Stem cell knowledge discovery big data platform

Abstract

The invention provides a stem cell knowledge discovery big data platform, which comprises a stem cell knowledge graph module for integrating stem cell scientific and technological literatures, scientific and technological information, scientific data and scientific and technological service resource scientific research data dispersed in different data sources, and associating knowledge to form a knowledge graph; the stem cell knowledge calculation environment module integrates algorithms, models, software and tools of stem cells and related fields, provides a knowledge calculation environment based on workflow, and realizes data management, data processing and data analysis functions; and the stem cell knowledge discovery service module provides stem cell scientific research big data integrated management, one-stop intelligent retrieval, accurate knowledge retrieval and knowledge navigation based on a knowledge graph, hotspot frontier detection and scientific research portrait functions. According to the method, domain data fusion and knowledge association are realized, data islands are broken, and comprehensive and efficient data acquisition, information push, knowledge discovery and intelligence support services are provided for stem cell scientific research activities and science and technology management.

Description

Technical Field

[0001] The present invention relates to the field of data discovery, and in particular to a stem cell knowledge discovery big data platform. Background Art

[0002] Stem cells are the hotspot and frontier of life science research today. Regenerative medicine with stem cell technology as the core is expected to become the third way to treat diseases after drug therapy and surgical treatment, and is pregnant with major scientific breakthroughs and huge industrial drive. In the field of stem cells, scientific research big data with scientific literature, scientific data, clinical trials, pharmaceutical products and scientific service resources as the core has shown a "blowout" growth. Scientific research is increasingly becoming a data-driven knowledge discovery activity. Big data-driven knowledge discovery and technological breakthroughs are becoming a new engine for scientific and technological innovation. The knowledge discovery platform that integrates scientific research big data and knowledge computing environment has become an important tool for scientific research activities.

[0003] The big data of scientific research in the field of stem cells is characterized by huge quantity, various types, complex relationships and scattered sources. How to automatically extract knowledge, structured organization, semantic association and knowledge calculation from massive multi-source heterogeneous data, and efficiently and accurately mine valuable knowledge from it is the key to knowledge discovery. At present, similar data integration knowledge discovery platforms in the world, such as the "Stem Cell Commons" of Harvard Stem Cell Institute (HSCI) and the "German Network for Bioinformatics Infrastructure (de.NBI)", mainly provide scientific data management, calculation and analysis services in the field. However, they have shortcomings such as single data type and lack of association, limited knowledge calculation algorithms, etc., which are difficult to meet the needs of knowledge discovery such as research hotspots, research priorities and development trend analysis in the field, as well as key technology mining and technology foresight based on big data knowledge calculation. Summary of the invention

[0004] In response to the new situation and new challenges of scientific research information application in the big data era, and to the needs of knowledge mining and knowledge discovery in the field of stem cells, the present invention comprehensively uses data mining and new generation artificial intelligence technology to build a stem cell knowledge discovery big data platform by integrating multi-source data in the field, mining the connotation of field knowledge, and integrating professional knowledge computing tools.

[0005] The technical solution adopted by the present invention is as follows: A stem cell knowledge discovery big data platform, comprising: The stem cell knowledge graph module integrates stem cell scientific literature, scientific information, scientific data and scientific service resource research data scattered in different data sources by using entity extraction, entity disambiguation, relationship extraction, knowledge linking, ontology and entity alignment, and automatically extracts, structurally organizes and knowledgeably associates the knowledge points contained therein to form a knowledge graph; The stem cell knowledge computing environment module integrates algorithms, models, software and tools in stem cells and related fields, provides a workflow-based knowledge computing environment for the stem cell knowledge discovery service module to call, and realizes data management, data processing and data analysis functions; The stem cell knowledge discovery service module, based on the stem cell knowledge graph module and the stem cell knowledge calculation module, provides integrated management of stem cell scientific research big data, "one-stop" intelligent retrieval, precise knowledge retrieval and knowledge navigation based on knowledge graphs, hot frontier detection, and scientific research profiling functions.

[0006] As a preferred solution, the stem cell knowledge graph module includes a basic data aggregation module, a knowledge connotation mining module and a knowledge semantic association module, wherein: The basic data aggregation module is used to aggregate multi-source heterogeneous scientific research big data in the field of stem cells, including papers and patents, fund projects, clinical trials, product regulations, and expert institutions, and establish a long-term update mechanism; The knowledge content mining module uses ontology-based knowledge extraction technology to conduct text mining on core scientific and technological literature in multi-source heterogeneous scientific research big data, obtain knowledge entities and the relationship between entities, so as to form a knowledge organization system of the stem cell knowledge map, and perform knowledge entity extraction, semantic annotation and data fusion on other scientific research data in the stem cell field based on the knowledge organization system to form complete instance data in the stem cell knowledge map; The knowledge semantic association module integrates scientometric indicators and text mining technology to establish semantic associations between various scientific and technological information and knowledge entities in the knowledge graph based on citations, acknowledgments, collaboration networks, and co-occurrence relationships of knowledge entities.

