Training an asymmetric signal perceptron through reinforcement in an artificial chemistry

State-of-the-art biochemical systems for medical applications and chemical computing are application-specific and cannot be reprogrammed or trained once fabricated. The implementation of adaptive biochemical systems that would offer flexibility through programmability and autonomous adaptation faces major challenges because of the large number of required chemical species as well as the timing-sensitive feedback loops required for learning. In this paper, we begin addressing these challenges with a novel chemical perceptron that can solve all 14 linearly separable logic functions. The system performs asymmetric chemical arithmetic, learns through reinforcement and supports both Michaelis–Menten as well as mass-action kinetics. To enable cascading of the chemical perceptrons, we introduce thresholds that amplify the outputs. The simplicity of our model makes an actual wet implementation, in particular by DNA-strand displacement, possible.     

What do molecules do when we are not looking? State sequence analysis for stochastic chemical systems

Many biomolecular systems depend on orderly sequences of chemical transformations or reactions. Yet, the dynamics of single molecules or small-copy-number molecular systems are significantly stochastic. Here, we propose state sequence analysis—a new approach for predicting or visualizing the behaviour of stochastic molecular systems by computing maximum probability state sequences, based on initial conditions or boundary conditions. We demonstrate this approach by analysing the acquisition of drug-resistance mutations in the human immunodeficiency virus genome, which depends on rare events occurring on the time scale of years, and the stochastic opening and closing behaviour of a single sodium ion channel, which occurs on the time scale of milliseconds. In both cases, we find that our approach yields novel insights into the stochastic dynamical behaviour of these systems, including insights that are not correctly reproduced in standard time-discretization approaches to trajectory analysis.     

Transient response characteristics in a biomolecular integral controller

The cellular behaviour of perfect adaptation is achieved through the use of an integral control element in the underlying biomolecular circuit. It is generally unclear how integral action affects the important aspect of transient response in these biomolecular systems, especially in light of the fact that it typically deteriorates the transient response in engineering contexts. To address this issue, the authors investigated the transient response in a computational model of a simple biomolecular integral control system involved in bacterial signalling. They find that the transient response can actually speed up as the integral gain parameter increases. On further analysis, they find that the underlying dynamics are composed of slow and fast modes and the speed‐up of the transient response is because of the speed‐up of the slow‐mode dynamics. Finally, they note how an increase in the integral gain parameter also leads to a decrease in the amplitude of the transient response, consistent with the overall improvement in the transient response. These results should be useful in understanding the overall effect of integral action on system dynamics, particularly for biomolecular systems.Inspec keywords: transient response, molecular biophysics, cellular biophysics, microorganisms, biocontrolOther keywords: transient response characteristics, biomolecular integral controller, cellular behaviour, integral control element, biomolecular circuit, integral action, computational model, biomolecular integral control system, bacterial signalling, integral gain parameter     

Engineering materials for artificial cells

The grand challenge of engineering a minimal artificial cell provides a controllable framework for studying the biochemical principles of life. Artificial cells contribute to an increased understanding of complex synthetic systems with life-like properties and provide opportunities to create autonomous cell-like materials. Recent efforts to develop life-like artificial cells by bottom-up approaches involve mimicking the behavior of lipid membranes to recapitulate fundamental cellular processes. This review describes the recent progress in engineering biomimetic artificial minimal cells and recently developed chemical strategies to drive de novo membrane formation from simple synthetic precursors. In the end, we briefly point out the challenges and possible future directions in the development of artificial cells.     

Modular dynamic biomolecular modelling with bond graphs: the unification of stoichiometry,thermodynamics, kinetics and data

Renewed interest in dynamic simulation models of biomolecular systems has arisen from advances in genome-wide measurement and applications of such models in biotechnology and synthetic biology. In particular, genome-scale models of cellular metabolism beyond the steady state are required in order to represent transient and dynamic regulatory properties of the system. Development of such whole-cell models requires new modelling approaches. Here, we propose the energy-based bond graph methodology, which integrates stoichiometric models with thermodynamic principles and kinetic modelling. We demonstrate how the bond graph approach intrinsically enforces thermodynamic constraints, provides a modular approach to modelling, and gives a basis for estimation of model parameters leading to dynamic models of biomolecular systems. The approach is illustrated using a well-established stoichiometric model of Escherichia coli and published experimental data.     

