Molecular convolutional neural networks with DNA regulatory circuits

SemiSynBio: A new era for neuromorphic computing

Neuromorphic computing has the potential to achieve the requirements of the next-generation artificial intelligence (AI) systems, due to its advantages of adaptive learning and parallel computing. Meanwhile, biocomputing has seen ongoing development with the rise of synthetic biology, becoming the driving force for new generation semiconductor synthetic biology (SemiSynBio) technologies. DNA-based biomolecules could potentially perform the functions of Boolean operators as logic gates and be used to construct artificial neural networks (ANNs), providing the possibility of executing neuromorphic computing at the molecular level. Herein, we briefly outline the principles of neuromorphic computing, describe the advances in DNA computing with a focus on synthetic neuromorphic computing, and summarize the major challenges and prospects for synthetic neuromorphic computing. We believe that constructing such synthetic neuromorphic circuits will be an important step toward realizing neuromorphic computing, which would be of widespread use in biocomputing, DNA storage, information security, and national defense.

Modeling One-Electron Oxidation Potentials and Hole Delocalization in Double-Stranded DNA by Multilayer and Dynamic Approaches

The number of innovative applications for DNA nowadays is growing quickly. Its use as a nanowire or electrochemical biosensor leads to the need for a deep understanding of the charge-transfer process along the strand, as well as its redox properties. These features are computationally simulated and analyzed in detail throughout this work by combining molecular dynamics, multilayer schemes, and the Marcus theory. One-electron oxidation potential and hole delocalization have been analyzed for six DNA double strands that cover all possible binary combinations of nucleotides. The results have revealed that the one-electron oxidation potential decreases with respect to the single-stranded DNA, giving evidence that the greater rigidity of a double helix induces an increase in the capacity of storing the positive charge generated upon oxidation. In addition, the hole is mainly stored in nucleobases with large reducer character, i.e., purines, especially when those are arranged in a stacked configuration in the same strand. From the computational point of view, the sampling needed to describe biological systems implies a significant computational cost. Here, we show that a small number of representative conformations generated by clustering analysis provides accurate results when compared with those obtained from sampling, reducing considerably the computational cost.

Engineering Modular DNA Reaction Networks for Signal Processing

Diversified molecular information-processing methods have significant implications for nanoscale manipulation and control, monitoring and disease diagnosis of organisms, and direct intervention in biological activities. However, as an effective approach for implementing multifunctional molecular information processing, DNA reaction networks (DRNs) with numerous functionally specialized molecular structures have challenged them on scale design, leading to increased network complexity, further causing problems such as signal leakage, attenuation, and cross-talk in network reactions. Our study developed a strategy for performing various signal-processing tasks through engineering modular DRNs. This strategy is based on a universal core unit with signal selection capability, and a time-adjustable signal self-resetting module is achieved by combing the core unit and self-resetting unit, which improves the time controllability of modular DRNs. In addition, multi-input and -output signal cross-catalytic and continuously adjustable signal delay modules were realized by combining core and threshold units, providing a flexible, precise method for modular DRNs to process the signal. The strategy simplifies the design of DRNs, helps generate design ideas for large-scale integrated DRNs with multiple functions, and provides prospects in biocomputing, gene regulation, and biosensing.

Bridging the Diagnostic Gap between Histopathologic and Hysteroscopic Chronic Endometritis with Deep Learning Models

Chronic endometritis (CE) is an inflammatory pathologic condition of the uterine mucosa characterized by unusual infiltration of CD138(+) endometrial stromal plasmacytes (ESPCs). CE is often identified in infertile women with unexplained etiology, tubal factors, endometriosis, repeated implantation failure, and recurrent pregnancy loss. Diagnosis of CE has traditionally relied on endometrial biopsy and histopathologic/immunohistochemistrical detection of ESPCs. Endometrial biopsy, however, is a somewhat painful procedure for the subjects and does not allow us to grasp the whole picture of this mucosal tissue. Meanwhile, fluid hysteroscopy has been recently adopted as a less-invasive diagnostic modality for CE. We launched the ARCHIPELAGO (ARChival Hysteroscopic Image-based Prediction for histopathologic chronic Endometritis in infertile women using deep LeArninG mOdel) study to construct the hysteroscopic CE finding-based prediction tools for histopathologic CE. The development of these deep learning-based novel models and computer-aided detection/diagnosis systems potentially benefits infertile women suffering from this elusive disease.