[0007] As a preferred solution, the data fusion of the knowledge connotation mining module also includes an entity alignment process, which specifically includes: By identifying, filtering and distinguishing entities from different sources, entities representing the same object in different data sources are merged into an entity with a unified identifier and added to the knowledge graph; When an entity belongs to two mutually exclusive categories at the same time or an attribute of an entity corresponds to multiple values, the reliability and richness of the data source and the frequency of occurrence of different information in each data source are considered separately according to each entity type to comprehensively determine the category or attribute corresponding value.

[0008] As a preferred solution, the stem cell knowledge computing environment module includes a data management module, a data processing module and a data analysis module, wherein: Data management module, which provides data access, data integration and data storage functions required for data analysis and knowledge discovery; Data processing module, providing management functions of algorithms, data, and tasks; The data analysis module is used to perform multi-dimensional, fine-grained, and multi-type visual analysis and display of the knowledge calculation results performed by the data processing module.

[0009] As a preferred solution, in the data management module, data access is used to acquire data. In addition to directly using the established stem cell knowledge graph data, data can be ingested from external systems; the data ingestion methods include local uploading and direct connection to the database; the data integration is used to clean, convert and integrate the acquired data according to rules, and data integration supports the function of creating a combined table, and multiple data sets are associated and then merged; the data storage supports previewing, exporting, appending and sharing operations on data.

[0010] As a preferred solution, in the stem cell knowledge discovery service module, multidimensional integrated indexing technology is used to index heterogeneous information in the knowledge graph and provide a unified data view, that is, metadata, knowledge entities, and knowledge organization systems of different data types are indexed separately and integrated, and a corresponding index update mechanism is adopted according to the data update frequency of different data sources; wherein the index fields include all commonly used fields for knowledge discovery, such as knowledge resource type, resource title, author, inventor, institution, applicant, institution type, publication year, source, keyword, classification, standard level, scientific instrument, experimental animal, experimental plan, experimental reagent, method and technology, cell, organ, disease, gene, scientific research activity, and scientific research output.

[0011] As a preferred solution, the integrated management of stem cell research big data provided by the stem cell knowledge discovery service module specifically includes: The relational database and graph database are used to integrate the management of stem cell research big data. The relational database is used to store the original basic data, and the graph database is used to store the knowledge graph data after knowledge processing.

[0012] As a preferred solution, the "one-stop" intelligent retrieval and accurate knowledge retrieval and knowledge navigation based on the knowledge graph provided by the stem cell knowledge discovery service module specifically include: "One-stop" intelligent search: Solr faceted search technology is used to perform faceted indexing of stem cell research big data and knowledge organization systems in the knowledge graph. Users can use Solr indexing to search all types of stem cell resources with one click, and obtain multiple types of scientific and technological information including stem cell news, papers and patents, fund projects, pharmaceutical products, policies and regulations, and industry intelligence; Accurate knowledge retrieval and knowledge navigation based on knowledge graph: Based on the constructed index, users can search and consult according to the provided stem cell knowledge points. They can quickly and comprehensively grasp the content of scientific and technological literature without reading the full text, thus achieving accurate knowledge retrieval and navigation.

[0013] As a preferred solution, the hotspot frontier detection provided by the stem cell knowledge discovery service module specifically includes: Based on the breakthrough direction of scientific research, combining text mining technology with scientometric indicators, calling the stem cell knowledge computing environment module, performing data collection and processing as well as feature extraction, and using corresponding algorithms to identify research groups, realize text mining of knowledge entities and scientific and technological information, generate a series of research topics, and complete hot spot identification and display.

[0014] As a preferred solution, the scientific research portraits provided by the stem cell knowledge discovery service module specifically include: Call the stem cell knowledge computing environment module to collect metadata of the literature in the knowledge graph, perform data preprocessing and feature extraction, and then identify research fields, evaluate scientific research influence, and analyze scientific research activity and trends. Integrate the above analysis results to build a comprehensive portrait of scientific researchers, and use charts, network diagrams or radar charts to visualize the scientific research portrait.

[0015] Compared with the existing technology, the beneficial effects of adopting the above technical solution are: the present invention realizes field data fusion and knowledge association, breaks data silos, and provides "comprehensive, professional, accurate and efficient" data acquisition, information push, knowledge discovery and intelligence support services for stem cell scientific research activities and scientific and technological management, promotes big data-driven knowledge discovery applications, promotes the integration of scientific research activities and informatization, and improves the level of scientific research informatization application. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 Schematic diagram of the overall framework of the stem cell knowledge discovery big data platform proposed in the present invention.

[0017] Figure 2 This is a schematic diagram of constructing a stem cell knowledge graph in one embodiment of the present invention.

[0018] Figure 3 Schematic diagram of entities in the stem cell knowledge graph in one embodiment of the present invention.

[0019] Figure 4 Schematic diagram of semantic relationships in a stem cell knowledge graph in one embodiment of the present invention.