Non‐normality can facilitate pulsing in biomolecular circuits

Non‐normality can underlie pulse dynamics in many engineering contexts. However, its role in pulses generated in biomolecular contexts is generally unclear. Here, the authors address this issue using the mathematical tools of linear algebra and systems theory on simple computational models of biomolecular circuits. They find that non‐normality is present in standard models of feedforward loops. They used a generalised framework and pseudospectrum analysis to identify non‐normality in larger biomolecular circuit models, finding that it correlates well with pulsing dynamics. Finally, they illustrate how these methods can be used to provide analytical support to numerical screens for pulsing dynamics as well as provide guidelines for design.Inspec keywords: linear algebra, feedforward, eigenvalues and eigenfunctions, network analysis, molecular biophysicsOther keywords: nonnormality, biomolecular circuits, pulse dynamics, engineering contexts, biomolecular contexts, linear algebra, systems theory, simple computational models, standard models, larger biomolecular circuit models     

Modulation of Biointeractions by Electrically Switchable Oligopeptide Surfaces: Structural Requirements and Mechanism

Understanding the dynamic behavior of switchable surfaces is of paramount importance for the development of controllable and tailor‐made surface materials. Herein, electrically switchable mixed self‐assembled monolayers based on oligopeptides have been investigated in order to elucidate their conformational mechanism and structural requirements for the regulation of biomolecular interactions between proteins and ligands appended to the end of surface tethered oligopeptides. The interaction of the neutravidin protein to a surface appended biotin ligand was chosen as a model system. All the considerable experimental data, taken together with detailed computational work, support a switching mechanism in which biomolecular interactions are controlled by conformational changes between fully extended (“ON” state) and collapsed (“OFF” state) oligopeptide conformer structures. In the fully extended conformation, the biotin appended to the oligopeptide is largely free from steric factors allowing it to efficiently bind to the neutravidin from solution. While under a collapsed conformation, the ligand presented at the surface is partially embedded in the second component of the mixed SAM, and thus sterically shielded and inaccessible for neutravidin binding. Steric hindrances aroused from the neighboring surface‐confined oligopeptide chains exert a great influence over the conformational behaviour of the oligopeptides, and as a consequence, over the switching efficiency. Our results also highlight the role of oligopeptide length in controlling binding switching efficiency. This study lays the foundation for designing and constructing dynamic surface materials with novel biological functions and capabilities, enabling their utilization in a wide variety of biological and medical applications.     

Biocomputing with Nanostructures on Lipid Bilayers

Biocomputation is the algorithmic manipulation of biomolecules. Nanostructures, most notably DNA nanostructures and nanoparticles, become active substrates for biocomputation when modified with stimuli‐responsive, programmable biomolecular ligands. This approach—biocomputing with nanostructures (“nano‐bio computing”)—allows autonomous control of matter and information at the nanoscale; their dynamic assemblies and beneficial properties can be directed without human intervention. Recently, lipid bilayers interfaced with nanostructures have emerged as a new biocomputing platform. This new nano‐bio interface, which exploits lipid bilayers as a chemical circuit board for information processing, offers a unique reaction space for realizing nanostructure‐based computation at a previously unexplored dimension. In this Concept, recent advances in nano‐bio computing are briefly reviewed and the newly emerging concept of biocomputing with nanostructures on lipid bilayers is introduced.     

Membrane Tension–Mediated Growth of Liposomes

Recent years have seen a tremendous interest in the bottom‐up reconstitution of minimal biomolecular systems, with the ultimate aim of creating an autonomous synthetic cell. One of the universal features of living systems is cell growth, where the cell membrane expands through the incorporation of newly synthesized lipid molecules. Here, the gradual tension‐mediated growth of cell‐sized (≈10 µm) giant unilamellar vesicles (GUVs) is demonstrated, to which nanometer‐sized (≈30 nm) small unilamellar vesicles (SUVs) are provided, that act as a lipid source. By putting tension on the GUV membranes through a transmembrane osmotic pressure, SUV–GUV fusion events are promoted and substantial growth of the GUV is caused, even up to doubling its volume. Thus, experimental evidence is provided that membrane tension alone is sufficient to bring about membrane fusion and growth is demonstrated for both pure phospholipid liposomes and for hybrid vesicles with a mixture of phospholipids and fatty acids. The results show that growth of liposomes can be realized in a protein‐free minimal system, which may find useful applications in achieving autonomous synthetic cells that are capable of undergoing a continuous growth–division cycle.     

Big data and machine learning: A roadmap towards smart plants

Industry 4.0 aims to transform chemical and biochemical processes into intelligent systems via the integration of digital components with the actual physical units involved. This process can be thought of as addition of a central nervous system with a sensing and control monitoring of components and regulating the performance of the individual physical assets (processes, units, etc.) involved. Established technologies central to the digital integrating components are smart sensing, mobile communication, Internet of Things, modelling and simulation, advanced data processing, storage and analysis, advanced process control, artificial intelligence and machine learning, cloud computing, and virtual and augmented reality. An essential element to this transformation is the exploitation of large amounts of historical process data and large volumes of data generated in real-time by smart sensors widely used in industry. Exploitation of the information contained in these data requires the use of advanced machine learning and artificial intelligence technologies integrated with more traditional modelling techniques. The purpose of this paper is twofold: a) to present the state-of-the-art of the aforementioned technologies, and b) to present a strategic plan for their integration toward the goal of an autonomous smart plant capable of self-adaption and self-regulation for short- and long-term production management.     

