Standardized sample library and heterogeneous test data platform interoperability processing method and system

Abstract

The application discloses a standardized sample library and heterogeneous test data platform interoperation processing method and system, the method comprising: mapping experimental event original data in each heterogeneous system into a standardized event primitive set containing event type, event timestamp, event object and event resource, and building a sample pedigree directed graph with sample entities and their intermediate forms as nodes and processing relationships as directed edges; then constructing each platform test result data into a heterogeneous observation graph, generating a candidate set of links under the constraint of feasibility, combining a consistency factor to establish a factor graph model and performing global solving through a constraint optimization algorithm to obtain the optimal link selection of the observation nodes; further detecting contradictory relationships and identifying uncertain subgraphs, implementing local freezing of the link relationships in the uncertain subgraphs, and generating standardized sample library records for other link relationships. The method can improve the accuracy and stability of cross-platform test result attribution.

Description

Technical Field

[0001] This application relates to the field of testing information processing technology, specifically a method and system for interoperability processing between a standardized sample library and a heterogeneous testing data platform. Background Technology

[0002] With the development of clinical gene testing services, laboratories typically operate multiple testing platforms simultaneously, such as next-generation sequencing, quantitative real-time polymerase chain reaction (qPCR), and digital polymerase chain reaction (DMR), which are connected to laboratory information management systems, laboratory information systems, data analysis platforms, and middleware systems. The same biological sample often undergoes multiple processing stages during testing, including sampling and receiving, aliquoting, nucleic acid extraction, library construction, instrumental testing, retesting, supplementary testing, confirmatory experiments, and report revision. The same target may also undergo initial testing, retesting, and confirmation on different platforms, thus creating cross-platform, cross-stage, and cross-system data flow relationships.

[0003] In existing technologies, different systems often have inconsistent recording standards for sample events, processing resources, and test results. This manifests in different data formats, inconsistent identification systems, and inconsistent result transmission sequences. Furthermore, some data only retains partial business identifiers or partial process information, making it difficult to maintain a continuous processing trajectory for the same sample across different systems. Especially after samples undergo processes such as dispensing, transformation, loading, and re-inspection, a single original sample may correspond to multiple intermediate forms and multiple result records. Relying solely on single pieces of information such as sample number, barcode, plate position, and batch number for association can easily lead to problems such as incorrect result attribution, confusion of samples from the same batch, and distortion of cross-platform result correspondence.

[0004] When test results exhibit temporal, resource usage, and re-verification relationships, existing processing methods typically lack a unified global verification mechanism, making it difficult to promptly identify contradictory or uncertain data relationships. Directly writing unverified results into a standardized sample library not only affects the accuracy of the sample lineage but may also adversely impact subsequent result integration, quality control, and report generation.

[0005] Therefore, how to achieve accurate correspondence and stable interoperability between test results and sample processing chain under the condition of parallel operation of multiple platforms and multiple systems has become a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0006] The purpose of this application is to provide a method and system for interoperability processing between a standardized sample library and a heterogeneous testing data platform, so as to solve the problems mentioned in the background art.

[0007] According to one aspect of this application, a method for interoperability processing between a standardized sample library and a heterogeneous testing data platform is provided, comprising the following steps: The raw data of experimental events in each heterogeneous system are mapped to event quadruple structures to obtain a standardized set of event primitives. The event quadruple structure includes event type, event timestamp, event object and event resource. A directed graph of sample genealogy is constructed based on the standardized event primitive set. The directed graph of sample genealogy uses sample entities and their intermediate forms as nodes and processing relationships as directed edges. The test results data from various platforms are used as observation nodes to construct a heterogeneous observation map. The observation nodes record platform type, target, timestamp and resource location information. Based on feasibility constraints, a candidate attachment set is generated for each of the observation nodes, and the candidate nodes in the candidate attachment set correspond to the sample nodes in the directed graph of the sample spectrum. A factor graph model is constructed based on the candidate attachment set and consistency factor. The factor graph model is then solved globally using a constrained optimization algorithm to obtain the optimal attachment selection for each observation node. Based on the global solution results, contradictory relationships are detected and uncertain subgraphs are identified. The attachment relationships within the uncertain subgraphs are locally frozen, and standardized sample library records are generated for other attachment relationships.

[0008] Preferably, the feasibility constraints include time feasibility constraints, platform feasibility constraints, resource feasibility constraints, target feasibility constraints, and lineage feasibility constraints; the time feasibility constraint is that the timestamp of the test result is later than the timestamp of the upstream processing event; the platform feasibility constraint is that the candidate node appears in the on-machine event record of the corresponding platform; the resource feasibility constraint is that the candidate node belongs to the same batch of samples loaded in the same plate position and well position within the corresponding time period; the target feasibility constraint is that the target of the observation node falls within the coverage area of ​​the target panel to which the candidate node belongs; the lineage feasibility constraint is that the test result corresponding to the observation node falls within the lineage branch of its possible source.

[0009] Preferably, the consistency factors include a temporal consistency factor, a lineage transmission consistency factor, a platform process compatibility consistency factor, a resource exclusivity consistency factor, and a logical continuity factor; the temporal consistency factor is determined based on the standardized time difference between the timestamp of the observation node and the timestamp of the sample node; the lineage transmission consistency factor is determined based on the reachability of the upstream ancestor node and the downstream descendant node of the sample node in the lineage graph after the connection; the platform process compatibility consistency factor is determined based on whether the candidate node has completely undergone all the necessary processing steps of the corresponding platform; the resource exclusivity consistency factor is determined based on whether the candidate node has a resource conflict with other observation nodes in the same resource location during the same time period; and the logical continuity factor is determined based on whether the time order of the re-examination observation node and the corresponding initial examination observation node meets the business rules.

[0010] Preferably, the factor graph model includes variable nodes, factor nodes, and edges; the variable nodes correspond to the affiliation selection of each observation node, and the value range is the sample nodes in its candidate attachment set and the null value option, where the null value option indicates that the observation node is not attached to any sample node; the factor nodes correspond to various consistency constraints; the edges connect the variable nodes to the factor nodes that constrain their values; the objective function of the global solution is to maximize the sum of the weighted consistency factors of all candidate attachment relationships.

