Determination method and device of variation evidence item and electronic equipment

Abstract

The invention discloses a variation evidence item determination method and device and electronic equipment, and belongs to the field of variation interpretation. The variation evidence item determination method comprises the steps of extracting judgment content of at least one evidence item based on a target principle, and obtaining at least one evidence item standard knowledge item; representing the evidence item standard knowledge items to obtain an evidence item standard knowledge graph; identifying information recorded in the target literature set, and matching the information with an evidence item standard knowledge graph to obtain an evidence item knowledge graph; and based on the evidence item knowledge graph, determining a target evidence item corresponding to the target variation. According to the variation evidence item determination method, the evidence item determination efficiency and the judgment accuracy and precision are remarkably improved.

Description

Method, device and electronic device for determining variant evidence items Technical Field

[0001] The present application belongs to the field of variant interpretation, and in particular relates to a method, device and electronic device for determining variant evidence items. Background Art

[0002] Variant interpretation is a medical evidence-based process that classifies detected sequence variants in combination with clinical phenotypes, clinical knowledge bases, personalized medication guidelines, and literature to determine their association with diseases, drugs, and individual characteristics and phenotypes. In related technologies, the literature recording the target variant is retrieved from a large amount of published literature mainly through manual retrieval and reading one by one. The specific content recording the target variant is then further extracted from the retrieved literature to determine one or at least one evidence item of the target variant, and the type of target variant is determined based on the evidence item. This method requires manual repeated retrieval, reading, and judgment, which consumes a lot of manpower and time, is costly, and has low efficiency. In addition, manual judgment is not only prone to omissions of information, but also relies on personal professional knowledge reserves. Different people have different understandings of the same content, which will cause differences in results and is greatly affected by subjective influences, resulting in low final accuracy.

[0003] Summary of the Invention

[0004] The present application aims to solve at least one of the technical problems existing in the prior art. To this end, the present application proposes a method, device and electronic device for determining variant evidence items, which improves the efficiency of determining evidence items and the accuracy of judgment results.

[0005] In a first aspect, the present application provides a method for determining a variant evidence item, the method comprising:

[0006] Extracting the judgment content of at least one evidence item based on the target principle to obtain at least one standard knowledge entry of the evidence item;

[0007] Representing the evidence item standard knowledge items to obtain the evidence item standard knowledge graph;

[0008] Identify the information recorded in the target document set, match it with the standard knowledge graph of evidence items, and obtain the knowledge graph of evidence items;

[0009] Based on the evidence item knowledge graph, a target evidence item corresponding to the target variation is determined.

[0010] According to the method for determining variant evidence items of this application, the judgment content of at least one evidence item is extracted using the target principle, and the obtained standard knowledge entry of at least one evidence item is represented to construct an evidence item standard knowledge graph. The information recorded in the target document set is then inferred based on the evidence item standard knowledge graph to obtain an evidence item knowledge graph. In subsequent applications, the target evidence items corresponding to the required target variants can be quickly and accurately obtained from massive amounts of literature, with a wide range of application scenarios. This eliminates the need for manual repeated searches, saving manpower and time costs, and significantly improving the efficiency of determining evidence items. Furthermore, it effectively avoids the influence of user subjective judgment on the final result, improves the uniformity of the judgment results, and has high judgment accuracy and precision.

[0011] According to one embodiment of the present application, representing the evidence item standard knowledge entries to obtain the evidence item standard knowledge graph includes:

[0012] Representing the evidence item standard knowledge entry with a triple as the smallest unit to obtain at least one evidence item standard knowledge representation entry;

[0013] Combine the at least one evidence item standard knowledge representation entry to obtain the evidence item standard knowledge graph.

[0014] According to one embodiment of the present application, the standard knowledge representation entry of the evidence item has the variant entity as the head node and the evidence item as the tail node.

[0015] According to one embodiment of the present application, in the evidence item knowledge graph, each node and edge except the tail node is represented by a specific title recorded in the target document set.

[0016] According to one embodiment of the present application, identifying the information recorded in the target document set, matching it with the standard knowledge graph of evidence items, and obtaining the knowledge graph of evidence items includes:

[0017] Acquire at least part of the entities and / or relationships included in the standard knowledge representation entry of the evidence item within the first target scope;

[0018] When the information recorded in the target document collection matches at least part of the entities and / or relationships, the standard knowledge representation entries of the evidence items are represented to construct the evidence item knowledge graph.

[0019] According to one embodiment of the present application, identifying the information recorded in the target document set, matching it with the standard knowledge graph of evidence items, and obtaining the knowledge graph of evidence items includes:

[0020] Based on the target document set, determining a sentence containing an entity and / or a relation of the first triple in a standard knowledge representation entry of an evidence item as an evidence sentence;

[0021] When other entities and / or relationships in the standard knowledge representation entries of the evidence item are obtained through contextual query around the evidence sentence, the standard knowledge representation entries of the evidence item corresponding to the other entities and / or relationships are transformed based on the specific names corresponding to the other entities and / or relationships recorded in the target document collection to generate the evidence item knowledge graph.

[0022] According to one embodiment of the present application, identifying the information recorded in the target document set, matching it with the standard knowledge graph of evidence items, and obtaining the knowledge graph of evidence items includes:

[0023] determining an evidence sentence based on the type of the standard knowledge representation entry of the evidence item;

[0024] The evidence item knowledge graph is generated based on the evidence sentence and the information recorded in the target document set.

[0025] According to one embodiment of the present application, extracting the judgment content of at least one evidence item based on the target principle to obtain at least one standard knowledge entry of the evidence item includes:

[0026] Based on the target principle, extracting the judgment content of the at least one evidence item based on entities and relationships between entities to obtain the at least one evidence item standard knowledge item;

[0027] The entity includes at least one of a variant entity, a disease entity, a patient entity, an indicator entity, a function entity and an evidence item.

[0028] In a second aspect, the present application provides a device for determining a variant evidence item, the device comprising:

[0029] A first processing module is configured to extract the judgment content of at least one evidence item based on a target principle to obtain at least one standard knowledge entry of the evidence item;

[0030] A second processing module is configured to represent the standard knowledge items of the evidence item and obtain a standard knowledge graph of the evidence item; the number of items included in the standard knowledge graph of the evidence item matches the standard knowledge items of the evidence item;

[0031] The third processing module is used to identify the information recorded in the target document set, match it with the standard knowledge graph of evidence items, and obtain the knowledge graph of evidence items;

[0032] The fourth processing module is used to determine the target evidence item corresponding to the target variation based on the evidence item knowledge graph.

[0033] According to the device for determining variant evidence items of the present application, the judgment content of at least one evidence item is extracted through the target principle, the obtained standard knowledge entry of at least one evidence item is represented to construct the standard knowledge graph of the evidence item, and the information recorded in the target document set is inferred based on the standard knowledge graph of the evidence item to obtain the knowledge graph of the evidence item. In the subsequent application process, the target evidence item corresponding to the required target variation can be quickly and accurately obtained from a large amount of literature, and the application scenarios are wide; there is no need for manual repeated retrieval, saving manpower and time costs, and significantly improving the efficiency of determining the evidence item; in addition, it effectively avoids the influence of the user's subjective judgment on the final result, improves the uniformity of the judgment result, and has high judgment accuracy and precision.

[0034] In a third aspect, the present application provides an electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein when the processor executes the computer program, the method for determining a variant evidence item as described in the first aspect above is implemented.

[0035] In a fourth aspect, the present application provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for determining a variant evidence item as described in the first aspect above.

[0036] In a fifth aspect, the present application provides a chip, which includes a processor and a communication interface, wherein the communication interface is coupled to the processor, and the processor is used to run a program or instruction to implement the method for determining a variant evidence item as described in the first aspect.

[0037] In a sixth aspect, the present application provides a computer program product, comprising a computer program, which, when executed by a processor, implements the method for determining a variant evidence item as described in the first aspect above.

[0038] The above one or at least one technical solution in the embodiments of the present application has at least one of the following technical effects:

[0039] The system extracts the judgment content of at least one evidence item based on the target principle, represents the obtained standard knowledge entry of at least one evidence item, and constructs a standard knowledge graph for the evidence item. Based on the standard knowledge graph, it then infers the information recorded in the target document set to obtain the evidence item knowledge graph. In subsequent applications, this method can quickly and accurately obtain the target evidence items corresponding to the target variants from a large amount of literature, with a wide range of applications. It eliminates the need for repeated manual searches, saving manpower and time, and significantly improving the efficiency of evidence item determination. Furthermore, it effectively avoids the influence of user subjective judgment on the final result, improves the uniformity of the judgment results, and achieves high judgment accuracy and precision.