Self-healing and repair of fabrics: A comprehensive review of the application toolkit

Self-healing fabrics respond to chemical and physical damage by restoring functional, structural, and morphological features. We present a comprehensive review of textile hybrids or composites capable of self-healing and repairing fabrics against damages across the micro- (µm), meso- (µm – mm), and macro-scale (>mm). The reviewed literature is organized in three sections presenting (i) the chemistry and fabrication principles of designing self-healing fabrics against increasing size scales of repair, (ii) stimuli-driven and autonomous healing, and (iii) the methods to characterize the recovery of wettability, barrier, morphological, mechanical, and other properties. The discussion of mainstream methods for developing self-healing fabrics focuses on coatings, composites, and specialized fabrication techniques required as the damage size grows from µm to mm to >mm. The section on stimuli-driven repair and autonomous recovery discusses the time scales associated with different damage repair, showing how external stimuli provide a higher driving force towards healing and accelerate material restoration than autonomous recovery. Finally, an array of optical, mechanical, and functional characterization techniques is discussed to evaluate the recovery yield and understand the repair mechanisms of the various fabrics. This review demonstrates the virtually limitless uses of next-generation self-healing systems, from separations to protective clothing, anti-fouling, and self-cleaning.     

Programmable DNA Nanoindicator‐Based Platform for Large‐Scale Square Root Logic Biocomputing

The prospect of programming molecular computing systems to realize complex autonomous tasks has advanced the design of synthetic biochemical logic circuits. One way to implement digital and analog integrated circuits is to use noncovalent hybridization and strand displacement reactions in cell‐free and enzyme‐free nucleic acid systems. To date, DNA‐based circuits involving tens of logic gates capable of implementing basic and complex logic functions have been demonstrated experimentally. However, most of these circuits are still incapable of realizing complex mathematical operations, such as square root logic operations, which can only be carried out with 4 bit binary numbers. A high‐capacity DNA biocomputing system is demonstrated through the development of a 10 bit square root logic circuit. It can calculate the square root of a 10 bit binary number (within the decimal integer 900) by designing DNA sequences and programming DNA strand displacement reactions. The input signals are optimized through the output feedback to improve performance in more complex logical operations. This study provides a more universal approach for applications in biotechnology and bioengineering.     

Inkjet printed surface enhanced Raman spectroscopy array on cellulose paper

A novel, ultra low-cost surface enhanced Raman spectroscopy (SERS) substrate has been developed by modifying the surface chemistry of cellulose paper and patterning nanoparticle arrays, all with a consumer inkjet printer. Micro/nanofabrication of SERS substrates for on-chip chemical and biomolecular analysis has been under intense investigation. However, the high cost of producing these substrates and the limited shelf life severely limit their use, especially for routine laboratory analysis and for point-of-sample analysis in the field. Paper-based microfluidic biosensing systems have shown great potential as low-cost disposable analysis tools. In this work, this concept is extended to SERS-based detection. Using an inexpensive consumer inkjet printer, cellulose paper substrates are modified to be hydrophobic in the sensing regions. Synthesized silver nanoparticles are printed onto this hydrophobic paper substrate with microscale precision to form sensing arrays. The hydrophobic surface prevents the aqueous sample from spreading throughout the paper and thus concentrates the analyte within the sensing region. A SERS fingerprint signal for Rhodamine 6G dye was observed for samples with as low as 10 femtomoles of analyte in a total sample volume of 1 μL. This extraordinarily simple technique can be used to construct SERS microarrays immediately before sample analysis, enabling ultra low-cost chemical and biomolecular detection in the lab as well as in the field at the point of sample collection.     

Modelling emergence of oscillations in communicating bacteria: a structured approach from one to many cells

Population-level measurements of phenotypic behaviour in biological systems may not necessarily reflect individual cell behaviour. To assess qualitative changes in the behaviour of a single cell, when alone and when part of a community, we developed an agent-based model describing the metabolic states of a population of quorum-coupled cells. The modelling is motivated by published experimental work of a synthetic genetic regulatory network (GRN) used in Escherichia coli cells that exhibit oscillatory behaviour across the population. To decipher the mechanisms underlying oscillations in the system, we investigate the behaviour of the model via numerical simulation and bifurcation analysis. In particular, we study the effect of an increase in population size as well as the spatio-temporal behaviour of the model. Our results demonstrate that oscillations are possible only in the presence of a high concentration of the coupling chemical and are due to a time scale separation in key regulatory components of the system. The model suggests that the population establishes oscillatory behaviour as the system''s preferred stable state. This is achieved via an effective increase in coupling across the population. We conclude that population effects in GRN design need to be taken into consideration and be part of the design process. This is important in planning intervention strategies or designing specific cell behaviours.     

