A standardized and automatic mapping method for non-standard biological gene circuit description

Abstract

The application discloses a kind of standardization and automatic mapping methods for non-standard biological gene circuit description, the method first adopts adaptive coding sniffing and DOM tree semantic defense split cleaning, extracts standardization triple;Second, based on feature library, entity classification and naming normalization are carried out, and general auxiliary element is isolated using factory mode instantiation, to prevent coordinate overwriting;Subsequently, by directed graph traversal, operon set and trans-regulation network are constructed, and anonymous promoter is automatically generated for topological chain break, and virtual expression cassette is constructed by triggering ghost filling mechanism for isolated regulation node;Finally, a two-layer adaptive spacing algorithm is used to map one-dimensional coordinates and output high-definition gene circuit diagram compatible with SBOL Visual specification.The application significantly improves the fault tolerance reconstruction capability of non-standard data, effectively solves the problem of graph element adhesion and topological collapse, and realizes the standardization of complex gene design visual presentation.

Description

Technical Field

[0001] This invention relates to the fields of synthetic biology, bioinformatics, computer-aided design, and scientific visualization, and particularly to a standardization and automatic drawing method for non-standard biological gene circuit descriptions. More specifically, this invention provides a technical solution for biological gene circuit text or semi-structured data with complex sources, heterogeneous formats, and inconsistent descriptions. This solution can automatically perform text cleaning, component identification, relation extraction, naming normalization, topology construction, missing data completion, layout optimization, and graphical output, enabling the automatic generation of gene circuit diagrams compatible with SBOL Visual semantics. Background Technology

[0002] In recent years, synthetic biology has gradually evolved from an "experiment-driven" approach to an engineering paradigm that follows a closed loop of "design-build-test-learn." Gene circuits, as the core design objects in synthetic biological systems, are widely used in applications such as metabolic pathway regulation, environmental perception, cell behavior control, in vivo diagnostics, gene expression regulation, and intelligent therapy. With the increasing complexity of designs, researchers are increasingly relying on standardized structural expression methods and graphical visualization techniques to manage and communicate gene circuit designs.

[0003] Within the existing technological framework, SBOL (Synthetic Biology Open Language) and its accompanying SBOLVisual provide a crucial standard foundation for the expression, exchange, and visualization of synthetic biology design objects. SBOL focuses on machine-readable representations of design objects, while SBOL Visual emphasizes standardized visual representations at the primitive level. Based on these standards, standardized expression and consistent presentation of objects such as promoters, ribosome binding sites, coding sequences, terminators, and regulatory relationships can be achieved.

[0004] However, in practical scientific research and engineering applications, the gene circuit information to be processed is often not naturally present in the standardized SBOL format, but is widely scattered throughout the main text of papers, experimental records, web page descriptions, Q&A texts, electronic documents, program comments, database export results, design communication records, and manually written structural descriptions. These inputs typically exhibit the following characteristics:

[0005] 1. Heterogeneous input format: It may be plain text, HTML text, tagged web page fragments, mixed-encoded files, or semi-structured paragraphs copied and pasted;

[0006] 2. Inconsistent naming conventions: The same component may have a standard name, alias, natural language name, abbreviation, or mixed uppercase and lowercase forms at the same time;

[0007] 3. Incomplete structural relationships: Some inputs only provide "who activates whom" or "who inhibits whom", without providing complete representation boxes or manipulator information;

[0008] 4. Inconsistent description granularity: Some inputs are descriptions of component sequence, some are descriptions of control relationships, and some are a mixture of both;

[0009] 5. Enhanced visualization requirements: Complex lines are prone to problems such as overlapping elements, conflicting control arcs, narrow canvas, and chaotic layout when automatically drawing.

[0010] Therefore, although existing SBOL-related tools provide a good framework for expressing standardized design objects, existing technologies still have significant shortcomings in the preprocessing and conversion stage from non-standard inputs to standard-compatible visual outputs. Summary of the Invention

[0011] The purpose of this invention is to address the problem in existing technologies where non-standard biological gene circuit descriptions are difficult to directly convert into standardized structural expressions and standardized graphical outputs. This invention proposes an SBOLVisual-compatible structural normalization and automatic drawing method for non-standard biological gene circuit descriptions. This method can automatically process gene circuit description information with complex origins, heterogeneous formats, inconsistent naming, and incomplete relationships, converting it into circuit structure representations and visual graphical outputs that conform to SBOL Visual semantics. This improves the automation level of gene circuit expression and drawing, the completeness of structural restoration, and the rendering stability of complex circuit diagrams.

[0012] To achieve the above objectives, the present invention adopts the following technical solution:

[0013] In a first aspect, the present invention provides a standardized and automated mapping method for describing non-standard biological gene circuits, comprising the following steps:

[0014] S1. Obtain multi-source non-standard gene circuit description text, then perform semantic anti-fragmentation cleaning to extract standardized entity relation triplet sequences;

[0015] S2. Based on multi-dimensional feature rules, classify, identify, and normalize the entities in the entity relationship triplet sequence, and instantiate them as physical memory objects.

