A method and system for design optimization of a two-component biosensor

Abstract

The present application relates to the technical field of biosensor modeling, and particularly relates to a design optimization method and system of a two-component biosensor. The present application constructs a reaction network graph based on the structured configuration data of the two-component biosensor; based on the reaction type and time scale attribute of each reaction edge in the reaction network graph, a corresponding modeling template is matched to generate a reaction flux expression or an algebraic relationship expression corresponding to each reaction edge; a state variable set, an algebraic variable set, a model parameter set, a state variable stoichiometric matrix, a reaction flux vector and an algebraic relationship set are obtained; based on the state variable stoichiometric matrix and the reaction flux vector, a common differential equation corresponding to the state variable is generated, and after substitution and elimination of the algebraic variable in the common differential equation, a reduced dimension mechanism model is obtained, and the design optimization of the two-component biosensor is realized.

Description

Technical Field

[0001] This invention relates to the field of biosensor modeling technology, and in particular to a design optimization method and system for a two-component biosensor. Background Technology

[0002] The two-component system is a core functional system in prokaryotes for adapting to changes in the external environment and completing signal sensing and intracellular transduction. Its core consists of two major functional elements: membrane-localized histidine kinases and intracellular regulatory proteins. In natural physiological processes, histidine kinases can specifically recognize inducing signals such as environmental stimuli and changes in metabolite concentration. Through autophosphorylation, transmolecular transfer of phosphate groups, and a phosphorylation cascade reaction of regulatory proteins, they precisely regulate the transcriptional activity of downstream target promoters, thereby initiating the specific expression of target genes. Based on this sensitive signal cascade response characteristic, the two-component system has been widely modified and reconstructed into intelligent biosensors that specifically respond to environmental signals and target metabolites. By outputting detection results through fluorescence, luminescence, colorimetry, or electrochemical signals, it achieves qualitative and quantitative identification of target analytes, possessing extremely high application value in environmental monitoring, microbial metabolic regulation, biofermentation detection, and pathogen screening.

[0003] The core performance characteristics of two-component biosensors, such as detection sensitivity, response threshold, dynamic detection range, and specificity, directly depend on key physiological processes within the system, including the expression ratio of internal components, phosphorylation dynamic equilibrium, protein complex binding efficiency, and promoter transcriptional activation capacity. To accurately guide the structural design, component ratio optimization, and performance iteration upgrades of two-component biosensors, clarifying the dynamic coupling relationship between inducer concentration, histidine kinase and response regulatory protein expression levels, protein phosphorylation homeostasis, promoter binding occupancy rate, and terminal detection signal output, and constructing accurate mechanistic dynamic models have become core technical means for sensor optimization design. Currently, mainstream sensor modeling systems typically describe the system operation process based on the law of mass action, Hill's function, first-order synthetic degradation kinetics, and ordinary differential equations. This effectively quantifies the impact of each core parameter on sensor performance, providing theoretical support for targeted sensor optimization and eliminating the blindness of traditional experimental trial and error.

[0004] In the engineering design of two-component biosensors, mechanistic modeling is the core bridge connecting the sensor's molecular structure, internal reaction mechanism, and terminal detection performance. The overall design scheme, component parameter ratios, and functional module modifications of the sensor all require a precise kinetic model as theoretical support. Modeling can quantify the impact of key parameters such as histidine kinase, expression levels of response regulatory proteins, phosphorylation cycle efficiency, and promoter binding ability on sensor sensitivity, response speed, and detection threshold. This helps researchers accurately select the optimal design scheme, avoiding the randomness of traditional pure experimental trial-and-error designs, and is the core foundation for achieving targeted modification, performance optimization, and standardized design of two-component sensors. However, current modeling methods for two-component biosensors still have certain shortcomings. The construction of existing two-component biosensor mechanistic models usually relies on researchers to complete these tasks manually, including manually determining state variables, manually writing reaction fluxes, manually compiling parameter tables, manually deriving ordinary differential equations, and writing simulation code. This approach has the following problems:

[0005] Current modeling processes typically require researchers to manually determine state variables, reaction fluxes, and parameters for a specific two-component system, and then derive a set of ordinary differential equations. For systems involving processes such as dimerization, complex formation, phosphorylation, dephosphorylation, and promoter binding, manual modeling is prone to problems such as omitted variables, duplicated parameters, or inconsistent equation dimensions, resulting in poor detection accuracy of the designed two-component biosensors. Furthermore, the lack of a unified data structure between different sensor models makes it difficult to directly transfer existing models to new two-component systems. Treating all intermediate complexes as independent state variables further increases model dimensionality and computational burden, requiring remodeling and re-exploration of design parameters for each new sensor development. This significantly prolongs the development cycle of novel two-component biosensors and increases experimental development costs. Summary of the Invention

[0006] Therefore, the technical problem to be solved by the present invention is to overcome the shortcomings of existing two-component biosensors, which rely on manual experience modeling, resulting in poor detection accuracy and difficulty in directly transferring different sensor models, leading to high R&D costs and long development cycles for two-component biosensors.

[0007] To address the aforementioned technical problems, this invention provides a design optimization method for a two-component biosensor, comprising: A reaction network graph is constructed based on structured configuration data of a two-component biosensor; the reaction network graph includes multiple molecular entity nodes and reaction edges between different molecular entity nodes; Based on the reaction type and time scale attributes of each reaction edge in the reaction network graph, the modeling template corresponding to each reaction edge is matched to generate the reaction flux expression or algebraic relation expression corresponding to each reaction edge. Based on the reaction network diagram and the reaction flux expression or algebraic relation expression corresponding to each reaction edge, a set of state variables, a set of algebraic variables, a set of model parameters, a stoichiometric matrix of state variables, and a set of reaction flux vectors and algebraic relations are generated. Based on the stoichiometric matrix of state variables and the reaction flux vector, ordinary differential equations corresponding to each state variable are generated. Using a set of algebraic relations, algebraic variables in the ordinary differential equations are replaced and eliminated to obtain the dimensionality-reduced mechanism model. Based on the dimensionality-reduced mechanistic model, the design optimization of a two-component biosensor is achieved.

[0008] Preferably, the structured configuration data includes molecular entity information, reaction type information, regulation direction information, input signal information, output signal information, and parameter constraint information.

[0009] Preferably, the method for constructing a reaction network graph based on structured configuration data of a two-component biosensor includes: Molecular entity nodes are generated based on structured configuration data. The molecular entity nodes include at least histidine kinase, response regulatory protein, phosphorylated histidine kinase, phosphorylated response regulatory protein, output promoter, reporter gene mRNA, immature reporter protein, and mature reporter protein. The reaction types between molecular entity nodes are obtained based on structured configuration data to construct reaction edges between molecular entity nodes; the edge attributes of the reaction edges include reaction type, reaction direction, reactants, products and parameter identifiers; A reaction network graph is constructed based on each molecular entity node and the reaction edges between each molecular entity node.