[0040] Furthermore, by setting the first target range, the matching range can be narrowed to increase the matching rate, thereby improving the construction efficiency of the evidence item knowledge graph without affecting the construction accuracy of the evidence item knowledge graph.

[0041] Furthermore, by transforming the standard knowledge representation entries of evidence items corresponding to sentences other than evidence sentences, and adding the transformed results to the evidence item knowledge graph, the coverage of the final evidence item knowledge graph can be effectively broadened, the association relationship between entities can be enriched, and the functional attributes of the evidence item knowledge graph can be enhanced, thereby further improving the accuracy and precision of the final judgment results.

[0042] Furthermore, by setting different ways of confirming evidence sentences according to standard knowledge representation entries of different types of evidence items, it is possible to quickly query related content, reduce noise, and improve the efficiency of constructing the knowledge graph of evidence items.

[0043] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become obvious from the description below, or will be learned through practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the description of the embodiments in conjunction with the following drawings, in which:

[0045] FIG1 is a flow chart of a method for determining a variation evidence item provided in an embodiment of the present application;

[0046] FIG2 is a schematic diagram of the structure of a device for determining a variation evidence item provided in an embodiment of the present application;

[0047] FIG3 is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0048] The following will be combined with the accompanying drawings in the embodiments of the present application to clearly describe the technical solutions in the embodiments of the present application. Obviously, the embodiments described are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field are within the scope of protection of this application.

[0049] The terms "first," "second," and the like in the specification and claims of this application are used to distinguish similar objects, and are not used to describe a specific order or precedence. It should be understood that the terms used in this manner are interchangeable where appropriate, so that the embodiments of this application can be implemented in an order other than that illustrated or described herein, and that the objects distinguished by "first," "second," and the like are generally of the same type, and do not limit the number of objects; for example, the first object can be one or more. In addition, the term "and / or" in the specification and claims refers to at least one of the connected objects, and the character " / " generally indicates that the objects connected are in an "or" relationship.

[0050] The following, in combination with the accompanying drawings, describes in detail the method for determining a variant evidence item, the device for determining a variant evidence item, the electronic device, and the readable storage medium provided in the embodiments of the present application through specific embodiments and their application scenarios.

[0051] The method for determining the variation evidence item may be applied to a terminal, and may be specifically executed by hardware or software in the terminal.

[0052] The terminal includes but is not limited to portable communication devices such as mobile phones or tablet computers. It should also be understood that in some embodiments, the terminal may not be a portable communication device, but a desktop computer.

[0053] It will be appreciated that the terminal may include one or more other physical user interface devices, such as a touch-sensitive surface, a physical keyboard, a mouse, and a joystick.

[0054] The method for determining a variant evidence item provided in an embodiment of the present application may be executed by an electronic device or a functional module or functional entity in the electronic device that can implement the method for determining a variant evidence item. The electronic devices mentioned in the embodiment of the present application include but are not limited to mobile phones, tablet computers, computers, cameras, and smart devices. The method for determining a variant evidence item provided in an embodiment of the present application is described below using an electronic device as an example of an execution subject.

[0055] As shown in FIG1 , the method for determining the variation evidence item includes steps 110 , 120 , 130 and 140 .

[0056] Step 110: extracting the judgment content of at least one evidence item based on the target principle to obtain at least one standard knowledge item of the evidence item;

[0057] In this step, the target principles may include: principles proposed by authoritative organizations or institutions in the industry or principles that are widely recognized, as well as user-defined principles.

[0058] Among them, the principles proposed by authoritative organizations or institutions in the industry or that are widely recognized may include but are not limited to: industry standards, expert consensus, guidelines, and expert recommendations, etc., and this application does not limit them.

[0059] Industry standards are relevant standards for the genetic testing industry.

[0060] For example, industry standards can be authoritative genetic variation classification standards and guidelines, such as the genetic variation classification standards and guidelines issued by the American College of Medical Genetics and Genomics (ACMG) in 2015 (Richards S, Aziz N, Bale S, et al. Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology [J]. Genetics in medicine, 2015, 17(5): 405-423.). This guideline defines 28 variant classification criteria, i.e., evidence items, and based on this, determines the joint standard rules for genetic variation classification, which ultimately classify sequence variants into types such as pathogenic, likely pathogenic, benign, likely benign, and of unknown significance.

[0061] Expert consensus refers to information such as the consensus reached by relevant experts in the industry.

[0062] Evidence items are criteria used to determine the type of target variation.

[0063] In some embodiments, step 110 may include:

[0064] Based on the target principle, extract the judgment content of at least one evidence item based on entities and relationships between entities to obtain at least one standard knowledge item of the evidence item;

[0065] The entity includes at least one of a variant entity, a disease entity, a patient entity, an indicator entity, a function entity, and an evidence item.

[0066] In this embodiment, the evidence item standard knowledge entry is composed of at least two entities and relationships between the entities.

[0067] The entity includes at least one of a variant entity, a disease entity, a patient entity, an indicator entity, a function entity, and an evidence item.

[0068] Among them, the variant entity is the site that differs from the reference genome.

[0069] The naming method can refer to the naming rules in the industry.

[0070] For example, for a certain differential site, the specific name of the variant entity can be MSH2c.2503A>C, MSH2p.Asn835His; among them, MSH2 is the gene name, c.2503A>C, and p.Asn835His respectively represent the changes caused by the differential site at the cDNA level and protein level.

[0071] Disease entities are medically recognized diseases.

[0072] Standardized names of disease entities can be obtained from various public databases, such as https: / / www.orpha.net / consor / cgi-bin / index.php and https: / / rarediseases.info.nih.gov / .

[0073] A patient entity is an individual or group of people who have been clinically diagnosed with a disease.

[0074] Become can include: patient, proband and case, etc.

[0075] The indicator entity is an evaluation indicator represented by a numerical value or a numerical range.

[0076] For example, it may be an OR value (odds ratio) or a CI (confidence interval).

[0077] Functional entities are the objects pointed to by functional indicators in experimental research.

[0078] For example, if the activity of a protein is reduced, the corresponding functional entity is the protein activity; or if the transcript of a gene is truncated, the corresponding functional entity is the transcript.

[0079] Evidence items are criteria used to determine the type of target variation.

[0080] In some embodiments, descriptions of each entity and definition requirements may also be written into the evidence item standard knowledge representation entry as attributes of the entity.

[0081] In this embodiment, the attributes corresponding to the disease entity may include: the correlation between the disease and the gene, the correlation between the disease and the phenotype, the disease penetrance and inheritance mode, etc.

[0082] The attributes corresponding to the indicator entity can be in a preset range.

[0083] During the actual execution process, a configuration library can be set to retrieve the attributes corresponding to the required entities. These attributes are the attribute contents that need to be obtained by querying or integrating public information.

[0084] For example, a configuration library may be set up for a disease entity, and the content recorded in the configuration library includes but is not limited to: the correlation between genes and diseases, the correlation between diseases and phenotypes, disease penetrance and inheritance mode, etc.

[0085] The configuration library may be derived from one or at least one public database, for example from https: / / www.omim.org / .

[0086] According to the method for determining variant evidence items provided in the embodiments of the present application, based on the target principle, the judgment content of at least one evidence item is extracted based on entities and the relationship between entities to obtain at least one standard knowledge entry for the evidence item. By structuredly representing the target principle to generate a standard knowledge entry for the evidence item, it is used for the subsequent construction of the knowledge graph, which helps to improve the accuracy and precision of the knowledge graph.

[0087] Step 120: Represent the evidence item standard knowledge items to obtain the evidence item standard knowledge graph;

[0088] In this step, the standard knowledge graph of evidence items is used to provide a logical path for evidence item judgment for variant interpretation, which can achieve rapid reasoning and response of knowledge.

[0089] It can be understood that at least one evidence item standard knowledge entry can be obtained through step 110, and the relevant evidence item standard knowledge entries in the at least one evidence item standard knowledge entry are represented by a knowledge graph to obtain the evidence item standard knowledge graph.

[0090] The number of entries included in the evidence item standard knowledge graph matches the evidence item standard knowledge entries.

[0091] In some embodiments, step 120 may include:

[0092] Representing the evidence item standard knowledge entry with a triple as the smallest unit to obtain at least one evidence item standard knowledge representation entry;

[0093] At least one evidence item standard knowledge representation entry is combined to obtain an evidence item standard knowledge graph.

[0094] In this embodiment, the triples are knowledge graph triples.

[0095] A triplet includes the nodes corresponding to two adjacent entities and the relationship between the two adjacent entities.