[0016] S3. Traverse the entity relation triple sequence, identify the sequential driving relationship and cross-domain control relationship between the physical memory objects based on the directed graph traversal algorithm, and dynamically construct a multi-scale manipulator set and an inverse control network composed of the physical memory objects; each manipulator in the multi-scale manipulator set is associated with a promoter and its downstream cascaded encoding sequence set, and the inverse control network contains a control arc with the encoding sequence as the starting node and the promoter as the target node;

[0017] S4. Standardize the expression box assembly of the manipulation subset: sequentially add promoter, ribosome binding site, coding sequence and terminator to generate assembled expression box; at the same time, based on the inverse control network, detect topology disconnection or isolated control nodes, and automatically trigger the replacement mechanism when detected, construct virtual expression box, and append the virtual expression box to the basic expression box sequence to obtain a complete rendering sequence;

[0018] S5. Using a two-layer adaptive spacing algorithm, all physical memory objects in the complete rendering sequence are mapped to a one-dimensional physical coordinate system. The rendering position of each biological element is solved. Then, based on the inverse regulation network, the corresponding regulation arc graphics are added according to the positions of the starting node and the target node of the regulation arc in the one-dimensional physical coordinate system. Finally, the drawing engine is driven to output a gene circuit map that conforms to the preset visual specifications.

[0019] Furthermore, the execution of semantic anti-split cleaning specifically includes:

[0020] Identify specific subscript labels representing biological nomenclature in multi-source non-standard gene circuit description text;

[0021] Without disrupting the overall text topology, the content enclosed by the specific subscript tag is converted into plain text format with a preset leading character;

[0022] The converted text is injected into the document object model parser to remove redundant rich text tags and extract the pure text that retains the core auxiliary semantics.

[0023] Based on a preset state machine cursor scanning algorithm, valid descriptive paragraphs are extracted, and continuous invisible whitespace characters are dynamically compressed to obtain standardized line-by-line descriptive text.

[0024] The standardized line-by-line description text is segmented line by line. Each line of text is cut into source nodes, relational verbs and target nodes according to a preset delimiter, thereby extracting the standardized entity relation triplet sequence.

[0025] Furthermore, the instantiation into a physical memory object specifically includes:

[0026] Assign initial visual attributes to the normalized entities;

[0027] When the instantiated entity is a general auxiliary element in the gene circuit, the instantiation factory pattern based on isolation scope is used to create the object;

[0028] The instantiation factory pattern generates a completely new data structure in memory each time it is called, and assigns a unique identifier containing a globally incrementing counter to the data structure so that general-purpose auxiliary elements of the same type have independent physical coordinates that are decoupled from each other in the addressing tree of the underlying drawing engine, thus preventing coordinate overwrite conflicts.

[0029] Furthermore, the construction of the multi-scale manipulation subset and the inverse control network specifically includes:

[0030] Initialize an empty submap table;

[0031] If the triple represents the physical relationship of the promoter-driven coding sequence, then the coding sequence is appended to the operator sequence governed by the promoter;

[0032] If the triple represents the extension relationship of the cascaded coded sequences, then the promoter to which the preceding coded sequence belongs is addressed in reverse in the currently constructed operator mapping table; if the addressing is successful, the following coded sequence is appended to the operator sequence governed by the promoter; if the addressing is unsuccessful, a topological break is determined, and an anonymous promoter is dynamically generated in memory as a virtual anchor point, and the preceding and following coded sequences are jointly governed by the anonymous promoter.

[0033] If a triple represents a non-physical connection relationship of a promoter that regulates a coding sequence, it is determined to be a cross-domain regulatory relationship and stored independently in the global inverse regulation list.

[0034] Furthermore, the automatic fill mechanism specifically includes:

[0035] Extract the set of names of the physical memory objects that have been incorporated into the basic representation box sequence;

[0036] Traverse the control arcs in the inverse control network, perform set difference operation, and verify whether the starting node and target node of the control arc exist in the name set;

[0037] If the check finds a missing free control node, then an isolated control node is determined to have occurred.

[0038] A virtual expression box is automatically constructed in memory for the free control node. The virtual expression box sequentially includes: a low-opacity placeholder promoter, an independent ribosome binding site, the free control node, and an independent terminator. The virtual expression box is appended to the end of the basic expression box sequence, and the manipulator boundary record is updated to form the complete rendering sequence.

[0039] Furthermore, the two-layer adaptive spacing algorithm includes:

[0040] Introduce a dual-layer spacing coordinate calculation engine;

[0041] Between adjacent primitives within the same manipulator, the physical offset coordinates are calculated using a preset internal compact spacing parameter;

[0042] At the physical boundaries of different operons, when the rendering cursor crosses the end of the previous operon and the beginning of the next operon, a step-wise external partition spacing parameter is forcibly triggered, and the external partition spacing parameter is greater than the internal compact spacing parameter, so as to decouple independent transcription units at the visual-physical level.

[0043] Secondly, the present invention provides a standardized and automated mapping system for describing non-standard biological gene circuits, comprising:

[0044] The data preprocessing and cleaning module is used to acquire multi-source non-standard gene circuit description text, perform semantic anti-fragmentation cleaning, and extract standardized entity relation triplet sequences.

[0045] The entity semantic recognition and instance isolation module is used to classify, identify and name normalize entities in the triple sequence based on multi-dimensional feature rules, and instantiate them as physical memory objects.

[0046] The topology network relationship extraction and dynamic clustering module is used to traverse the triple sequence, identify the sequential driving relationship and cross-domain regulation relationship between entities based on the directed graph traversal algorithm, and dynamically construct multi-scale manipulation subsets and inverse regulation networks.

[0047] The standard expression box assembly and fault-tolerant module is used to assemble the manipulation subset into standardized expression boxes and trigger an automatic fault-infilling mechanism to generate virtual expression boxes when a topology break or isolated control node is detected.

