A molecular multi-gene classifier for disease diagnostics

DNA-Programmed Reaction to Evaluate Specific IgE for Allergy Point-of-Care Testing

A DNA-programmed reaction to evaluate non-nucleic acids inputs with computation speed (≈30 min) and sensitivity (sub-picomolar) suitable for analyzing physiologically relevant biomarkers in a one-pot format and point-of-care testing setting is reported. Specifically, a DNA programme based on the proximity-activation exponential amplification reaction (PEAR) is designed to evaluate specific IgE (sIgE) against Der p 2 implicated in dust mite allergy which affects millions worldwide. In this work, we tailored the molecular components of the input-to-oligo barcode conversion module as an AND gate to detect inputs with binding specificity to Der p 2 antigen and is of an IgE isotype. In addition, an in situ biotinylation method is developed to generate amplified oligo barcodes amendable for direct visualization on a lateral flow format. As a proof-of-concept demonstration of its potential clinical utility, 21 clinical samples are evaluated by the as-developed sIgE PEAR programme using the dual readout modality of real-time fluorescence measurement for precise input quantification and simple lateral flow yes/no answer.

Implementing complex nucleic acid circuits in living cells

Synthetic nucleic acid-based computing has demonstrated complex computational capabilities in vitro. However, translating these circuits into living cells remains challenging because of instability and cellular interference. We introduce an allosteric strand exchange (ASE) strategy for complex intracellular computing. Leveraging conformational cooperativity to regulate strand exchange, ASE offers a modular platform for designing intracellular circuits with flexible programmability. We engineer a scalable circuit architecture based on ASE that can execute AND and OR logic and scale to an eight-input expression. We demonstrate ASE-based circuits can detect messenger RNAs with high specificity in mammalian cells via AND logic computation. The capacity of ASE-based circuits to accept messenger RNAs as inputs enables integration of endogenous cellular information for efficient multi-input information processing, demonstrated by a multi-input molecular classifier monitoring key cell reprogramming events. Reprogramming ASE-based circuit to interface with CRISPR-Cas9 enables programmable control of Cas9-targeting activity for gene editing, highlighting their potential for advancing intracellular biocomputation.

Accelerating Toehold-Mediated DNA Strand Displacement Reaction using Polyquaternium

Toehold-mediated strand displacement (TMSD) reaction is a widely used programming language in DNA nanotechnology, but its performance is significantly limited by slow kinetics, especially for low-concentration reactants. Herein, we report on polyquaternium-2 (PQ2) as an effective and efficient accelerator of TMSD reaction. We show that PQ2 could drastically increase the reaction constant of 1-nt TMSD by 105-fold. Significant acceleration of TMSD reactions with sub-nanomolar input has been demonstrated in various TMSD-based catalytic DNA amplifiers. By stabilizing DNA reactants and increasing their effective local concentrations, PQ2 enables much faster reaction kinetics in response to picomolar inputs while eliminating the dependence on toehold length, mitigating the inhibitory effect of secondary structures, maintaining single-base discriminating power, and protecting TMSD system in serum. Also, it improves cascaded signal transmission over an 11-layer circuit with 26 rounds of TMSD reactions, with a half-completion time of only 5.3 minutes. The simple-to-use and low-cost PQ2 offers a promising solution for uncovering the full potential of DNA nanotechnology and will facilitate more efficient and versatile TMSD-based applications from sensitive biosensing to high-performance molecular computing.

Circular RNA-Based Molecular Computation Enhances Plasma Biomarker Detection in Biliary Tract Cancer

Molecular diagnosis of biliary tract cancer (BTC) remains a significant clinical challenge due to the lack of sensitive and specific diagnostic tools. Although large panels based on multi-biomarkers have demonstrated potential in enhancing diagnostic accuracy, their clinical application is hindered by complexity and high costs. To overcome these limitations, we have developed a DNA-based molecular computation system that integrates the detection of plasma circular RNAs (circRNAs) with molecular computation, enabling intelligent and rapid diagnosis of BTC. By identifying and validating a specific and small set of circRNA biomarkers that are differentially expressed in BTC patients, we designed a DNA computation framework that directly translates biomarker levels into diagnostic outcomes without the need of external analysis and interpretation. In a validation cohort of 70 individuals, our approach achieved a diagnostic accuracy of 83%, demonstrating its potential as a cost-effective and efficient tool for rapid diagnosis of BTC. This work highlights the potential of molecular computation in enhancing integrated and rapid cancer diagnostics, paving the way for clinical implementation.

A Multi-Input Molecular Classifier Based on Digital DNA Strand Displacement for Disease Diagnostics

DNA-based molecular computing systems for biomarkers have emerged as powerful tools for intelligent diagnostics. However, with the variety of feature biomarkers expanding, current molecular computing systems suffer from the use of a large number of oligonucleotides and limited encoding capability. Here, the study develops an alternative molecular computing approach termed Digital DNA Strand Displacement (DDSD) which recognizes targets and operates target valence through DNA polymerase-based extension and strand release. DDSD significantly reduced the number of used oligonucleotide species, provided robust molecular classifiers. In clinical blood samples, a 96% accuracy rate is achieved with a DDSD-based binary classifier for distinguishing bacterial and viral infections, a 100% accuracy rate is achieved with a multiclass classifier for identifying pathogen types, surpassing existing classifier systems. Moreover, DDSD can be readily expanded. Cascade DDSD is developed, enabling simultaneous computing of up to 14 valence states with a maximum valence of 25. Multiway junction DDSD is implemented to achieve high-valence computing by compact DNA nanostructures rather than split DNA computing units, reducing the potential leakage. The implementation of DDSD enhances the capability of valence-based intelligent molecular diagnostics and multiplexed biomarker detection.