Scaling up of a Self-Confined Catalytic Hybridization Circuit for Robust microRNA Imaging

The precise regulation of cellular behaviors within a confined, crowded intracellular environment is highly amenable in diagnostics and therapeutics. While synthetic circuitry system through a concatenated chemical reaction network has rarely been reported to mimic dynamic self-assembly system. Herein, a catalytic self-defined circuit (CSC) for the hierarchically concatenated assembly of DNA domino nanostructures is engineered. By incorporating pre-sealed symmetrical fragments into the preying hairpin reactants, the CSC system allows the hierarchical DNA self-assembly via a microRNA (miRNA)-powered self-sorting catalytic hybridization reaction. With minimal strand complexity, this self-sustainable CSC system streamlined the circuit component and achieved localization-intensified cascaded signal amplification. Profiting from the self-adaptively concatenated hybridization reaction, a reliable and robust method has been achieved for discriminating carcinoma tissues from the corresponding para-carcinoma tissues. The CSC-sustained self-assembly strategy provides a comprehensive and smart toolbox for organizing various hierarchical DNA nanostructures, which may facilitate more insights for clinical diagnosis and therapeutic assessment.

A spatially localized DNA linear classifier for cancer diagnosis

Molecular computing is an emerging paradigm that plays an essential role in data storage, bio-computation, and clinical diagnosis with the future trends of more efficient computing scheme, higher modularity with scaled-up circuity and stronger tolerance of corrupted inputs in a complex environment. Towards these goals, we construct a spatially localized, DNA integrated circuits-based classifier (DNA IC-CLA) that can perform neuromorphic architecture-based computation at a molecular level for medical diagnosis. The DNA-based classifier employs a two-dimensional DNA origami as the framework and localized processing modules as the in-frame computing core to execute arithmetic operations (e.g. multiplication, addition, subtraction) for efficient linear classification of complex patterns of miRNA inputs. We demonstrate that the DNA IC-CLA enables accurate cancer diagnosis in a faster (about 3 h) and more effective manner in synthetic and clinical samples compared to those of the traditional freely diffusible DNA circuits. We believe that this all-in-one DNA-based classifier can exhibit more applications in biocomputing in cells and medical diagnostics.

Programmable Entropy-Driven Circuit-Cascaded Self-Feedback DNAzyme Network for Ultra-Sensitive Fluorescence and Photoelectrochemical Dual-Mode Biosensing

Inspired by natural DNA networks, programmable artificial DNA networks have become an attractive tool for developing high-performance biosensors. However, there is still a lot of room for expansion in terms of sensitivity, atom economy, and result self-validation for current microRNA sensors. In this protocol, miRNA-122 as a target model, an ultrasensitive fluorescence (FL) and photoelectrochemical (PEC) dual-mode biosensing platform is developed using a programmable entropy-driven circuit (EDC) cascaded self-feedback DNAzyme network. The well-designed EDC realizes full utilization of the DNA strands and improves the atomic economy of the signal amplification system. The unique and rational design of the double-CdSe quantum-dot-released EDC substrate and the cascaded self-feedback DNAzyme amplification network significantly avoids high background signals and enhances sensitivity and specificity. Also, the enzyme-free, programmable EDC cascaded DNAzyme network effectively avoids the risk of signal leakage and enhances the accuracy of the sensor. Moreover, the introduction of superparamagnetic Fe3O4@SiO2-cDNA accelerates the rapid extraction of E2-CdSe QDs and E3-CdSe QDs, which greatly improves the timeliness of sensor signal reading. In addition to the strengths of linear range (6 orders of magnitude) and stability, the biosensor design with dual signal reading makes the test results self-confirming.

Arbitrary Digital DNA Computing: A Programmable Molecular Perceptron Driven by Lambda Exonuclease for Lighting up Concatenated Circuits

DNA circuits, as a type of biochemical system, have the capability to synchronize the perception of molecular information with a chemical reaction response and directly process the molecular characteristic information in biological activities, making them a crucial area in molecular digital computing and smart bioanalytical applications. Instead of cascading logic gates, the traditional research approach achieves multiple logic operations which limits the scalability of DNA circuits and increases the development costs. Based on the interface reaction mechanism of Lambda exonuclease, the molecular perceptron proposed in this study, with the need for only adjusting weight and bias parameters to alter the corresponding logic expressions, enhances the versatility of the molecular circuits. We also establish a mathematical model and an improved heuristic algorithm for solving weights and bias parameters for arbitrary logic operations. The simulation and FRET experiment results of a series of logic operations demonstrate the universality of molecular perceptron. We hope the proposed molecular perceptron can introduce a new design paradigm for molecular circuits, fostering innovation and development in biomedical research related to biosensing, targeted therapy, and nanomachines.