[0048] The standardized rendering and adaptive coordinate system module is used to map the assembled expression box and the virtual expression box to a one-dimensional physical coordinate system, solve the rendering position based on the adaptive spacing algorithm, and drive the drawing engine to output a gene circuit map that conforms to the preset visual specifications.

[0049] Thirdly, the present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the above-described method.

[0050] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the above-described method.

[0051] The core idea of ​​this invention is to first acquire biological gene circuit description data to be processed. This data can originate from natural language text, semi-structured paragraphs, web page fragments, experimental records, program output, or explanatory content with HTML tags. Then, preprocessing operations are performed on the input data, including encoding recognition, text cleaning, HTML tag removal, subscript information conversion, and extraction of target circuit description paragraphs, to form a unified and parsable intermediate text. Next, the biological element entities and their relationships in the intermediate text are identified, the sequential and regulatory relationships between elements are extracted, and heterogeneous nomenclature, aliases, abbreviations, and element names in natural language form are normalized to map them to standard element types compatible with SBOL Visual primitives.

[0052] After completing component identification and naming normalization, this invention further automatically constructs expression boxes and manipulator topologies based on component sequence relationships, and generates corresponding activation arcs or suppression arcs based on regulatory relationships. For cases where endpoints are missing, structures are incomplete, or regulatory relationships are isolated in the input, this invention generates placeholder expression boxes or anonymous nodes through an automatic completion mechanism to ensure the connectivity and renderability of the overall circuit topology. Simultaneously, to avoid coordinate overlap, layout conflicts, and graphic misalignment issues caused by multiple expression boxes sharing the same primitive instance, this invention introduces an object uniqueness processing mechanism to generate independent instances for ribosome binding sites, terminators, and other auxiliary primitives, thereby improving the stability and accuracy of the automatic drawing process.

[0053] During the graph generation stage, this invention performs layout optimization based on the circuit topology, number of operons, regulatory span, and total element length. It automatically inserts intervals between different operons and dynamically adjusts the canvas size according to the circuit complexity, ultimately outputting a gene circuit diagram compatible with SBOL Visual semantics. The output not only visually displays the structural composition and regulatory relationships within the gene circuit but also facilitates subsequent design verification, result presentation, engineering documentation, and further standardized data exchange.

[0054] Compared with the prior art, the present invention has at least the following beneficial effects:

[0055] Firstly, this invention can adapt to non-standard inputs from various sources and formats. By using adaptive encoding sniffing and semantic anti-split cleaning, it can retain biological subscript semantics while removing redundant tags such as HTML. Combined with state machine cursor truncation and dynamic whitespace compression, it can stably extract standardized triples from dirty data such as paper text, web page fragments, and experimental records, thereby improving the usability of the original descriptive data.

[0056] Secondly, this invention achieves automatic classification of promoters / CDS based on a binary decision tree of a multi-dimensional feature lexicon, and combines a regularized naming normalization algorithm to eliminate node redundancy and logical chain breaks caused by synonyms, thereby improving the accuracy of component identification and structure recovery.

[0057] Thirdly, this invention can solve the problems of line breakage and unrenderable conditions caused by incomplete input through the ghost fill mechanism.

[0058] Fourth, this invention addresses the coordinate overwriting and graphic overlap issues caused by multiple identical auxiliary elements (RBS, terminator) sharing a global object by adopting an instance isolation mechanism based on the factory pattern: each call generates an independent data structure and assigns a globally unique identifier, enabling the underlying drawing engine to assign independent physical coordinates to each instance, thus achieving object uniqueness; in addition, this invention achieves layout optimization through a two-layer adaptive spacing engine, improving the stability and readability of automatic drawing of complex gene circuit diagrams.

[0059] In summary, this invention can achieve automated closed-loop processing from non-standard description to SBOL Visual-compatible graphical output, and has strong practical value and promotional significance. Attached Figure Description

[0060] Figure 1 This is a diagram of the overall architecture of the present invention.

[0061] Figure 2 This is a data flow diagram of the present invention.

[0062] Figure 3 This is a flowchart illustrating the automatic clustering and virtual replacement mechanism for multi-scale topological networks in Embodiments 3 and 4 of the present invention.

[0063] Figure 4 This is a comparison diagram of the beneficial effects of factory pattern instantiation isolation and double-layer adaptive spacing technology in Embodiments 2 and 5 of the present invention. Detailed Implementation

[0064] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0065] In preferred implementation scenarios, the method provided by this invention can be executed using standard computer processing equipment (such as a computing terminal or cloud server containing a central processing unit, memory, and non-volatile storage media). The system underlyingly supports mainstream interpreted programming language runtime environments and integrates a Document Object Model (DOM) parsing engine and a vector drawing core library supporting the Synthetic Biology Open Language (SBOL) Visual standard.

[0066] like Figure 1 As shown, the method and system architecture provided by this invention include the following modules:

[0067] Data preprocessing and cleaning module (corresponding to Implementation Example 1): Responsible for handling multi-source heterogeneous input files (such as TXT / HTML), and initiating adaptive encoding sniffing to shield underlying platform differences. Under the premise of strictly protecting biologically specific naming semantics (such as subscript tag protection), it uses state machine truncation and DOM tree parsing techniques to dimensionality-reducing and cleaning dirty data carrying complex rich text tags into standardized linear plain text.