[0010] Preferably, the method for matching the corresponding modeling template based on the reaction type and time scale attributes of each reaction edge in the reaction network graph includes: When the reaction type of the reaction edge is transcription, translation or degradation, a first-order synthetic degradation template is matched; When the reaction type of the reaction edge is binding, dissociation, dimerization or complex formation, the characteristic time scale in the time scale attribute of the reaction edge is compared with the system reference time scale to determine whether the reaction edge is a fast equilibrium reaction. When the reaction edge is a fast equilibrium reaction and can generate a unique algebraic relation expression, the fast equilibrium algebraic template is matched; When the reaction edge is not a fast equilibrium reaction, or cannot generate a unique algebraic relation expression, match the mass action template; When the reaction type of the reaction edge belongs to the inducer-regulated histidine kinase autophosphorylation, the Hill function input gating template is matched; When the reaction type of the reaction edge belongs to promoter occupation, match the promoter occupation template; When the reaction type of the reaction edge belongs to reporter protein maturation, the reporter protein maturation template is matched.

[0011] Preferably, when the modeling template corresponding to the reaction edge is a first-order synthetic degradation template, the reaction flux expression corresponding to the reaction edge is: , When the modeling template corresponding to the reaction edge is the mass action template, the expression for the reaction flux corresponding to the reaction edge is: , When the modeling template corresponding to the reaction edge is a fast equilibrium algebraic template, the algebraic relation expression corresponding to the reaction edge is: , When the modeling template corresponding to the reaction edge is a Hill function input-gated template, the reaction flux expression corresponding to the reaction edge is: , When the modeling template corresponding to the reaction edge is a promoter-occupied template, the reaction flux expression corresponding to the reaction edge is: , When the modeling template corresponding to the reaction edge is a reporter protein maturation template, the reaction flux expression corresponding to the reaction edge is: , in, The reaction flux corresponding to the reaction edge. Let be the first-order reaction rate constant corresponding to the reaction edge. The reactants corresponding to the reaction edge, The first corresponding to the reaction edge One reactant, The mass effect reaction rate constant corresponding to the reaction edge. This represents the equilibrium constant corresponding to the fast equilibrium algebraic template. This represents the reactant index corresponding to the reaction edge. This represents the quantity of reactants corresponding to the reaction edge. Indicates concentration. The reactants corresponding to the reaction edge, Indicates the first reaction edge corresponding to the reaction edge The reaction order corresponding to each reactant. This indicates the concentration of dimers, complexes, or bound states formed after a rapid equilibrium reaction. , Indicates the input signal concentration as The autophosphorylation rate constant at that time, This represents the basic autophosphorylation rate constant under conditions of no input signal or low input signal. This represents the maximum autophosphorylation rate constant. Indicates the input signal concentration. Represents the Hill coefficient. This represents the concentration of the input signal at which the half-maximum input response is generated. This indicates the concentration of unphosphorylated histidine kinase. , Indicates promoter occupancy. This indicates the concentration of regulatory factors involved in promoter binding. This represents the promoter binding coordination coefficient. This indicates the concentration of the regulatory factor when the promoter reaches a half-occupancy state. Indicates the total promoter concentration. This represents the basic transcription rate constant when there are no activating regulators at the response edge. This represents the maximum transcription rate constant when the response-side promoter is fully occupied. This represents the reporter protein maturation rate constant. This indicates the concentration of immature reporter proteins.

[0012] Preferably, the reaction network graph is either a directed graph or a directed multigraph.

[0013] Preferably, the method for generating the set of state variables, the set of algebraic variables, the set of model parameters, the stoichiometric matrix of state variables, and the set of reaction flux vectors and algebraic relations based on the reaction network diagram and the reaction flux expression or algebraic relation expression corresponding to each reaction edge includes: Based on the reaction type and time scale attributes of each molecular entity node in the reaction network diagram, the molecular entity nodes are classified to obtain the set of state variables and the set of algebraic variables. The parameter identifiers in the reaction flux expressions or algebraic relation expressions corresponding to each reaction edge are summarized and duplicated to obtain the model parameter set. Based on the stoichiometric relationship between reactants and products in each reaction edge, a stoichiometric matrix of state variables is generated. Based on the order of the reaction edges in the reaction network diagram, the reaction flux expressions corresponding to each reaction edge are arranged to form a reaction flux vector; the algebraic relation expressions are summarized to form an algebraic relation set.

[0014] Preferably, the method for generating ordinary differential equations corresponding to each state variable based on the stoichiometric matrix of state variables and the reaction flux vector, replacing and eliminating algebraic variables in the ordinary differential equations, and combining the set of algebraic relations to obtain the dimensionality-reduced mechanistic model includes: The ordinary differential equations corresponding to each state variable are: , in, This represents a state variable vector composed of various state variables. This represents a vector of algebraic variables composed of various algebraic variables. Indicates time, Represents the stoichiometry matrix of state variables. Represents the reaction flux vector. Represents the input variable vector. Represents the model parameter vector; Substituting the algebraic expressions in the set of algebraic relations into the ordinary differential equations corresponding to each state variable, and replacing and eliminating algebraic variables in the system of ordinary differential equations, yields the dimensionality-reduced mechanism model.

[0015] Preferably, after obtaining the dimensionality-reduced mechanism model, a consistency check is performed on the dimensionality-reduced mechanism model. If the check fails, an error message or correction suggestion is generated.

[0016] This invention also provides a design optimization system for a two-component biosensor, comprising: A reaction network graph construction module is used to construct a reaction network graph based on the structured configuration data of a two-component biosensor; the reaction network graph includes multiple molecular entity nodes and reaction edges between different molecular entity nodes; The template rule base matching module is used to match the modeling template corresponding to each reaction edge based on the reaction type and time scale attributes of each reaction edge in the reaction network graph, and generate the reaction flux expression or algebraic relation expression corresponding to each reaction edge. The model element generation module is used to generate ordinary differential equations corresponding to each state variable based on the stoichiometric matrix of state variables and the reaction flux vector. It then uses a set of algebraic relations to replace and eliminate algebraic variables in the ordinary differential equations to obtain the dimensionality-reduced mechanism model. The equation generation and dimensionality reduction module is used to generate a set of ordinary differential equations consisting of ordinary differential equations corresponding to each molecular entity node based on the stoichiometric matrix of state variables and the reaction flux vector. It also performs algebraic dimensionality reduction processing on the ordinary differential equations corresponding to the molecular entity nodes that are algebraic variables in the reaction network diagram to obtain the dimensionality-reduced mechanism model. The design optimization module is used to optimize the design of two-component biosensors based on the dimensionality-reduced mechanistic model.