A novel activation function based on DNA enzyme-free hybridization reaction and its implementation on nonlinear molecular learning systems

With the advent of the post-Moore's Law era, the development of traditional silicon-based computers has reached its limit, and there is an urgent need to develop new computing technologies to meet the needs of science, technology, and daily life. Due to its super-strong parallel computing capability and outstanding data storage capacity, DNA computing has become an important branch and hot research topic of new computer technology. DNA enzyme-free hybridization reaction technology is widely used in DNA computing, showing excellent performance in computing power and information processing. Studies have shown that DNA molecules not only have the computing function of electronic devices, but also exhibit certain human brain-like functions. In the field of artificial intelligence, activation functions play an important role as they enable artificial intelligence systems to fit and predict non-linear and complex variable relationships. Due to the difficulty of implementing activation functions in DNA computing, DNA circuits cannot easily achieve all the functions of artificial intelligence. DNA circuits need to rely on electronic computers to complete the training and learning process. Based on the parallel computing characteristics of DNA computing and the kinetic features of DNA molecule displacement reactions, this paper proposes a new activation function. This activation function can not only be easily implemented by DNA enzyme-free hybridization reaction reactions, but also has good nesting properties in DNA circuits, and can be cascaded with other DNA reactions to form a complete DNA circuit. This paper not only provides the mathematical analysis of the proposed activation function, but also provides a detailed analysis of its kinetic features. The activation function is then nested into a nonlinear neural network for DNA computing. This system is capable of fitting and predicting a certain nonlinear function.

Plug-and-Play Module for Reversible and Continuous Control of DNA Strand Displacement Kinetics

Regulatory modules for controlling the kinetics of toehold-mediated strand displacement (TMSD) play critical roles in designing dynamic and dissipative DNA chemical reaction networks (CRNs) but are hardwired into sequence designs. Herein, we introduce antitoehold (At), a plug-and-play module for reversible and continuous tuning of TMSD kinetics by temporarily occupying the toehold domain via a metastable duplex and base stacking. We demonstrate that kinetic control can be readily activated or deactivated in real time for any TMSD by simply adding At or anti-At. Continuous tuning of TMSD kinetics can also be achieved by altering the concentration of At. Moreover, the simple addition of At could readily reprogram existing TMSDs into a pulse-generation DNA CRN with continuous tunability. Our At approach also offers a new way for engineering continuously tunable DNA hybridization probes, which may find practical uses for discriminating clinically important mutations. Because of the simplicity, we anticipate that At will find wide applications for engineering DNA CRNs with diverse dynamic and dissipative behaviors, and DNA hybridization probes with tunable affinity and selectivity.

DNA as a universal chemical substrate for computing and data storage

DNA computing and DNA data storage are emerging fields that are unlocking new possibilities in information technology and diagnostics. These approaches use DNA molecules as a computing substrate or a storage medium, offering nanoscale compactness and operation in unconventional media (including aqueous solutions, water-in-oil microemulsions and self-assembled membranized compartments) for applications beyond traditional silicon-based computing systems. To build a functional DNA computer that can process and store molecular information necessitates the continued development of strategies for computing and data storage, as well as bridging the gap between these fields. In this Review, we explore how DNA can be leveraged in the context of DNA computing with a focus on neural networks and compartmentalized DNA circuits. We also discuss emerging approaches to the storage of data in DNA and associated topics such as the writing, reading, retrieval and post-synthesis editing of DNA-encoded data. Finally, we provide insights into how DNA computing can be integrated with DNA data storage and explore the use of DNA for near-memory computing for future information technology and health analysis applications.

Synthetic biological neural networks: From current implementations to future perspectives

Artificial neural networks, inspired by the biological networks of the human brain, have become game-changing computing models in modern computer science. Inspired by their wide scope of applications, synthetic biology strives to create their biological counterparts, which we denote synthetic biological neural networks (SYNBIONNs). Their use in the fields of medicine, biosensors, biotechnology, and many more shows great potential and presents exciting possibilities. So far, many different synthetic biological networks have been successfully constructed, however, SYNBIONN implementations have been sparse. The latter are mostly based on neural networks pretrained in silico and being heavily dependent on extensive human input. In this paper, we review current implementations and models of SYNBIONNs. We briefly present the biological platforms that show potential for designing and constructing perceptrons and/or multilayer SYNBIONNs. We explore their future possibilities along with the challenges that must be overcome to successfully implement a scalable in vivo biological neural network capable of online learning.

Lipid vesicle-based molecular robots

A molecular robot, which is a system comprised of one or more molecular machines and computers, can execute sophisticated tasks in many fields that span from nanomedicine to green nanotechnology. The core parts of molecular robots are fairly consistent from system to system and always include (i) a body to encapsulate molecular machines, (ii) sensors to capture signals, (iii) computers to make decisions, and (iv) actuators to perform tasks. This review aims to provide an overview of approaches and considerations to develop molecular robots. We first introduce the basic technologies required for constructing the core parts of molecular robots, describe the recent progress towards achieving higher functionality, and subsequently discuss the current challenges and outlook. We also highlight the applications of molecular robots in sensing biomarkers, signal communications with living cells, and conversion of energy. Although molecular robots are still in their infancy, they will unquestionably initiate massive change in biomedical and environmental technology in the not too distant future.