[0011] Preferably, identifying the uncertain subgraph includes: traversing all connection relationships to detect contradictory situations, constructing a contradictory relationship graph based on the detected contradictory relationships, performing connected component analysis on the contradictory relationship graph, and marking connected components whose ratio of the number of contradictory edges to the number of observed nodes in the corresponding connected component exceeds a preset density threshold or whose average consistency score is lower than a preset score threshold as uncertain subgraphs.

[0012] Preferably, partially freezing the connections within the uncertain subgraph includes: preventing all connections within the uncertain subgraph from entering the standard release process, and instead placing them in a confirmation queue; re-evaluating the deterministic state of the subgraph after some connections within the uncertain subgraph are confirmed through manual confirmation or upstream data supplementation; and if the confirmed connections reduce the uncertainty of the entire subgraph to below a threshold, then lifting the freeze on the subgraph and incorporating it into the standard release process.

[0013] Preferably, the generated result record includes: generating a standardized record for observation nodes that have been confirmed to be attached but do not belong to the uncertain subgraph; the standardized record includes sample node identifier, phylogenetic location information, attachment evidence summary, consistency score range and uncertainty marker; generating a record to be confirmed for observation nodes that are within the uncertain subgraph and marking the uncertainty marker as to be confirmed.

[0014] In another aspect, this application also provides an interoperable processing system for a standardized sample library and a heterogeneous testing data platform, comprising: The event standardization module is used to map the raw data of experimental events in various heterogeneous systems into event quadruple structures to obtain a set of standardized event primitives. The event quadruple structure includes event type, event timestamp, event object, and event resource. The genealogy construction module is used to construct a directed graph of sample genealogy based on the standardized event primitive set. The directed graph of sample genealogy uses sample entities and their intermediate forms as nodes and processing relationships as directed edges. The observation graph construction module is used to construct a heterogeneous observation graph by using the test result data from various platforms as observation nodes. The observation nodes record platform type, target, timestamp and resource location information. The candidate set generation module is used to generate a candidate attachment set for each observation node based on feasibility constraints, wherein the candidate nodes in the candidate attachment set correspond to the sample nodes in the directed graph of the sample spectrum. The global solution module is used to construct a factor graph model based on the candidate attachment set and the consistency factor, and to perform a global solution on the factor graph model using a constrained optimization algorithm to obtain the optimal attachment selection for each observation node. The results processing module is used to detect contradictory relationships and identify uncertain subgraphs based on the global solution results, locally freeze the attachment relationships within the uncertain subgraphs, and generate standardized sample library records for other attachment relationships.

[0015] In another aspect, this application also provides an electronic device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the interoperability processing method between the standardized sample library and the heterogeneous test data platform as described above.

[0016] In another aspect, this application provides a storage medium storing computer program instructions that can be executed by a processor to implement the interoperability processing method between the standardized sample library and the heterogeneous test data platform as described above.

[0017] This invention maps experimental events in heterogeneous systems into standardized event primitives that include event type, event timestamp, event object, and event resource. Based on this, a directed graph of sample hierarchy is constructed. Then, the test results from multiple platforms are constructed into observation nodes. Combined with feasibility constraints, a candidate attachment set is generated. A factor graph model is established using a consistency factor for global solution. Thus, even under complex processes such as sample packaging, extraction, library construction, computerization, retesting, and confirmation, the accurate attribution of test results across systems and platforms to sample nodes can be achieved. Meanwhile, by identifying contradictory relationships in the global solution results, areas with concentrated conflicts are marked as uncertain subgraphs and locally frozen. This not only suppresses the spread of incorrect attribution downstream and prevents uncertain data from overwriting confirmed data, but also ensures the normal publication of data in non-conflict areas, improving the accuracy, stability, and traceability of the standardized sample library records, while taking into account both data quality control and overall processing efficiency. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0019] Other features, objects, and advantages of this application will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings: Figure 1 This is a schematic diagram illustrating an interoperability processing method between a standardized sample library and a heterogeneous testing data platform, provided as an embodiment of this application.

[0020] Figure 2 This is a schematic diagram of the process for generating a limited candidate attachment set provided in an embodiment of this disclosure.

[0021] Figure 3 This is a schematic diagram of the optimal attachment selection process provided in the embodiments of this disclosure.

[0022] Figure 4 This is a schematic diagram illustrating the differentiated processing provided in the embodiments of this disclosure.

[0023] Figure 5 This is a schematic diagram of the structure of an interoperable processing system for a standardized sample library and a heterogeneous test data platform, provided in an embodiment of this application.

[0024] Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0025] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0026] According to embodiments of this disclosure, a method for interoperability processing between a standardized sample library and a heterogeneous testing data platform is provided. This method is applicable to scenarios in clinical gene testing laboratories where the same biological sample needs to undergo multiple experimental processing steps, and the test results are distributed across multiple heterogeneous information systems. The laboratory simultaneously runs testing technology platforms such as next-generation sequencing (NGS), quantitative polymerase chain reaction (qPCR), and digital polymerase chain reaction (ddPCR), each platform connecting to a laboratory information management system (LIMS), a laboratory information system (LIS), a data analysis platform, and middleware systems. The same original sample may undergo multiple steps, including sampling and receiving, aliquoting, nucleic acid extraction, library construction, instrumental testing, retesting, supplementary testing, confirmation experiments, and report revision. Furthermore, the same target may undergo initial testing, retesting, and confirmation on different platforms sequentially. The data formats, identification systems, and result return timings of the heterogeneous systems differ, and the standardized sample library needs to accurately attribute test results from multiple systems to the corresponding sample nodes.

[0027] The following detailed description, in conjunction with specific embodiments, illustrates the implementation process of the interoperability processing method between the standardized sample library and the heterogeneous testing data platform described in this application. It should be noted that this embodiment is merely for explaining this application and is not intended to limit the scope of protection of this application. Any conventional adjustments or substitutions made by those skilled in the art to the steps without departing from the concept of this application should be included within the scope of protection of this application.