[0096] The standard knowledge representation entry of an evidence item is a representation obtained by connecting the relevant partial evidence item standard knowledge entries with triples as the smallest unit.

[0097] Taking the genetic variation classification standards and guidelines as an example, the specific content of the evidence item standard knowledge representation entry is as follows:

[0098] The first evidence item standard knowledge representation entries may include: pathogenicity supporting evidence (PP1, PP5), benign strong evidence (BS4), and benign supporting evidence (BP5, BP6).

[0099] The standard knowledge representation entries of the second evidence item may include: strong evidence of pathogenicity (PS3, PS4), supporting evidence of pathogenicity (PP4), and strong evidence of benignity (BS3).

[0100] The standard knowledge representation entries of the third evidence item may include: strong pathogenicity evidence (PS2), moderate pathogenicity evidence (PM3, PM6) and benign supporting evidence (BP2).

[0101] The representation items are designed in a targeted manner according to the characteristics of each evidence item, providing a reasonable and efficient path for the reasoning of evidence items.

[0102] An evidence item standard knowledge representation entry may include one or more triples.

[0103] In the case of including multiple triples, the multiple triples are connected based on a certain structure.

[0104] The structures of knowledge representation entries of different evidence items can be different.

[0105] It can be understood that different evidence item standard knowledge representation entries can be obtained based on different evidence item standard knowledge entries.

[0106] The evidence item standard knowledge graph is composed of multiple evidence item standard knowledge representation entries.

[0107] After obtaining at least one different evidence item standard knowledge representation entry, at least one different evidence item standard knowledge representation entry is associated and combined to obtain the evidence item standard knowledge graph.

[0108] In the actual implementation process, when representing the standard knowledge items of each evidence item, they can be processed one by one or in batches.

[0109] In some embodiments, the standard knowledge representation entry of the evidence item has the variant entity as the head node and the evidence item as the tail node.

[0110] In this embodiment, the variant entity is a site that differs from the reference genome.

[0111] Evidence items are criteria used to determine the type of variation.

[0112] The relationship between nodes is the relationship between the entities corresponding to two adjacent nodes.

[0113] In some embodiments, the evidence item standard knowledge entry may further include at least one intermediate node.

[0114] In this embodiment, the intermediate node corresponds to the intermediate entity, and the intermediate entity may include: variation entity, disease entity, patient entity, indicator entity, and function entity, etc.

[0115] For example, the first evidence item standard knowledge representation entry may include: a head node representing the variant entity, a tail node representing the evidence item, and an edge representing the relationship between the variant entity and the evidence item;

[0116] The second evidence item standard knowledge representation entry may include: a head node represented by a variant entity, one or more intermediate nodes represented by intermediate entities, a tail node represented by an evidence item, and edges represented by relationships between the nodes, and each intermediate node has only one parent node and / or child node;

[0117] The third evidence item standard knowledge representation entry may include: a head node represented by a variant entity, intermediate nodes represented by one or more intermediate entities, a tail node represented by an evidence item, and edges represented by the relationships between the nodes, wherein at least one intermediate node has two or more parent nodes and / or child nodes.

[0118] Among them, the parent node refers to the adjacent node along the direction of the path head node; the child node refers to the adjacent node along the direction of the path tail node.

[0119] It is understandable that the structures of different evidence item standard knowledge representation entries may be different, and the differences may include: nodes and the number of nodes, edges and the number of edges, and the number of triples.

[0120] During actual execution, each intermediate node may be sequentially determined as the first intermediate node, the second intermediate node, etc. in the direction from the head node to the tail node, and correspond to the first intermediate entity and the second intermediate entity.

[0121] For example, the standard knowledge representation entry A of the evidence item includes: a head node and a tail node; the standard knowledge representation entry B of the evidence item includes: a head node, a first intermediate node and a tail node; the standard knowledge representation entry C of the evidence item includes: a head node, a first intermediate node, a second intermediate node and a tail node.

[0122] In the direction from the head node to the tail node, the relationship between the mutated entity and the corresponding entity of the next node is the first relationship, represented as the first edge, and so on.

[0123] For example, for a standard knowledge representation entry A of a certain evidence item that only contains a head node and a tail node, the first relationship is the relationship between the variant entity and the evidence item, corresponding to the first edge in the standard knowledge representation entry of the evidence item;

[0124] For another evidence item standard knowledge representation entry B containing a head node, a first intermediate node, and a tail node, the first relationship is the relationship between the variant entity and the first intermediate entity, and the second relationship is the relationship between the first intermediate entity and the evidence item, which correspond to the first edge and the second edge in the evidence item standard knowledge representation entry, respectively;

[0125] For another evidence item standard knowledge entry C containing a head node, a first intermediate node, a second intermediate node, and a tail node, the first relationship is the relationship between the variant entity and the first intermediate entity, the second relationship is the relationship between the first intermediate entity and the second intermediate entity, and the third relationship is the relationship between the second intermediate entity and the evidence item, which correspond to the first edge, the second edge, and the third edge in the standard knowledge representation entry of the evidence item, respectively.

[0126] From the head node to the tail node, the first triplet and the second triplet are sequentially constructed according to the combination of each node and relationship.

[0127] For example, for a certain evidence item, the standard knowledge representation entry A only contains the head node and the tail node, and only contains the first triple, i.e., the head node, the tail node, and the first edge;

[0128] For another evidence item standard knowledge representation entry B including a head node, a first intermediate node, and a tail node, the first triplet includes the head node, the first intermediate node, and the first edge, and the second triplet includes the first intermediate node, the tail node, and the second edge.

[0129] Different standard knowledge entries of evidence items record different information, so the structure of the standard knowledge representation entries of the evidence items obtained after final representation may also be different.

[0130] For example, when the evidence item standard knowledge entry X consists of a variant entity, an evidence item, and the relationship between the variant entity and the evidence item, the evidence item standard knowledge graph is written with the variant entity as the head node, the evidence item as the tail node, and the relationship between the variant entity and the evidence item as the edge.

[0131] For example, when the standard knowledge entry Y of the evidence item is composed of a variant entity, a disease entity, an evidence item, the relationship between the variant entity and the disease entity, and the relationship between the disease entity and the evidence item, the variant entity is used as the head node, the disease entity is used as the middle node, the evidence item is used as the tail node, and the relationship between the variant entity and the disease entity, and the relationship between the disease entity and the evidence item are written into the standard knowledge graph of the evidence item as edges.

[0132] For example, taking pathogenicity supporting evidence PP1 as an example, it is defined in the genetic variation classification standards and guidelines as "the variant co-segregates with the disease in multiple affected family members, and the gene where the variant is located is considered to be pathogenic" (Cosegregation with disease in multiple affected family members in a gene definitively known to cause the disease).

[0133] Based on the definition, we can extract the PP1 evidence item standard knowledge entry. The entities it contains include variant entities and PP1 evidence items, and the relationships between variant entities and PP1 evidence items. We represent the PP1 evidence item standard knowledge entry as a triple, with the variant entity as the head node, the PP1 evidence item as the tail node, and the relationship between the variant entity and the PP1 evidence item as the edge. This PP1 evidence item standard knowledge representation entry is obtained. This evidence item standard knowledge representation entry belongs to the first evidence item standard knowledge representation entry. The corresponding first evidence item standard knowledge representation entry for PP1 is written into the evidence item standard knowledge graph.

[0134] Taking PS4 with strong evidence of pathogenicity as an example, it is defined in the genetic variation classification standards and guidelines as "The prevalence of the variant in affected individuals is significantly increased compared with the prevalence in controls."

[0135] According to the definition, the PS4 evidence item standard knowledge items are extracted. The entity content contained in them includes variant entities, indicator entities, and PS4 evidence items. The relationships contained in them include the relationship between variant entities and indicator entities, and the relationship between indicator entities and PS4 evidence items. The PS4 evidence item standard knowledge items are represented as triples, with the variant entity as the head node, the indicator entity as the first intermediate node, and the PS4 evidence item as the tail node. The relationship between the variant entity and the indicator entity is the first edge, and the relationship between the indicator entity and the PS4 evidence item is the second edge. This results in the PS4 evidence item standard knowledge representation items.

[0136] The indicator entity can be an OR value, CI, or a combination thereof. Furthermore, specific thresholds are set for the OR and CI values, for example, an OR value greater than 5 and a CI value range excluding 1. In other words, only when the indicator entity exists and meets the threshold requirements will the PS4 evidence item be inferred according to the path indicated by the standard knowledge representation entry of the PS4 evidence item.