[0017] Compared with the prior art, the above-described technical solution of the present invention has the following advantages:

[0018] This invention discloses a design optimization method and system for a two-component biosensor. By constructing a reaction network diagram through structured configuration data standardization, it unifies the reaction data structure of all two-component systems, breaking down barriers between different sensor models. This enables the model architecture to be reusable and transferable across systems, eliminating the need to repeatedly build modeling frameworks for new sensors. Furthermore, based on the reaction type of each reaction edge, it automatically matches a dedicated modeling template, standardizing the generation of reaction flux expressions or algebraic relation expressions corresponding to each reaction edge. This completely eliminates human modeling errors such as variable omissions, parameter duplication, and inconsistent equation dimensions, ensuring the accuracy and reliability of the mechanistic model from the source. Simultaneously, it intelligently divides the set of state variables into sets of algebraic variables, abandoning the traditional modeling method of all state variables. Algebraic expressions replace differential equations, completing model elimination and dimensionality reduction. While fully preserving the bioreaction kinetics and without sacrificing detection accuracy, it significantly reduces redundant dimensions of the model, effectively lowering the computational burden of model simulation. Finally, based on the standardized state variable stoichiometric matrix and flux vector, the equation system is constructed and intelligent dimensionality reduction is achieved, outputting a high-precision, low-dimensional, and universal mechanism model. This eliminates the need for repeated manual debugging and correction, greatly shortening the modeling and R&D cycle of novel two-component biosensors, significantly reducing experimental R&D costs, and improving the detection accuracy and design optimization effect of two-component biosensors from the bottom layer of the model. Attached Figure Description

[0019] To make the content of this invention easier to understand, the invention will be further described in detail below with reference to specific embodiments and accompanying drawings, wherein:

[0020] Figure 1 This is a flowchart illustrating the design optimization method for a two-component biosensor according to the present invention.

[0021] Figure 2 This is a structural diagram of a design optimization system for a two-component biosensor according to the present invention. Detailed Implementation

[0022] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, so that those skilled in the art can better understand and implement the present invention. However, the embodiments described are not intended to limit the present invention.

[0023] Reference Figure 1 As shown, this embodiment provides a design optimization method for a two-component biosensor, including: Step S1: Construct a reaction network diagram based on the structured configuration data of the two-component biosensor; In this embodiment, specifically, structured configuration data Includes: molecular entity information Reaction type information Information on the direction of regulation Input signal information Output signal information Parameter constraint information Time scale information .

[0024] In this embodiment, specifically, the reaction network graph can be either a directed graph or a directed multigraph.

[0025] Step S11: Obtain structured configuration data for the two-component biosensor, including: Acquire molecular entity information: Acquire basic information of the two-component biosensor input by the user or imported from the file. The basic information includes histidine kinase NarX, response regulatory protein NarL, inducer, output promoter PyeaR, and reporter gene sfGFP. Obtain information on the direction of regulation: Obtain the direction of regulation of histidine kinase autophosphorylation, dephosphorylation or phosphate group transfer by inducers; Obtain output signal information: Obtain the activation or inhibition relationship between response regulatory proteins or phosphorylated response regulatory proteins and the output promoter; Obtain parameter constraint information: Obtain parameter constraint information corresponding to each reaction process. The parameter constraint information includes parameter name, parameter initial value, parameter upper and lower bounds, and parameter unit. Acquire timescale information: Acquire the fast equilibrium marker, characteristic timescale, reaction rate parameters, system reference timescale, and preset timescale ratio threshold corresponding to binding, dissociation, dimerization, complex formation, or other candidate fast reactions; The characteristic timescale is calculated based on the reaction rate parameters corresponding to the reaction edges.

[0026] Information on molecular entities, reaction types, regulatory directions, input signals, output signals, parameter constraints, and time scales is encapsulated into structured configuration data.

[0027] The structured configuration data is stored in at least one of the following formats: tabular files, JSON files, XML files, graphical configuration files, or database records.

[0028] Output signals include fluorescence intensity, luminescence intensity, colorimetric signals, or electrochemical signals.

[0029] Step S12: Generate molecular entity nodes based on structured configuration data. The molecular entity nodes include at least histidine kinase, response regulatory protein, phosphorylated histidine kinase, phosphorylated response regulatory protein, output promoter, reporter gene mRNA, immature reporter protein, and mature reporter protein.

[0030] The node attributes of the molecular entity node include at least the node name and node type; the node type is used to characterize the biological category of the molecular entity and includes at least one of the following: input signal, histidine kinase, response regulatory protein, phosphorylated protein, promoter, mRNA, immature reporter protein, mature reporter protein, dimer, complex, and promoter-bound state.

[0031] Step S13: Obtain the reaction types between molecular entity nodes based on structured configuration data to construct reaction edges between molecular entity nodes; the edge attributes of the reaction edges include reaction type, reaction direction, reactants, products, parameter identifiers, and time scale attributes; the time scale attributes include at least the characteristic time scale. The reaction types include: transcription, translation, degradation, binding, dissociation, phosphorylation, dephosphorylation, promoter occupation, or reporter protein maturation reactions.

[0032] Step S14: Construct a reaction network graph based on each molecular entity node and the reaction edges between each molecular entity node. ,in, Represents the set of molecular entity nodes. This represents the set of reaction edges.

[0033] In this embodiment, preferably, after the reaction network graph is constructed, isolated nodes, duplicate edges and missing reactants in the reaction network graph are checked.

[0034] After constructing the reaction network graph, a structural consistency check is performed on the reaction network graph. The structural consistency check includes at least one of the following: isolated node check, duplicate edge check, reactant and product integrity check, and edge attribute integrity check.

[0035] For any anomalies detected, anomaly correction processing is performed based on the structured configuration data and preset verification rules, specifically including: For isolated nodes, determine whether the isolated node is marked as an input signal node, an output signal associated node, or a user-specified reserved node; if so, the isolated node is reserved; if not, supplement the corresponding reaction associated edges according to the structured configuration data, or mark the molecular entity nodes without valid reaction relationships as invalid nodes and remove them. For repeated edges, compare the reactants, products, reaction type, reaction direction, modeling template type, and parameter identifiers of each reaction edge; when all edge attributes are the same, the corresponding reaction edge is identified as a repeated edge and merged; when at least one edge attribute is different, the corresponding reaction edge is retained as a different reaction edge in the directed multigraph. In cases where reactants, products, or reaction relationships are missing, the structured configuration data is backtracked; when the missing content can be uniquely identified based on the structured configuration data, molecular entity nodes, reaction relationship edges, or edge attributes are added; when the missing content cannot be uniquely identified, an anomaly message or correction suggestion is generated. After completing the anomaly correction process, the reaction network diagram is checked for structural consistency again. If the check passes, the validated reaction network diagram is output. If the check fails, the anomalies that failed the check and the corresponding correction suggestions are output.