[0020] Figure 5 Schematic diagram of the stem cell knowledge computing environment module framework in one embodiment of the present invention.

[0021] Figure 6 It is a schematic diagram of algorithm assembly in one embodiment of the present invention.

[0022] Figure 7 This is the integrated management of stem cell research big data in one embodiment of the present invention.

[0023] Figure 8 This is a schematic diagram of a “one-stop” search in an embodiment of the present invention.

[0024] Fig. 9 This is a schematic diagram of precise knowledge retrieval and knowledge navigation based on knowledge graph in one embodiment of the present invention.

[0025] Fig.10 This is a schematic diagram of the hot frontiers in the field of stem cells in one embodiment of the present invention.

[0026] Fig.11 This is a schematic diagram of the scientific research results of scientists in the field of stem cells in one embodiment of the present invention. DETAILED DESCRIPTION

[0027] Embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar modules or modules with the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present application, and cannot be construed as limitations on the present application. On the contrary, the embodiments of the present application include all changes, modifications and equivalents that fall within the spirit and connotation of the appended claims.

[0028] In view of the problems in the prior art, the present invention proposes a stem cell knowledge discovery big data platform, which mainly includes three parts: stem cell knowledge graph module, stem cell knowledge computing environment module and stem cell knowledge discovery service module. The overall architecture is as follows: Figure 1 The following three parts are explained one by one: 1. Stem cell knowledge graph module The stem cell knowledge graph module is the data foundation of the entire platform. Please refer to Figure 2In this embodiment, knowledge extraction, knowledge mining, knowledge fusion and visualization technologies are comprehensively used to timely, accurately and normatively integrate scientific research big data such as scientific literature, scientific information, scientific data and scientific service resources on stem cells in "political, industrial, academic, research, medical and application" scattered in different data sources, and automatically extract, structure and organize the knowledge points contained therein to form a knowledge graph, so as to realize the effective integration of "multi-form, multi-granularity and multi-dimensional" data, information and knowledge, and provide high-quality data support for knowledge discovery.

[0029] Furthermore, the stem cell knowledge graph module proposed in this embodiment includes a basic data aggregation module, a knowledge connotation mining module and a knowledge semantic association module, wherein: The basic data aggregation module is used to aggregate multi-source heterogeneous scientific research big data such as papers and patents, fund projects, clinical trials, product regulations, expert institutions, etc. in the field of stem cells, and establish a long-term update mechanism. In one embodiment, relational databases, graph databases, and Solr indexing technologies are used to effectively store and manage basic data of scientific research big data, knowledge graph triple data, and comprehensive service data for knowledge discovery big data platforms. Different processing is performed for different data types before subsequent knowledge mining. For example, D2R mapping and graph mapping can be performed for structured data, and wrapper processing is used for semi-structured data, while information extraction can be performed directly for text data.

[0030] The use of knowledge mining technology can be used to mine domain knowledge entities from the perspectives of scientific researchers such as scientific instruments, animal models, experimental techniques, cell organs, and disease genes, and reveal the knowledge connotation contained in the scientific and technological information in the field of stem cells in a multi-dimensional and fine-grained manner. In this embodiment, the knowledge connotation mining module uses ontology-based knowledge extraction technology to obtain deep knowledge related to stem cells from multi-source heterogeneous data sources. First, refer to the UMLS super vocabulary, DisGeNET, StemCellCommons and other field knowledge ontologies and knowledge organization systems (such as thesauri, classification systems, semantic dictionaries, etc.), and use the mature biomedical field knowledge extraction tool SemRep to perform text mining on some core scientific and technological literature to obtain knowledge entities and the relationship between entities to form a knowledge organization system of the stem cell knowledge map. Then, according to the knowledge organization system, other scientific research data in the field of stem cells are subjected to knowledge entity extraction, semantic annotation, data fusion and data verification to form complete instance data in the knowledge map.

[0031] After forming the instance data in the knowledge graph, it is also necessary to establish semantic associations between the instances. In this embodiment, the knowledge semantic association module integrates scientometric indicators and text mining technology, and establishes semantic associations between various types of scientific and technological information, knowledge entities, etc. in the knowledge graph based on citations, acknowledgments, collaboration networks, and co-occurrence relationships of knowledge entities. The stem cell knowledge graph defines 58 semantic relationships in three categories: "document-document, document-knowledge entity, knowledge entity-knowledge entity", among which the core is the knowledge semantic association between "knowledge entity-knowledge entity". This embodiment uses SemRep to extract the semantic relationship between knowledge entities and knowledge entities from stem cell scientific research data, and presents the semantic association between knowledge entities in the form of SPO semantic network.

[0032] In the process of building the knowledge graph by cooperating with each module in the stem cell knowledge graph module, entity alignment is an indispensable part, which is the process of whether two entities in the same or different data sources point to the same object in the real world. The goal of entity alignment in the knowledge graph is to be able to link multiple data from different sources with high quality, create a unified knowledge representation specification from the top level, and thus help the computer understand the underlying data. Entity alignment is the basis of knowledge fusion and the prerequisite for whether the knowledge graph is standardized and scalable. Existing knowledge graph entity alignment technology has many problems and challenges in terms of data quality, matching efficiency and other aspects.