[0068] The entity semantic recognition and instance isolation module (corresponding to Implementation Example 2) is responsible for the underlying information extraction task, mapping the cleaned text into triples. Based on a preset feature word decision tree, it performs accurate classification and regularized naming normalization of biological primitives (such as promoters, coding sequences, etc.), and innovatively uses the Factory Pattern to instantiate high-frequency auxiliary components into independent physical memory objects with unique identifiers, fundamentally eliminating coordinate overwrite conflicts.

[0069] Topology network relationship extraction and dynamic clustering module (corresponding to Implementation Example 3): The core graph theory algorithm hub of this system, responsible for identifying physical sequential driving and serial extension relationships between entities, and dynamically growing a multi-scale operand set. When a topology addressing chain breaks, it automatically creates an "anonymous promoter" as an anchor point, and simultaneously extracts cross-domain inverse regulation relationships (such as inhibition / activation) into independent logical network layers.

[0070] Standard expression box assembly and fault-tolerant module (corresponding to Example 4): Following the central dogma of biology, it automatically traverses the set of operands and precisely assembles the necessary auxiliary components (such as ribosome binding sites (RBS) and terminators). For isolated control nodes caused by extremely incomplete inputs, the system automatically triggers the "Ghost Placeholder" mechanism to generate a virtual expression box with a valid physical placeholder out of thin air, achieving a self-consistent closed loop of topological logic.

[0071] Standardized rendering and adaptive coordinate system module (corresponding to Example 5): Responsible for mapping abstract primitive sequences to a one-dimensional physical coordinate system. Based on the pioneering dual-layer spacing (internal compactness and external decoupling) adaptive algorithm, it solves the optimal rendering position, combined with a dynamic canvas scaling mechanism to prevent primitive compression and deformation, and finally drives the underlying drawing engine to output high-definition gene circuit maps (such as SVG or PNG format) conforming to the SBOLVisual specification.

[0072] like Figure 2 As shown, in the entire data flow lifecycle, external non-standard inputs are first converted into standardized in-memory intermediate data structures (from cleaned triples to entity instances to manipulator mapping tables to sequence arrays with padding and coordinates), and finally output as visual graphic files through computational geometric transformations. This achieves a highly fault-tolerant, end-to-end automated mapping from "non-standard machine-unreadable language" to "standardized high-fidelity engineering drawings".

[0073] Example 1: Error-tolerant cleaning and multi-dimensional extraction algorithm for non-standard gene text

[0074] In practical scientific research and engineering communication, the gene circuit description files to be processed often originate from exported experimental records, copied and pasted from web pages, or document transfers between different operating systems, which are prone to character encoding errors and rich text tag contamination. To address this technical pain point, this embodiment discloses a fault-tolerant cleaning and multi-dimensional extraction method for non-standard gene text, including the following steps:

[0075] Step 1.1: Multi-encoding sniffing and adaptive polling reading

[0076] To address the potential UnicodeDecodeError exceptions triggered by documents from multiple platforms, the system has constructed an encoding probe pool containing various mainstream character sets (such as utf-8, utf-8-sig, gbk, gb2312, shift_jis, etc.). When reading input files, the system does not rely on a single static encoding but instead initiates an adaptive polling decoder. This decoder attempts the encoding specifications in the probe pool one by one according to a preset priority. Once lossless decoding is completed under a specific encoding, polling is immediately interrupted, and the byte stream is converted into a standard Unicode string. This fallback mechanism significantly improves the system's throughput and fault tolerance for cross-system legacy data.

[0077] Step 1.2: Semantic anti-splitting protection and DOM tree dimensionality reduction and cleaning

[0078] For semi-structured text containing HTML or similar markup language tags, directly using brute-force regular expressions to remove them can disrupt the critical semantics of gene nomenclature (e.g., incorrectly truncating or combining the promoter P_lac into Plac). Therefore, before constructing the DOM parse tree, the system prioritizes implementing a semantic anti-splitting protection mechanism:

[0079] Specifically, the system schedules a preprocessing regular expression engine to anchor specific subscript tags (such as...) within the global text domain. After capturing this feature structure, the system converts the content wrapped in the tag into plain text format with leading underscores (i.e., without disrupting the overall text topology) without compromising the overall text topology. Mapped to _lac). After semantic preservation transformation, the system injects the entire string into the Document Object Model (DOM) parser. The parser then traverses the entire node tree, stripping away all redundant formatting tags (such as...). , , <font>etc.), and finally extract the peer pure text that retains the semantics of key affiliated names.

[0080] Step 1.3: State machine cursor interception and dynamic whitespace character compression

[0081] The dimension-reduced plain text usually still contains a large amount of contextual redundant information. The system adopts a state machine-based cursor scanning algorithm to lock the core valid payload. The system scans the text stream line by line. When the cursor detects a preset starting anchor feature string (for example, taking "line structure description" as the line start), the capturing flag of the state machine flips to the active state, and starts to push subsequent lines into the valid buffer queue; when consecutive blank newlines are detected in the active state ( ) or the end of the file, the structure description block is regarded as ended, triggering the state machine to sleep and stop interception.

[0082] For each line of valid data in the buffer queue, the system further invokes a dynamic compression filtering pipeline: first, remove the invisible control characters at the beginning and end of the line, then use a regular operator (such as r'\s+') to forcibly collapse multiple consecutive spaces or tabs in the line caused by manual input habits into a single standard space, to ensure that the subsequent relation extraction engine can cut the triple data with precise word segmentation granularity.

[0083] Embodiment 2: Semantic recognition and instance isolation of multi-scale biological element objects

[0084] After completing text cleaning and obtaining the triple sequence set composed of "Node A - Relation Verb - Node B", the system needs to parse these non-standardized string nodes into standard objects carrying definite physical and visual attributes.