[0028] like Figure 1 As shown in the figure, this application discloses a schematic diagram of an interoperability processing method between a standardized sample library and a heterogeneous testing data platform, including the following method steps: S1, map the raw data of experimental events in each heterogeneous system into an event quadruple structure to obtain a standardized set of event primitives. The event quadruple structure includes event type, event timestamp, event object and event resource. S2, construct a directed graph of sample hierarchy based on the set of standardized event primitives. The directed graph of sample hierarchy uses sample entities and their intermediate forms as nodes and processing relationships as directed edges. S3, using the test result data from each platform as observation nodes, constructs a heterogeneous observation map, where the observation nodes record platform type, target, timestamp, and resource location information; S4, Based on feasibility constraints, generate a candidate attachment set for each observation node, wherein the candidate nodes in the candidate attachment set correspond to the sample nodes in the directed graph of the sample spectrum. S5. Construct a factor graph model based on the candidate attachment set and consistency factor, and use a constrained optimization algorithm to solve the factor graph model globally to obtain the optimal attachment selection for each observation node. S6. Based on the global solution results, detect contradictory relationships and identify uncertain subgraphs. Implement local freezing of the attachment relationships within the uncertain subgraphs and generate standardized sample library records for other attachment relationships.

[0029] In some embodiments, for step S1, experimental event-related data is extracted from various heterogeneous systems, and data from different sources is uniformly mapped to a finite set of experimental event primitives. The extracted data source types include system log records, status bit change records, task execution records, file metadata information, etc. The extracted raw data is mapped to a unified event quadruple structure, which includes four fields: event type, event timestamp, event object, and event resource.

[0030] Event types are represented by standard values ​​from a finite enumeration set, exemplarily including sampling receipt events, barcode assignment events, aliquoting events, nucleic acid extraction events, library construction events, plate well loading events, sequencing events, raw result generation events, result review events, and report release events. Event timestamps use a unified timestamp format to record the actual time of the event. Event objects identify the sample entity or sample derivative affected by the event, which can be the original sample, aliquoted sample, extraction product, or library construction product. Event resources record the physical or logical resource information occupied by the event, which can be plate position, well position, reagent batch, instrument operation number, sequencing channel, index combination, etc.

[0031] It should be noted that the above mapping normalizes semantically equivalent but formatted raw records from different systems into a single structure. For example, a nucleic acid extraction completion event represented by a task status change record in the LIMS system and the same event represented by an operation log record in the LIS system are both represented by a standard four-element structure after mapping: event type is nucleic acid extraction event, event object is the corresponding extraction product identifier, and event resource is the corresponding instrument operation number. The standardized event primitive set is output after mapping. In some embodiments, when multiple raw records are generated in different systems for the same experimental processing step, deduplication and merging are performed using the event object identifier and event timestamp as the joint key, retaining the record with the most complete information.

[0032] In some embodiments, for step S2, a directed graph of sample hierarchy is constructed based on a standardized set of event primitives. The directed graph of sample hierarchy uses sample entities and their intermediate forms as nodes, and sample processing relationships as directed edges. Nodes are classified according to the processing stage of the sample: original sample nodes correspond to the initial sample entity after the sampling and receiving event; packaged sample nodes correspond to the sub-sample entities generated by the packaged event; extracted product nodes correspond to the nucleic acid products generated by the nucleic acid extraction event; library construction product nodes correspond to the library entities generated by the library construction event; and plate well loading nodes correspond to the plate well loading event, recording the physical position of the sample in the experimental plate.

[0033] The source-destination relationships between nodes are established based on the object identifiers and resource identifiers in the event primitives. For the packaging event, split edges are established between the original sample node and each packaged sample node, with the direction of the split edges pointing from the original sample node to the packaged sample node, indicating that the sample splits from one entity into multiple sub-entities. For the nucleic acid extraction event, transformation edges are established between the packaged sample node and the extraction product node, indicating that the sample changes from one form to another. For the library construction event, transformation edges are established between the extraction product node and the library construction product node.

[0034] For plate-hole loading events, a loading edge is established between the library construction product node or extraction product node and the plate-hole loading node. This loading edge records the relationship of the sample entering a specific physical location. For machine-on events, the corresponding machine-on event record is associated with the plate-hole loading node. The construction process follows the timestamp order of the event primitives, retaining all branch relationships in phylogenetic structures with branches to reflect the true sample evolution path. Each edge includes the timestamp information of the corresponding event primitive, used for subsequent calculation of the temporal consistency factor.

[0035] The phylogenetic tree is organized around the experimental processing chain, rather than using the business primary key as the index. This allows subsequent detection results entering the system to still regress to the correct sample node using the context information of their respective processing chain, improving the stability of attribution determination in cases of incomplete or ambiguous primary key identifiers.

[0036] Optionally, the constructed sample phylogenetic graph is completed. Based on phylogenetic transitivity rules, potential missing nodes are inferred. For example, if an extraction product node has a corresponding data entry event but no preceding library construction event node, the missing library construction product node is added and marked as a hypothetical node. Cycle detection is performed on the phylogenetic graph. Since the sample processing flow does not physically generate cycles, detected cycles are retained according to timestamp order, with the earliest record being preserved and duplicate records deleted.

[0037] In some embodiments, for step S3, the test result data from each platform first enters the heterogeneous observation graph as observation nodes. A buffer layer is established between the test result data and the standardized sample library to prevent the test results from contaminating the existing data in the standardized sample library before undergoing global consistency verification.

[0038] Each observation node records the platform type identifier, target, timestamp, well / plate position run number, basic quality indicators, target category, and result value type. The basic quality indicators vary depending on the platform type, including different quality control parameters. For example, basic quality indicators for NGS platforms include sequencing depth and base quality value; for qPCR platforms, they include cycle threshold and amplification efficiency; and for ddPCR platforms, they include effective droplet count and positive droplet ratio. Multiple observation nodes within the same testing batch are associated according to plate / well position information and run batch information to form an observation batch structure. Multiple testing results of the same target at different time sequences are chained together to form an observation time-series chain.