[0033] Since the data for constructing the knowledge graph is scientific research data from different data sources, this embodiment improves the quality of entity alignment through standardized data preprocessing, that is, by identifying, screening and distinguishing entities from different sources, the entities representing the same object in different data sources are merged into an entity with a unified identifier and added to the knowledge graph. For example, unique identifiers such as DOI, ISSN, ISBN, and ORCID are used to remove duplicates of journal papers, journals, books, and researchers, and titles, authors, and publication years are used to determine the same journal paper. Titles, authors, graduate schools, years, and instructors are used to distinguish degree theses, and titles and dates are used to distinguish scientific reports. Compared with scientific and technological literature, it is more difficult to align projects, institutions, experts, and concepts / terms due to the existence of homonyms, multiple name deformations, synonyms, and other problems. Therefore, it is further proposed in this embodiment that, based on standardized data preprocessing, personalized data processing rules are further designed with the help of existing standard libraries and vocabularies. For example, researchers can be initially distinguished by email, name, and affiliated institutions, and institutions can be initially distinguished by name and location. Rule-based algorithms can also be used, with strong rules set for the same ORCID and the same email as the same person, and weak rules for similarity calculations for Chinese and English name variants, first-level institutions, second-level institutions, cooperative relationships, research fields (keywords, subject terms), etc. Projects are differentiated by formulating personalized rules based on funding numbers, project names, and funding years. Concepts / terms are prototyped, translated into Chinese and English, and mapped using existing word lists such as UMLS, STKOS, Word Net, etc. Synonyms and subject word associations are obtained through co-occurrence calculations and cluster analysis.

[0034] Furthermore, when integrating knowledge entities from different data sources to construct a knowledge graph, there may be a situation where an entity belongs to two mutually exclusive categories at the same time or an attribute of an entity corresponds to multiple values. Therefore, it is necessary to clarify which category or value the entity belongs to and make attribute decisions. In this embodiment, it is proposed to perform comprehensive calculations based on factors such as the reliability and richness of the data source and the frequency of occurrence of different information in each data source according to each entity type to ensure that the category or value of the entity can be accurately divided.

[0035] Please refer to Figure 3Based on the proposed stem cell knowledge graph module, 16 categories of scientific research big data are integrated to construct a stem cell knowledge graph, including: ① scientific and technological literature such as papers, patents, reports, journals, monographs, etc.; ② scientific and technological information such as news, stem cell research and development trends, experts, projects, policies and regulations; ③ scientific data such as clinical trials, pharmaceutical products, scientific experiments; ④ scientific and technological service resources such as scientific instruments, experimental animals and experimental reagents. More than 400,000 scientific research data in the field of stem cells are integrated from the above 16 categories of 103 authoritative core data sources. Based on the aggregated data, more than 20,000 knowledge entities are mined from the field of scientific research data using knowledge extraction technology from the perspectives of scientific instruments, experimental animals, experimental plans, experimental reagents, methods and techniques, cells, organs, diseases, genes, etc.

[0036] In practical applications, please refer to Figure 4 , we can also use scientometric indicators and text mining technology to further realize the multi-perspective knowledge association between scientific research data and knowledge entities in the knowledge graph, and link domain expert institutions, scientific and technological achievements, products and services and other resources.

[0037] 2. Stem Cell Knowledge Computing Environment Module There are many complex key technical issues in the field of stem cell research, and these issues often involve large amounts of data and complex biological processes. How to efficiently analyze and predict solutions to these technical issues and how to foresee future technological development directions have also become difficult problems. The present invention proposes a stem cell knowledge computing environment module to perform data mining and knowledge computing on multi-source heterogeneous stem cell big data. Combined with the artificial intelligence, machine deep learning, and CRISPR gene editing technologies provided by the data processing module, it can construct a visual cell tissue prediction model and other knowledge discovery tools to predict changes in cell layout for cancer and other diseases. It can solve some complex key stem cell technical issues and reveal a large amount of rich basic cell biology information, accelerate progress in stem cell research, cancer research, and drug development, and provide support for more accurate technology foresight. As the computing core of the entire platform, the stem cell knowledge computing environment module integrates algorithms, models, software, and tools in stem cells and related fields, and provides a workflow-based knowledge computing environment for the stem cell knowledge discovery service module to call, realize data management, data processing, and data analysis functions. The framework of the stem cell knowledge computing environment module is as follows: Figure 5 shown.