[0137] This evidence item standard knowledge representation entry consists of two triples, includes an intermediate node, and has only one parent node and one child node. It belongs to the second evidence item standard knowledge representation entry. The second evidence item standard knowledge representation entry corresponding to PS4 is written into the evidence item standard value knowledge graph.

[0138] Taking strong pathogenicity evidence PS2 as an example, its definition in the genetic variation classification standards and guidelines is "a de novo variant detected in a patient with no family history, both parents confirmed negative, and the patient's family history meets the inheritance pattern of a de novo variant" (De novo (both maternity and paternity confirmed) in a patient with the disease and no family history). Based on this definition, standard knowledge entries for PS2 evidence items are extracted. The entities contained in them include: variant entity, patient entity, and PS2 evidence item. The relationships contained in them include the relationship between the variant entity and the patient entity, and the relationship between the patient entity and the PS2 evidence item. The standard knowledge representation entries for PS2 evidence items are represented as triples, with the variant entity as the head node, the patient entity as the first intermediate node, and the PS2 evidence item as the tail node. The relationship between the variant entity and the patient entity is the first edge, and the relationship between the patient entity and the PS2 evidence item is the second edge. This results in standard knowledge representation entries for PS2 evidence items.

[0139] According to the genetic variation classification standards and guidelines, for new variants, when the parent-child relationship is confirmed, it meets the evidence item PS2; when the parent-child relationship cannot be confirmed, it meets the evidence item PM6 (Assumed de novo, but without confirmation of paternity and maternity).

[0140] The same method was used to extract and represent the standard knowledge representation of PM6 evidence items and integrate them with the standard knowledge representation of PS2 evidence items. This is reflected in the presence of two child nodes after the patient entity in the standard knowledge representation of PS2 and PM6 evidence items, namely, the PS2 evidence item as the last node and the PM6 evidence item as the last node, which belongs to the third standard knowledge representation of evidence items.

[0141] Write the third evidence item standard knowledge representation entries corresponding to PS2 and PM6 into the evidence item standard value knowledge graph.

[0142] According to the method for determining the variant evidence item provided in the embodiment of the present application, the variant entity is used as the head node, the evidence item is used as the tail node, and the relationship between the nodes is used as the edge to connect the nodes in sequence to generate a standard knowledge entry for the evidence item, which has high accuracy and precision.

[0143] Step 130: Identify the information recorded in the target document set, match it with the standard knowledge graph of evidence items, and obtain the knowledge graph of evidence items;

[0144] In this step, the target document set may be any relevant one or more document sets.

[0145] The types of documents may include, but are not limited to, monographs, newspaper articles, journal articles, dissertations, research reports, standards, patents, monographs, and extracted documents from collections of papers, and this application does not limit these.

[0146] The information recorded in the target document set can be any one or more nodes or one or more relationships (including variant entities, disease entities, patient entities, indicator entities, and functional entities, etc.) in the evidence item knowledge graph other than evidence.

[0147] During the actual implementation process, one or more specific documents can be obtained through manual selection, for example, it can be any one or more documents; it can also be one or more documents that record the target variation that are manually retrieved; it can also be one or more documents that record related diseases that are manually retrieved; when the target document set consists of multiple articles, the target document set can also be a combination of the above multiple situations; this application does not limit the content of the target document set.

[0148] In some embodiments, when the target document set consists of multiple articles, all published documents may also be collected without any restrictions on the collection channels and methods. For example, the documents may be obtained from the copyright holders.

[0149] It is understandable that the acquired evidence item knowledge graphs are different depending on the target document set.

[0150] During the actual execution process, the number of entries included in the evidence item standard knowledge graph matches the evidence item standard knowledge entries; and the number of evidence item standard knowledge entries should be greater than or equal to the entries included in the evidence item knowledge graph finally obtained.

[0151] During the actual implementation process, the nodes and relationships involved in the standard knowledge graph of the evidence item can be matched with the content in the target document set. If the match is successful, the nodes and relationships in the standard knowledge graph of the evidence item are transcribed using the expressions in the target document set to obtain the evidence item knowledge graph.

[0152] For example, for a standard knowledge entry of an evidence item, it can be identified and extracted based on the entities and edges recorded in the standard knowledge entry of the evidence item. When all entities and edges can be matched, the information in the matching target document set is used to represent the entities and edges to obtain the evidence item knowledge graph.

[0153] It should be noted that the evidence item knowledge graph constructed based on step 130 changes based on the changes in the actual recorded information in the target document set. The evidence item knowledge graph may include one or more variations (variations of the header nodes) to objectively present all identifiable and matching content recorded in the target document set.

[0154] For different information recorded in the target document set, different evidence item knowledge graphs can be constructed.

[0155] During the actual execution process, multiple evidence item knowledge graphs can be constructed based on the method of steps 110 to 130.

[0156] The multiple evidence item knowledge graphs constructed can be stored in a local or cloud database and can be called when needed.

[0157] In some embodiments, in the evidence item knowledge graph, each node except the tail node and the edge are represented by a specific title recorded in the target document set.

[0158] In this embodiment, the evidence item knowledge graph is composed of evidence item representation entries, and the evidence item representation entries are converted from evidence item standard knowledge representation entries.

[0159] It can be understood that if the record information of all documents in the target document set does not match the information of any evidence item standard knowledge representation entry, it is considered that it does not meet the requirements of all evidence items, and the evidence item representation entry contained in the evidence item knowledge graph is 0.

[0160] It should be noted that for the same target variation, the knowledge graph of the evidence item may be different in different situations, and this difference is closely related to the target document set.

[0161] For example, when the information recorded in the target document set does not match any standard knowledge entry of the evidence item, the evidence item representation entry in the evidence item knowledge graph is 0.

[0162] For example, when the target document set is updated (such as added or changed), the information recorded therein matches a standard knowledge entry of a certain evidence item, that is, it meets the requirements of the evidence item, then the corresponding evidence item standard knowledge representation entry can be converted into an evidence item representation entry and written into the evidence item knowledge graph.

[0163] When constructing evidence item knowledge graphs for at least one different information, the evidence item knowledge graphs corresponding to the different information may be different.

[0164] For example, if the information A recorded in the target document set is consistent with the standard knowledge entry X of the evidence item, that is, it meets the requirements of the evidence item, the corresponding standard knowledge representation entry X of the evidence item can be converted into an evidence item representation entry and written into the evidence item knowledge graph.

[0165] For example, if the information B recorded in the target document set does not match the standard knowledge entry X of the same evidence item, that is, it does not meet the requirements of the evidence item, then the evidence item knowledge graph does not contain the evidence item representation entry corresponding to the standard knowledge entry X of the evidence item.

[0166] In some embodiments, when there is at least one piece of information, the specific construction process can be to operate on the at least one piece of information one by one, or to operate on the at least one piece of information in batches. A specific example of operating on the at least one piece of information in batches can be to generate an evidence item knowledge graph for all variations recorded in the target document set.

[0167] It is understandable that in the document collection, the specific names of various entities have the characteristics of diverse description forms, and there are also a large number of aliases and non-standard descriptions.

[0168] The specific names of the entities have been described in the above embodiments and will not be repeated here.

[0169] Step 140: Determine the target evidence item corresponding to the target variation based on the evidence item knowledge graph.

[0170] In this step, the target variation is a variation to be interpreted; for example, it can be a variation of interest or an evidence item to be determined, etc., which is not limited in this application.

[0171] The target evidence item refers to the evidence item determined based on the constructed evidence item knowledge graph, including: the specific evidence items included in the constructed evidence item knowledge graph, and null values.

[0172] Among them, the specific evidence item is the evidence item that meets the requirements in the constructed evidence item knowledge graph;

[0173] A null value means there is no matching evidence item.

[0174] According to the method for determining variant evidence items of the present application, the judgment content of at least one evidence item is extracted through the target principle, the obtained standard knowledge entry of at least one evidence item is represented to construct the standard knowledge graph of the evidence item, and the information recorded in the target document set is inferred based on the standard knowledge graph of the evidence item to obtain the knowledge graph of the evidence item. In the subsequent application process, the target evidence items corresponding to the required target variants can be quickly and accurately obtained from massive documents, and the application scenarios are wide; there is no need for manual repeated retrieval, saving manpower and time costs, and significantly improving the efficiency of determining evidence items; in addition, it effectively avoids the influence of user subjective judgment on the final result, improves the uniformity of the judgment result, and has high judgment accuracy and precision.

[0175] In some embodiments, the method may further include: updating the target document set.

[0176] In this embodiment, an update mode may be set, such as real-time update or periodic update, which is not limited in this application.