[0039] Specifically, the inspection result data from LIMS, LIS, data analysis platform and middleware system are uniformly parsed to extract fields such as platform type identifier, target, timestamp, plate position and hole position operation number information, basic quality indicators, target category and result value type, and observation nodes are generated according to a unified data structure.

[0040] Using the batch information as the main index and the plate position and hole position information as the positioning basis, batch association relationships are established for observation nodes belonging to the same inspection batch to form an observation batch structure; for observation nodes with the same target and a sequential relationship in time, a time sequence association relationship is established according to the timestamp order to form an observation time sequence chain.

[0041] Through the above processing, the test results data from heterogeneous sources and formats are organized into a unified heterogeneous observation map, providing standardized input for the subsequent generation of candidate attachment sets based on time feasibility constraints, platform feasibility constraints, resource feasibility constraints, target feasibility constraints, and lineage feasibility constraints.

[0042] In some embodiments, for step S4, a finite candidate attachment set is generated for each observation node, representing the set of sample phylogenetic graph nodes to which the observation node may belong. The generation of the candidate attachment set is based on the following computable technical constraints.

[0043] like Figure 2 As shown, Figure 2 This is a schematic diagram of the process for generating a limited candidate attachment set according to an embodiment of this disclosure. In step S201, a time feasibility constraint is imposed. Based on the physical causal relationship of the experimental processing, the timestamp of the test result must be later than the timestamp of its upstream processing event.

[0044] For each observation node, calculate its timestamp. With the event timestamps of each sample node in the phylogenetic diagram Time difference between ,filter and Within the preset time window Sample nodes within the time frame are considered as time-feasible candidates. The value is set differently depending on the platform type, and the specific value is determined based on the laboratory's business process and sample turnover cycle.

[0045] Specifically, for each sample node in the phylogenetic graph, the timestamp of the most recent processing event corresponding to that node is obtained as... Then compare it with the timestamp of the observed node. If This indicates that the observation node was generated earlier than the candidate sample node was processed, violating the causal time sequence, and the candidate node is directly excluded. If A value indicating an excessively long time interval indicates extremely low credibility for the candidate relationship, and it is thus excluded. After time feasibility screening, candidate nodes that meet the conditions are retained for the next constraint screening.

[0046] In step S202, platform feasibility constraints are implemented. Candidate nodes must actually have entered the platform's processing flow. Admission criteria for each platform type are maintained, defined as the sequence of event types that should exist in the phylogenetic tree for that platform. Each candidate sample node in the candidate attachment set is traversed, and the phylogenetic tree is checked to see if there are any records of ddPCR events entering the corresponding platform on the downstream path of that candidate node. For example, for the ddPCR platform, if a candidate sample node does not have any ddPCR events on its downstream path in the phylogenetic tree, then that node does not meet the ddPCR platform's admission criteria and is removed from the candidate attachment set. This constraint ensures that test results can only be attributed to sample nodes that have actually undergone the corresponding platform's processing flow.

[0047] In step S203, resource feasibility constraints are implemented. Based on the physical exclusivity of experimental plate well locations, the same well location in the same run can only correspond to a limited set of samples within the same time period. For each observation node, based on its recorded well location / plate location run number information, the loading events corresponding to the same run batch and the same well location in the phylogenetic tree are queried to obtain the set of sample nodes loaded at that resource location within the corresponding time period. These sample nodes are then considered as resource feasibility candidates. If an observation node lacks well location / plate location information, this constraint is skipped, and all candidate nodes filtered by the preceding constraints are retained.

[0048] In step S204, target feasibility constraints are implemented. The target panel, site set, and project type of the candidate node must be compatible with the upstream detection application or review path. In the genealogy diagram, the original sample node is usually associated with detection project application information, which is passed downstream along the genealogy edge to each packaging node and product node. The target information of the observation node is matched and verified with the detection application information inherited by the candidate node. Target compatibility judgment is based on the target panel coverage relationship: if the target of the observation node falls within the coverage of the target panel to which the candidate node belongs, it is determined to be compatible; otherwise, the candidate node is removed from the candidate attachment set. For observation nodes in the review path, it is also necessary to verify whether their targets are consistent with the targets in the initial detection stage.

[0049] In step S205, spectral feasibility constraints are applied. When an observation node lacks the plate location / aperture location information and operation number information needed to determine a spectral branch, this constraint is skipped, and candidate nodes filtered by previous constraints are retained. The test result can only fall within the spectral branch of its possible origin. Based on the plate location / aperture location information and operation number information of the observation node, the upstream processing node corresponding to the resource is traced backward in the spectral graph. Then, starting from this upstream node, the spectral edge is traced upward to the original sample node to determine the complete spectral branch corresponding to the resource location. It is checked whether each candidate node in the candidate attachment set falls within this spectral branch, and candidate nodes that do not meet the spectral reachability requirement are eliminated. The purpose of this constraint is to further narrow down the candidate range using the structural information of the spectral graph, avoiding incorrectly assigning the test result to a spectral branch unrelated to its physical origin.

[0050] The five constraints described above are applied sequentially, with the input for each step being the output of the previous step. Each constraint applies time, process, resources, project, and graph structure considerations to the candidate pool, and the pool is narrowed down as relevant information becomes available.

[0051] In step S206, after comprehensive constraint screening, a candidate attachment set is output. Each candidate node in the candidate attachment set is accompanied by a label indicating the constraints it satisfies. When the candidate attachment set is empty, it indicates that the observation node cannot establish a candidate relationship with any known sample node, and the manual confirmation process begins. When the candidate attachment set contains only one candidate node, the attribution relationship is initially determined, but its consistency still needs to be verified through subsequent global solution. When the candidate attachment set contains multiple candidate nodes, there is ambiguity in attribution, which needs to be resolved through subsequent global solution.