[0036] Step S2: Based on the reaction type of each reaction edge in the reaction network diagram, match the corresponding modeling template to generate the reaction flux expression or algebraic relation expression for each reaction edge; Specifically, the reaction type, reactant node, product node, control direction, parameter identifier, and time scale attribute of each reaction edge are read. For reaction edges that can directly determine the modeling template based on the reaction type, the corresponding modeling template is matched based on the reaction type. For reaction edges that can be matched with the mass action template or the fast equilibrium algebra template, the fast equilibrium condition is first determined, and then the corresponding modeling template is matched based on the determination result.

[0037] In this embodiment, preferably, when the reaction type of the reaction edge belongs to transcription, translation or degradation, a first-order synthetic degradation template is matched; When the modeling template corresponding to the reaction edge is a first-order synthesis-degradation template, the reaction flux expression corresponding to the reaction edge is: , in, The reaction flux corresponding to the reaction edge Let be the first-order reaction rate constant corresponding to the reaction edge. The reactants corresponding to the reaction edge, Indicates concentration.

[0038] For example, when the reaction type of a reaction edge is degradation, the modeling template corresponding to the reaction edge is a first-order synthetic degradation template, and the reaction flux expression corresponding to this reaction edge is: , in, The reaction type of the reaction edge belongs to the reaction flux during degradation. This identifies the degradation parameters of the reaction edge.

[0039] When the reaction type of the reaction edge is binding, dissociation, dimerization or complex formation, the characteristic time scale of the reaction edge is compared with the system reference time scale to determine whether the reaction edge is a fast equilibrium reaction. when If the reaction edge is determined to be a fast equilibrium reaction, then the reaction edge is determined to be a fast equilibrium reaction; otherwise, the reaction edge is not a fast equilibrium reaction. The characteristic time scale of the reaction edge. As a reference time scale for the system, The preset time scale ratio threshold; For a first-order reaction, the characteristic timescale is expressed as: , in, Indicates the first For each first-order reaction, the reaction rate constant is used; for non-first-order reactions, the local characteristic time scale is calculated based on the kinetic expression and parameter values ​​of the corresponding reaction, or the characteristic time scale is read from the pre-set characteristic time scale in the structured configuration data.

[0040] When the reaction edge is a fast equilibrium reaction and can generate a unique algebraic relation expression, the fast equilibrium algebraic template is matched; In this embodiment, after determining whether the completed reaction edge is a fast equilibrium reaction, the determination result is used as the edge attribute of the reaction edge.

[0041] When the modeling template corresponding to the reaction edge is a fast equilibrium algebraic template, the algebraic relation expression corresponding to the reaction edge is: , in, This indicates the concentration of dimers, complexes, or bound states formed after a rapid equilibrium reaction. This represents the equilibrium constant corresponding to the fast equilibrium algebraic template. This represents the quantity of reactants corresponding to the reaction edge. This represents the reactant index corresponding to the reaction edge. Indicates the first reaction edge corresponding to the reaction edge The reaction order corresponding to each reactant.

[0042] When the reaction edge is not a fast equilibrium reaction, or cannot generate a unique algebraic relation expression, match the mass action template; When the modeling template corresponding to the reaction edge is the mass action template, the expression for the reaction flux corresponding to the reaction edge is: , in, is the mass-effect reaction rate constant corresponding to the reaction edge.

[0043] For example: when the reaction type of the reaction edge is a binding reaction. A represents the first reactant corresponding to the reaction edge. This indicates the second reactant corresponding to the reaction edge. This indicates the product corresponding to the reaction edge.

[0044] Using the mass action template, the expression for the forward reaction flux corresponding to this reaction edge is: , This represents the forward reaction flux corresponding to the reaction edge. This is the identifier for the binding parameters of the reaction edge.

[0045] When the reaction type of the reaction edge belongs to the inducer-regulated histidine kinase autophosphorylation, the Hill function input gating template is matched; When the modeling template corresponding to the reaction edge is a Hill function input-gated template, the reaction flux expression corresponding to the reaction edge is: , In the formula, , This indicates the concentration of unphosphorylated histidine kinase. Indicates the input signal concentration as The autophosphorylation rate constant at that time, This represents the basic autophosphorylation rate constant under conditions of no input signal or low input signal. This represents the maximum autophosphorylation rate constant. Indicates the input signal concentration. Represents the Hill coefficient. This represents the input signal concentration at which the half-maximum input response is generated. The input signal concentration.

[0046] When the reaction type of the reaction edge belongs to promoter occupation, match the promoter occupation template; When the modeling template corresponding to the reaction edge is a promoter-occupied template, the reaction flux expression corresponding to the reaction edge is: , In the formula, Indicates the total promoter concentration. This represents the basic transcription rate constant when there are no activating regulators at the response edge. This represents the maximum transcription rate constant when the response-side promoter is fully occupied. , Indicates promoter occupancy. This indicates the concentration of regulatory factors involved in promoter binding. This represents the promoter binding coordination coefficient. This indicates the concentration of the regulatory factor when the promoter reaches a half-occupancy state.

[0047] When the reaction type of the reaction edge belongs to reporter protein maturation, the reporter protein maturation template is matched; When the modeling template corresponding to the reaction edge is a reporter protein maturation template, the reaction flux expression corresponding to the reaction edge is: , in, This represents the reporter protein maturation rate constant. This indicates the concentration of immature reporter proteins.

[0048] Based on the matched modeling template, reactant nodes, product nodes, control directions, and parameter identifiers, dynamic reaction flux expressions or algebraic relation expressions corresponding to each reaction edge are automatically generated, and these expressions are bound to the corresponding reaction edges. According to the order of the dynamic reaction edges, the dynamic reaction flux expressions are arranged to form a reaction flux vector; the algebraic relation expressions corresponding to rapid equilibrium reactions, dimerization equilibrium, complex formation equilibrium, and promoter occupation processes are summarized to form an algebraic relation set. The reaction flux vector is used to generate ordinary differential equations corresponding to state variables, and the algebraic relation set is used to generate algebraic expressions corresponding to algebraic variables and perform subsequent model dimensionality reduction.