Regulation of transmembrane current through modulation of biomimetic lipid membrane composition

Ion transport through biological channels is influenced not only by the structural properties of the channels themselves but also by the composition of the phospholipid membrane, which acts as a scaffold for these nanochannels. Drawing inspiration from how lipid membrane composition modulates ion currents, as seen in the activation of the K+ channel in Streptomyces A (KcsA) by anionic lipids, we propose a biomimetic nanochannel system that integrates DNA nanotechnology with two-dimensional graphene oxide (GO) nanosheets. By modifying the length of the multibranched DNA nanowires generated through the hybridization chain reaction (HCR) and varying the concentration of the linker strands that integrate these DNA nanowire structures with the GO membrane, the composition of the membrane can be effectively adjusted, consequently impacting ion transport. This method provides a strategy for developing devices with highly efficient and tunable ion transport, suitable for applications in mass transport, environmental protection, biomimetic channels, and biosensors.

Logical Analysis of Multiple miRNAs with Isothermal Molecular Classifiers Based on LATE-RCA

Logical analysis of multiple-miRNA expression information and immediate output of diagnostic results facilitates early cancer detection. In this work, we constructed an isothermal molecular classifier capable of performing computations on multiple miRNAs and directly providing diagnosis results. First, we developed linear-after-the-exponential rolling circle amplification (LATE-RCA), a nearly linear isothermal amplification that does not destroy the original quantitative information about miRNAs. By designing different numbers of weighted coding sequences on the circular template, we naturally implemented multiplication in the LATE-RCA process. Summation, subtraction, and reporting were then carried out by strand displacement reactions. The entire workflow of the classifier was validated using synthetic gastric cancer and healthy miRNA samples with an accuracy of 100%, demonstrating its robustness and accuracy. Compared with existing molecular classifiers, our approach performs under isothermal conditions, streamlines computational procedures, and simplifies probe design. We believe that this isothermal molecular classifier has promising prospects in personalized precision medicine.

Autonomous walking dynamics of a nanorobot on a nanopore track driven by salt concentration gradients

The walking of a nanorobot with DNA legs requires a preset track to serve as footholds for the DNA legs and a track-matched driving mechanism to propel the nanorobot. Recently, a newly suggested track formed by multiple nanopores has garnered attraction due to its chemical stability. The nanorobot can be powered by biased leg-nanopore interactions along the walking direction. Here, we propose utilizing a salt concentration gradient along the nanopore track to induce an interaction bias based on different local ion screening effects on the nanopore charges. The nanorobot walking behaviors under different salt concentration gradients are studied through a combination of computational simulations and theoretical analyses. We find that the walking properties (such as velocity and directionality) highly depend on the local interactions experienced by the lagging leg. Under strong leg-nanopore attraction, the lagging leg needs a drag force provided by the leading leg to leave from the nanopore, while under weak attraction, the lagging leg can leave the nanopore without the assistance of the leading leg. Therefore, different walking modes can be observed under various ion conditions, leading to the complicated walking dynamics of the nanorobot driven by the salt concentration gradients.

Enabling global-scale nucleic acid repositories through versatile, scalable biochemical selection from room-temperature archives

Conventional collection, preservation, and retrieval of nucleic acid specimens, particularly unstable RNA, require costly cold-chain infrastructure and rely on inefficient robotic sample handling, hindering downstream analyses. These generate critical bottlenecks for global pathogen surveillance and genomic biobanking efforts, prohibiting large-scale nucleic acid sample collection and analyses that are needed to empower pathogen tracing, as well as rare disease diagnostics1. Here, we introduce a scalable nucleic acid storage system that enables rapid and precise retrieval on pooled nucleic acid samples-stored at room-temperature with minimal physical footprint2,3-using versatile database-like queries on barcoded, encapsulated samples. Queries can incorporate numerical ranges, categorical filters, and combinations thereof, which is a significant advancement beyond previous demonstrations limited to single-sample retrieval or Boolean classifiers. We apply our system to a pool of ninety-six mock SARS-CoV-2 genomic samples identified with theoretical patient data including patient age, geographic location, and diagnostic state, allowing rapid, multiplexed nucleic acid sample retrieval in a scalable manner to empower genomic analyses. By avoiding expensive and cumbersome freezer storage and retrieval systems, our approach in principle scales to millions of samples without loss of fidelity or throughput, thereby supporting the development of large-scale pathogen and genomic repositories in under-resourced or isolated regions of the US and worldwide.

Transfer learning Bayesian optimization for competitor DNA molecule design for use in diagnostic assays

With the rise in engineered biomolecular devices, there is an increased need for tailor-made biological sequences. Often, many similar biological sequences need to be made for a specific application meaning numerous, sometimes prohibitively expensive, lab experiments are necessary for their optimization. This paper presents a transfer learning design of experiments workflow to make this development feasible. By combining a transfer learning surrogate model with Bayesian optimization, we show how the total number of experiments can be reduced by sharing information between optimization tasks. We demonstrate the reduction in the number of experiments using data from the development of DNA competitors for use in an amplification-based diagnostic assay. We use cross-validation to compare the predictive accuracy of different transfer learning models, and then compare the performance of the models for both single objective and penalized optimization tasks.