[0085] Step 2.1: Binary decision tree classification based on feature lexicon

[0086] The system performs rule feature-based classification judgment on the extracted graphic element names to distinguish whether they belong to a promoter or a conventional coding sequence / gene. In the specific implementation, the system converts the string copy of the node name to lowercase, and introduces a precompiled comprehensive feature matching library. The judgment conditions include but are not limited to:

[0087] 1. Character prefix matching judgment (for example, the string starts with the letter p or p_);

[0088] 2. Explicit Chinese identifier judgment (for example, the string contains the word "promoter" embedded therein);

[0089] 3. Classic synthetic biology motif library identification (traversal comparison of strong characteristic sequence libraries such as t7, lac, ara, bad, trc, sp6, etc.).

[0090] If any of the above conditions are met, the system determines that the node is a startup subtype object and pre-assigns it the corresponding visual attributes that conform to the SBOL Visual specification (such as a specific blue polyline primitive); if no conditions are met, the system will classify it as a CDS type object by default and give it a right-angle polygon appearance and italic label style.

[0091] Step 2.2: Regularized naming normalization processing

[0092] To prevent the same biological element from being mistakenly identified as different physical entities in subsequent graph-theoretic topology due to minor differences in writing habits (such as "P_lac", "Plac", "lac promoter", "Lac-promoter"), the system introduces a mandatory naming normalization dimensionality reduction algorithm.

[0093] For nodes identified as promoters, the system executes a serialization string cleaning pipeline: First, redundant descriptive suffixes or prefixes in the name are removed (such as "promoter", "prom", etc.); second, residual connectors or spaces are uniformly converted to standardized underscores ("_") using regular expressions ([\s-]+); finally, residual underscore markers at the edges are removed, and core identifiers are converted to all uppercase (e.g., forced output in P_LAC format). This operation constructs an absolutely unique primary key namespace, completely eliminating the risk of redundancy and logical chain breaks in primitive nodes caused by synonyms and heterogeneities.

[0094] Step 2.3: Factory instantiation isolation to prevent global memory conflicts

[0095] In the standardized graphical transformation of synthetic biology circuits, expression box sequences often require the frequent insertion of common auxiliary elements, such as ribosome binding sites (RBS) and terminators. In the existing rendering mechanisms of underlying plotting engines (such as dnaplotlib), if multiple expression boxes reference the same configuration dictionary or object instance, when the engine writes the calculated physical coordinates (start / end), the coordinate sequence of the later-rendered element will directly overwrite the coordinates of the earlier-rendered element, resulting in severe rendering ghosting and positional drift.

[0096] To overcome this technical limitation, this system innovatively adopts the Factory Pattern based on isolated scope. The system abandons globally unique static configuration objects, instead encapsulating independent constructors such as `get_u0()` and `get_t0()`. Each time these constructors are called, they not only allocate a completely new dictionary data structure in memory, but more importantly, they integrate a protected globally incrementing counter (such as `_rbs_counter` and `_term_counter`). By pushing the dynamic fingerprint of the counter into the object's unique identifier (the `Name` attribute, such as automatically generated `rbs_1`, `rbs_2`), from a macroscopic biological semantic perspective, they all belong to isomorphic RBSs. However, in the computer's underlying memory management and graphics engine addressing tree, they are independent physical entities with decoupled coordinates. This mechanism completely eliminates the memory pointer pollution and coordinate overwriting risks during the automatic assembly of large and complex gene circuits, laying a solid foundation for subsequent high-fault-tolerant topology assembly.

[0097] like Figure 4 As shown in the upper part, Embodiment 2 vividly and intuitively compares the significant improvement of the factory pattern isolation technology in resolving primitive visual conflicts. In traditional rendering methods, due to the system's global reference to the shared counter (such as <_rbs_counter_shared_counter>), when the same auxiliary element appears multiple times in the gene circuit (such as "RBS ->CDS->RBS"), it will cause underlying memory conflicts and coordinate overwriting. This causes the subsequently rendered elements to directly cover the physical coordinate range of the previous element (for example, both coordinate domains are incorrectly locked at "start: 100, end: 110"), ultimately resulting in severe graphic overlap. To overcome this technical pain point, Embodiment 2 of this invention innovatively adopts a factory pattern based on isolation scope for instantiation isolation. This approach generates an independent counter (such as <_rbs_counter_independent counter>) and a unique identifier (such as RBS_1 and RBS_2) for each element instance, enabling the underlying graphics engine to assign absolutely independent one-dimensional coordinate system addresses to different instances (e.g., assigning "start: 100, end: 110" and "start: 125, end: 135" respectively). This improvement in the underlying architecture fundamentally eliminates the hidden dangers of memory pointer pollution and coordinate conflicts, ensuring that each independent biological primitive can obtain a clear, unobstructed, and accurate visualization.

[0098] Example 3: Automatic Clustering and Cross-Domain Extraction of Complex Controlled Topology Networks

[0099] After isolating and initializing primitive instances, the core technical challenge of the system is how to reconstruct a multi-level topological network conforming to the biological central dogma from fragmented binary relationships (such as "A drives B", "C tandems D", "E inhibits F"). This embodiment discloses a dynamic growth model of transcription units (Operon) based on directed graph traversal and an algorithm for extracting trans-regulation relationships. Figure 3 As shown.