[0049] Step S3: Based on the reaction network diagram and the reaction flux expression or algebraic relation expression corresponding to each reaction edge, generate the set of state variables, the set of algebraic variables, the set of model parameters, the stoichiometric matrix of state variables, and the reaction flux vector and algebraic relation set; Step S31: Based on the reaction type and time scale attributes of each molecular entity node in the reaction network diagram, classify the molecular entity nodes to obtain the set of state variables. } and set of algebraic variables , Indicates the first One state variable, The total number of state variables. Indicates the first algebraic variables, The total number of algebraic variables; In this embodiment, the classification method can optionally treat each molecular entity node in the reaction edge that is a fast equilibrium reaction as an algebraic variable, and the molecular entity nodes of the other reaction edges as state variables. For molecular entity nodes that are external inputs and whose concentration or intensity is directly given by user input, file import, or preset input function, they are classified as external input variables and are not included in the set of state variables or algebraic variables; when the molecular entity corresponding to the external input is generated, consumed, or dynamically changed in the reaction network, it is classified as a state variable. For observations used solely to convert the concentration of mature reporter proteins into fluorescent, luminescent, colorimetric, or electrochemical signals, they are classified as observation variables and not included in the set of state variables or algebraic variables. When the variable type label of a molecular entity node is inconsistent with its associated reaction type, time scale attribute, or algebraic relation expression, a variable classification anomaly prompt or correction suggestion is generated, and the variable type of the molecular entity node is re-determined according to the preset verification rules.

[0050] The set of state variables includes variables that need to be described dynamically by ordinary differential equations. When molecular entity nodes participate in transcription, translation, degradation, phosphorylation, dephosphorylation, phosphate group transfer, or reporter protein maturation reactions, and it is necessary to retain the transient process in which their concentration changes over time, the molecular entity nodes are classified as state variables. The set of algebraic variables includes dimer, complex, or promoter-bound states calculated by fast equilibrium relations or algebraic expressions.

[0051] Step S32: Summarize and remove duplicate parameter identifiers from the reaction flux expressions or algebraic relation expressions corresponding to each reaction edge to obtain the model parameter set; Step S33: Generate the stoichiometric matrix of state variables based on the stoichiometric relationships of reactants and products in each reaction edge: , in, This indicates that the b-th reaction pair is the first... Stoichiometric coefficients of each state variable Indicates the total number of state variables. Indicates the total number of reaction edges; when the first... The state variable in the th... When consumed as a reactant in a reaction zone It is a negative value; when the first The state variable in the th... When produced as a product in a reaction edge It is a positive value; when the first The state variable does not participate in the first... When there is a reaction edge, It is zero.

[0052] Step S34: According to the order of the reaction edges in the reaction network diagram, arrange the reaction flux expressions corresponding to each reaction edge to form a reaction flux vector; summarize the algebraic relation expressions to form an algebraic relation set.

[0053] The set of state variables, the set of algebraic variables, the set of model parameters, the stoichiometric matrix of state variables, the reaction flux vector, and the set of algebraic relations are linked to form a unified model element data structure. This model element data structure is used for generating the ordinary differential equation system of state variables, generating the expression of algebraic variables, model dimensionality reduction, model consistency verification, and model code output.

[0054] Step S4: Based on the stoichiometric matrix of state variables and the reaction flux vector, generate the ordinary differential equations corresponding to each state variable. Using the set of algebraic relations, replace and eliminate the algebraic variables in the ordinary differential equations to obtain the dimensionality-reduced mechanism model. The ordinary differential equations corresponding to each state variable are: , in, This represents a state variable vector composed of various state variables. This represents a vector of algebraic variables composed of various algebraic variables. Indicates time, Represents the stoichiometry matrix of state variables. Represents the reaction flux vector. Represents the input variable vector. Represents the model parameter vector; Each reaction flux in the reaction flux vector is calculated based on state variables, algebraic variables, input variables, and model parameters; when the corresponding reaction flux expression does not contain algebraic variables, the reaction flux is calculated based on state variables, input variables, and model parameters.

[0055] Substituting the algebraic expressions in the set of algebraic relations into the ordinary differential equations corresponding to each state variable, and replacing and eliminating algebraic variables in the system of ordinary differential equations, yields the dimensionality-reduced mechanism model.

[0056] For molecular entity nodes that are algebraic variables, the corresponding algebraic expression is generated based on their upstream variables, combined with the equilibrium constant or the occupancy function; the algebraic expression is generated based on the upstream variables of the algebraic variables, combined with the equilibrium constant, the dissociation equilibrium constant or the promoter occupancy function. No. The algebraic relation expression corresponding to each algebraic variable is expressed as: , in, Indicates the first algebraic variables, Indicates the first One state variable, Indicates the first An algebraic function of an algebraic variable with respect to the state variable, input variable, and model parameters. This represents the total number of algebraic variables.

[0057] Summarizing the algebraic relation expressions corresponding to each algebraic variable, we get: , in, This represents a state variable vector composed of various state variables. This represents a vector of algebraic variables composed of various algebraic variables. It represents the algebraic mapping relationship composed of the algebraic functions corresponding to each algebraic variable.

[0058] For example, when the dimerization process of phosphorylation-regulating proteins meets the rapid equilibrium condition, the phosphorylation-regulating protein dimer can be represented as: , Promoter occupancy can be expressed as: , in, This indicates that phosphorylation regulates the concentration of protein dimers. This indicates that phosphorylation regulates the concentration of protein monomers. This represents the dissociation equilibrium constant of the dimerization reaction. Indicates promoter occupancy. This indicates the concentration of regulatory factors involved in promoter binding. This represents the promoter binding coordination coefficient. This indicates the concentration of the regulatory factor when the promoter reaches a half-occupancy state.

[0059] When the algebraic relation expression cannot uniquely determine the value of the corresponding algebraic variable, or when it is necessary to retain the transient change process of the corresponding molecular entity node, the molecular entity node is not subjected to algebraic elimination, and the molecular entity node is reclassified as a state variable, or an algebraic relation anomaly prompt is generated.

[0060] By substituting the algebraic expressions from the set of algebraic relations into the ordinary differential equations corresponding to each state variable, and by replacing and eliminating algebraic variables in the system of ordinary differential equations, the dimensionality-reduced mechanistic model is obtained. , By substituting the algebraic relation expression, the algebraic variables appearing in the ordinary differential equations of the state variables are eliminated, so that the dimensionality-reduced mechanism model only contains the ordinary differential equations corresponding to the state variables.