[0052] In some embodiments, for step S5, such as Figure 3 As shown, Figure 3 This is a schematic diagram of the optimal attachment selection process provided in an embodiment of this disclosure. In step S301, a consistency factor is defined for each candidate attachment relationship between an observation node and a sample node.

[0053] Time Consistency Factor This reflects the degree of matching between the timestamps of observed nodes and the timelines of candidate sample node spectral systems. The forward time difference is calculated. Calculated according to piecewise functions:

[0054] in For the timestamp of the observation node, For sample node event timestamps , , This is a preset time threshold parameter. , , The corresponding factor decay value and satisfying The values ​​of each parameter are determined based on the platform type and business characteristics. This violates the causal time constraint. The value is zero. The smaller the time difference, the closer the temporal correlation between the observed node and the candidate sample node, and the higher the confidence level of the attribution; the larger the time difference, the more interference factors may exist in the middle, and the lower the confidence level of the attribution accordingly.

[0055] Lineage transmission consistency factor This reflects whether the candidate connection relationships are consistent with the transitivity constraints of the phylogenetic graph. The position of the candidate sample node in the phylogenetic graph is determined, and starting from that node, the graph is traversed upstream and downstream to check whether the edge relationships between nodes are complete.

[0056] Specifically, if after an observation node is attached to a sample node, it can be traced back to the original sample node without interruption along the lineage edges from that sample node, and the edge relationships of the downstream descendant nodes of that sample node are also complete, then... The value is 1.0. If there are speculative nodes in the traceability path, a deduction is made based on the proportion of speculative nodes to the total number of nodes in the path. If there is a break in the traceability path, Choose the lower value. This factor ensures that the attachment relationship is consistent with the physical evolution path of the sample.

[0057] Platform process compatibility consistency factor This reflects whether candidate sample nodes have completed all necessary processing steps for their respective platforms in the correct order. For the target platform type, query the necessary processing step sequence for that platform. Taking the NGS platform as an example, the necessary processing step sequence typically includes nucleic acid extraction, library construction, plate loading, and sequencing. Check whether the candidate sample nodes in the phylogenetic tree have completed all steps in the order of this sequence. (The last sentence appears to be incomplete and possibly refers to a separate process: "Completely undergoing all necessary steps...") The value is 1.0. When steps are missing, the deduction is made according to the proportion of the number of missing steps to the total number of necessary steps.

[0058] Resource Exclusivity Consistency Factor This factor reflects the physical exclusivity constraint of resources. Only one sample can occupy the same resource location at any given time. For observation nodes and candidate sample nodes in candidate attachment relationships, the resource occupancy record corresponding to that sample node is queried to check whether the resource location has been exclusively occupied by other observation nodes at the same time. The value is 1.0 when there is no resource conflict and zero when there is a conflict. This factor participates in the global solution as a hard constraint; candidate relationships with resource conflicts will be directly excluded.

[0059] Logical continuity factor This reflects the logical continuity requirement of retesting and confirmatory tests. In the testing process, retest results should originate from the same lineage as the initial test results, and confirmatory tests should be conducted after the initial screening and retesting.

[0060] The specific calculation method is as follows: If the current observation node is marked as a re-inspection type, check whether the candidate sample node has a corresponding initial inspection observation node that has been successfully attached, and the timestamp of the re-inspection observation node is later than the timestamp of the initial inspection observation node; if the current observation node is marked as a confirmation experiment type, check whether the candidate sample node has both initial inspection and re-inspection observation nodes that have been successfully attached, and whether the time order of the three satisfies the logical relationship of initial inspection first, re-inspection in the middle, and confirmation last. The value is 1.0 when the rule is met, and a lower value is used when it is violated. For observation nodes that are neither re-inspection nor confirmation types, the default value of this factor is 1.0.

[0061] Technology coexistence factor This reflects the compatibility of cross-platform test results on the same sample node. Some testing platforms have technically mutually exclusive relationships; for example, the same nucleic acid extraction product cannot be used simultaneously on multiple platforms with sample consumption conflicts within a specific time period. The value is calculated based on whether the candidate connection relationship will introduce cross-platform interference; a value of 1.0 is taken when there is no interference, and a value of zero is taken when interference exists.

[0062] In step S302, a factor graph model is constructed based on the candidate attachment set and consistency factors, converting the observation nodes, sample nodes, and candidate attachment relationships into a factor graph representation. The factor graph is a bipartite graph structure containing two types of nodes: variable nodes and factor nodes. Variable nodes correspond to the affiliation choices of each observation node; the value range of each variable node is the sample nodes in its candidate attachment set plus a null value option, indicating that the observation node is not attached to any sample node. Factor nodes correspond to various consistency constraints; each factor node connects to one or more variable nodes, indicating that there are constraint relationships between these variable nodes.

[0063] Specifically, the temporal consistency factor, lineage transmission consistency factor, and platform process compatibility consistency factor are unary factor nodes, each connecting to a single variable node; the resource exclusivity consistency factor, logical continuity factor, and technical coexistence factor are binary or multi-factor nodes, connecting multiple variable nodes that share the same resource location or the same lineage branch.

[0064] In step S303, a constrained optimization algorithm is used to globally solve the factor graph model to obtain the optimal attachment selection for each observation node. The objective function is to maximize the sum of weighted consistency factors for all candidate attachment relationships.

[0065] in The total number of observation nodes. For observation nodes The candidate hook set, For binary decision variables, For overall consistency score:

[0066] in These correspond to the consistency factors mentioned above, in turn. For the corresponding weight parameters, satisfying and The initial values ​​of the weight parameters are preset by domain experts based on the importance of each constraint dimension, and the range of values ​​is [insert range here]. Optionally, after accumulating historical labeled data, the weight parameters can be statistically calibrated by minimizing the attachment error rate.

[0067] The constraints include: each observation node can be attached to at most one candidate sample node, i.e. This constraint ensures that each test result belongs to only one sample node; observation nodes competing for the same resource location cannot be simultaneously attached to different sample nodes corresponding to that resource, reflecting the physical exclusivity of the resource. The candidate relations are not involved in the solution, that is... This constraint will directly exclude candidate relationships that have been confirmed to have resource conflicts.