High-Speed Sequential DNA Computing Using a Solid-State DNA Origami Register

DNA computing leverages molecular reactions to achieve diverse information processing functions. Recently developed DNA origami registers, which could be integrated with DNA computing circuits, allow signal transmission between these circuits, enabling DNA circuits to perform complex tasks in a sequential manner, thereby enhancing the programming space and compatibility with various biomolecules of DNA computing. However, these registers support only single-write operations, and the signal transfer involves cumbersome and time-consuming register movements, limiting the speed of sequential computing. Here, we designed a solid-state DNA origami register that compresses output data from a 3D solution to a 2D surface, establishing a rewritable register suitable for solid-state storage. We developed a heterogeneous integration architecture of liquid-state circuits and solid-state registers, reducing the register-mediated signal transfer time between circuits to less than 1 h, thereby achieving fast sequential DNA computing. Furthermore, we designed a trace signal amplifier to read surface-stored signals back into solution. This compact approach not only enhances the speed of sequential DNA computing but also lays the foundation for the visual debugging and automated execution of DNA molecular algorithms.

Refined design of a DNA logic gate for implementing a DNA-based three-level circuit

DNA computing circuits are favored by researchers because of their high density, high parallelism, and biocompatibility. However, compared with electronic circuits, current DNA circuits have significant errors in understanding the OFF state and logic "0". Nowadays, DNA circuits only have two input states: logic "0" and logic "1", where logic "0" also means the OFF state. Corresponding to an electronic circuit, it is more like an on-off switch than a logic circuit. To correct this conceptual confusion, we propose a three-level circuit. The circuit divides the input signal into three cases: "none", logic "0" and logic "1". In subsequent experiments, 34 input combinations of the primary AND gate, OR gate as well as secondary AND-OR and OR-AND cascade circuits were successfully implemented to perform the operation, which distinguished the OFF state and logic "0" correctly. Based on this, we successfully implemented a more complex voting operation with only 12 strands. We believe that our redefinition of the OFF state and logic "0" will promote tremendous developments in DNA computing circuits.

SARS-CoV-2 human challenge reveals biomarkers that discriminate early and late phases of respiratory viral infections

Blood transcriptional biomarkers of acute viral infections typically reflect type 1 interferon (IFN) signalling, but it is not known whether there are biological differences in their regulation that can be leveraged for distinct translational applications. We use high frequency sampling in the SARS-CoV-2 human challenge model to show induction of IFN-stimulated gene (ISG) expression with different temporal and cellular profiles. MX1 gene expression correlates with a rapid and transient wave of ISG expression across all cell types, which may precede PCR detection of replicative infection. Another ISG, IFI27, shows a delayed but sustained response restricted to myeloid cells, attributable to gene and cell-specific epigenetic regulation. These findings are reproducible in experimental and naturally acquired infections with influenza, respiratory syncytial virus and rhinovirus. Blood MX1 expression is superior to IFI27 expression for diagnosis of early infection, as a correlate of viral load and for discrimination of virus culture positivity. Therefore, MX1 expression offers potential to stratify patients for antiviral therapy or infection control interventions. Blood IFI27 expression is superior to MX1 expression for diagnostic accuracy across the time course of symptomatic infection and thereby, offers higher diagnostic yield for respiratory virus infections that incur a delay between transmission and testing.

Engineering a dual-loop molecular circuit with buffering capability to solve molecular information tasks

Molecular circuits, as an effective strategy for implementing artificial biochemical networks, have been widely constructed to process molecular-level information tasks both in vivo and in vitro. However, the complex and diverse structures of molecular devices, along with inflexible signal output methods, pose significant challenges for molecular circuits to handle complex molecular information tasks. In response to the growing field of molecular circuits, we design an exonuclease-driven fan-out molecular device (FMD) with a programmable cascade approach capable of receiving uniform signal types and transmitting multifunctional signals. Combined with the buffering reaction proposed here, the approach expands the dynamic properties of biochemical networks. Unlike the conventional delay strategy, the buffering process not only withstands transient changes in transmission signals, but also delays the transmission of lossless signals. Furthermore, we construct a dual-loop molecular circuit with adjustable buffering modes, thereby enabling signal amplification, time delay, and a differentiated output. Finally, we develop a method to obtain the colorimetric output of dual pulse signals driven by a dual-loop molecular circuit with buffering and hence precisely classify multiple signals. This work promises programmable and multifunctional molecular circuits in nanomachines, molecular computing, and biomedical applications.

Predicting DNA Reactions with a Quantum Chemistry-Based Deep Learning Model

In this study, a deep learning model based on quantum chemistry is introduced to enhance the accuracy and efficiency of predicting DNA reaction parameters. By integrating quantum chemical calculations with self-designed descriptor matrices, the model offers a comprehensive description of energy variations and considers a broad range of relevant factors. To overcome the challenge of limited labeled data, an active learning method is employed. The results demonstrate that this model outperforms existing methods in predicting DNA hybridization free energies and strand displacement rate constants, thus advancing the understanding of DNA molecular interactions, and aiding in the precise design and optimization of DNA-based systems.

Engineering Sequestration-Based Biomolecular Classifiers with Shared Resources

Constructing molecular classifiers that enable cells to recognize linear and nonlinear input patterns would expand the biocomputational capabilities of engineered cells, thereby unlocking their potential in diagnostics and therapeutic applications. While several biomolecular classifier schemes have been designed, the effects of biological constraints such as resource limitation and competitive binding on the function of those classifiers have been left unexplored. Here, we first demonstrate the design of a sigma factor-based perceptron as a molecular classifier working based on the principles of molecular sequestration between the sigma factor and its antisigma molecule. We then investigate how the output of the biomolecular perceptron, i.e., its response pattern or decision boundary, is affected by the competitive binding of sigma factors to a pool of shared and limited resources of core RNA polymerase. Finally, we reveal the influence of sharing limited resources on multilayer perceptron neural networks and outline design principles that enable the construction of nonlinear classifiers using sigma-based biomolecular neural networks in the presence of competitive resource-sharing effects.