Programmable Primer Switching for Regulating Enzymatic DNA Circuits

Developing DNA strand displacement reactions (SDRs) offers crucial technical support for regulating artificial nucleic acid circuits and networks. More recently, enzymatic SDR-based DNA circuits have gained significant attention because of their modular design, high orthogonality signaling, and extremely fast reaction rates. Typical enzymatic SDRs are regulated by relatively long primers (20-30 nucleotides) that hybridize to form stable double-stranded structures, facilitating enzyme-initiated events. Implementing more flexible primer-based enzymatic SDR regulations remains challenging due to the lack of convenient and simple primer control mechanism, which consequently limits the development of enzymatic DNA circuits. In this study, we propose an approach, termed primer switching regulation, that implements programmable and flexible regulations of enzymatic circuits by introducing switchable wires into the enzymatic circuits. We applied this method to generate diverse enzymatic DNA circuits, including cascading, fan-in/fan-out, dual-rail, feed-forward, and feedback functions. Through this method, complex circuit functions can be implemented by just introducing additional switching wires without reconstructing the basic circuit frameworks. The method is experimentally demonstrated to provide flexible and programmable regulations to control enzymatic DNA circuits and has future applications in DNA computing, biosensing, and DNA storage.

Simulation-Assisted DNA Nanodevice Serve as a General Optical Platform for Multiplexed Analysis of Micrornas

Small frame nucleic acids (FNAs) serve as excellent carrier materials for various functional nucleic acid molecules, showcasing extensive potential applications in biomedicine development. The carrier module and function module combination is crucial for probe design, where an improper combination can significantly impede the functionality of sensing platforms. This study explores the effect of various combinations on the sensing performance of nanodevices through simulations and experimental approaches. Variances in response velocities, sensitivities, and cell uptake efficiencies across different structures are observed. Factors such as the number of functional molecules loaded, loading positions, and intermodular distances affect the rigidity and stability of the nanostructure. The findings reveal that the structures with full loads and moderate distances between modules have the lowest potential energy. Based on these insights, a multisignal detection platform that offers optimal sensitivity and response speed is developed. This research offers valuable insights for designing FNAs-based probes and presents a streamlined method for the conceptualization and optimization of DNA nanodevices.

How to make DNA data storage more applicable

The storage of digital data is becoming a worldwide problem. DNA has been recognized as a biological solution due to its ability to store genetic information without alteration over long periods. The first demonstrations of high-capacity long-lasting DNA digital data storage have been shown. However, high storage costs and slow retrieval of the data must be overcome to make DNA data storage more applicable and marketable. Herein, we discuss the issues and recent advances in DNA data storage methods and highlight pathways to make this technology more applicable to real-world digital data storage. We envision that a combination of molecular biology, nanotechnology, novel polymers, electronics, and automation with systematic development will allow DNA data storage sufficient for everyday use.

pH-Controlled Resettable Modular DNA Strand-Displacement Circuits

Sophisticated dynamic molecular systems with diverse functions have been fabricated by using the fundamental tool of toehold-mediated strand displacement (TMSD) in the field of dynamic DNA nanotechnology. However, simple approaches to reset these TMSD-based dynamic systems are lacking due to the difficulty in creating kinetically favored pathways to implement the backward resetting reactions. Here, we develop a facile proton-driven strategy to achieve complete resetting of a modular DNA circuit by integrating a pH-responsive intermolecular CG-C+ triplex DNA and an i-motif DNA into the conventional DNA substrate. The pH-programmed strategy allows modular DNA components to specifically associate/dissociate to promote the forward/backward TMSD reactions, thereby enabling the modular DNA circuit to be repeatedly operated at a constant temperature without generating any DNA waste products. Leveraging this tractable approach, we further constructed two resettable DNA logic gates used for logical computation and two resettable catalytic DNA systems with good performance in signal transduction and amplification.

Toward three-dimensional DNA industrial nanorobots

Nanoscale industrial robots have potential as manufacturing platforms and are capable of automatically performing repetitive tasks to handle and produce nanomaterials with consistent precision and accuracy. We demonstrate a DNA industrial nanorobot that fabricates a three-dimensional (3D), optically active chiral structure from optically inactive parts. By making use of externally controlled temperature and ultraviolet (UV) light, our programmable robot, ~100 nanometers in size, grabs different parts, positions and aligns them so that they can be welded, releases the construct, and returns to its original configuration ready for its next operation. Our robot can also self-replicate its 3D structure and functions, surpassing single-step templating (restricted to two dimensions) by using folding to access the third dimension and more degrees of freedom. Our introduction of multiple-axis precise folding and positioning as a tool/technology for nanomanufacturing will open the door to more complex and useful nano- and microdevices.