[0068] In one embodiment, a constrained integer programming method is used to transform the above problem into an integer linear programming problem for solution. For small-scale problems, a branch and bound method is used for exact solution. In another embodiment, for large-scale problems, resource exclusivity constraints and logical continuity constraints are relaxed into the objective function using Lagrange multipliers. Subgradient optimization is used to iteratively update the multiplier values, solving the relaxed simplified problem in each iteration until convergence or the preset upper limit of the number of iterations is reached. When the number of observation nodes and candidate connections is small, an exact solution method is preferred; when the problem size is large, an approximate solution method is used.

[0069] The solution outputs the optimal connection selection for each observation node and the overall consistency score. If the overall consistency score is lower than a preset deterministic threshold, the solution is invalid. The connection is marked as an unstable connection, and then proceeds to the subsequent uncertain subgraph identification process. The value can be determined according to the quality control requirements.

[0070] In some embodiments, after the global solution in step S6, the determinism of some connection relationships is still low. Forcibly assigning a unique affiliation will cause errors to propagate downstream along the spectrum, and the entire batch blocking will prevent high-deterministic data from being released in a timely manner. This step forms an uncertain subgraph by identifying local areas where contradictions are concentrated, and only freezes the standardized release of this local area, while the release of other areas continues as normal. This limits the scope of error propagation to within the uncertain subgraph, while maintaining the normal throughput of most data.

[0071] Specifically, such as Figure 4 As shown, Figure 4 This is a schematic diagram illustrating the differentiated processing provided in an embodiment of this disclosure. In step S401, contradictory relationships are detected and uncertain subgraphs are identified based on the global solution results.

[0072] Iterate through all the attachments in the global solution results and check for the following contradictions: the same resource location is attached to different sample nodes in an exclusive manner by multiple observation nodes, i.e., there is a resource exclusive conflict; the overall consistency score of the same observation node on all candidate relationships is lower than the deterministic threshold. This means that the observation node cannot find a credible target to which it belongs; there is a logical continuity conflict between the attachment results of multiple observation nodes within the same phylogenetic branch, for example, the re-examination result is attached to a phylogenetic branch different from the initial examination result.

[0073] A contradiction association graph is constructed based on the detected contradictions. Nodes in the graph correspond to observation nodes involved in the contradiction. The rules for establishing contradiction edges are as follows: if two observation nodes compete for the same resource position, a resource conflict contradiction edge is established between them; if two observation nodes belong to the same lineage branch and there is a logical continuity conflict between their attachment results, a logical conflict contradiction edge is established between them; if two observation nodes share the same candidate sample node and the resource capacity of that candidate node is insufficient to accommodate both simultaneously, a capacity conflict contradiction edge is established between them. Connectivity component analysis is performed on the contradiction association graph, and each connected component corresponds to a contradiction cluster.

[0074] For each cluster of contradictions, calculate two indices: contradiction density and contradiction density. Defined as the number of contradictory edges within a contradictory cluster. Number of nodes within the contradictory cluster The ratio, i.e. Average consistency score Defined as the arithmetic mean of the overall consistency scores of all connections within a contradictory cluster. When Exceeding the preset density threshold or Below the preset score threshold When this happens, the contradictory cluster is marked as an uncertain subgraph. and The value can be determined according to the laboratory's quality control requirements.

[0075] The two judgment conditions mentioned above characterize uncertainty from different dimensions: To measure the concentration of contradictory relationships, high density indicates severe internal conflict within the contradictory cluster; To measure the overall reliability of the linkage, a low score indicates that the linkages within the contradictory cluster are generally unreliable.

[0076] In step S402, the attachment relationships within the uncertain subgraph are partially frozen, and standardized sample library records are generated for other attachment relationships.

[0077] For observation nodes that do not belong to any uncertain subgraph, a standardized sample library record is generated based on the global solution results, and the process is initiated into the standard publication process. Specifically, the test result data of the observation node is associated with its optimally attached sample node to generate a standardized record containing complete attribution information.

[0078] For regions marked as uncertain subgraphs, all connections within those regions are set to a frozen state and added to the confirmation queue.

[0079] Specifically, the freezing granularity is the entire uncertain subgraph. Within this subgraph, all connections are interrelated through contradictory edges. If only one connection is frozen while the others are allowed, the allowed connections may generate new attribution errors due to contradictory dependencies with the frozen connections. By using the entire contradictory cluster as the freezing unit, it ensures that all connections within the contradictory cluster do not propagate potential errors downstream until the entire cluster is confirmed. Furthermore, since contradictory clusters typically involve only a few long-tailed outliers, the freezing scope is limited and does not affect the timely release of most normal sample data.

[0080] In some embodiments, the lifting of the freeze is triggered as follows: after some relationships within a subgraph are manually reviewed and confirmed, the confirmation result is re-added as a known constraint to the globally consistent solution model, and the local solution is re-executed for the subgraph. If the contradiction density of the subgraph decreases after the re-solution... The following and the average consistency score rose to Once the above steps are completed, the frozen state will be lifted, and the connection relationships within the subgraph will be incorporated into the standard release process.

[0081] Alternatively, the upstream system may supplement the missing event primitive data, completing the missing nodes or edge relationships in the genealogical graph and thus eliminating some contradictory relationships in the contradictory cluster. Optionally, if the freeze time exceeds a preset period, an overdue reminder may be generated to notify relevant personnel for manual intervention.

[0082] In step S403, based on the global solution results and the frozen state, result records are generated for each observation node. Specifically, standardized sample library records are generated for observation nodes that have been confirmed to be attached but do not belong to the uncertain subgraph; unconfirmed records are generated for observation nodes that are within the uncertain subgraph. Each record contains five fields: sample node identifier, phylogenetic location information, attachment evidence summary, consistency score interval, and uncertainty marker.