DNA circuit-based immunoassay for ultrasensitive protein pattern classification

Cytokines are important immune modulators, and pivotal biomarkers for the diagnostic of various diseases. In standard analytical procedure, each protein is detected individually, using for instance gold standard ELISA protocols or nucleic acid amplification-based immunoassays. In recent years, DNA nanotechnology has been employed for creating sophisticated biomolecular systems that perform neuromorphic computing on molecular inputs, opening the door to concentration pattern recognition for biomedical applications. This work introduces immuno-PUMA (i-PUMA), an isothermal amplification-based immunoassay for ultrasensitive protein detection. The assay couples the convenience of supported format of an ELISA protocol with the computing capabilities of a DNA/enzyme circuit. We demonstrate a limit of detection of 2.1 fM, 8.7 fM and 450 aM for IL12, IL4 and IFNγ cytokines, respectively, outperforming the traditional ELISA format. i-PUMA's versatility extends to molecular computation, allowing the creation of 2-input perceptron-like classifiers for IL12 and IL4, with tunable weight sign and amplitude. Overall, i-PUMA represents a sensitive, low-cost, and versatile immunoassay with potential applications in multimarker-based sample classification, complementing existing molecular profiling techniques.

Strong sequence-dependence in RNA/DNA hybrid strand displacement kinetics

Strand displacement reactions underlie dynamic nucleic acid nanotechnology. The kinetic and thermodynamic features of DNA-based displacement reactions are well understood and well predicted by current computational models. By contrast, understanding of RNA/DNA hybrid strand displacement kinetics is limited, restricting the design of increasingly complex RNA/DNA hybrid reaction networks with more tightly regulated dynamics. Given the importance of RNA as a diagnostic biomarker, and its critical role in intracellular processes, this shortfall is particularly limiting for the development of strand displacement-based therapeutics and diagnostics. Herein, we characterise 22 RNA/DNA hybrid strand displacement systems, alongside 11 DNA/DNA systems, varying a range of common design parameters including toehold length and branch migration domain length. We observe  that differences in stability between RNA-DNA hybrids and DNA-DNA duplexes have large effects on strand displacement rates, with rates for equivalent sequences differing by up to 3 orders of magnitude. Crucially, however, this effect is strongly sequence-dependent, with RNA invaders strongly favoured in a system with RNA strands of high purine content, and disfavoured in a system when the RNA strands have low purine content. These results lay the groundwork for more general design principles, allowing for creation of de novo reaction networks with novel complexity while maintaining predictable reaction kinetics.

Rewireable Building Blocks for Enzyme-Powered DNA Computing Networks

Neural networks enable the processing of large, complex data sets with applications in disease diagnosis, cell profiling, and drug discovery. Beyond electronic computers, neural networks have been implemented using programmable biomolecules such as DNA; this confers unique advantages, such as greater portability, electricity-free operation, and direct analysis of patterns of biomolecules in solution. Analogous to bottlenecks in electronic computers, the computing power of DNA-based neural networks is limited by the ability to add more computing units, i.e., neurons. This limitation exists because current architectures require many nucleic acids to model a single neuron. Each additional neuron compounds existing problems such as long assembly times, high background signal, and cross-talk between components. Here, we test three strategies to solve this limitation and improve the scalability of DNA-based neural networks: (i) enzymatic synthesis for high-purity neurons, (ii) spatial patterning of neuron clusters based on their network position, and (iii) encoding neuron connectivity on a universal single-stranded DNA backbone. We show that neurons implemented via these strategies activate quickly, with a high signal-to-background ratio and process-weighted inputs. We rewired our modular neurons to demonstrate basic neural network motifs such as cascading, fan-in, and fan-out circuits. Finally, we designed a prototype two-layer microfluidic device to automate the operation of our circuits. We envision that our proposed design will help scale DNA-based neural networks due to its modularity, simplicity of synthesis, and compatibility with various neural network architectures. This will enable portable computing power for applications in portable diagnostics, compact data storage, and autonomous decision making for lab-on-a-chips.

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.

Precise control of transmembrane current via regulating bionic lipid membrane composition

The transport of ions through biological ion channels is regulated not only by their structural characteristics but also by the composition of the phospholipid membrane, which serves as a carrier for nanochannels. Inspired by the modulation of ion currents by lipid membrane composition, exemplified by the activation of the K+ channel of Streptomyces A by anionic lipids, we present a biomimetic nanochannel system based on combining DNA nanotechnology with two-dimensional graphene oxide (GO) nanosheets. By designing multibranched DNA nanowires, we assemble programmable DNA scaffold networks (DSNs) on the GO surface to precisely control membrane composition. Modulating the DSN layers from one to five enhances DNA composition, yielding a maximum 12-fold enhancement in ion current, primarily due to charge effects. Incorporating DNAzymes facilitates reversible modulation of membrane composition, enabling cyclic conversion of ion current. This approach offers a pathway for creating devices with highly efficient, tunable ion transport, applicable in diverse fields like mass transport, environmental protection, biomimetic channels, and biosensors.