[0177] The evidence item standard knowledge graph includes at least one evidence item standard knowledge representation entry.

[0178] The evidence item knowledge graph includes at least one evidence item representation entry.

[0179] It should be noted that the number of evidence item representation entries included in the evidence item knowledge graph can be any non-negative integer value, for example, it can be 0, 1 or more, and this application does not limit it.

[0180] When the information corresponding to the target variation recorded in the target document set matches the evidence item standard knowledge representation entry in the evidence item standard knowledge graph, the evidence item standard knowledge representation entry is represented based on the specific title recorded in the target document set to construct the evidence item knowledge graph.

[0181] In other embodiments, when the record information of all documents in the target document set does not match the information of any evidence item standard knowledge representation entry, it is considered that the requirements of all evidence items are not met.

[0182] In some embodiments, step 130 may further include:

[0183] Acquire at least part of the entities and / or relations included in the standard knowledge representation entries of the evidence items within the first target scope;

[0184] When the information recorded in the target document set matches at least some entities and / or relationships, the standard knowledge representation entries of the evidence items are represented to construct an evidence item knowledge graph.

[0185] In this embodiment, the first target range is a preset range, which can be user-defined, such as being defined as within 7 sentences of the same paragraph.

[0186] At least some of the entities and / or relationships may be all entities and / or relationships included in the standard knowledge representation entry of the evidence item within the first target scope, and these all entities and / or relationships do not include the tail node.

[0187] For example, all entities and relationships of the evidence item standard knowledge representation entry are queried within a preset scope; if all entities and relationships are included, the evidence item standard knowledge representation entry is converted into an evidence item representation entry and written into the evidence item knowledge graph.

[0188] According to the method for determining variant evidence items provided in the embodiments of the present application, by setting a first target range, the matching range can be narrowed to increase the matching rate, thereby improving the construction efficiency of the evidence item knowledge graph without affecting the construction accuracy of the evidence item knowledge graph.

[0189] In some embodiments, step 130 may include:

[0190] Based on the target document set, a sentence containing the entity and / or relation of the first triple in the standard knowledge representation entry of the evidence item is determined as an evidence sentence;

[0191] When other entities and / or relationships in the standard knowledge representation entries of the evidence items are obtained through contextual queries around the evidence sentence, the standard knowledge representation entries of the evidence items corresponding to the other entities and / or relationships are transformed based on the specific names corresponding to the other entities and / or relationships recorded in the target document collection to generate an evidence item knowledge graph.

[0192] In this embodiment, the first triplet is the triplet where the head node is located.

[0193] The entities and / or relations of the first triplet may include multiple modes such as only entities, only relations, and a combination of relations and entities.

[0194] Among them, the entities and / or relationships of the first triplet can be: a variant entity, or a first intermediate entity, or a combination of a variant entity and a first intermediate entity, or a first relationship, or a combination of the first relationship and a variant entity, or a combination of the first relationship and the first intermediate entity.

[0195] For example, the entity may be a variant entity and / or a first intermediate entity, and the relationship may be a first relationship.

[0196] Other entities and / or relations are relations and / or entities (except for evidence items) that are not used in the determination of the evidence sentence.

[0197] Other entities and / or relations may include relations and / or entities not used in the first triple in the standard knowledge representation entry of the evidence item; or may include relations and / or entities not used in other triples in the standard knowledge representation entry of the evidence item.

[0198] For example, the standard knowledge representation entry A of the evidence item includes the first triple. When the evidence sentence is determined using only the variant entity of the first triple, it is still necessary to query in the context of the evidence sentence whether there is matching content for the first relation that is not used in the first triple.

[0199] For example, the standard knowledge representation entry B of the evidence item includes: the first triplet and the second triplet. When the evidence sentence is determined using only the first relationship of the first triplet, it is necessary to query whether there is matching content in the context of the evidence sentence among the variant entity, the first intermediate entity, and the second relationship.

[0200] If the query obtains matching content, the standard knowledge representation entries of the evidence items corresponding to the variant entity, the first intermediate entity and the second relationship with matching content are transformed to obtain the evidence item knowledge entries corresponding to the variant entity, the first intermediate entity and the second relationship, and these evidence item knowledge entries are added to the evidence item knowledge graph.

[0201] According to the method for determining variant evidence items provided in the embodiments of the present application, by converting the standard knowledge representation entries of evidence items corresponding to sentences other than evidence sentences, and adding the converted results to the evidence item knowledge graph, the coverage of the final evidence item knowledge graph can be effectively broadened, the association relationships between entities can be enriched, and the functional attributes of the evidence item knowledge graph can be enhanced, thereby further improving the accuracy and precision of the final judgment results.

[0202] In some embodiments, querying the context around the evidence sentence to obtain other entities and / or relationships in the standard knowledge representation entry of the evidence item may include: querying the context around the evidence sentence to obtain other entities and / or relationships in the standard knowledge representation entry of the evidence item within the second target scope.

[0203] In this embodiment, the second target range may be a preset range. For example, the second target range may be set to the evidence sentence and 3 or 5 sentences before and after the evidence sentence in the same paragraph, etc., which is not limited in this application.

[0204] For example, during the actual execution process, you can also search for the first sentence, the second sentence, and the third sentence in the same paragraph context in sequence, and stop searching when relevant content is matched.

[0205] In some embodiments, the document information (such as the document unique identifier PMID, document title, document publication date, etc.), evidence sentences, and target range statements that record the variant entity (head node) can also be used as attributes of the variant entity and written into the evidence item knowledge graph.

[0206] Optionally, the variant entities and their attributes included in the literature collection are aggregated to generate a corresponding knowledge base for easy retrieval and can be used for query and verification.

[0207] During the actual execution process, the sentence of the entity and / or relationship corresponding to the triple of the head node can be determined as the evidence sentence, and other entities and / or relationships in the standard knowledge representation entry of the evidence item can be queried in the context around the evidence sentence. If other entities and / or relationships are included, the standard knowledge representation entry of the evidence item is converted into an evidence item representation entry and written into the evidence item knowledge graph.

[0208] It should be noted that when determining that a sentence containing the entity and / or relationship of the first triple in the standard knowledge representation entry of the evidence item is an evidence sentence and when querying other entities and / or relationships in the standard knowledge representation entry of the evidence item in the context around the evidence sentence, both involve operations on at least one of the entities and relationships. In the actual execution process, the entities and relationships need to be identified and extracted respectively.

[0209] In this application, there is no limitation on the method of identifying the entity.

[0210] For example, recognition can be performed by searching a specific vocabulary, or by using a pre-trained model.

[0211] First, use a specific vocabulary for entity recognition

[0212] When using a specific vocabulary for entity recognition, the source of the vocabulary may include but is not limited to literature and public databases.

[0213] For example, for the patient entity, common descriptions can be collected from the literature to build a vocabulary.

[0214] Descriptions of patients can include: patient, case, individual, and object, etc.

[0215] In some embodiments, the vocabulary can also be enriched according to part-of-speech conversion rules.

[0216] In some embodiments, synonyms, near-synonyms, and antonyms may also be used to expand the vocabulary.

[0217] Second, recognition through pre-training model

[0218] When using a pre-trained model for entity recognition, there is no restriction on the type of the pre-trained model, such as a neural network model.

[0219] Taking variant entities as an example, we can use documents with variant annotation information to train the model to obtain a pre-trained model that can identify variant entities, and then use this pre-trained model to identify variant entities in other documents.

[0220] In some embodiments, after identifying entities and / or relationships, the identified specific titles may be aligned.

[0221] In this embodiment, the identified specific titles may be aligned to align the identified entity titles to the corresponding entity standard terms.

[0222] For example, the standard terminology for variant entities can refer to the naming conventions published by the Human Genome Variation Society (HGVS, https: / / www.hgvs.org / ). There are no restrictions on the entity alignment method, and regular expressions can be used, for example.

[0223] In this application, there is no limitation on the method of relationship identification.

[0224] For example, the identification can be performed by establishing a vocabulary library for the related keywords.

[0225] Specifically, sources of relationship keywords may include but are not limited to literature and public databases.

[0226] Taking the PS3 evidence item "in vivo and in vitro functional experiments have clearly shown that mutations can lead to impaired gene function" as an example, we can collect common descriptions of the keyword "impaired" in scientific research literature to establish a vocabulary library, which may include impaired, aberrant, defective, and disrupt.

[0227] In some embodiments, the vocabulary can also be enriched according to part-of-speech conversion rules.

[0228] In some embodiments, synonyms, near-synonyms, and antonyms may also be used to expand the vocabulary.