[0100] Step 3.1: A graph theory-based model of the dynamic growth of transcription units (Operon)

[0101] In synthetic biology, a standard transcription unit (operon) typically consists of a promoter and multiple downstream tandem genes (CDS). The system initializes a dynamic hash map (operon_map) in memory to manage the association between promoters and CDS. The system traverses the cleaned entity triplet sequence and executes the following branch decisions:

[0102] Physically driven relation clustering (Promoter → CDS):

[0103] When the system detects that the upstream node type is a promoter and the downstream node type is a CDS, it determines that it is a direct physical drive. The system uses the normalized name of the promoter as the hash key and appends the instance objects of the downstream CDS to the value list corresponding to the key in an array append manner. To prevent node redundancy caused by duplicate text descriptions, the system performs anti-duplicate verification (deduplication logic) before pushing, ensuring that the same CDS remains unique under the jurisdiction of a single promoter.

[0104] Serial extension relationship clustering (CDS → CDS):

[0105] When the system detects that both upstream and downstream nodes are CDS (Channel-Derivative) nodes, it determines that this is a tandem extension phenomenon of multi-gene co-transcription. Since the triplet lacks direct promoter information at this point, the system triggers a reverse addressing algorithm:

[0106] The system iterates through all constructed operator arrays in the current operon_map, searching for whether the upstream CDS already exists in the governing sequence of a certain promoter. If the addressing is successful, the downstream CDS is appended to the end of the governing array of that promoter, thus realizing the dynamic backward growth of the operator.

[0107] Exception Handling and Anonymous Anchor Creation: If reverse addressing fails (i.e., the upstream CDS has not yet been assigned to any known promoter), the system determines that there is a "forward addressing failure". To ensure the coherence of subsequent one-dimensional sequence assembly, the system dynamically generates an anonymous promoter in memory as a virtual anchor (e.g., named __anon_prom_for_XXX), and assigns the two concatenated CDS to the jurisdiction of this anonymous anchor. This mechanism effectively solves the "isolated fragment" problem that often occurs in non-standard input.

[0108] Step 3.2: Stripping and Cross-Domain Extraction of Trans-regulation Relationships

[0109] In addition to cis-linking based on the physical double helix backbone of DNA, gene circuits also include trans-regional regulatory logic mediated by protein macromolecules.

[0110] When the system detects that the upstream node is a CDS and the downstream node is a promoter, it determines that it is a logical regulatory signal (not a physical connection) across transcription units. At this time, the system no longer pushes it into the physical control array of operon_map, but instead separates it and stores it independently in the global regulatory relationship set (reg list).

[0111] During the decoupling process, the system dynamically assigns different sets of vector attributes to the control arc by semantically parsing the relational verbs (e.g., by matching feature words such as "suppress" or "activate" using regular expressions). For example, for the "suppress" relation, the system configures it with a flat-headed endpoint, red color (0.95, 0.3, 0.25), and a relatively high arc curvature parameter (arc_height = 25); for the "activate" relation, it configures an arrowhead endpoint, green color (0.38, 0.82, 0.32), and a relatively low arc curvature parameter. This process achieves perfect data decoupling between the physical layer and the logical layer in the topology.

[0112] Example 4: Standardized Expression Box Assembly and Isolated Component Replacement Mechanism

[0113] After the topological clustering in Example 3, the system has obtained logical block information, but the underlying vector plotting engine (such as dnaplotlib) requires a strict, biologically compliant one-dimensional linear sequence (Design List) as input.

[0114] Step 4.1: The standardized sequence weaving system following the central dogma initializes a global one-dimensional array (design) and a boundary cursor recorder (operon_boundaries). The system iterates through each transcription unit in the operon_map:

[0115] First, the starter object is pushed into the design array. If the starter is the "anonymous starter" generated in Example 3, the system instantiates it as a gray placeholder primitive with very low transparency (e.g., alpha = 0.2) to visually weaken its presence, but retains its space placeholder function as a physical anchor point.

[0116] Subsequently, all CDSs under the jurisdiction of this promoter are traversed. Before each CDS is pushed into the design array, the system forcibly calls the aforementioned instantiation factory function (get_u0()) to weave a physically isolated ribosome binding site (RBS) in front of the CDS.

[0117] After all CDS processing for the current transcription unit is complete, the system calls the factory function (get_t0()) again to weave a terminator at the end of the unit.

[0118] After completing the weaving of a single transcription unit, the system records the start and end indices of the unit in the design array into operon_boundaries to prepare for subsequent spacing decoupling.

[0119] Step 4.2: Implicit Node Scanning and Ghost Placeholder Mechanism

[0120] This is the core fault-tolerance mechanism of this invention for handling extreme non-standard input. In practical applications, user-input text may be highly incomplete, for example, only the sentence "Gene A represses promoter B" is entered, without defining the complete expression box of "Gene A" anywhere in the text. When rendering such isolated nodes with regulatory arcs, traditional graphics engines cannot find the start and end coordinates of the node on the physical skeleton, which can lead to null pointer exceptions or memory crashes. Therefore, this invention introduces an implicit node scanning and ghost padding mechanism, including the following steps:

[0121] The system first extracts the set of all component names that have obtained valid physical placeholders (i.e., occupied_names) from the current design array.

[0122] The system iterates through the set of control relationships (reg) and performs a set difference operation to verify the starting node and target node of each control arc. If a node is found not to exist in the occupied_names set, it is determined to be a detached control node.