One-Electron Oxidation Potentials and Hole Delocalization in Heterogeneous Single-Stranded DNA

The study of DNA processes is essential to understand not only its intrinsic biological functions but also its role in many innovative applications. The use of DNA as a nanowire or electrochemical biosensor leads to the need for a deep investigation of the charge transfer process along the strand as well as of the redox properties. In this contribution, the one-electron oxidation potential and the charge delocalization of the hole formed after oxidation are computationally investigated for different heterogeneous single-stranded DNA strands. We have established a two-step protocol: (i) molecular dynamics simulations in the frame of quantum mechanics/molecular mechanics (QM/MM) were performed to sample the conformational space; (ii) energetic properties were then obtained within a QM1/QM2/continuum approach in combination with the Marcus theory over an ensemble of selected geometries. The results reveal that the one-electron oxidation potential in the heterogeneous strands can be seen as a linear combination of that property within the homogeneous strands. In addition, the hole delocalization between different nucleobases is, in general, small, supporting the conclusion of a hopping mechanism for charge transport along the strands. However, charge delocalization becomes more important, and so does the tunneling mechanism contribution, when the reducing power of the nucleobases forming the strand is similar. Moreover, charge delocalization is slightly enhanced when there is a correlation between pairs of some of the interbase coordinates of the strand: twist/shift, twist/slide, shift/slide, and rise/tilt. However, the internal structure of the strand is not the predominant factor for hole delocalization but the specific sequence of nucleotides that compose the strand.

Multifunctional dumbbell probes-based logic circuits: microRNAs logic detection and tumor cells identification

BACKGROUND

The powerful logic processing capability of DNA logic circuits over multiple input signals perfectly meets the demands of multi-biomarker-based clinical diagnostics. As important biomarkers for cancer diagnosis and treatment, the orthogonal differential expression of microRNAs (miRNAs) in different diseases and different cancer cells makes the precise logical detection of multiple miRNAs particularly critical.

RESULTS

Therefore, we constructed two fundamental "AND" and "OR" logic gates and one "AND-OR" logic gate on the basis of our proposed multifunctional dumbbell probes. These logic gates allowed for the logical profiling of multiple cancer-associated miRNAs. In addition, by making simple adjustments to the functional modules of multifunctional dumbbell probes, the three logic gates we proposed could be easily transformed without the use of sophisticated probe design. Remarkably, these logic gates, in particular the "AND-OR" logic gate, were able to compute several miRNAs simultaneously, demonstrating excellent cell identification capabilities.

SIGNIFICANCE

Overall, this work provided a new idea for accurately distinguishing multiple cell types and showed great application prospects.

Nucleic Acid-to-Small Molecule Converter through Amplified Hairpin DNA Circuits

Many microRNAs (miRNAs) are characteristically found in cancer cells, making miRNAs promising marker biomolecules for cancer diagnosis and therapeutics. However, it is challenging to use miRNA as a cancer signature because it is difficult to convert the nucleic acid sequence information into molecular functionality. To address this challenge, we realize nucleic acid-to-small molecule converters using hairpin DNA circuits. Harnessing a Staudinger reduction as a trigger for the conversion, we constructed hybridization chain reaction (HCR) and catalytic hairpin assembly (CHA) circuits that respond to oncogenic miR-21. Fluorophore and dye molecules were released in response to miR-21 through the HCR, providing fluorogenic and chromogenic readouts. Selective cytotoxicity in miR-21-abundant cells was realized by the CHA to release the anticancer drug SN-38. This would be the first example of selective activation of a small-molecule prodrug triggered by oncogenic miRNA in human living cells.

Neural network execution using nicked DNA and microfluidics

DNA has been discussed as a potential medium for data storage. Potentially it could be denser, could consume less energy, and could be more durable than conventional storage media such as hard drives, solid-state storage, and optical media. However, performing computations on the data stored in DNA is a largely unexplored challenge. This paper proposes an integrated circuit (IC) based on microfluidics that can perform complex operations such as artificial neural network (ANN) computation on data stored in DNA. We envision such a system to be suitable for highly dense, throughput-demanding bio-compatible applications such as an intelligent Organ-on-Chip or other biomedical applications that may not be latency-critical. It computes entirely in the molecular domain without converting data to electrical form, making it a form of in-memory computing on DNA. The computation is achieved by topologically modifying DNA strands through the use of enzymes called nickases. A novel scheme is proposed for representing data stochastically through the concentration of the DNA molecules that are nicked at specific sites. The paper provides details of the biochemical design, as well as the design, layout, and operation of the microfluidics device. Benchmarks are reported on the performance of neural network computation.