[0229] In some embodiments, step 130 may further include:

[0230] Determine the evidence sentence based on the type of the standard knowledge representation entry of the evidence item;

[0231] Based on the evidence sentences and the information recorded in the target document set, a knowledge graph of evidence items is generated.

[0232] In this embodiment, based on the type, at least one evidence item standard knowledge representation entry can be divided into a first evidence item standard knowledge representation entry, a second evidence item standard knowledge representation entry, a third evidence item standard knowledge representation entry, etc.

[0233] The entities and / or relations of the first triple include at least one of a first relation, a variant entity and a first intermediate entity.

[0234] For the first evidence item standard knowledge representation entry, the first relation can be used to determine the evidence sentence;

[0235] For the second evidence item standard knowledge representation entry, the evidence sentence may be determined using at least the first intermediate entity, and the evidence sentence may also be determined using the first relation or the variant entity located at the head node;

[0236] For the third evidence item standard knowledge representation entry, the first relationship can be used to determine the evidence sentence. According to the method for determining variant evidence items provided in the embodiment of the present application, by setting different ways to confirm evidence sentences according to the standard knowledge representation entries of different types of evidence items, it is possible to quickly query related content, reduce noise, and improve the efficiency of constructing the evidence item knowledge graph.

[0237] The following describes the process of constructing the knowledge graph of evidence items corresponding to relevant information through a specific embodiment.

[0238] The standard knowledge representation entries of evidence item 1 include: variant entity (head node), evidence item 1 (tail node), and relationship 1 between the variant entity and the evidence item (first edge).

[0239] The pre-trained model is used to identify relevant information for document 1 in the target document set, and the sentence containing relevant information is determined to be the evidence sentence. Within the scope of the context 1 of the evidence sentence, the relationship 1 between the variant entity and the evidence item is identified through the relationship keyword library of relationship 1, and the matching relationship keyword 1 is queried.

[0240] According to the standard knowledge representation entry of evidence item 1, it can be deduced that the record in document 1 meets the requirements of evidence item 1. Then the standard knowledge representation entry of evidence item 1 is transformed to form an evidence item representation entry with the variant entity as the head node, evidence item 1 as the tail node, and keyword 1 as the first edge. This evidence item representation entry is written into the evidence item knowledge graph of the target variant as part of the evidence item knowledge graph of the relevant information.

[0241] Similar methods can be used to obtain relevant information and other evidence items for other parts of document 1 and other documents in the document collection.

[0242] All relevant information evidence item representation entries obtained by traversing the document set can constitute the evidence item knowledge graph of the relevant information.

[0243] In actual application, the method provided in this application can be used to construct an evidence item knowledge graph for all or part of the evidence items based on industry standards or guidelines.

[0244] When only some of the evidence items are judged by constructing an evidence item knowledge graph, other evidence items can be judged using other methods.

[0245] During the actual implementation process, based on the genetic variation classification standards and guidelines, evidence item standard knowledge entries can be extracted for evidence items that need to be judged based on at least one of the patient and / or family records, classification result records, and functional verification records.

[0246] The evidence items of the standard knowledge items extracted include: strong evidence of pathogenicity (PS2, PS3, PS4), moderate evidence of pathogenicity (PM3, PM6), supporting evidence of pathogenicity (PP1, PP4, PP5), strong evidence of benignity (BS3, BS4) and supporting evidence of benignity (BP2, BP5, BP6).

[0247] The aforementioned standard knowledge entries of each evidence item are represented by triples as the smallest unit to form standard knowledge representation entries of each evidence item.

[0248] All the aforementioned evidence item standard knowledge representation entries are associated to form an evidence item standard knowledge graph.

[0249] For specific information, entities and / or relationships in the standard knowledge graph of evidence items are identified and extracted in the selected target document set.

[0250] If the entities and relationships in the standard knowledge representation entry of a certain evidence item all have matching content, it can be inferred that the content recorded in the target document set meets the requirements of the evidence item; then the entities and relationships in the standard knowledge representation entry of the evidence item are represented using the specific names recorded in the target document set and converted into the evidence item representation entry of the evidence item.

[0251] Traverse the selected target document set, associate all evidence items representing the target variation obtained, and form a knowledge graph of evidence items corresponding to the relevant information.

[0252] The method described in this embodiment can be used to accurately and efficiently obtain document record information and determine whether the information meets the evidence item requirements, and directly obtain the evidence item conclusion of the target variation in the selected target document set.

[0253] In some embodiments, for the evidence items recorded in the genetic variation classification standards and guidelines that need to be judged based on at least one of functional inference, computer prediction, and population records, the information source for the judgment can also be selected from one or more bioinformatics software prediction results, public database records, or a combination thereof.

[0254] Among them, functional inference is the inference of the functional impact of the mutant protein based on the known mutation records of the target gene.

[0255] The target gene is the gene where the mutation is located.

[0256] The known variants can be one or more and can be pathogenic variants and / or benign variants.

[0257] The known variation record includes: the sequence, position, segment, type and proportion of each type of the known variation.

[0258] Computer prediction refers to the use of bioinformatics to evaluate the impact of mutations on genes or gene products.

[0259] Impacts include harmful impacts and harmless impacts.

[0260] The evaluation includes at least one of: protein conservation evaluation, nucleic acid conservation evaluation and shear hazard evaluation.

[0261] Population records refer to the presence of variants in large populations.

[0262] The presence includes allele frequency, and may further include the number of homozygotes and / or hemizygotes.

[0263] In this embodiment, the evidence items judged based on at least one of functional inference, computer prediction and population records may include: very strong evidence of pathogenicity (PVS1), strong evidence of pathogenicity (PS1), moderate evidence of pathogenicity (PM1, PM2, PM4, PM5), supporting evidence of pathogenicity (PP2, PP3), independent evidence of benignity (BA1), strong evidence of benignity (BS1, BS2) and supporting evidence of benignity (BP1, BP3, BP4, BP7).

[0264] In some embodiments, functional inference evidence items may also include: very strong evidence of pathogenicity (PVS1), strong evidence of pathogenicity (PS1), moderate evidence of pathogenicity (PM1, PM4, PM5), supporting evidence of pathogenicity (PP2), and supporting evidence of benignity (BP1, BP3).

[0265] In some embodiments, the computer prediction evidence items may include: pathogenicity supporting evidence (PP3) and benignity supporting evidence (BP4, BP7).

[0266] In some embodiments, the crowd record evidence items may include pathogenic moderate evidence (PM2), benign independent evidence (BA1), and benign strong evidence (BS1, BS2).

[0267] In some embodiments, the evidence items obtained in the above embodiments can also be combined to classify the target variation according to the genetic variation classification standards and the joint standard rules for genetic variation classification formulated in the guidelines, and the classification result is selected from one of the five categories: pathogenic, possibly pathogenic, benign, possibly benign, and of unknown significance.

[0268] The following uses a specific variation as an example to illustrate the construction of the evidence item knowledge graph and its application in variation interpretation.

[0269] For example, in the actual execution process, the evidence item content of the target mutation c.417C>T of the TGFBI gene can be searched and judged in the target document Romero P, Moraga M, Herrera L. Different phenotypes of lattice corneal dystrophy type I in patients with 417C>T(R124C)and 1762A>G(H572R)mutations in TGFBI(BIGH3)[J]. Molecular Vision, 2010, 16: 1601.

[0270] In the target document, by matching the relational keyword library, we can identify the first relation in the standard knowledge representation entry of the PS2 evidence item, thereby determining the evidence sentence as "Paternity was confirmed, indicating the R124C mutation was a de novo mutation.", where "de novo" is the first relation in the standard knowledge representation entry of the PS2 evidence item.

[0271] Taking this evidence sentence as the center, within the scope of one sentence before and after it, the variant entity, patient entity and second relationship in the standard knowledge representation entry of PS2 evidence item are identified.

[0272] In the evidence sentence, the pre-trained model is used to identify the specific name of the target variant "R124C".

[0273] In the second sentence of the evidence sentence, “Individual I-1 of Family Three does not carry that mutation.”, the patient entity “individual” is identified through keyword library matching.

[0274] In the evidence sentence, the second relation “paternity was confirmed” is identified through relation keyword library matching.

[0275] By reasoning based on the standard knowledge representation items of PS2 evidence items, it can be found that the content recorded in this document is consistent with the PS2 evidence items, and the PS2 evidence items can be given.

[0276] The variant entity in the PS2 evidence item standard knowledge representation entry is described using "R124C" or the standard name of the target variant "TGFBI c.417C>T", the first relationship is described using "de novo", the patient entity is described using "individual", and the second relationship is described using "paternity was confirmed". The PS2 evidence item standard knowledge representation entry is converted into a PS2 evidence item knowledge representation entry.