[0038] The data management module, as the foundation of the knowledge computing environment, mainly provides data access, integration and storage, and other data management functions necessary for intelligence analysis and knowledge discovery. In addition to directly using the stem cell knowledge graph data, the data access part can also ingest external data of the system. In one embodiment, when ingesting external data, two methods can be used: local upload or direct connection to the database. The local upload method supports multiple data types such as Excel, Txt, CSV, and Xml, and the direct connection to the database supports mainstream databases such as MySQL, Oracle, SqlServer, HBase, MongoDb, and Hive. The data integration part can realize the functions of cleaning, converting, and integrating the data accessed by the data according to the rules. In this embodiment, 20 data cleaning and conversion rules such as string search and replacement, time and date format conversion, and null value filling are provided. At the same time, this part also provides the function of creating a combined table, which can associate multiple data sets and then merge them. The data storage part supports operations such as previewing, exporting, appending, and sharing of data; among them, the data of the shared partition can be used by other users.

[0039] The data processing module mainly provides functions such as algorithm management, data management, and task management. Algorithm management realizes the unified management of system algorithms, data management realizes the control of file data and table data; task management realizes task scheduling and task monitoring. In this embodiment, the data processing module integrates 30 algorithms such as machine learning, recommendation system, and natural language processing. It should be noted that the algorithms integrated in the data processing module proposed in this embodiment can be expanded and added according to needs.

[0040] The data analysis module is used to perform multi-dimensional, fine-grained, and multi-type visual analysis and display of the knowledge calculation results performed by the data processing module. In this embodiment, data visualization supports the display of various chart effects, such as bar charts, line charts, pie charts, scatter plots, heat maps, maps, radar charts, funnel charts, word clouds, relationship diagrams, etc. In addition, the charts can also be displayed centrally through the dashboard. The data displayed in the charts in the dashboard will be dynamically updated as the data in the analyzed table changes. The dashboard supports sharing in multiple ways such as links, WeChat, Weibo, and QQ space. The platform also supports the direct generation of analysis reports from analysis results. The analysis report module realizes the operation functions such as group management of analysis reports, creation, sharing, and export of analysis reports.

[0041] The present invention is based on the ZStack cloud environment and builds a general knowledge computing framework and data processing module based on Spark. The knowledge computing framework integrates 26 data cleaning, processing and fusion rules and 30 general data mining algorithms such as natural language processing, classification regression, recommendation, and result evaluation. On this basis, combined with the knowledge graph, 20 stem cell-related knowledge computing models are further integrated, providing 12 visualization analysis functions such as bar charts, line charts, pie charts, funnel charts, radar charts, word cloud charts, etc. At the same time, a template-based report automatic generation tool is developed to form a complete stem cell knowledge computing environment.

[0042] In traditional data computing, the data source is relatively single, the format is relatively regular, the association is relatively simple, and the data computing technology is relatively mature. However, in a computing environment oriented to knowledge discovery, not only are the data sources diverse, but the data format is complex and has many associations. It is a typical multi-source heterogeneous data. Traditional data mining algorithms cannot be directly applied to multi-source heterogeneous data. The knowledge computing environment module proposed in the present invention proposes to use an open computing interface to solve the above-mentioned multi-source heterogeneous data knowledge computing problem. In addition to providing built-in data cleaning, text mining, machine learning and other algorithms for users to use directly, the knowledge computing environment also has good scalability, supports users to customize the development of new algorithms, and can combine new algorithms with built-in algorithms to create new business models to meet specific knowledge discovery needs.

[0043] To better introduce the working process of the stem cell computing environment proposed by the present invention, please refer to Figure 6 After the knowledge computing environment receives the intelligence analysis algorithm process submitted by the user, Oozie combines multiple jobs in order and submits them to Yarn, which then allocates and schedules resources and starts the cluster for computing. Algorithm resources include algorithm libraries and algorithm customization functions. In addition to the many algorithms built into the platform, users are also allowed to write and upload customized algorithms. These programs can be written in languages ​​supported by the Spark platform, such as Scala, Java, R or Python. After development, they are packaged into jar packages and uploaded to the platform, and the input and output parameters are customized to meet personalized user needs.

[0044] 3. Stem cell knowledge discovery service module The stem cell knowledge discovery service module is based on the stem cell knowledge graph module and the stem cell knowledge calculation module. It combines the specific needs of stem cell scientific research informatization to provide integrated management of stem cell scientific research big data, "one-stop" intelligent retrieval, precise knowledge retrieval and knowledge navigation based on knowledge graphs, hot frontier detection, and scientific research portrait functions. It can provide basic data, information products, and personalized knowledge discovery and intelligence support for the institute's scientific and technological innovation and scientific research management.

[0045] Since the data objects of user-oriented knowledge services include scientific research big data such as various scientific and technological literature, clinical trials, scientific data and scientific and technological service resources, as well as knowledge entities in knowledge graphs, newly generated explicit knowledge in knowledge computing environments, etc., their storage forms are divided into relational data tables and RDF triples in graph databases. It is necessary to adopt a unified knowledge view to shield the differences in the form of data objects and provide a unified data retrieval interface to the knowledge discovery service platform. Multidimensional indexing is a technology for multi-level and multi-angle indexing of multi-form and complex data. It can be used to integrate heterogeneous information sources and provide a unified data retrieval interface. It has been widely used in knowledge fusion. In this embodiment, a unified data view based on multidimensional indexing is proposed, that is, metadata, knowledge entities, and knowledge organization systems of different data types are indexed separately and integrated, wherein the index fields include all commonly used fields of knowledge discovery, such as knowledge resource type, resource title, author, inventor, institution, applicant, institution type, publication year, source, keyword, classification, standard level, scientific instrument, experimental animal, experimental plan, experimental reagent, method technology, cell, organ, disease, gene, scientific research activity, scientific research output, etc. At the same time, it is also proposed to adopt corresponding index update mechanisms according to the data update frequency of different data sources, such as daily update, weekly update, monthly update, etc.