Self-Assembly of DNA Nanostructures in Different Cations

The programmable nature of DNA allows the construction of custom-designed static and dynamic nanostructures, and assembly conditions typically require high concentrations of magnesium ions that restricts their applications. In other solution conditions tested for DNA nanostructure assembly, only a limited set of divalent and monovalent ions are used so far (typically Mg2+ and Na+ ). Here, we investigate the assembly of DNA nanostructures in a wide variety of ions using nanostructures of different sizes: a double-crossover motif (76 bp), a three-point-star motif (~134 bp), a DNA tetrahedron (534 bp) and a DNA origami triangle (7221 bp). We show successful assembly of a majority of these structures in Ca2+ , Ba2+ , Na+ , K+ and Li+ and provide quantified assembly yields using gel electrophoresis and visual confirmation of a DNA origami triangle using atomic force microscopy. We further show that structures assembled in monovalent ions (Na+ , K+ and Li+ ) exhibit up to a 10-fold higher nuclease resistance compared to those assembled in divalent ions (Mg2+ , Ca2+ and Ba2+ ). Our work presents new assembly conditions for a wide range of DNA nanostructures with enhanced biostability.

Enzyme-Free Nucleic Acid Circuits for Fold-Change Detection

Fold-change detection is widespread in sensory systems of various organisms. Dynamic DNA nanotechnology provides an important toolbox for reproducing structures and responses of cellular circuits. In this work, we construct an enzyme-free nucleic acid circuit based on the incoherent feed-forward loop using toehold-mediated DNA strand displacement reactions and explore its dynamic behaviors. The mathematical model based on ordinary differential equations is used to evaluate the parameter regime required for fold-change detection. After selecting appropriate parameters, the constructed synthetic circuit exhibits approximate fold-change detection for multiple rounds of inputs with different initial concentrations. This work is anticipated to shed new light on the design of DNA dynamic circuits in the enzyme-free environment.

Accelerating DNA computing via freeze-thaw cycling

DNA computing harnesses the immense potential of DNA molecules to enable sophisticated and transformative computational processes but is hindered by low computing speed. Here, we propose freeze-thaw cycling as a simple yet powerful method for high-speed DNA computing without complex procedures. Through iterative cycles, we achieve a substantial 20-fold speed enhancement in basic strand displacement reactions. This acceleration arises from the utilization of eutectic ice phase as a medium, temporarily increasing the effective local concentration of molecules during each cycle. In addition, the acceleration effect follows the Hofmeister series, where kosmotropic anions such as sulfate (SO42-) reduce eutectic phase volume, leading to a more notable enhancement in strand displacement reaction rates. Leveraging this phenomenon, freeze-thaw cycling demonstrates its generalizability for high-speed DNA computing across various circuit sizes, achieving up to a remarkable 120-fold enhancement in reaction rates. We envision its potential to revolutionize molecular computing and expand computational applications in diverse fields.

Programmable Molecular Signal Transmission Architecture and Reactant Regeneration Strategy Driven by EXO λ for DNA Circuits

The characteristics of DNA hybridization enable molecular computing through strand displacement reactions, facilitating the construction of complex DNA circuits, which is an important way to realize information interaction and processing at a molecular level. However, signal attenuation in the cascade and shunt process hinders the reliability of the calculation results and further expansion of the DNA circuit scale. Here, we demonstrate a novel programmable exonuclease-assisted signal transmission architecture, where DNA strand with toehold employed to inhibit the hydrolysis process of EXO λ is applied in DNA circuits. We construct a series circuit with variable resistance and a parallel circuit with constant current source, ensuring excellent orthogonal properties between input and output sequences while maintaining low leakage (<5%) during the reaction. Additionally, a simple and flexible exonuclease-driven reactant regeneration (EDRR) strategy is proposed and applied to construct parallel circuits with constant voltage sources that could amplify the output signal without extra DNA fuel strands or energy. Furthermore, we demonstrate the effectiveness of the EDRR strategy in reducing signal attenuation during cascade and shunt processes by constructing a four-node DNA circuit. These findings offer a new approach to enhance the reliability of molecular computing systems and expand the scale of DNA circuits in the future.

A molecular assessment of the practical potential of DNA-based computation

The immense information density of DNA and its potential for massively parallelized computations, paired with rapidly expanding data production and storage needs, have fueled a renewed interest in DNA-based computation. Since the construction of the first DNA computing systems in the 1990s, the field has grown to encompass a diverse array of configurations. Simple enzymatic and hybridization reactions to solve small combinatorial problems transitioned to synthetic circuits mimicking gene regulatory networks and DNA-only logic circuits based on strand displacement cascades. These have formed the foundations of neural networks and diagnostic tools that aim to bring molecular computation to practical scales and applications. Considering these great leaps in system complexity as well as in the tools and technologies enabling them, a reassessment of the potential of such DNA computing systems is warranted.