Modelling and analysis of the sugar cataract development process using stochastic hybrid systems

Modelling and analysis of biochemical systems such as sugar cataract development (SCD) are critical because they can provide new insights into systems, which cannot be easily tested with experiments; however, they are challenging problems due to the highly coupled chemical reactions that are involved. The authors present a stochastic hybrid system (SHS) framework for modelling biochemical systems and demonstrate the approach for the SCD process. A novel feature of the framework is that it allows modelling the effect of drug treatment on the system dynamics. The authors validate the three sugar cataract models by comparing trajectories computed by two simulation algorithms. Further, the authors present a probabilistic verification method for computing the probability of sugar cataract formation for different chemical concentrations using safety and reachability analysis methods for SHSs. The verification method employs dynamic programming based on a discretisation of the state space and therefore suffers from the curse of dimensionality. To analyse the SCD process, a parallel dynamic programming implementation that can handle large, realistic systems was developed. Although scalability is a limiting factor, this work demonstrates that the proposed method is feasible for realistic biochemical systems.     

Machine learning-guided synthesis of advanced inorganic materials

Synthesis of materials with minimum number of trials is of paramount importance towards the acceleration of advanced materials development. The enormous complexity involved in existing multi-variable synthesis methods leads to high uncertainty, numerous trials and exorbitant cost. Recently, machine learning (ML) has demonstrated tremendous potential for material discovery and property enhancement. Here, we extend the application of ML to guide material synthesis process through the establishment of the methodology including model construction, optimization, and progressive adaptive model (PAM). Two representative multi-variable systems are studied. A classification ML model on chemical vapor grown MoS₂ is developed, capable of optimizing the synthesis conditions to achieve a higher success rate. And a regression model is constructed on the hydrothermal-grown carbon quantum dots, to enhance the process-related properties such as the photoluminescence quantum yield. The importance of synthesis parameters on experimental outcomes is particularly extracted from the constructed ML models. Furthermore, off-line analysis shows that enhancement of the experimental outcome with minimized number of trials can be achieved with the effective feedback loops in PAM, suggesting the great potential of involving ML to guide new material synthesis at the beginning stage. This work serves as a proof of concept for using ML in facilitating the synthesis of inorganic materials, thereby revealing the feasibility and remarkable capability of ML in opening up a new promising window for accelerating material development.     

Deep Reinforcement Learning-Based Job Shop Scheduling ofSmartManufacturing

Industry 4.0 production environments and smart manufacturing systems integrate both the physical and decision-making aspects of manufacturing operations into autonomous and decentralized systems. One of the key aspects of these systems is a production planning, specifically, Scheduling operations on the machines. To cope with this problem, this paper proposed a Deep Reinforcement Learning with an Actor-Critic algorithm (DRLAC). We model the Job-Shop Scheduling Problem (JSSP) as a Markov Decision Process (MDP), represent the state of a JSSP as simple Graph Isomorphism Networks (GIN) to extract nodes features during scheduling, and derive the policy of optimal scheduling which guides the included node features to the best next action of schedule. In addition, we adopt the Actor-Critic (AC) network’s training algorithm-based reinforcement learning for achieving the optimal policy of the scheduling. To prove the proposed model’s effectiveness, first, we will present a case study that illustrated a conflict between two job scheduling, secondly, we will apply the proposed model to a known benchmark dataset and compare the results with the traditional scheduling methods and trending approaches. The numerical results indicate that the proposed model can be adaptive with real-time production scheduling, where the average percentage deviation (APD) of our model achieved values between 0.009 and 0.21 compared with heuristic methods and values between 0.014 and 0.18 compared with other trending approaches.     

含间隙多约束碰撞振动系统稳定性分析

针对一类单自由度含间隙多约束碰撞振动系统,通过在碰撞面处建立系统的Poincaré 映射,推导系统的Jacobi 矩阵,将连续动力系统转换为离散动力系统,并利用Gram-Schmidt 正交化和范式归一化计算得到系统的Lyapunov 指数谱。通过数值模拟,计算系统混沌吸引子与周期吸引子的收敛序列,结合系统相图、单参分岔图及Lyapunov 指数 谱,分析系统周期运动稳定性及各类分岔现象,通过控制系统参数双向变化发现相邻周期运动间存在的周期共存现 象,验证该计算方法的有效性和正确性,研究成果可为后续针对该系统的混沌判断及混沌控制提供理论依据。