[0046] The stem cell knowledge discovery service module mainly provides the following services: (1) Integrated management of stem cell research big data The present invention uses the relational database MySQL and the graph database Neo4j to perform integrated management of stem cell research big data. The management interface is as follows: Figure 7 As shown. Relational databases store original basic data, while graph databases store knowledge graph data after knowledge processing. Neo4j is a NoSQL type database. Compared with traditional SQL type relational databases, its storage structure is not a two-dimensional list structure, but a graph structure. It can flexibly expand data structure and type, and is widely used in knowledge graph data management with triples as the core.

[0047] (2) “One-stop” intelligent search Based on the multi-dimensional indexing technology adopted by the platform, the knowledge organization system in the stem cell research big data and knowledge graph is faceted indexed. Users can use Solr index to perform "one-stop" intelligent retrieval of all types of stem cell resources. Please refer to Figure 8 Through one-click search, you can obtain various types of scientific and technological information such as stem cell news, papers and patents, fund projects, medical products, policies and regulations, and industry intelligence.

[0048] (3) Accurate knowledge retrieval and knowledge navigation based on knowledge graph Based on the constructed multidimensional index and stem cell knowledge graph, such as Fig. 9 As shown, users can search and consult based on the provided stem cell knowledge points. They can quickly and comprehensively grasp the content of scientific and technological literature without having to read the full text, thus achieving accurate knowledge retrieval and navigation.

[0049] (4) Hotspot frontier detection The research content in the field of stem cells is extensive and constantly changing. How to accurately capture and understand current research hotspots, priorities and future development trends in order to effectively allocate scientific research resources and accurately position scientific research directions has become a difficult problem. Triggered from different levels such as international R&D priorities, Chinese R&D priorities, and key breakthrough directions of scientific research of the Chinese Academy of Sciences, the present invention uses bibliometrics, scientometrics, knowledge metrology and other methods to statistically analyze and mine a large amount of scientific and technological information in the field of stem cells, including scientific and technological policies, scientific research projects, scientific papers, patents, standards, new drug instructions, etc. It can dynamically discover specific research hotspots, research priorities and development trends in the field of stem cells, and provide a scientific basis and reference for "benchmarking and catching up". Specifically, by calling the stem cell knowledge computing environment module, data collection and processing and feature extraction are carried out, and the corresponding algorithm is used to identify the research group, realize text mining of knowledge entities and scientific and technological information, generate a series of research topics, complete hotspot identification and display, and the display interface is as follows: Fig.10 The hotspot frontier detection process mainly includes data collection, preprocessing, feature extraction, hotspot identification and trend analysis steps, which are mainly implemented through the stem cell knowledge computing environment. The specific steps include: S1. Data collection: Retrieve literature data in related fields from the stem cell knowledge graph, including the title, abstract, keywords, author, publication year, number of citations, and other information of the literature.

[0050] S2. Data preprocessing: Data cleaning and text processing of literature data, including word segmentation, removal of stop words, stemming or lemmatization, etc.

[0051] S3. Use BERT to generate vectors and calculate the similarity between documents.

[0052] S4. Hotspot identification: Use the clustering algorithm K-means to group the literature and identify the research group. By building a co-occurrence network, identify important nodes, use the topic model LDA to mine potential topics in the literature set, complete hotspot detection, and provide it to users.

[0053] (5) Scientific research portrait Call the stem cell knowledge computing environment module, use knowledge graph and portrait technology to conduct scientific research portraits of hot frontier topics, researchers and research institutions from the perspectives of papers, patents, projects, news and research hotspots, and display pages such as Fig.11 The specific steps of scientific research portrait are as follows: S1. Knowledge graph retrieval: Collect metadata of documents in the knowledge graph, such as author, title, abstract, keywords, publication year, journal information, number of citations, etc. S2. Data preprocessing: data cleaning and text processing, including word segmentation, stop word removal, stemming or lemmatization, etc.

[0054] S3. Feature extraction: Use BERT to generate a vector representation of the document content and calculate the similarity between documents.

[0055] S4. Research field identification: Use the topic model LDA to identify the research field and expertise of researchers. Determine research topics and keywords through methods such as co-word analysis and co-citation analysis.

[0056] S5. Research influence assessment: Calculate the citation indicators of researchers, such as total citations, average citations, citation rate, etc. Analyze the academic influence of researchers, such as through impact factors, journal rankings and other indicators.