An enzymatically activated AND-gate DNA logic circuit for tumor cells recognition via multi-microRNAs detection

The DNA-based logic circuit, constructed to mimic biochemical reaction networks, is highly significant in detecting biomarkers at the molecular level. The differences in the expression levels of microRNAs (miRNAs) within different types of cells provide hope for distinguishing cell subtypes. However, reliance on a single miRNA often leads to unreliable results. Herein, we constructed an enzyme-triggered cascade logic circuit based on the AND gate, which is capable of generating corresponding fluorescence signals in the presence of target miRNAs. The introduction of apurinic/apyrimidinic (AP) sites effectively reduces the likelihood of false signal generation. Amplification of the fluorescence signal relies on the catalytic hairpin assembly and the repetitive reuse of the multicomponent nucleic acid enzyme (MNAzyme). We demonstrated that the logic circuit can not only distinguish cancer cells from normal cells but also identify different types of cancer cells. The programmability of the logic circuits and the simplicity of the assay system allow us to modify the functional sequences to recognize different types of biomarkers, thus providing a reference for the identification of various cell subtypes.

DNA Framework-Based Programmable Atom-Like Nanoparticles for Non-Coding RNA Recognition and Differentiation of Cancer Cells

The cooperative diagnosis of non-coding RNAs (ncRNAs) can accurately reflect the state of cell differentiation and classification, laying the foundation of precision medicine. However, there are still challenges in simultaneous analyses of multiple ncRNAs and the integration of biomarker data for cell typing. In this study, DNA framework-based programmable atom-like nanoparticles (PANs) are designed to develop molecular classifiers for intra-cellular imaging of multiple ncRNAs associated with cell differentiation. The PANs-based molecular classifier facilitates signal amplification through the catalytic hairpin assembly. The interaction between PAN reporters and ncRNAs enables high-fidelity conversion of ncRNAs expression level into binding events, and the assessment of in situ ncRNAs levels via measurement of the fluorescent signal changes of PAN reporters. Compared to non-amplified methods, the detection limits of PANs are reduced by four orders of magnitude. Using human gastric cancer cell lines as a model system, the PANs-based molecular classifier demonstrates its capacity to measure multiple ncRNAs in living cells and assesses the degree of cell differentiation. This approach can serve as a universal strategy for the classification of cancer cells during malignant transformation and tumor progression.

Coupling Exponential to Linear Amplification for Endpoint Quantitative Analysis

Exponential DNA amplification techniques are fundamental in ultrasensitive molecular diagnostics. These systems offer a wide dynamic range, but the quantification requires real-time monitoring of the amplification reaction. Linear amplification schemes, despite their limited sensitivity, can achieve quantitative measurement from a single end-point readout, suitable for low-cost, point-of-care, or massive testing. Reconciling the sensitivity of exponential amplification with the simplicity of end-point readout would thus break through a major design dilemma and open a route to a new generation of massively scalable quantitative bioassays. Here a hybrid nucleic acid-based circuit design is introduced to compute a logarithmic function, therefore providing a wide dynamic range based on a single end-point measurement. CELIA (Coupling Exponential amplification reaction to LInear Amplification) exploits a versatile biochemical circuit architecture to couple a tunable linear amplification stage - optionally embedding an inverter function - downstream of an exponential module in a one-pot format. Applied to the detection of microRNAs, CELIA provides a limit of detection in the femtomolar range and a dynamic range of six decades. This isothermal approach bypasses thermocyclers without compromising sensitivity, thereby opening the way to applications in various diagnostic assays, and providing a simplified, cost-efficient, and high throughput solution for quantitative nucleic acid analysis.

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.

Fluorophore and nanozyme-functionalized DNA walking: A dual-mode DNA logic biocomputing platform for microRNA sensing in clinical samples

Inspired by the programmability and modifiability of nucleic acids, point-of-care (POC) diagnostics for nucleic acid target detection is evolving to become more diversified and intelligent. In this study, we introduce a fluorescent and photothermal dual-mode logic biosensing platform that integrates catalytic hairpin assembly (CHA), toehold-mediated stand displacement reaction (SDR) and a DNA walking machine. Dual identification and signal reporting modules are incorporated into DNA circuits, orchestrated by an AND Boolean logic gate operator and magnetic beads (MBs). In the presence of bispecific microRNAs (miRNAs), the AND logic gate activates, driving the DNA walking machine, and facilitating the collection of hairpin DNA stands modified with FAM fluorescent group and CeO2@Au nanoparticles. The CeO2@Au nanoparticles, served as a nanozyme, can oxidize TMB into oxidation TMB (TMBox), enabling a near-infrared (NIR) laser-driven photothermal effect following the magnetic separation of MBs. This versatile platform was employed to differentiate between plasma samples from breast cancer patients, lung cancer patients, and healthy donors. The thermometer-readout transducers, derived from the CeO2@Au@DNA complexes, provided reliable results, further corroborated by fluorescence assays, enhancing the confidence in the diagnostics compared to singular detection method. The dual-mode logic biosensor can be easily customized to various nucleic acid biomarkers and other POC signal readout modalities by adjusting recognition sequences and modification strategies, heralding a promising future in the development of intelligent, flexible diagnostics for POC testing.