[0277] It can be understood that, in this embodiment, the target document set only includes one selected document (ie, the target document), and the evidence item knowledge graph of the target variation TGFBI c.417C>T is composed of only one PS2 evidence item knowledge representation entry.

[0278] In other embodiments, the target variation TGFBI c.417C>T may also be searched in an established target document collection, which includes over 60,000 documents.

[0279] Among them, the supplementary literature is Romero P, Vogel M, Diaz JM, et al. Anticipation in familial lattice corneal dystrophy type I with R124C mutation in the TGFBI (BIGH3) gene [J]. Molecular Vision, 2008, 14: 829.

[0280] In this supplementary document, by matching the relational keyword library, the first relation in the standard knowledge representation entry of the PP1 evidence item can be identified, thereby determining the evidence sentence as "The R124C mutation in TGFBI cosegregated with LCD type I in the investigated family.", where "cosegregated" is the first relation in the standard knowledge representation entry of the PP1 evidence item.

[0281] In the evidence sentence, the pre-trained model is used to identify the specific name of the target variant "R124C".

[0282] By reasoning based on the standard knowledge representation items of the PP1 evidence item, it can be known that the content recorded in this document is consistent with the PP1 evidence item, and the PP1 evidence item can be given.

[0283] The variant entity in the PP1 evidence item standard knowledge representation entry is described using "R124C" or the standard name of the target variant "TGFBI c.417C>T", and the first relationship is described using "cosegregated", and the PP1 evidence item standard knowledge representation entry is converted into a PP1 evidence item knowledge representation entry.

[0284] It can be understood that, in this embodiment, the output PP1 evidence item knowledge representation entry can be combined with the output PS2 evidence item knowledge representation entry in the previous case to form a more comprehensive evidence item knowledge graph of the target variation TGFBI c.417C>T.

[0285] In addition, in the established literature collection, no matching content is identified for the exemplified PS4 evidence item, so no PS4 evidence item knowledge representation entry is output, and there is no PS4 evidence item knowledge representation entry in the evidence item knowledge graph of the target variant TGFBI c.417C>T.

[0286] In other embodiments, 22,000 hereditary tumor-related documents collected from public databases can be determined as the target document set, and manually interpreted and reviewed by interpretation experts.

[0287] By comparing the judgment results of pathogenicity supporting evidence PP1, strong pathogenicity evidence PS4, and strong pathogenicity evidence PS2 obtained using the method of this application with the aforementioned manual interpretation results, the F1 values ​​of each evidence item were 89.36%, 62.07%, and 80%, respectively.

[0288] For other evidence items of genetic variation classification standards and guidelines, as well as evidence items based on other industry standards or guidelines, the same method as this application can be adopted to extract evidence item standard entries, construct evidence item standard knowledge graphs, and construct evidence item knowledge graphs to achieve judgment of evidence items.

[0289] During the research and development process, the inventors discovered that in the actual practice of variant interpretation, it is necessary to comprehensively and extensively collect information on the target variant, sequentially determine whether each piece of information meets the requirements of each evidence item, and then determine the classification of the target variant based on all the evidence items. Of the 28 evidence items, nearly half of the evidence items rely on literature records for judgment. In other words, first, it is necessary to retrieve documents that record the target variant from a large amount of published literature. Second, it is necessary to further extract the specific content of the target variant from the retrieved documents. Then, based on the specific content, it is determined whether it meets one or at least one of the 28 evidence items. Finally, one or more evidence items for the target variant are determined.

[0290] In related technologies, evidence items are primarily determined through manual retrieval and individual reading and judgment. This method is not suitable for large volumes of rapidly updated documents. Furthermore, the process of determining whether each piece of information meets the requirements of each evidence item is cumbersome. Relying solely on manual judgment not only easily leads to omissions of information but also relies on individual expertise. Different people may interpret the same content differently, leading to discrepancies in the results. Furthermore, when different people retrieve the same document at different times, repeated manual retrieval, reading, and judgment are required, which is inefficient.

[0291] In this application, the judgment content of at least one evidence item is extracted through the target principle, and the obtained at least one evidence item standard knowledge entry is represented to construct an evidence item standard knowledge graph. Then, the information recorded in the target document set is inferred based on the evidence item standard knowledge graph to obtain the evidence item knowledge graph. The evidence item knowledge graph can be updated accordingly based on changes in the document, and is suitable for massive and rapidly updated documents.

[0292] In addition, in this application, the evidence items of the target variation are determined through the constructed evidence item knowledge graph, which can comprehensively, quickly, accurately and efficiently obtain the document record information and judge whether the information meets the evidence item requirements without the need for subjective judgment by the user, thereby avoiding the influence of subjective judgment on the judgment results and effectively avoiding the omission of information caused by manual methods; at the same time, it also lowers the judgment threshold, making the operation of determining the evidence items simpler and more convenient, and significantly improving the accuracy of the judgment results.

[0293] According to the method for determining variant evidence items provided in the embodiment of the present application, the judgment content of at least one evidence item is extracted through the target principle, the obtained standard knowledge entry of at least one evidence item is represented to construct the standard knowledge graph of the evidence item, and the information recorded in the target document set is inferred based on the standard knowledge graph of the evidence item to obtain the knowledge graph of the evidence item. In the subsequent application process, the target evidence items corresponding to the required target variants can be quickly and accurately obtained from massive documents, and the application scenarios are wide; there is no need for manual repeated retrieval, which saves manpower and time costs and significantly improves the efficiency of determining the evidence items; in addition, it effectively avoids the influence of the user's subjective judgment on the final result, improves the uniformity of the judgment result, and has high judgment accuracy and precision.

[0294] The method for determining a variant evidence item provided in the embodiment of the present application can be executed by a device for determining a variant evidence item. In the embodiment of the present application, the device for determining a variant evidence item executing the method for determining a variant evidence item is used as an example to illustrate the device for determining a variant evidence item provided in the embodiment of the present application.

[0295] The embodiment of the present application also provides a device for determining a variation evidence item.

[0296] As shown in FIG. 2 , the apparatus for determining a variation evidence item includes: a first processing module 210 , a second processing module 220 , a third processing module 230 and a fourth processing module 240 .

[0297] A first processing module 210 is configured to extract the judgment content of at least one evidence item based on a target principle to obtain at least one standard knowledge item of the evidence item;

[0298] The second processing module 220 is used to represent the evidence item standard knowledge items and obtain the evidence item standard knowledge graph;

[0299] The third processing module 230 is used to identify the information recorded in the target document set, match it with the standard knowledge graph of evidence items, and obtain the knowledge graph of evidence items;

[0300] The fourth processing module 240 is used to determine the target evidence item corresponding to the target variation based on the evidence item knowledge graph.

[0301] According to the device for determining variant evidence items provided in the embodiment of the present application, the judgment content of at least one evidence item is extracted through the target principle, the obtained standard knowledge entry of at least one evidence item is represented to construct the standard knowledge graph of the evidence item, and the information recorded in the target document set is inferred based on the standard knowledge graph of the evidence item to obtain the knowledge graph of the evidence item. In the subsequent application process, the target evidence items corresponding to the required target variants can be quickly and accurately obtained from massive documents, and the application scenarios are wide; there is no need for manual repeated retrieval, which saves manpower and time costs and significantly improves the efficiency of determining the evidence items; in addition, it effectively avoids the influence of the user's subjective judgment on the final result, improves the uniformity of the judgment result, and has high judgment accuracy and precision.

[0302] In some embodiments, the second processing module 220 may also be used to:

[0303] Representing the evidence item standard knowledge entry with a triple as the smallest unit to obtain at least one evidence item standard knowledge representation entry;

[0304] At least one evidence item standard knowledge representation entry is combined to obtain an evidence item standard knowledge graph.

[0305] In some embodiments, the third processing module 230 may also be used to:

[0306] Acquire at least part of the entities and / or relations included in the standard knowledge representation entries of the evidence items within the first target scope;

[0307] When the information recorded in the target document set matches at least some entities and / or relationships, the standard knowledge representation entries of the evidence items are represented to construct an evidence item knowledge graph.

[0308] In some embodiments, the third processing module 230 may also be used to:

[0309] Based on the target document set, a sentence containing the entity and / or relation of the first triple in the standard knowledge representation entry of the evidence item is determined as an evidence sentence;

[0310] When other entities and / or relationships in the standard knowledge representation entries of the evidence items are obtained through contextual queries around the evidence sentence, the standard knowledge representation entries of the evidence items corresponding to the other entities and / or relationships are transformed based on the specific names corresponding to the other entities and / or relationships recorded in the target document collection to generate an evidence item knowledge graph.