[0057] S6. Research activity and trend analysis: Analyze the changes in the number of publications and citations of researchers over time to evaluate their research activity and research trends. Identify emerging research directions of researchers through methods such as emergent word analysis.

[0058] S7. Portrait construction and visualization: Integrate the above analysis results to construct a comprehensive portrait of scientific researchers. Use charts, network diagrams, radar charts, etc. to visualize the scientific research portrait.

[0059] (6) “Stem Cell Assistant” WeChat APP In practical applications, the stem cell knowledge discovery big data platform proposed in the present invention also provides a mobile version for providing a simple and fast entry on the mobile terminal to guide the stem cell knowledge discovery big data platform. The WeChat APP backend is developed based on the Python language, designed in a RESTFUL style, and uses the Flask lightweight web framework to build a logical system. The backend system uses the ORM method to simplify the interaction between the system and the database, and uses Redis as a cache to reduce the concurrent access pressure on the database. WeChat APP supports multi-angle retrieval of stem cell knowledge graph data and visual navigation services. In addition, WeChat APP also takes advantage of the mobile terminal and uses machine learning algorithms to analyze user behavior data to divide user groups and perform corresponding scientific research labeling. Based on user tags and reading time, collection behavior, and forwarding behavior, collaborative filtering recommendations are performed to provide accurate knowledge services for scientific researchers. It shares a database with the knowledge discovery big data platform, and all data are highly consistent with the stem cell knowledge discovery big data platform, and synchronous updates are achieved.

[0060] The stem cell knowledge discovery big data platform proposed in this invention can solve the following problems in the field of stem cells: 1) Analysis of hot topics, research priorities and development trends in stem cell research The research content in the field of stem cells is extensive and constantly changing. How to accurately capture and understand the current research hotspots, key points and future development trends in order to effectively allocate scientific research resources and accurately locate scientific research directions has become a difficult problem. The present invention can dynamically discover specific research hotspots, key points and development trends in the field of stem cells by statistically analyzing and mining a large amount of scientific and technological information in the field of stem cells, including scientific and technological policies, scientific research projects, scientific papers, patents, standards, new drug instructions, etc., and provide a scientific basis and reference for "benchmarking and catching up".

[0061] 2) Analysis and foresight of key stem cell technologies based on big data knowledge computing There are many complex key technical problems in the field of stem cell research, and these problems often involve large amounts of data and complex biological processes. How to efficiently analyze and predict solutions to these technical problems and how to foresee future technological development directions have also become difficult problems. The present invention conducts data mining and knowledge calculation on multi-source heterogeneous stem cell big data, combines artificial intelligence, machine deep learning, and CRISPR gene editing and other technologies, builds a visual cell organization prediction model and other knowledge discovery tools, and predicts changes in cell layout for cancer and other diseases. It can solve some complex key stem cell technical problems and reveal a large amount of rich basic cell biology information, accelerate progress in stem cell research, cancer research, and drug development, and provide support for more accurate technology foresight.

[0062] 3) Integrated management of stem cell research data In the process of stem cell research activities, it is often necessary to use and analyze a large number of "fine-grained" data objects, including knowledge units in scientific literature, experimental data in scientific instruments from generation to analysis results, the quarantine experiment process of scientific research experimental animals and the continuous tracking of experimental data, etc. These data are often scattered in different systems. How to manage these data sources and improve the efficiency of scientific research and information services has gradually become a thorny problem in this field. The present invention can connect data islands from different sources by constructing an integrated stem cell research data management platform, effectively improving scientific research efficiency and information service level.

[0063] For those skilled in the art, the specific meanings of the above terms in the present invention can be understood in specific situations; the drawings in the embodiments are used to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings herein can be arranged and designed in various different configurations.

[0064] Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limitations on the present application. Ordinary technicians in this field can change, modify, replace and modify the above embodiments within the scope of the present application.

Claims

1. Stem cell knowledge discovery big data platform, characterized by: include: The stem cell knowledge graph module integrates stem cell scientific literature, scientific information, scientific data and scientific service resource research data scattered in different data sources by using entity extraction, entity disambiguation, relationship extraction, knowledge linking, ontology and entity alignment, and automatically extracts, structurally organizes and knowledgeably associates the knowledge points contained therein to form a knowledge graph; The stem cell knowledge computing environment module integrates algorithms, models, software and tools in stem cells and related fields, provides a workflow-based knowledge computing environment for the stem cell knowledge discovery service module to call, and realizes data management, data processing and data analysis functions; The stem cell knowledge discovery service module, based on the stem cell knowledge graph module and the stem cell knowledge calculation module, provides integrated management of stem cell scientific research big data, "one-stop" intelligent retrieval, precise knowledge retrieval and knowledge navigation based on knowledge graphs, hot frontier detection, and scientific research profiling functions.