Recent advances in enzyme-free and enzyme-mediated single-nucleotide variation assay in vitro

Single-nucleotide variants (SNVs) are the most common type variation of sequence alterations at a specific location in the genome, thus involving significant clinical and biological information. The assay of SNVs has engaged great awareness, because many genome-wide association studies demonstrated that SNVs are highly associated with serious human diseases. Moreover, the investigation of SNV expression levels in single cells are capable of visualizing genetic information and revealing the complexity and heterogeneity of single-nucleotide mutation-related diseases. Thus, developing SNV assay approaches in vitro, particularly in single cells, is becoming increasingly in demand. In this review, we summarized recent progress in the enzyme-free and enzyme-mediated strategies enabling SNV assay transition from sensing interface to the test tube and single cells, which will potentially delve deeper into the knowledge of SNV functions and disease associations, as well as discovering new pathways to diagnose and treat diseases based on individual genetic profiles. The leap of SNV assay achievements will motivate observation and measurement genetic variations in single cells, even within living organisms, delve into the knowledge of SNV functions and disease associations, as well as open up entirely new avenues in the diagnosis and treatment of diseases based on individual genetic profiles.

Machine Learning-Assistant Colorimetric Sensor Arrays for Intelligent and Rapid Diagnosis of Urinary Tract Infection

Urinary tract infections (UTIs), which can lead to pyelonephritis, urosepsis, and even death, are among the most prevalent infectious diseases worldwide, with a notable increase in treatment costs due to the emergence of drug-resistant pathogens. Current diagnostic strategies for UTIs, such as urine culture and flow cytometry, require time-consuming protocols and expensive equipment. We present here a machine learning-assisted colorimetric sensor array based on recognition of ligand-functionalized Fe single-atom nanozymes (SANs) for the identification of microorganisms at the order, genus, and species levels. Colorimetric sensor arrays are built from the SAN Fe1-NC functionalized with four types of recognition ligands, generating unique microbial identification fingerprints. By integrating the colorimetric sensor arrays with a trained computational classification model, the platform can identify more than 10 microorganisms in UTI urine samples within 1 h. Diagnostic accuracy of up to 97% was achieved in 60 UTI clinical samples, holding great potential for translation into clinical practice applications.

Synthetic Genes For Dynamic Regulation Of DNA-Based Receptors

We present a strategy to control dynamically the loading and release of molecular ligands from synthetic nucleic acid receptors using in vitro transcription. We demonstrate this by engineering three model synthetic DNA-based receptors: a triplex-forming DNA complex, an ATP-binding aptamer, and a hairpin strand, whose ability to bind their specific ligands can be cotranscriptionally regulated (activated or inhibited) through specific RNA molecules produced by rationally designed synthetic genes. The kinetics of our DNA sensors and their genetically generated inputs can be captured using differential equation models, corroborating the predictability of the approach used. This approach shows that highly programmable nucleic acid receptors can be controlled with molecular instructions provided by dynamic transcriptional systems, illustrating their promise in the context of coupling DNA nanotechnology with biological signaling.

DNA-based molecular classifiers for the profiling of gene expression signatures

Although gene expression signatures offer tremendous potential in diseases diagnostic and prognostic, but massive gene expression signatures caused challenges for experimental detection and computational analysis in clinical setting. Here, we introduce a universal DNA-based molecular classifier for profiling gene expression signatures and generating immediate diagnostic outcomes. The molecular classifier begins with feature transformation, a modular and programmable strategy was used to capture relative relationships of low-concentration RNAs and convert them to general coding inputs. Then, competitive inhibition of the DNA catalytic reaction enables strict weight assignment for different inputs according to their importance, followed by summation, annihilation and reporting to accurately implement the mathematical model of the classifier. We validated the entire workflow by utilizing miRNA expression levels for the diagnosis of hepatocellular carcinoma (HCC) in clinical samples with an accuracy 85.7%. The results demonstrate the molecular classifier provides a universal solution to explore the correlation between gene expression patterns and disease diagnostics, monitoring, and prognosis, and supports personalized healthcare in primary care.

Advances in Machine Learning Processing of Big Data from Disease Diagnosis Sensors

Exploring accurate, noninvasive, and inexpensive disease diagnostic sensors is a critical task in the fields of chemistry, biology, and medicine. The complexity of biological systems and the explosive growth of biomarker data have driven machine learning to become a powerful tool for mining and processing big data from disease diagnosis sensors. With the development of bioinformatics and artificial intelligence (AI), machine learning models formed by data mining have been able to guide more sensitive and accurate molecular computing. This review presents an overview of big data collection approaches and fundamental machine learning algorithms and discusses recent advances in machine learning and molecular computational disease diagnostic sensors. More specifically, we highlight existing modular workflows and key opportunities and challenges for machine learning to achieve disease diagnosis through big data mining.

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.

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.

Reservoir Computing With Dynamic Reservoir using Cascaded DNA Memristors

This article proposes molecular and DNA memristors where the state is defined by a single output variable. In past molecular and DNA memristors, the state of the memristor was defined based on two output variables. These memristors cannot be cascaded because their input and output sizes are different. We introduce a different definition of state for the molecular and DNA memristors. This change allows cascading of memristors. The proposed memristors are used to build reservoir computing (RC) models that can process temporal inputs. An RC system consists of two parts: reservoir and readout layer. The first part projects the information from the input space into a high-dimensional feature space. We also study the input-state characteristics of the cascaded memristors and show that the cascaded memristors retain the memristive behavior. The cascade connections in a reservoir can change dynamically; this allows the synthesis of a dynamic reservoir as opposed to a static one in the prior work. This reduces the number of memristors significantly compared to a static reservoir. The inputs to the readout layer correspond to one molecule per state instead of two; this significantly reduces the number of molecular and DSD reactions for the readout layer. A DNA RC system consisting of DNA memristors and a DNA readout layer is used to detect seizures from intra-cranial electroencephalogram (iEEG). We also demonstrate that a DNA RC system consisting of three cascaded DNA memristors and a DNA readout layer can be used to solve the time-series prediction task. The proposed approach can reduce the number of DNA strand displacement (DSD) reactions by three to five times compared to prior approaches.