[0083] Specifically, the sample node identifier records the standardized sample node number corresponding to the record, which uniquely identifies a sample node in the phylogenetic tree. The phylogenetic location information records the specific location of the record in the sample phylogenetic tree, including the processing stage, node type, and the phylogenetic depth of the node from the original sample node. The attached evidence summary records the scores of various consistency factors and the overall consistency score, enabling the subsequent system to understand the basis for determining the attachment relationship.

[0084] The consistency score range is determined based on the overall consistency score and its fluctuation range during the global solution process. The fluctuation range is obtained by comparing the score difference between the optimal solution and the second-best solution. The uncertainty flag is either confirmed or pending confirmation. Confirmed indicates that the record comes from a non-frozen area and the overall consistency score is higher than the deterministic threshold, while pending confirmation indicates that the record comes from a frozen area or the overall consistency score is lower than the deterministic threshold.

[0085] Standardized sample library records with a confirmed status are pushed to the standard consumption queue of the downstream system, while records awaiting confirmation are pushed to the pending confirmation queue along with an explanation of the reason for the freeze. The downstream system performs differentiated processing based on the uncertainty marker: for confirmed records, the downstream system can directly use their test result data; for records awaiting confirmation, the downstream system does not overwrite existing confirmed records to avoid data write-back conflicts caused by uncertain data replacing confirmed data.

[0086] Optionally, a corresponding evidence chain document is generated for each standardized record. The evidence chain document records the complete derivation process of the attachment relationship, including the source of the event primitives used and their original system identifiers, the generation process of the candidate attachment set and the screening results of each constraint, the calculation basis and specific values ​​of each consistency factor, the selection process of the attachment relationship in the global optimization solution, and the score comparison of competing candidates. The evidence chain document supports subsequent manual review and source tracing, enabling reviewers to trace the basis for determining the attribution of each standardized record.

[0087] Optionally, a result feedback interface is provided to receive processing status feedback from downstream systems. When a downstream system detects an anomaly in a confirmed record, a re-evaluation process is triggered. In some embodiments, for multiple test results received at different time points for the same sample node, the time sequence relationship of each result and the corresponding consistency score are retained in the record, enabling the downstream system to distinguish between initial test results, retest results, and confirmation test results.

[0088] It should be noted that although the operations of the method of this application are described in a specific order in the accompanying drawings, this does not require or imply that these operations must be performed in that specific order, or that all the operations shown must be performed to achieve the desired result. On the contrary, the steps depicted in the flowchart can be performed in a different order. Additionally or alternatively, certain steps may be omitted, multiple steps may be combined into one step, and / or one step may be broken down into multiple steps.

[0089] Please see Figure 5 , Figure 5 This application provides an embodiment of an interoperable processing system for a standardized sample library and a heterogeneous testing data platform. The system embodiment is related to… Figure 1 Corresponding to the illustrated method embodiments, this system can be specifically applied to various electronic devices. The system specifically includes: The event standardization module 501 is used to map the raw data of experimental events in various heterogeneous systems into event quadruple structures to obtain a set of standardized event primitives. The event quadruple structure includes event type, event timestamp, event object and event resource. The genealogy construction module 502 is used to construct a directed graph of sample genealogy based on the standardized event primitive set. The directed graph of sample genealogy uses sample entities and their intermediate forms as nodes and processing relationships as directed edges. The observation graph construction module 503 is used to construct a heterogeneous observation graph by using the test result data from various platforms as observation nodes. The observation nodes record platform type, target, timestamp and resource location information. The candidate set generation module 504 is used to generate a candidate attachment set for each observation node based on feasibility constraints, wherein the candidate nodes in the candidate attachment set correspond to the sample nodes in the directed graph of the sample spectrum. The global solution module 505 is used to construct a factor graph model based on the candidate attachment set and the consistency factor, and to perform a global solution on the factor graph model using a constraint optimization algorithm to obtain the optimal attachment selection for each observation node. The result processing module 506 is used to detect contradictory relationships and identify uncertain subgraphs based on the global solution results, locally freeze the attachment relationships within the uncertain subgraphs, and generate standardized sample library records for other attachment relationships.

[0090] Each processing unit and / or module in the embodiments of this application can be implemented by an analog circuit that implements the functions described in the embodiments of this application, or by software that executes the functions described in the embodiments of this application.

[0091] Based on the same inventive concept, this application also provides an electronic device. The method corresponding to the electronic device can be the method in the foregoing embodiments, and its problem-solving principle is similar to that method. The electronic device provided in this application includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the methods and / or technical solutions of the foregoing embodiments of this application.

[0092] Figure 6The diagram illustrates the structure of an electronic device suitable for implementing the methods and / or technical solutions in the embodiments of this application. The electronic device includes a central processing unit (CPU) 601, which can perform various appropriate actions and processes based on a program stored in a read-only memory (ROM) 602 or a program loaded from a storage section 608 into a random access memory (RAM) 603. The RAM 603 also stores various programs and data required for system operation. The CPU 601, ROM 602, and RAM 603 are interconnected via a bus 604. An input section 606, an output section 607, a communication section 609, and an input / output (I / O) interface 605 are also connected to the bus 604.

[0093] In particular, the methods and / or embodiments in this application can be implemented as computer software programs. For example, the embodiments disclosed in this application include a computer program product comprising a computer program carried on a storage medium, the computer program containing program code for performing the methods shown in the flowchart. When the computer program is executed by the central processing unit (CPU) 601, it performs the functions defined in the methods of this application.

[0094] Another embodiment of this application provides a computer-readable storage medium having computer program instructions stored thereon, which can be executed by a processor to implement the methods and / or technical solutions of any one or more embodiments of this application described above.