[0123] For each detached control node, the system forcibly constructs a "minimalist virtual expression box" in memory. This expression box is instantiated according to the structure of [gray transparent placeholder initiator] -> [independent RBS] -> [detached node itself] -> [independent terminating element] and appended to the end of the global design array. At the same time, the set of manipulator boundaries is updated synchronously to ensure that subsequent algorithms can correctly trigger the external separation spacing.

[0124] Through this supplementary mechanism, the system provides all abstract control logic nodes with legitimate physical space carriers, fundamentally eliminating the problem of control arc rendering failure caused by the rendering engine not being able to find node coordinates, and realizing the self-consistent closed loop of topological logic.

[0125] Example 5: Adaptive Dual-Layer Spacing Coordinate System and SBOL Visual Standard Rendering

[0126] After completing the one-dimensional serialization and reassembly of all primitives, the system enters the final coordinate system mapping and visual rendering pipeline.

[0127] Step 5.1: One-dimensional coordinate adaptive extrapolation based on dual-gap

[0128] The system must map abstract object sequences into drawing instructions with precise one-dimensional spatial coordinates (start and end). Simply arranging elements at fixed intervals causes different transcription units (operons) to appear visually stuck together, making them difficult to distinguish. Therefore, this invention proposes a dual-layer spacing coordinate calculation engine, including:

[0129] Internal compact spacing (PART_GAP): The system sets the physical distance between adjacent elements (such as between RBS and CDS) within the same expression box to a very small constant (e.g., 2 units) to ensure that functional units appear visually compact.

[0130] External decoupling gap (OPERON_GAP): When the system traverses the design array to calculate coordinates, it synchronously monitors the current index value. Once the current cursor index hits the operon_boundaries boundary set recorded in step 4.1 (and is not the start of the sequence), the system will forcibly trigger a leap in coordinate offset increment (e.g., a sudden increase of 40 units). Through this step-size intervention, the algorithm forcibly separates different transcription units in the spatial coordinate dimension without changing the data structure, achieving visual decoupling and anti-crowding optimization of the multi-module layout of complex gene circuits.

[0131] Step 5.2: Dynamic Canvas Size Mapping and High-Fidelity Atlas Rendering

[0132] To address the primitive compression and deformation problem caused by extremely long sequences, the system abandons the fixed canvas size of traditional drawing tools and instead establishes an adaptive resolution mapping function.

[0133] The system extracts the end coordinate value of the last element in the design array as the physical total length (total_width). Based on this parameter, the theoretical optimal width of the matplotlib drawing board is dynamically derived using a linear scaling formula (e.g., fig_width = max(7, total_width / 200*7)), and the aspect ratio is locked.

[0134] Finally, the system sends the sequence of flat elements with absolute coordinates, along with a set of control arcs linked to start and end anchor points, to a standard SBOL graphics renderer (such as the renderDNA interface based on dnaplotlib). The rendering engine performs vector strokes and fills, disables the coordinate axes (axis('off')), and ultimately outputs high-resolution PNG or SVG files that conform to cross-platform publishing standards.

[0135] like Figure 4 The second half of this document, Example 5, elaborates on the technological breakthrough brought about by the two-layer spacing engine in decoupling complex operon layouts. Existing conventional rendering tools are often limited by a globally singular fixed spacing strategy (such as maintaining a uniform 2u spacing between all adjacent elements). When rendering complex gene circuits containing multiple transcription units, this easily leads to severe visual adhesion and crowding of different operon blocks (such as [Op1_Parts] and [Op2_Parts]), significantly reducing the hierarchy and readability of engineering drawings. To address this layout defect, Example 5 of this invention introduces a two-layer adaptive spacing mechanism. This mechanism, while maintaining a compact arrangement (2u spacing) of elements within a single operon, can accurately identify the physical boundaries of transcription units and forcibly inject a step-like external partition spacing (i.e., 40u OPERON_GAP) between the terminator of the previous operon (such as [Terminator_Op1]) and the promoter of the next operon (such as [Promoter_Op2]). With this innovative coordinate deduction strategy, the system successfully stretched and separated logically independent transcription units at the physical visual level with great tension, completely eliminating visual congestion and giving the final output gene circuit map excellent modular decoupling effect and professional-grade information expression capability.< / font>

Claims

1. A standardized and automated mapping method for describing non-standard biological gene circuits, characterized in that, Includes the following steps: S1. Obtain multi-source non-standard gene circuit description text, then perform semantic anti-fragmentation cleaning to extract standardized entity relation triplet sequences; S2. Based on multi-dimensional feature rules, classify, identify, and normalize the entities in the entity relationship triplet sequence, and instantiate them as physical memory objects. S3. Traverse the entity relation triple sequence, identify the sequential driving relationship and cross-domain control relationship between the physical memory objects based on the directed graph traversal algorithm, and dynamically construct a multi-scale manipulator set and an inverse control network composed of the physical memory objects; each manipulator in the multi-scale manipulator set is associated with a promoter and its downstream cascaded encoding sequence set, and the inverse control network contains a control arc with the encoding sequence as the starting node and the promoter as the target node; S4. Standardize the expression box assembly of the manipulation subset: sequentially add promoter, ribosome binding site, coding sequence and terminator to generate assembled expression box; at the same time, based on the inverse control network, detect topology disconnection or isolated control nodes, and automatically trigger the replacement mechanism when detected, construct virtual expression box, and append the virtual expression box to the basic expression box sequence to obtain a complete rendering sequence; S5. Using a two-layer adaptive spacing algorithm, all physical memory objects in the complete rendering sequence are mapped to a one-dimensional physical coordinate system. The rendering position of each biological element is solved. Then, based on the inverse regulation network, the corresponding regulation arc graphics are added according to the positions of the starting node and the target node of the regulation arc in the one-dimensional physical coordinate system. Finally, the drawing engine is driven to output a gene circuit map that conforms to the preset visual specifications.