Holistic Prediction of AuNP Aggregation in Diverse Aqueous Suspensions Based on Machine Vision and Dark-Field Scattering Imaging

The localized surface-plasmon resonance of the AuNP in aqueous media is extremely sensitive to environmental changes. By measuring the signal of plasmon scattering light, the dark-field microscopic (DFM) imaging technique has been used to monitor the aggregation of AuNPs, which has attracted great attention because of its simplicity, low cost, high sensitivity, and universal applicability. However, it is still challenging to interpret DFM images of AuNP aggregation due to the heterogeneous characteristics of the isolated and discontinuous color distribution. Herein, we introduce machine vision algorithms for the training of DFM images of AuNPs in different saline aqueous media. A visual deep learning framework based on AlexNet is constructed for studying the aggregation patterns of AuNPs in aqueous suspensions, which allows for rapid and accurate identification of the aggregation extent of AuNPs, with a prediction accuracy higher than 0.96. With the aid of machine learning analysis, we further demonstrate the prediction ability of various aggregation phenomena induced by both cation species and the concentration of the external saline solution. Our results suggest the great potential of machine vision frameworks in the accurate recognition of subtle pattern changes in DFM images, which can help researchers build predictive analytics based on DFM imaging data.

A geometric framework for reaction enumeration in computational nucleic acid devices

Cascades of DNA strand displacement reactions enable the design of potentially large circuits with complex behaviour. Computational modelling of such systems is desirable to enable rapid design and analysis. In previous work, the expressive power of graph theory was used to enumerate reactions implementing strand displacement across a wide range of complex structures. However, coping with the rich variety of possible graph-based structures required enumeration rules with complicated side-conditions. This paper presents an alternative approach to tackle the problem of enumerating reactions at domain level involving complex structures by integrating with a geometric constraint solving algorithm. The rule sets from previous work are simplified by replacing side-conditions with a general check on the geometric plausibility of structures generated by the enumeration algorithm. This produces a highly general geometric framework for reaction enumeration. Here, we instantiate this framework to solve geometric constraints by a structure sampling approach in which we randomly generate sets of coordinates and check whether they satisfy all the constraints. We demonstrate this system by applying it to examples from the literature where molecular geometry plays an important role, including DNA hairpin and remote toehold reactions. This work therefore enables integration of reaction enumeration and structural modelling.

A user-driven machine learning approach for RNA-based sample discrimination and hierarchical classification

RNA-based sample discrimination and classification can be used to provide biological insights and/or distinguish between clinical groups. However, finding informative differences between sample groups can be challenging due to the multidimensional and noisy nature of sequencing data. Here, we apply a machine learning approach for hierarchical discrimination and classification of samples with high-dimensional miRNA expression data. Our protocol comprises data preprocessing, unsupervised learning, feature selection, and machine-learning-based hierarchical classification, alongside open-source MATLAB code.

Moving dynamics of a nanorobot with three DNA legs on nanopore-based tracks

DNA nanorobots have garnered increasing attention in recent years due to their unique advantages of modularity and algorithm simplicity. To accomplish specific tasks in complex environments, various walking strategies are required for the DNA legs of the nanorobot. In this paper, we employ computational simulations to investigate a well-designed DNA-legged nanorobot moving along a nanopore-based track on a planar membrane. The nanorobot consists of a large nanoparticle as the robot core and three single-stranded DNAs (ssDNAs) as the robot legs. The nanopores linearly embedded in the membrane serve as the toeholds for the robot legs. A charge gradient along the pore distribution mainly powers the activation of the nanorobot. The nanorobot can move in two modes: a walking mode, where the robot legs sequentially enter the nanopores, and a jumping mode, where the robot legs may skip a nanopore to reach the next one. Moreover, we observe that the moving dynamics of the nanorobot on the nanopore-based tracks depends on pore-pore distance, pore charge gradient, external voltage, and leg length.

DNA-based programmable gate arrays for general-purpose DNA computing

The past decades have witnessed the evolution of electronic and photonic integrated circuits, from application specific to programmable1,2. Although liquid-phase DNA circuitry holds the potential for massive parallelism in the encoding and execution of algorithms3,4, the development of general-purpose DNA integrated circuits (DICs) has yet to be explored. Here we demonstrate a DIC system by integration of multilayer DNA-based programmable gate arrays (DPGAs). We find that the use of generic single-stranded oligonucleotides as a uniform transmission signal can reliably integrate large-scale DICs with minimal leakage and high fidelity for general-purpose computing. Reconfiguration of a single DPGA with 24 addressable dual-rail gates can be programmed with wiring instructions to implement over 100 billion distinct circuits. Furthermore, to control the intrinsically random collision of molecules, we designed DNA origami registers to provide the directionality for asynchronous execution of cascaded DPGAs. We exemplify this by a quadratic equation-solving DIC assembled with three layers of cascade DPGAs comprising 30 logic gates with around 500 DNA strands. We further show that integration of a DPGA with an analog-to-digital converter can classify disease-related microRNAs. The ability to integrate large-scale DPGA networks without apparent signal attenuation marks a key step towards general-purpose DNA computing.