[0311] In some embodiments, the third processing module 230 may also be used to:

[0312] Determine the evidence sentence based on the type of the standard knowledge representation entry of the evidence item;

[0313] Based on the evidence sentences and the information recorded in the target document set, a knowledge graph of evidence items is generated.

[0314] In some embodiments, the first processing module 210 may also be used to:

[0315] Based on the target principle, extract the judgment content of at least one evidence item based on entities and relationships between entities to obtain at least one standard knowledge item of the evidence item;

[0316] The entity includes at least one of a variant entity, a disease entity, a patient entity, an indicator entity, a function entity, and an evidence item.

[0317] The device for determining the variation evidence item in the embodiment of the present application can be an electronic device or a component in the electronic device, such as an integrated circuit or a chip. The electronic device can be a terminal or other device other than a terminal. For example, the electronic device can be a mobile phone, a tablet computer, a laptop computer, a PDA, an in-vehicle electronic device, a mobile Internet device (MID), an augmented reality (AR) / virtual reality (VR) device, a robot, a wearable device, an ultra-mobile personal computer (UMPC), a netbook or a personal digital assistant (PDA), etc. It can also be a server, a network attached storage (NAS), a personal computer (PC), a television (TV), a teller machine or a self-service machine, etc., and the embodiment of the present application does not make specific limitations.

[0318] The apparatus for determining a variation evidence item in the embodiment of the present application may be a device having an operating system. The operating system may be an Android operating system, an iOS operating system, or other possible operating systems, which are not specifically limited in the embodiment of the present application.

[0319] The apparatus for determining a variant evidence item provided in the embodiment of the present application can implement each process implemented in the embodiment of the method for determining a variant evidence item in FIG1 , and will not be described again here to avoid repetition.

[0320] In some embodiments, as shown in Figure 3, the embodiment of the present application also provides an electronic device 300, including a processor 301, a memory 302, and a computer program stored on the memory 302 and executable on the processor 301. When the program is executed by the processor 301, each process of the embodiment of the method for determining the above-mentioned variant evidence item is implemented, and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.

[0321] It should be noted that the electronic devices in the embodiments of the present application include the mobile electronic devices and non-mobile electronic devices mentioned above.

[0322] An embodiment of the present application also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the various processes of the above-mentioned embodiment of the method for determining the variant evidence item are implemented, and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.

[0323] The processor is the processor in the electronic device described in the above embodiment. The readable storage medium includes a computer readable storage medium, such as a computer read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0324] An embodiment of the present application also provides a computer program product, including a computer program, which implements the above-mentioned method for determining the variation evidence item when executed by a processor.

[0325] The processor is the processor in the electronic device described in the above embodiment. The readable storage medium includes a computer readable storage medium, such as a computer read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0326] An embodiment of the present application further provides a chip, which includes a processor and a communication interface, wherein the communication interface is coupled to the processor, and the processor is used to run programs or instructions to implement the various processes of the above-mentioned embodiment of the method for determining the variant evidence item, and can achieve the same technical effect. To avoid repetition, it will not be repeated here.

[0327] It should be understood that the chip mentioned in the embodiments of the present application can also be called a system-level chip, a system chip, a chip system or a system-on-chip chip, etc.

[0328] It should be noted that, in this article, the terms "comprise", "include" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the statement "comprises a ..." does not exclude the presence of other identical elements in the process, method, article or device comprising the element. In addition, it should be noted that the scope of the methods and devices in the embodiments of the present application is not limited to performing functions in the order shown or discussed, and may also include performing functions in a substantially simultaneous manner or in the opposite order according to the functions involved. For example, the described method may be performed in an order different from that described, and various steps may also be added, omitted, or combined. In addition, the features described with reference to certain examples may be combined in other examples.

[0329] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus the necessary general hardware platform, and of course can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art can be embodied in the form of a computer software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), including a number of instructions for enabling a terminal (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0330] The embodiments of the present application are described above in conjunction with the accompanying drawings, but the present application is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of this application, ordinary technicians in this field can also make many forms without departing from the purpose of this application and the scope of protection of the claims, all of which are within the protection of this application.

[0331] Throughout this specification, reference to terms such as "one embodiment," "some embodiments," "illustrative embodiments," "examples," "specific examples," or "some examples" means that the specific features, structures, materials, or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or at least one embodiment or example.

[0332] Although the embodiments of the present application have been shown and described, those skilled in the art will appreciate that various changes, modifications, substitutions, and variations may be made to the embodiments without departing from the principles and intent of the present application, and that the scope of the present application is defined by the claims and their equivalents.

Claims

1. A method for determining a variant evidence item, characterized in that: include: Extracting the judgment content of at least one evidence item based on the target principle to obtain at least one standard knowledge item of the evidence item; Representing the evidence item standard knowledge items to obtain the evidence item standard knowledge graph; Identify the information recorded in the target document set, match it with the standard knowledge graph of evidence items, and obtain the knowledge graph of evidence items; Based on the evidence item knowledge graph, a target evidence item corresponding to the target variation is determined.

2. The method for determining a variant evidence item according to claim 1, characterized in that: The step of representing the evidence item standard knowledge items to obtain the evidence item standard knowledge graph includes: Representing the evidence item standard knowledge entry with a triple as the smallest unit, and obtaining at least one evidence item standard knowledge representation entry; Combine the at least one evidence item standard knowledge representation entry to obtain the evidence item standard knowledge graph.

3. The method for determining a variant evidence item according to claim 1, characterized in that: The standard knowledge representation entry of the evidence item has the variant entity as the head node and the evidence item as the tail node.

4. The method for determining a variant evidence item according to any one of claims 1 to 3, characterized in that: In the evidence item knowledge graph, each node and edge except the tail node is represented by a specific title recorded in the target document set.

5. The method for determining a variant evidence item according to any one of claims 1 to 3, characterized in that: The information recorded in the target document set is identified, matched with the standard knowledge graph of the evidence item, and the knowledge graph of the evidence item is obtained, including: Acquire at least part of the entities and / or relations included in the standard knowledge representation entry of the evidence item within the first target scope; In the case where the information recorded in the target document collection matches at least part of the entities and / or relationships, the standard knowledge representation entries of the evidence items are represented to construct the knowledge graph of the evidence items.

6. The method for determining a variant evidence item according to any one of claims 1 to 3, characterized in that: The information recorded in the target document set is identified, matched with the standard knowledge graph of the evidence item, and the knowledge graph of the evidence item is obtained, including: Based on the target document set, determining a sentence containing an entity and / or a relation of the first triple in a standard knowledge representation entry of an evidence item as an evidence sentence; When other entities and / or relationships in the standard knowledge representation entries of the evidence item are obtained through context query around the evidence sentence, the standard knowledge representation entries of the evidence items corresponding to the other entities and / or relationships are transformed based on the specific titles corresponding to the other entities and / or relationships recorded in the target document collection to generate the evidence item knowledge graph.

7. The method for determining a variant evidence item according to any one of claims 1 to 3, characterized in that: The information recorded in the target document set is identified, matched with the standard knowledge graph of the evidence item, and the knowledge graph of the evidence item is obtained, including: Determining an evidence sentence based on the type of the standard knowledge representation entry of the evidence item; The evidence item knowledge graph is generated based on the evidence sentence and the information recorded in the target document set.

8. The method for determining a variant evidence item according to any one of claims 1 to 3, characterized in that: The step of extracting the judgment content of at least one evidence item based on the target principle to obtain at least one evidence item standard knowledge item includes: Based on the target principle, extracting the judgment content of the at least one evidence item based on entities and relationships between entities to obtain the at least one evidence item standard knowledge item; Wherein, the entity includes at least one of a variant entity, a disease entity, a patient entity, an indicator entity, a functional entity and an evidence item.

9. A device for determining a variant evidence item, characterized in that: include: A first processing module is used to extract the judgment content of at least one evidence item based on the target principle to obtain at least one evidence item standard knowledge item; A second processing module is used to represent the standard knowledge items of the evidence items and obtain a standard knowledge graph of the evidence items; the number of items included in the standard knowledge graph of the evidence items matches the standard knowledge items of the evidence items; The third processing module is used to identify the information recorded in the target document set, match it with the standard knowledge graph of the evidence item, and obtain the knowledge graph of the evidence item; The fourth processing module is used to determine the target evidence item corresponding to the target variation based on the evidence item knowledge graph.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the method for determining the variant evidence item as described in any one of claims 1-8 is implemented.

Images