2. The standardization and automatic mapping method for describing non-standard biological gene circuits according to claim 1, characterized in that, Performing semantic anti-split cleaning specifically includes: Identify specific subscript labels representing biological nomenclature in multi-source non-standard gene circuit description text; Without disrupting the overall text topology, the content enclosed by the specific subscript tag is converted into plain text format with a preset leading character; The converted text is injected into the document object model parser to remove redundant rich text tags and extract the pure text that retains the core auxiliary semantics. Based on a preset state machine cursor scanning algorithm, valid descriptive paragraphs are extracted, and continuous invisible whitespace characters are dynamically compressed to obtain standardized line-by-line descriptive text. The standardized line-by-line description text is segmented line by line. Each line of text is cut into source nodes, relational verbs and target nodes according to a preset delimiter, thereby extracting the standardized entity relation triplet sequence.

3. The standardization and automatic mapping method for describing non-standard biological gene circuits according to claim 1, characterized in that, Instantiation into a physical memory object specifically includes: Assign initial visual attributes to the normalized entities; When the instantiated entity is a general auxiliary element in the gene circuit, the instantiation factory pattern based on isolation scope is used to create the object; The instantiation factory pattern generates a completely new data structure in memory each time it is called, and assigns a unique identifier containing a globally incrementing counter to the data structure so that general-purpose auxiliary elements of the same type have independent physical coordinates that are decoupled from each other in the addressing tree of the underlying drawing engine, thus preventing coordinate overwrite conflicts.

4. The standardization and automatic mapping method for describing non-standard biological gene circuits according to claim 1, characterized in that, The construction of multi-scale manipulation subsets and inverse control networks specifically includes: Initialize an empty submap table; If the triple represents the physical relationship of the promoter-driven coding sequence, then the coding sequence is appended to the operator sequence governed by the promoter; If the triple represents the extension relationship of the cascaded coded sequences, then the promoter to which the preceding coded sequence belongs is addressed in reverse in the currently constructed operator mapping table; if the addressing is successful, the following coded sequence is appended to the operator sequence governed by the promoter; if the addressing is unsuccessful, a topological break is determined, and an anonymous promoter is dynamically generated in memory as a virtual anchor point, and the preceding and following coded sequences are jointly governed by the anonymous promoter. If a triple represents a non-physical connection relationship of a promoter that regulates a coding sequence, it is determined to be a cross-domain regulatory relationship and stored independently in the global inverse regulation list.

5. The standardization and automatic mapping method for describing non-standard biological gene circuits according to claim 1, characterized in that, The specific steps involved in triggering the fill mechanism to construct the virtual expression box are: Extract the set of names of the physical memory objects that have been incorporated into the basic representation box sequence; Traverse the control arcs in the inverse control network, perform set difference operation, and verify whether the starting node and target node of the control arc exist in the name set; If the check finds a missing free control node, then an isolated control node is determined to have occurred. A virtual expression box is automatically constructed in memory for the free control node. The virtual expression box sequentially includes: a low-opacity placeholder promoter, an independent ribosome binding site, the free control node, and an independent terminator. The virtual expression box is appended to the end of the basic expression box sequence, and the manipulator boundary record is updated to form the complete rendering sequence.

6. The standardization and automatic mapping method for describing non-standard biological gene circuits according to claim 1, characterized in that, The two-layer adaptive spacing algorithm includes: Introducing a dual-layer spacing coordinate calculation engine; Between adjacent primitives within the same manipulator, the physical offset coordinates are calculated using a preset internal compact spacing parameter; At the physical boundaries of different operons, when the rendering cursor crosses the end of the previous operon and the beginning of the next operon, a step-wise external partition spacing parameter is forcibly triggered, and the external partition spacing parameter is greater than the internal compact spacing parameter, so as to decouple independent transcription units at the visual-physical level.

7. A standardization and automatic mapping system for describing non-standard biological gene circuits, implementing the method as described in any one of claims 1-6, characterized in that, include: The data preprocessing and cleaning module is used to acquire multi-source non-standard gene circuit description text, perform semantic anti-fragmentation cleaning, and extract standardized entity relation triplet sequences. The entity semantic recognition and instance isolation module is used to classify, identify and name normalize entities in the entity relationship triple sequence based on multi-dimensional feature rules, and instantiate them as physical memory objects. The topology network relationship extraction and dynamic clustering module is used to traverse the entity relationship triple sequence, identify the sequential driving relationship and cross-domain control relationship between physical memory objects based on the directed graph traversal algorithm, and dynamically construct multi-scale manipulation subsets and inverse control networks. The standard expression box assembly and fault-tolerant module is used to perform standardized expression box assembly on the manipulation subset, and triggers an automatic fault-infilling mechanism when a topology break or isolated control node is detected to generate a virtual expression box to obtain a complete rendering sequence. The standardized rendering and adaptive coordinate system module is used to map the complete rendering sequence to a one-dimensional physical coordinate system using a two-layer adaptive spacing algorithm, solve the rendering position for each biological element, add regulatory arc graphics based on the inverse regulatory network, and drive the drawing engine to output a gene circuit map that conforms to the preset visual specifications.

8. A computer device, characterized in that, It includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of the method as described in any one of claims 1-6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1-6.

Images