[0061] When no algebraic dimensionality reduction is performed, and both state variables and algebraic variables are treated as independent dynamic variables, the number of ordinary differential equations can be expressed as follows: , After algebraic variable substitution and elimination, the number of ordinary differential equations in the dimension-reduced mechanistic model is expressed as follows: , The reduction in the number of ordinary differential equations caused by the algebraic dimensionality reduction process is expressed as follows: , in, Indicates the total number of state variables. Represents the total number of algebraic variables. This represents the number of ordinary differential equations without algebraic dimensionality reduction. This represents the number of ordinary differential equations in the dimensionality-reduced mechanistic model. This represents the difference in ordinary differential equations between the undimension-reduced and dimensionality-reduced mechanistic models.

[0062] Step S5: Based on the dimensionality-reduced mechanistic model, optimize the design of the two-component biosensor.

[0063] Based on the dimensionality-reduced mechanistic model, the adjustable design parameters such as inducer concentration, expression levels of various proteins, and kinetic rate constants within the structured configuration data are iterated and adjusted. The model is solved in batches to obtain the time-series output results of mature reporter protein concentration under different parameter combinations. The detection response amplitude, response duration, and steady-state signal-to-noise ratio are used as optimization evaluation indicators. Within the feasible range of parameter values, the optimal parameter combination that meets the preset detection performance threshold is iteratively screened to complete the structural and response parameter design optimization of the two-component biosensor system.

[0064] After obtaining the dimensionality-reduced mechanism model, a consistency check is performed on the dimensionality-reduced mechanism model. If the check fails, an error message or correction suggestion is generated.

[0065] The model consistency verification includes at least one of the following: variable integrity verification, parameter integrity verification, equation dimension verification, input-output mapping verification, and non-negativity constraint verification. The reduced-dimensionality mechanism model is subjected to a variable and parameter integrity check to determine whether there are any missing, duplicate, or undefined states, algebraic variables, model parameters, parameter units, parameter boundaries, and initial values. The structural consistency of the reduced mechanism model is checked to determine whether the number of ordinary differential equations is consistent with the number of state variables, and whether there are isolated nodes, broken reaction paths, input signals that cannot reach the output variables, or state variables and parameters that do not meet the non-negativity constraints in the reaction network graph. Equation dimension verification is represented as: , in: This represents the number of ordinary differential equations in the dimensionality-reduced mechanistic model; Represents a set of state variables; This indicates the number of state variables in the set of state variables; Chemometric matrix representing state variables number of rows; This represents the total number of state variables.

[0066] Nonnegativity constraints are expressed as: , , in: (0) indicates the first The initial concentration of each state variable, Indicates the total number of state variables. Indicates the first Each model parameter This represents the set of model parameter indices that need to satisfy the non-negativity condition.

[0067] Output the dimensionality-reduced mechanistic model as executable model code or machine-readable model file, which can be used for numerical simulation, parameter identification, response prediction, or sensor performance analysis.

[0068] The machine-readable model file can be at least one of JSON, XML, SBML, MATLAB, and Python files. Each machine-readable model file in any format stores model structure information, variable set, parameter set, reaction flux set, ordinary differential equation system, algebraic expression, input-output mapping relationship, and solver configuration information. The executable model code includes state variable definitions, parameter definitions, algebraic variable calculation functions, reaction flux calculation functions, right-hand side functions of ordinary differential equations, input-output mapping functions, and numerical solver calling interfaces; The executable model code or machine-readable model file is output to a local storage unit, a remote server, a database, or a simulation computing platform, and the model metadata is output synchronously. The model metadata includes model version, generation time, configuration data source, template matching record, and consistency verification result.

[0069] In this embodiment, the two-component biosensor includes an activation-type two-component biosensor and an inhibition-type two-component biosensor.

[0070] The method of this invention can convert the structural information of a two-component biosensor into a solvable mechanistic model, reducing the workload of manually deriving equations and writing model code, and improving the consistency and reusability of model generation.

[0071] Based on Example 1, this Example 2 uses the NarX-NarL nitrate two-component biosensor as an example to illustrate the method of the present invention.

[0072] The structured configuration data includes histidine kinase NarX, response regulator NarL, input signal nitrate, output promoter PyeaR, and reporter gene sfGFP.

[0073] The generated set of state variables includes:

[0074] in, It is narX mRNA. It is the NarX protein. It is narL mRNA. It is the NarL protein. For phosphorylated NarX, For phosphorylated NarL, It is sfGFP mRNA. It is an immature sfGFP. It is a mature sfGFP.

[0075] In this embodiment, the phosphorylated NarL dimer and PyeaR promoter occupancy rates are used as algebraic variables, and the resulting set of algebraic variables includes: , in, This indicates the concentration of phosphorylated NarL dimer. This indicates the PyeaR promoter occupancy rate.

[0076] When the phosphorylated NarL dimerization process meets the rapid equilibrium condition, the concentration of phosphorylated NarL dimer is... Represented as: , in, This indicates the concentration of phosphorylated NarL monomer. This represents the dissociation equilibrium constant of the phosphorylated NarL dimerization reaction.

[0077] PyeaR promoter occupancy is expressed as: , in, This represents the promoter binding coordination coefficient. This represents the concentration of phosphorylated NarL dimer when the PyeaR promoter reaches a half-occupancy state.

[0078] Substituting the above algebraic relational expressions into the reaction flux expressions and ordinary differential equations corresponding to the state variables, and eliminating the independent dynamic equations corresponding to phosphorylated NarL dimers and promoter occupancy, we obtain the dimension-reduced mechanistic model.

[0079] Without algebraic dimensionality reduction and treating both algebraic variables as independent dynamic variables in the model, the unreduced mechanistic model consists of 11 ordinary differential equations. After algebraic variable substitution and elimination, the dimensionality-reduced mechanistic model consists of 9 ordinary differential equations, namely: , , , in, The number of ordinary differential equations in the undimension-reduced mechanistic model. This represents the number of ordinary differential equations in the dimensionality-reduced mechanistic model. This represents the difference in ordinary differential equations between the undimension-reduced and dimensionality-reduced mechanistic models.

[0080] Nitrate regulates the autophosphorylation rate of NarX Represented as: , in, This represents the nitrate concentration.