Intramolecular and intermolecular hole delocalization rules the reducer character of isolated nucleobases and homogeneous single-stranded DNA

The use of DNA strands as nanowires or electrochemical biosensors requires a deep understanding of charge transfer processes along the strand, as well as of the redox properties. These properties are computationally assessed in detail throughout this study. By applying molecular dynamics and hybrid QM/continuum and QM/QM/continuum schemes, the vertical ionization energies, adiabatic ionization energies, vertical attachment energies, one-electron oxidation potentials, and delocalization of the hole generated upon oxidation have been determined for nucleobases in their free form and as part of a pure single-stranded DNA. We show that the reducer ability of the isolated nucleobases is explained by the intramolecular delocalization of the positively charged hole, while the enhancement of the reducer character when going from aqueous solution to the strand correlates very well with the intermolecular hole delocalization. Our simulations suggest that the redox properties of DNA strands can be tuned by playing with the balance between intramolecular and intermolecular charge delocalization.

Construction of DNA-based molecular circuits using normally open and normally closed switches driven by lambda exonuclease

Building synthetic molecular circuits is an important way to realize ion detection, information processing, and molecular computing. However, it is still challenging to implement the NOT logic controlled by a single molecule input in synthetic molecular circuits wherein the presence or absence of the molecule represents the ON or OFF state of the input. Here, based on lambda exonuclease (λ exo), for the first time, we propose the normally open (NO) and normally closed (NC) switching strategy with a unified signal transmission mechanism to build molecular circuits. Specifically, the opposite logic can be output with or without a single signal, and the state of the switch can be adjusted by the addition order and time interval of the upstream signal and switch signal, which endows the switch with time-responsive characteristics. In addition, a time-delay relay with the function of delayed disconnection is developed to realize quantitative control of outputs, which has the potential to meet the automation control need of the system. Finally, digital square and square root circuits are constructed by cascading the NO and NC switches, which demonstrates the versatility of switches. Our design can be extended to time logic and complex digital computing circuits for use in information processing and nanomachines.

Self-assembly of DNA nanostructures in different cations

The programmable nature of DNA allows the construction of custom-designed static and dynamic nanostructures, and assembly conditions typically require high concentrations of magnesium ions which restricts their applications. In other solution conditions tested for DNA nanostructure assembly, only a limited set of divalent and monovalent ions have been used so far (typically Mg 2+ and Na + ). Here, we investigate the assembly of DNA nanostructures in a wide variety of ions using nanostructures of different sizes: a double-crossover motif (76 bp), a three-point-star motif (∼134 bp), a DNA tetrahedron (534 bp) and a DNA origami triangle (7221 bp). We show successful assembly of a majority of these structures in Ca 2+ , Ba 2+ , Na + , K + and Li + and provide quantified assembly yields using gel electrophoresis and visual confirmation of a DNA origami triangle using atomic force microscopy. We further show that structures assembled in monovalent ions (Na + , K + and Li + ) exhibit up to a 10-fold higher nuclease resistance compared to those assembled in divalent ions (Mg 2+ , Ca 2+ and Ba 2+ ). Our work presents new assembly conditions for a wide range of DNA nanostructures with enhanced biostability.

A survey on molecular-scale learning systems with relevance to DNA computing

DNA computing has emerged as a promising alternative to achieve programmable behaviors in chemistry by repurposing the nucleic acid molecules into chemical hardware upon which synthetic chemical programs can be executed. These chemical programs are capable of simulating diverse behaviors, including boolean logic computation, oscillations, and nanorobotics. Chemical environments such as the cell are marked by uncertainty and are prone to random fluctuations. For this reason, potential DNA-based molecular devices that aim to be deployed into such environments should be capable of adapting to the stochasticity inherent in them. In keeping with this goal, a new subfield has emerged within DNA computing, focusing on developing approaches that embed learning and inference into chemical reaction systems. If realized in biochemical contexts, such molecular machines can engender novel applications in fields such as biotechnology, synthetic biology, and medicine. Therefore, it would be beneficial to review how different ideas were conceived, how the progress has been so far, and what the emerging ideas are in this nascent field of 'molecular-scale learning'.

Unlocking Genetic Profiles with a Programmable DNA-Powered Decoding Circuit

Human genetic architecture provides remarkable insights into disease risk prediction and personalized medication. Advances in genomics have boosted the fine-mapping of disease-associated genetic variants across human genome. In healthcare practice, interpreting intricate genetic profiles into actionable medical decisions can improve health outcomes but remains challenging. Here an intelligent genetic decoder is engineered with programmable DNA computation to automate clinical analyses and interpretations. The DNA-based decoder recognizes multiplex genetic information by one-pot ligase-dependent reactions and interprets implicit genetic profiles into explicit decision reports. It is shown that the DNA decoder implements intended computation on genetic profiles and outputs a corresponding answer within hours. Effectiveness in 30 human genomic samples is validated and it is shown that it achieves desirable performance on the interpretation of CYP2C19 genetic profiles into drug responses, with accuracy equivalent to that of Sanger sequencing. Circuit modules of the DNA decoder can also be readily reprogrammed to interpret another pharmacogenetics genes, provide drug dosing recommendations, and implement reliable molecular calculation of polygenic risk score (PRS) and PRS-informed cancer risk assessment. The DNA-powered intelligent decoder provides a general solution to the translation of complex genetic profiles into actionable healthcare decisions and will facilitate personalized healthcare in primary care.

