied to identify and suppress errors that complicate the compositions of larger systems. This approach leads to highly repetitive and modular genetics, in stark contrast to the encoding of natural regulatory networks. The use of a hardware-independent language and the creation of additional UCFs will allow a single design to be transformed into DNA for different organisms, genetic endpoints, operating conditions, and gate technologies.

http://science.sciencemag.org/content/sci/352/6281/aac7341.full.pdf

Genetic circuit design automation

In the last two decades, the vast majority of microfluidic systems have been built in poly(dimethylsiloxane) (PDMS) by soft lithography, a technique based on PDMS micromolding. A long list of key PDMS properties have contributed to the success of soft lithography: PDMS is biocompatible, elastomeric, transparent, gas-permeable, water-impermeable, fairly inexpensive, copyright-free, and rapidly prototyped with high precision using simple procedures. However, the fabrication process typically involves substantial human labor, which tends to make PDMS devices difficult to disseminate outside of research labs, and the layered molding limits the 3D complexity of the devices that can be produced. 3D-printing has recently attracted attention as a way to fabricate microfluidic systems due to its automated, assembly-free 3D fabrication, rapidly decreasing costs, and fast-improving resolution and throughput. Resins with properties approaching those of PDMS are being developed. Here we review past and recent efforts in 3D-printing of microfluidic systems. We compare the salient features of PDMS molding with those of 3D-printing and we give an overview of the critical barriers that have prevented the adoption of 3D-printing by microfluidic developers, namely resolution, throughput, and resin biocompatibility. We also evaluate the various forces that are persuading researchers to abandon PDMS molding in favor of 3D-printing in growing numbers.

/pdf/the-upcoming-3d-printing-revolution-in-microfluidics-2n25a4i4t7.pdf

The upcoming 3D-printing revolution in microfluidics

3D printing has the potential to significantly change the field of microfluidics. The ability to fabricate a complete microfluidic device in a single step from a computer model has obvious attractions, but it is the ability to create truly three dimensional structures that will provide new microfluidic capability that is challenging, if not impossible to make with existing approaches. This critical review covers the current state of 3D printing for microfluidics, focusing on the four most frequently used printing approaches: inkjet (i3DP), stereolithography (SLA), two photon polymerisation (2PP) and extrusion printing (focusing on fused deposition modeling). It discusses current achievements and limitations, and opportunities for advancement to reach 3D printing's full potential.

/pdf/3d-printed-microfluidic-devices-enablers-and-barriers-1g6kxbpnfb.pdf

3D printed microfluidic devices: enablers and barriers

Cells are able to navigate environments, communicate, and build complex patterns by initiating gene expression in response to specific signals. Engineers need to harness this capability to program cells to perform tasks or build chemicals and materials that match the complexity seen in nature. This review describes new tools that aid the construction of genetic circuits. We show how circuit dynamics can be influenced by the choice of regulators and changed with expression “tuning knobs.” We collate the failure modes encountered when assembling circuits, quantify their impact on performance, and review mitigation efforts. Finally, we discuss the constraints that arise from operating within a living cell. Collectively, better tools, well-characterized parts, and a comprehensive understanding of how to compose circuits are leading to a breakthrough in the ability to program living cells for advanced applications, from living therapeutics to the atomic manufacturing of functional materials.

/pdf/principles-of-genetic-circuit-design-1stvh1u9g2.pdf

Principles of genetic circuit design

As a data-driven science, genomics largely utilizes machine learning to capture dependencies in data and derive novel biological hypotheses. However, the ability to extract new insights from the exponentially increasing volume of genomics data requires more expressive machine learning models. By effectively leveraging large data sets, deep learning has transformed fields such as computer vision and natural language processing. Now, it is becoming the method of choice for many genomics modelling tasks, including predicting the impact of genetic variation on gene regulatory mechanisms such as DNA accessibility and splicing. This Review describes different deep learning techniques and how they can be applied to extract biologically relevant information from large, complex genomic data sets.

Deep learning: new computational modelling techniques for genomics

For genetically engineered circuits, the movement of RNA polymerase across the DNA during transcription needs to be tightly controlled. A large library of strong terminators will make circuit design easier and more efficient.

Characterization of 582 natural and synthetic terminators and quantification of their design constraints

Genetic circuits perform computational operations based on interactions between freely diffusing molecules within a cell. When transcription factors are combined to build a circuit, unintended interactions can disrupt its function. Here, we apply 'part mining' to build a library of 73 TetR-family repressors gleaned from prokaryotic genomes. The operators of a subset were determined using an in vitro method, and this information was used to build synthetic promoters. The promoters and repressors were screened for cross-reactions. Of these, 16 were identified that both strongly repress their cognate promoter (5- to 207-fold) and exhibit minimal interactions with other promoters. Each repressor-promoter pair was converted to a NOT gate and characterized. Used as a set of 16 NOT/NOR gates, there are >10(54) circuits that could be built by changing the pattern of input and output promoters. This represents a large set of compatible gates that can be used to construct user-defined circuits.

/pdf/genomic-mining-of-prokaryotic-repressors-for-orthogonal-10ss1hdvdh.pdf

Genomic mining of prokaryotic repressors for orthogonal logic gates.

Genomic mining of prokaryotic repressors for orthogonal logic gates

Genetic circuits require many regulatory parts in order to implement signal processing or execute algorithms in cells. A potentially scalable approach is to use dCas9, which employs small guide RNAs (sgRNAs) to repress genetic loci via the programmability of RNA:DNA base pairing. To this end, we use dCas9 and designed sgRNAs to build transcriptional logic gates and connect them to perform computation in living cells. We constructed a set of NOT gates by designing five synthetic Escherichia coli σ70 promoters that are repressed by corresponding sgRNAs, and these interactions do not exhibit crosstalk between each other. These sgRNAs exhibit high on-target repression (56- to 440-fold) and negligible off-target interactions (< 1.3-fold). These gates were connected to build larger circuits, including the Boolean-complete NOR gate and a 3-gate circuit consisting of four layered sgRNAs. The synthetic circuits were connected to the native E. coli regulatory network by designing output sgRNAs to target an E. coli transcription factor (malT). This converts the output of a synthetic circuit to a switch in cellular phenotype (sugar utilization, chemotaxis, phage resistance).

/pdf/multi-input-crispr-cas-genetic-circuits-that-interface-host-433e9ohais.pdf

Alec A. K. Nielsen

Papers

Genetic circuit design automation

Characterization of 582 natural and synthetic terminators and quantification of their design constraints

Genomic mining of prokaryotic repressors for orthogonal logic gates.

Genomic mining of prokaryotic repressors for orthogonal logic gates

Multi-input CRISPR/Cas genetic circuits that interface host regulatory networks