[0095] The flowcharts or block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of electronic devices, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-specific system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0096] The above description represents the preferred embodiments of the present invention. It should be noted that, for those skilled in the art, various improvements and modifications can be made without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for interoperability processing between a standardized sample library and a heterogeneous testing data platform, characterized in that, include: The raw data of experimental events in each heterogeneous system are mapped to event quadruple structures to obtain a standardized set of event primitives. The event quadruple structure includes event type, event timestamp, event object and event resource. A directed graph of sample genealogy is constructed based on the standardized event primitive set. The directed graph of sample genealogy uses sample entities and their intermediate forms as nodes and processing relationships as directed edges. The test results data from various platforms are used as observation nodes to construct a heterogeneous observation map. The observation nodes record platform type, target, timestamp and resource location information. Based on feasibility constraints, a candidate attachment set is generated for each of the observation nodes, and the candidate nodes in the candidate attachment set correspond to the sample nodes in the directed graph of the sample spectrum. A factor graph model is constructed based on the candidate attachment set and consistency factor. The factor graph model is then solved globally using a constrained optimization algorithm to obtain the optimal attachment selection for each observation node. Based on the results of the global solution, contradictory relationships are detected and uncertain subgraphs are identified. The attachment relationships within the uncertain subgraphs are locally frozen, and standardized sample library records are generated for other attachment relationships.

2. The method for interoperability processing between a standardized sample library and a heterogeneous testing data platform according to claim 1, characterized in that, The feasibility constraints include time feasibility constraints, platform feasibility constraints, resource feasibility constraints, target feasibility constraints, and spectrum feasibility constraints; the time feasibility constraint is that the timestamp of the test result is later than the timestamp of the upstream processing event. The platform feasibility constraint is that the candidate node appears in the on-machine event record of the corresponding platform; the resource feasibility constraint is that the candidate node belongs to the same batch of sample nodes loaded in the same plate position and well position within the corresponding time period; the target feasibility constraint is that the target of the observation node falls within the coverage of the target panel to which the candidate node belongs; the lineage feasibility constraint is that the test result corresponding to the observation node falls within the lineage branch of its possible source.

3. The method for interoperability processing between a standardized sample library and a heterogeneous testing data platform according to claim 1, characterized in that, The consistency factors include temporal consistency factors, lineage transmission consistency factors, platform process compatibility consistency factors, resource exclusivity consistency factors, and logical continuity factors. The temporal consistency factor is determined based on the standardized time difference between the timestamp of the observation node and the timestamp of the sample node. The lineage transmission consistency factor is determined based on the reachability of the upstream ancestor node and downstream descendant node of the sample node in the lineage graph after it is attached. The platform process compatibility consistency factor is determined based on whether the candidate node has completely undergone all the necessary processing steps of the corresponding platform. The resource exclusivity consistency factor is determined based on whether the candidate node has resource conflicts with other observation nodes in the same resource location during the same time period. The logical continuity factor is determined based on whether the time order of the re-inspection observation node and the corresponding initial inspection observation node meets the business rules.

4. The method of claim 3, wherein the standardized sample library and the heterogeneous testing data platform are interoperable. The factor graph model includes variable nodes, factor nodes, and edges; the variable nodes correspond to the affiliation selection of each observation node, and their values ​​range from sample nodes in its candidate attachment set to null values, where a null value indicates that the observation node is not attached to any sample node; the factor nodes correspond to various consistency constraints; edges connect variable nodes to factor nodes that constrain their values; the objective function of the global solution is to maximize the sum of weighted consistency factors of all candidate attachment relationships.

5. The method of claim 1, wherein the standardized sample library and the heterogeneous testing data platform are interoperable. Identifying the uncertain subgraph includes: traversing all connection relationships to detect contradictory situations, constructing a contradictory relationship graph based on the detected contradictory relationships, performing connected component analysis on the contradictory relationship graph, and marking connected components whose ratio of the number of contradictory edges to the number of observed nodes in the corresponding connected component exceeds a preset density threshold or whose average consistency score is lower than a preset score threshold as uncertain subgraphs.

6. The method for interoperability processing between a standardized sample library and a heterogeneous testing data platform according to claim 1, characterized in that, The partial freezing of the connections within the uncertain subgraph includes: preventing all connections within the uncertain subgraph from entering the standard release process and instead placing them in a confirmation queue; re-evaluating the deterministic state of the uncertain subgraph after some connections within the uncertain subgraph are confirmed through manual confirmation or upstream data supplementation; and releasing the frozen state of the uncertain subgraph and incorporating it into the standard release process if the confirmed connections reduce the uncertainty of the entire uncertain subgraph to below a threshold.

7. The method for interoperability processing between a standardized sample library and a heterogeneous testing data platform according to claim 1, characterized in that, The generated result records include: generating standardized records for observation nodes that have been confirmed to be attached but do not belong to the uncertain subgraph; the standardized records include sample node identifiers, phylogenetic location information, attachment evidence summaries, consistency score intervals, and uncertainty markers; generating unconfirmed records for observation nodes that are within the uncertain subgraph and marking the uncertainty markers as unconfirmed.

8. A standardized sample library interoperability processing system with a heterogeneous assay data platform, characterized by, include: The event standardization module is used to map the raw data of experimental events in various heterogeneous systems into event quadruple structures to obtain a set of standardized event primitives. The event quadruple structure includes event type, event timestamp, event object, and event resource. The genealogy construction module is used to construct a directed graph of sample genealogy based on the standardized event primitive set. The directed graph of sample genealogy uses sample entities and their intermediate forms as nodes and processing relationships as directed edges. The observation graph construction module is used to construct a heterogeneous observation graph by using the test result data from various platforms as observation nodes. The observation nodes record platform type, target, timestamp and resource location information. The candidate set generation module is used to generate a candidate attachment set for each observation node based on feasibility constraints, wherein the candidate nodes in the candidate attachment set correspond to the sample nodes in the directed graph of the sample spectrum. The global solution module is used to construct a factor graph model based on the candidate attachment set and the consistency factor, and to perform a global solution on the factor graph model using a constrained optimization algorithm to obtain the optimal attachment selection for each observation node. The results processing module is used to detect contradictory relationships and identify uncertain subgraphs based on the results of the global solution, to locally freeze the attachment relationships within the uncertain subgraphs, and to generate standardized sample library records for other attachment relationships.

9. An electronic device, the electronic device comprising: include: At least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-7.

10. A storage medium having stored thereon computer program instructions, characterized in that, The computer program instructions can be executed by a processor to implement the method as described in any one of claims 1-7.

Images