Modular reconfiguration of DNA origami assemblies using tile displacement

The power of natural evolution lies in the adaptability of biological organisms but is constrained by the time scale of genetics and reproduction. Engineeringartificial molecular machines should not only include adaptability as a core feature but also apply it within a larger design space and at a faster time scale. A lesson from engineering electromechanical robots is that modular robots can perform diverse functions through self-reconfiguration, a large-scale form of adaptation. Molecular machines made of modular, reconfigurable components may form the basis for dynamic self-reprogramming in future synthetic cells. To achieve modular reconfiguration in DNA origami assemblies, we previously developed a tile displacement mechanism in which an invader tile replaces another tile in an array with controlled kinetics. Here, we establish design principles for simultaneous reconfigurations in tile assemblies using complex invaders with distinct shapes. We present toehold and branch migration domain configurations that expand the design space of tile displacement reactions by two orders of magnitude. We demonstrate the construction of multitile invaders with fixed and variable sizes and controlled size distributions. We investigate the growth of three-dimensional (3D) barrel structures with variable cross sections and introduce a mechanism for reconfiguring them into 2D structures. Last, we show an example of a sword-shaped assembly transforming into a snake-shaped assembly, illustrating two independent tile displacement reactions occurring concurrently with minimum cross-talk. This work serves as a proof of concept that tile displacement could be a fundamental mechanism for modular reconfiguration robust to temperature and tile concentration.

Multifunctional Exo III-assisted scalability strategy for constructing DNA molecular logic circuits

The construction of logic circuits is critical to DNA computing. Simple and effective scalability methods have been the focus of attention in various fields related to constructing logic circuits. We propose a double-stranded separation (DSS) strategy to facilitate the construction of complex circuits. The strategy combines toehold-mediated strand displacement with exonuclease III (Exo III), which is a multifunctional nuclease. Exo III can quickly recognize an apurinic/apyrimidinic (AP) site. DNA oligos with an AP site can generate an output signal by the strand displacement reaction. However, in contrast to traditional strand displacement reactions, the double-stranded waste from the strand displacement can be further hydrolysed by the endonuclease function of Exo III, thus generating an additional output signal. The DSS strategy allows for the effective scalability of molecular logic circuits, enabling multiple logic computing capabilities simultaneously. In addition, we succeeded in constructing a logic circuit with dual logic functions that provides foundations for more complex circuits in the future and has a broad scope for development in logic computing, biosensing, and nanomachines.

Programming DNA Reaction Networks Using Allosteric DNA Hairpins

The construction of DNA reaction networks with complex functions using various methods has been an important research topic in recent years. Whether the DNA reaction network can perform complex tasks and be recycled directly affects the performance of the reaction network. Therefore, it is very important to design and implement a DNA reaction network capable of multiple tasks and reversible regulation. In this paper, the hairpin allosteric method was used to complete the assembly task of different functional nucleic acids. In addition, information conversion of the network was realized. In this network, multiple hairpins were assembled into nucleic acid structures with different functions to achieve different output information through the cyclic use of trigger strands. A method of single-input dual-output information conversion was proposed. Finally, the network with signal amplification and reversible regulation was constructed. In this study, the reversible regulation of different functional nucleic acids in the same network was realized, which shows the potential of this network in terms of programmability and provides new ideas for constructing complex and multifunctional DNA reaction networks.

Molecular Computation for Molecular Classification

DNA as an informational polymer has, for the past 30 years, progressively become an essential molecule to rationally build chemical reaction networks endowed with powerful signal-processing capabilities. Whether influenced by the silicon world or inspired by natural computation, molecular programming has gained attention for diagnosis applications. Of particular interest for this review, molecular classifiers have shown promising results for disease pattern recognition and sample classification. Because both input integration and computation are performed in a single tube, at the molecular level, this low-cost approach may come as a complementary tool to molecular profiling strategies, where all biomarkers are quantified independently using high-tech instrumentation. After introducing the elementary components of molecular classifiers, some of their experimental implementations are discussed either using digital Boolean logic or analog neural network architectures.

DNA strand displacement based computational systems and their applications

DNA computing has become the focus of computing research due to its excellent parallel processing capability, data storage capacity, and low energy consumption characteristics. DNA computational units can be precisely programmed through the sequence specificity and base pair principle. Then, computational units can be cascaded and integrated to form large DNA computing systems. Among them, DNA strand displacement (DSD) is the simplest but most efficient method for constructing DNA computing systems. The inputs and outputs of DSD are signal strands that can be transferred to the next unit. DSD has been used to construct logic gates, integrated circuits, artificial neural networks, etc. This review introduced the recent development of DSD-based computational systems and their applications. Some DSD-related tools and issues are also discussed.