Protein-DNA Conjugates with a Discrete Number of Oligonucleotide Strands for Highly Reproducible Protein Quantification by the DNA Proximity Assay

Protein-oligonucleotide conjugates are increasingly used as detection probes in biological applications such as proximity sensing and spatial biology. The preparation of high-quality conjugate probes as starting reagents is critical for achieving good and consistent performance, which we demonstrate via the DNA proximity assay (DPA) for the one-pot quantification of protein targets. We first established a complete conjugation and anion-exchange chromatography purification workflow to reproducibly obtain pure subpopulations of protein probes carrying a discrete number of oligonucleotide strands. A systematic study using the purified conjugate sub-populations confirmed that the order of conjugate (number of oligonucleotides per protein) and its purity (the absence of the unconjugated antibody) were important for ensuring optimal and reproducible assay performance. The streamlined workflow was then successfully used to conjugate a pair of universal DPA initiator oligonucleotides onto a wide range of binders including antibodies, nanobodies, and antigens which enabled the versatile detection of different types of proteins such as cytokines, total antibodies, and specific antibody isotypes. The good assay robustness (the inter-assay coefficient of variation lower than 5%) and linear calibration curve was achieved across all targets with just a single mix-and-incubate reaction step and a short reaction time of 30 min. We anticipate the streamlined protein-oligonucleotide probe preparation workflow developed in this work to have broad utility across applications leveraging the specificity of protein bio-recognition with the programmability of DNA hybridization.

Design and Simulation of a Multilayer Chemical Neural Network That Learns via Backpropagation

The design and implementation of adaptive chemical reaction networks, capable of adjusting their behavior over time in response to experience, is a key goal for the fields of molecular computing and DNA nanotechnology. Mainstream machine learning research offers powerful tools for implementing learning behavior that could one day be realized in a wet chemistry system. Here we develop an abstract chemical reaction network model that implements the backpropagation learning algorithm for a feedforward neural network whose nodes employ the nonlinear "leaky rectified linear unit" transfer function. Our network directly implements the mathematics behind this well-studied learning algorithm, and we demonstrate its capabilities by training the system to learn a linearly inseparable decision surface, specifically, the XOR logic function. We show that this simulation quantitatively follows the definition of the underlying algorithm. To implement this system, we also report ProBioSim, a simulator that enables arbitrary training protocols for simulated chemical reaction networks to be straightforwardly defined using constructs from the host programming language. This work thus provides new insight into the capabilities of learning chemical reaction networks and also develops new computational tools to simulate their behavior, which could be applied in the design and implementations of adaptive artificial life.

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.

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.

DNA-framework-based multidimensional molecular classifiers for cancer diagnosis

A molecular classification of diseases that accurately reflects clinical behaviour lays the foundation of precision medicine. The development of in silico classifiers coupled with molecular implementation based on DNA reactions marks a key advance in more powerful molecular classification, but it nevertheless remains a challenge to process multiple molecular datatypes. Here we introduce a DNA-encoded molecular classifier that can physically implement the computational classification of multidimensional molecular clinical data. To produce unified electrochemical sensing signals across heterogeneous molecular binding events, we exploit DNA-framework-based programmable atom-like nanoparticles with n valence to develop valence-encoded signal reporters that enable linearity in translating virtually any biomolecular binding events to signal gains. Multidimensional molecular information in computational classification is thus precisely assigned weights for bioanalysis. We demonstrate the implementation of a molecular classifier based on programmable atom-like nanoparticles to perform biomarker panel screening and analyse a panel of six biomarkers across three-dimensional datatypes for a near-deterministic molecular taxonomy of prostate cancer patients.

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'.

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.

FRET cascade miRNA addition probe from non-crosstalk DNA photonic wire assisted with clustering algorithm for early diagnosis of acute myocardial infarction

Early and accurate diagnosis of acute myocardial infarction (AMI) can significantly reduce patient mortality. A variety of miRNAs are found to dysregulate in AMI patients, but the up- or down-regulation of a specific miRNA may not be evident in the early stage, making it difficult to achieve accurate diagnosis. Here, proposing the design that DNA photonic wire (PW) with no spectral crosstalk would make an excellent template for miRNA conjoint analysis, we report the construction of a miRNA addition probe for the additive analysis of two up-regulated miRNAs (miR-133a and miR-208a) for early diagnosis of AMI in clinical serum samples. A three-dye non-crosstalk DNA PW is built to form the two-step fluorescence resonance energy transfer (FRET) cascade system, in which three paths can blocking the FRET cascade for separate or additive analysis of the two miRNAs. K-Means clustering algorithm is further utilized to classify the output signals of the miRNA addition probe, achieving a 100% accurate diagnosis of early AMI in both the training (n = 40) and validation (n = 19) cohorts of clinical serum samples.

A review of sensors for classification and subtype discrimination of cancer: Insights into circulating tumor cells and tumor-derived extracellular vesicles

Liquid biopsy can reflect the state of tumors in vivo non-invasively, thus providing a strong basis for the early diagnosis, individualized treatment monitoring and prognosis of tumors. Circulating tumor cells (CTCs) and tumor-derived extracellular vesicles (tdEVs) contain information-rich components, such as nucleic acids and proteins, and they are essential markers for liquid biopsies. Their capture and analysis are of great importance for the study of disease occurrence and development and, consequently, have been the subject of many reviews. However, both CTCs and tdEVs carry the biological characteristics of their original tissue, and few reviews have focused on their function in the staging and classification of cancer. In this review, we focus on state-of-the-art sensors based on the simultaneous detection of multiple biomarkers within CTCs and tdEVs, with clinical applications centered on cancer classification and subtyping. We also provide a thorough discussion of the current challenges and prospects for novel sensors with the ultimate goal of cancer classification and staging. It is hoped that these most advanced technologies will bring new insights into the clinical practice of cancer screening and diagnosis.

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.