[0081] The automatically generated partial ordinary differential equations are as follows: , , , , in, Indicates time; This indicates the concentration of mRNA encoding histidine kinase NarX; The parameter representing the transcription rate of the narX gene; This represents the effective transcriptional activity coefficient of the promoter corresponding to the narX gene; This represents the first-order degradation rate constant of narX mRNA; This indicates the concentration of NarL, a protein that regulates phosphorylation. This represents the reaction flux from phosphorylated histidine kinase NarX to the reaction regulator protein NarL; This indicates the reaction flux during the dephosphorylation of the phosphorylation regulator protein NarL; This represents the first-order degradation rate constant of the phosphorylation regulator protein NarL; This indicates the mRNA concentration of the reporter gene sfGFP; This parameter represents the maximum transcriptional rate driven by the PyeaR promoter to transcribe the reporter gene sfGFP. Indicates the PyeaR promoter occupancy rate; This represents the first-order degradation rate constant of sfGFP mRNA; This indicates the concentration of the immature reporter protein sfGFP; This represents the translation rate constant for the conversion of sfGFP mRNA into immature sfGFP. This represents the maturation rate constant for the transformation of immature sfGFP into mature sfGFP; This represents the first-order degradation rate constant of immature sfGFP; This indicates the concentration of the mature reporter protein sfGFP; This represents the first-order degradation rate constant of mature sfGFP.

[0082] The output model code or machine-readable model file is represented by the following output signal:

[0083] in, For fluorescence output signal, This is the scaling factor. This is a background signal.

[0084] Based on the dimension reduction mechanism model of the NarX-NarL nitrate two-component biosensor that has passed consistency verification, the design of the sensor is optimized.

[0085] In this embodiment, at least one of NarX expression level, NarL expression level, PyeaR promoter transcription intensity, and sfGFP translation efficiency is set as an adjustable design parameter, and corresponding feasible value ranges are set for each.

[0086] Based on a preset parameter search strategy, multiple candidate design parameter combinations are generated within the feasible value range of each design parameter. For each candidate design parameter combination, the dimensionality reduction mechanism model is numerically solved at multiple preset nitrate concentrations to obtain the simulation trajectory of each state variable changing over time. The values ​​of phosphorylated NarL dimer concentration and PyeaR promoter occupancy rate at the corresponding time points are calculated based on algebraic relational expressions.

[0087] Based on the mapping relationship between the concentration of mature sfGFP and the fluorescence output signal, the dynamic fluorescence response and steady-state fluorescence response under different nitrate concentrations were obtained, and at least one of the following sensor performance evaluation indicators was calculated: response amplitude, response time, detection threshold, dynamic detection range, and sensitivity.

[0088] Based on the preset optimization objectives and constraints, the sensor performance evaluation indicators corresponding to each candidate design parameter combination are compared and screened to obtain the candidate design parameter combination with the optimal objective function value or that meets the preset performance threshold.

[0089] The NarX expression level, NarL expression level, PyeaR promoter transcription intensity, and sfGFP translation efficiency obtained from the screening were matched with the parameter calibration values ​​corresponding to the candidate promoter, ribosome binding site, or gene copy number, respectively, to determine the component configuration scheme and reaction parameter configuration scheme of the NarX-NarL nitrate two-component biosensor.

[0090] refer to Figure 2 This embodiment three provides a design optimization system for a two-component biosensor, including: A reaction network graph construction module is used to construct a reaction network graph based on the structured configuration data of a two-component biosensor; the reaction network graph includes multiple molecular entity nodes and reaction edges between different molecular entity nodes; The template rule base matching module is used to match the modeling template corresponding to each reaction edge based on the reaction type and time scale attributes of each reaction edge in the reaction network graph, and generate the reaction flux expression or algebraic relation expression corresponding to each reaction edge. The model element generation module is used to generate ordinary differential equations corresponding to each state variable based on the stoichiometric matrix of state variables and the reaction flux vector. It then uses a set of algebraic relations to replace and eliminate algebraic variables in the ordinary differential equations to obtain the dimensionality-reduced mechanism model. The equation generation and dimensionality reduction module is used to generate a set of ordinary differential equations consisting of ordinary differential equations corresponding to each molecular entity node based on the stoichiometric matrix of state variables and the reaction flux vector. It also performs algebraic dimensionality reduction processing on the ordinary differential equations corresponding to the molecular entity nodes that are algebraic variables in the reaction network diagram to obtain the dimensionality-reduced mechanism model. The consistency verification module is used to perform consistency verification on the dimensionality-reduced mechanism model; The design optimization module is used to optimize the design of two-component biosensors based on the dimensionality-reduced mechanistic model.

[0091] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0092] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0093] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0094] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0095] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A method for designing and optimizing a two-component biosensor, characterized in that, include: A reaction network graph is constructed based on structured configuration data of a two-component biosensor; the reaction network graph includes multiple molecular entity nodes and reaction edges between different molecular entity nodes; Based on the reaction type and time scale attributes of each reaction edge in the reaction network graph, the modeling template corresponding to each reaction edge is matched to generate the reaction flux expression or algebraic relation expression corresponding to each reaction edge. Based on the reaction network diagram and the reaction flux expression or algebraic relation expression corresponding to each reaction edge, a set of state variables, a set of algebraic variables, a set of model parameters, a stoichiometric matrix of state variables, and a set of reaction flux vectors and algebraic relations are generated. Based on the stoichiometric matrix of state variables and the reaction flux vector, ordinary differential equations corresponding to each state variable are generated. Using a set of algebraic relations, algebraic variables in the ordinary differential equations are replaced and eliminated to obtain the dimensionality-reduced mechanism model. Based on the dimensionality-reduced mechanistic model, the design optimization of a two-component biosensor is achieved.

2. The design and optimization method for a two-component biosensor according to claim 1, characterized in that, Structured configuration data includes information on molecular entities, reaction types, regulatory directions, input signals, output signals, and parameter constraints.

3. The design optimization method for a two-component biosensor according to claim 1, characterized in that, Methods for constructing response network graphs based on structured configuration data from two-component biosensors include: Molecular entity nodes are generated based on structured configuration data. The molecular entity nodes include at least histidine kinase, response regulatory protein, phosphorylated histidine kinase, phosphorylated response regulatory protein, output promoter, reporter gene mRNA, immature reporter protein, and mature reporter protein. The reaction types between molecular entity nodes are obtained based on structured configuration data to construct reaction edges between molecular entity nodes; the edge attributes of the reaction edges include reaction type, reaction direction, reactants, products and parameter identifiers; A reaction network graph is constructed based on each molecular entity node and the reaction edges between each molecular entity node.

4. The design optimization method for a two-component biosensor according to claim 3, characterized in that, Based on the reaction type and time scale attributes of each reaction edge in the reaction network graph, methods for matching corresponding modeling templates include: When the reaction type of the reaction edge is transcription, translation or degradation, a first-order synthetic degradation template is matched; When the reaction type of the reaction edge is binding, dissociation, dimerization or complex formation, the characteristic time scale in the time scale attribute of the reaction edge is compared with the system reference time scale to determine whether the reaction edge is a fast equilibrium reaction. When the reaction edge is a fast equilibrium reaction and can generate a unique algebraic relation expression, the fast equilibrium algebraic template is matched; When the reaction edge is not a fast equilibrium reaction, or cannot generate a unique algebraic relation expression, match the mass action template; When the reaction type of the reaction edge belongs to the inducer-regulated histidine kinase autophosphorylation, the Hill function input gating template is matched; When the reaction type of the reaction edge belongs to promoter occupation, match the promoter occupation template; When the reaction type of the reaction edge belongs to reporter protein maturation, the reporter protein maturation template is matched.

5. The design optimization method for a two-component biosensor according to claim 4, characterized in that, When the modeling template corresponding to the reaction edge is a first-order synthesis-degradation template, the reaction flux expression corresponding to the reaction edge is: ， When the modeling template corresponding to the reaction edge is the mass action template, the expression for the reaction flux corresponding to the reaction edge is: ， When the modeling template corresponding to the reaction edge is a fast equilibrium algebraic template, the algebraic relation expression corresponding to the reaction edge is: ， When the modeling template corresponding to the reaction edge is a Hill function input-gated template, the reaction flux expression corresponding to the reaction edge is: ， When the modeling template corresponding to the reaction edge is a promoter-occupied template, the reaction flux expression corresponding to the reaction edge is: ， When the modeling template corresponding to the reaction edge is a reporter protein maturation template, the reaction flux expression corresponding to the reaction edge is: ， in, The reaction flux corresponding to the reaction edge. Let be the first-order reaction rate constant corresponding to the reaction edge. The reactants corresponding to the reaction edge, The first corresponding to the reaction edge One reactant, The mass effect reaction rate constant corresponding to the reaction edge. This represents the equilibrium constant corresponding to the fast equilibrium algebraic template. This represents the reactant index corresponding to the reaction edge. This represents the quantity of reactants corresponding to the reaction edge. Indicates concentration. The reactants corresponding to the reaction edge, Indicates the first reaction edge corresponding to the reaction edge The reaction order corresponding to each reactant. This indicates the concentration of dimers, complexes, or bound states formed after a rapid equilibrium reaction. , Indicates the input signal concentration as The autophosphorylation rate constant at that time, This represents the basic autophosphorylation rate constant under conditions of no input signal or low input signal. This represents the maximum autophosphorylation rate constant. Indicates the input signal concentration. Represents the Hill coefficient. This represents the concentration of the input signal at which the half-maximum input response is generated. This indicates the concentration of unphosphorylated histidine kinase. , Indicates promoter occupancy. This indicates the concentration of regulatory factors involved in promoter binding. This represents the promoter binding coordination coefficient. This indicates the concentration of the regulatory factor when the promoter reaches a half-occupancy state. Indicates the total promoter concentration. This represents the basic transcription rate constant when there are no activating regulators at the response edge. This represents the maximum transcription rate constant when the response-side promoter is fully occupied. This represents the reporter protein maturation rate constant. This indicates the concentration of immature reporter proteins.

6. The design optimization method for a two-component biosensor according to claim 1, characterized in that, The reaction network diagram can be either a directed graph or a directed multigraph.

7. The design optimization method for a two-component biosensor according to claim 1, characterized in that, Methods for generating sets of state variables, algebraic variables, model parameters, state variable stoichiometric matrices, reaction flux vectors, and algebraic relation sets based on reaction network diagrams and reaction flux expressions or algebraic relation expressions corresponding to each reaction edge include: Based on the reaction type and time scale attributes of each molecular entity node in the reaction network diagram, the molecular entity nodes are classified to obtain the set of state variables and the set of algebraic variables. The parameter identifiers in the reaction flux expressions or algebraic relation expressions corresponding to each reaction edge are summarized and duplicated to obtain the model parameter set. Based on the stoichiometric relationship between reactants and products in each reaction edge, a stoichiometric matrix of state variables is generated. Based on the order of the reaction edges in the reaction network diagram, the reaction flux expressions corresponding to each reaction edge are arranged to form a reaction flux vector; the algebraic relation expressions are summarized to form an algebraic relation set.

8. The design optimization method for a two-component biosensor according to claim 7, characterized in that, The method for generating ordinary differential equations (ODEs) for each state variable based on the stoichiometric matrix of state variables and the reaction flux vector, and then substituting and eliminating algebraic variables in the ODEs, combined with a set of algebraic relations, to obtain the dimensionality-reduced mechanistic model includes: The ordinary differential equations corresponding to each state variable are: ， in, This represents a state variable vector composed of various state variables. This represents a vector of algebraic variables composed of various algebraic variables. Indicates time, Represents the stoichiometry matrix of state variables. Represents the reaction flux vector. Represents the input variable vector. Represents the model parameter vector; Substituting the algebraic expressions in the set of algebraic relations into the ordinary differential equations corresponding to each state variable, and replacing and eliminating algebraic variables in the system of ordinary differential equations, yields the dimensionality-reduced mechanism model.

9. The design optimization method for a two-component biosensor according to claim 1, characterized in that, After obtaining the dimensionality-reduced mechanism model, a consistency check is performed on the dimensionality-reduced mechanism model. If the check fails, an error message or correction suggestion is generated.

10. A design optimization system for a two-component biosensor, characterized in that, include: A reaction network graph construction module is used to construct a reaction network graph based on the structured configuration data of a two-component biosensor; the reaction network graph includes multiple molecular entity nodes and reaction edges between different molecular entity nodes; The template rule base matching module is used to match the modeling template corresponding to each reaction edge based on the reaction type and time scale attributes of each reaction edge in the reaction network graph, and generate the reaction flux expression or algebraic relation expression corresponding to each reaction edge. The model element generation module is used to generate ordinary differential equations corresponding to each state variable based on the stoichiometric matrix of state variables and the reaction flux vector. It then uses a set of algebraic relations to replace and eliminate algebraic variables in the ordinary differential equations to obtain the dimensionality-reduced mechanism model. The equation generation and dimensionality reduction module is used to generate a set of ordinary differential equations consisting of ordinary differential equations corresponding to each molecular entity node based on the stoichiometric matrix of state variables and the reaction flux vector. It also performs algebraic dimensionality reduction processing on the ordinary differential equations corresponding to the molecular entity nodes that are algebraic variables in the reaction network diagram to obtain the dimensionality-reduced mechanism model. The design optimization module is used to optimize the design of two-component biosensors based on the dimensionality-reduced mechanistic model.

Images