2. The stem cell knowledge discovery big data platform according to claim 1, characterized in that: The stem cell knowledge graph module includes a basic data aggregation module, a knowledge connotation mining module and a knowledge semantic association module, wherein: The basic data aggregation module is used to aggregate multi-source heterogeneous scientific research big data in the field of stem cells, including papers and patents, fund projects, clinical trials, product regulations, and expert institutions, and establish a long-term update mechanism; The knowledge content mining module uses ontology-based knowledge extraction technology to conduct text mining on core scientific and technological literature in multi-source heterogeneous scientific research big data, obtain knowledge entities and the relationship between entities, so as to form a knowledge organization system of the stem cell knowledge map, and perform knowledge entity extraction, semantic annotation and data fusion on other scientific research data in the stem cell field based on the knowledge organization system to form complete instance data in the stem cell knowledge map; The knowledge semantic association module integrates scientometric indicators and text mining technology to establish semantic associations between various scientific and technological information and knowledge entities in the knowledge graph based on citations, acknowledgments, collaboration networks, and co-occurrence relationships of knowledge entities.

3. The stem cell knowledge discovery big data platform according to claim 2, characterized in that: The data fusion of the knowledge connotation mining module also includes an entity alignment process, which specifically includes: By identifying, filtering and distinguishing entities from different sources, entities representing the same object in different data sources are merged into an entity with a unified identifier and added to the knowledge graph; When an entity belongs to two mutually exclusive categories at the same time or an attribute of an entity corresponds to multiple values, the reliability and richness of the data source and the frequency of occurrence of different information in each data source are considered separately according to each entity type to comprehensively determine the category or attribute corresponding value.

4. The stem cell knowledge discovery big data platform according to claim 1, characterized in that: The stem cell knowledge computing environment module includes a data management module, a data processing module and a data analysis module, wherein: Data management module, which provides data access, data integration and data storage functions required for data analysis and knowledge discovery; Data processing module, providing management functions of algorithms, data, and tasks; The data analysis module is used to perform multi-dimensional, fine-grained, and multi-type visual analysis and display of the knowledge calculation results performed by the data processing module.

5. The stem cell knowledge discovery big data platform according to claim 4, characterized in that: In the data management module, data access is used to acquire data. In addition to directly using the established stem cell knowledge graph data, data can be ingested from external systems, where data ingestion methods include local uploading and direct connection to the database; the data integration is used to clean, convert and integrate the acquired data according to rules, and data integration supports the function of creating a combined table, linking multiple data sets and then merging them; the data storage supports previewing, exporting, appending and sharing operations on data.

6. The stem cell knowledge discovery big data platform according to claim 1, characterized in that: In the stem cell knowledge discovery service module, multidimensional integrated indexing technology is used to index heterogeneous information in the knowledge graph and provide a unified data view, that is, metadata, knowledge entities, and knowledge organization systems of different data types are indexed separately and integrated, and a corresponding index update mechanism is adopted according to the data update frequency of different data sources; wherein, the index fields include all commonly used fields for knowledge discovery, such as knowledge resource type, resource title, author, inventor, institution, applicant, institution type, publication year, source, keyword, classification, standard level, scientific instrument, experimental animal, experimental plan, experimental reagent, method and technology, cell, organ, disease, gene, scientific research activity, and scientific research output.

7. The stem cell knowledge discovery big data platform according to claim 6, characterized in that: The integrated management of stem cell research big data provided by the stem cell knowledge discovery service module specifically includes: The relational database and graph database are used to integrate the management of stem cell research big data. The relational database is used to store the original basic data, and the graph database is used to store the knowledge graph data after knowledge processing.

8. The stem cell knowledge discovery big data platform according to claim 6, characterized in that: The "one-stop" intelligent retrieval and accurate knowledge retrieval and knowledge navigation based on the knowledge graph provided by the stem cell knowledge discovery service module specifically include: "One-stop" intelligent search: Solr faceted search technology is used to perform faceted indexing of stem cell research big data and knowledge organization systems in the knowledge graph. Users can use Solr indexing to search all types of stem cell resources with one click, and obtain multiple types of scientific and technological information including stem cell news, papers and patents, fund projects, pharmaceutical products, policies and regulations, and industry intelligence; Accurate knowledge retrieval and knowledge navigation based on knowledge graph: Based on the constructed index, users can search and consult according to the provided stem cell knowledge points. They can quickly and comprehensively grasp the content of scientific and technological literature without reading the full text, thus achieving accurate knowledge retrieval and navigation.

9. The stem cell knowledge discovery big data platform according to claim 1 or 6, characterized in that: The hotspot frontier detection provided by the stem cell knowledge discovery service module specifically includes: Based on the breakthrough direction of scientific research, combining text mining technology with scientometric indicators, calling the stem cell knowledge computing environment module, performing data collection and processing as well as feature extraction, and using corresponding algorithms to identify research groups, realize text mining of knowledge entities and scientific and technological information, generate a series of research topics, and complete hot spot identification and display.

10. The stem cell knowledge discovery big data platform according to claim 1 or 6, characterized in that: The scientific research portraits provided by the stem cell knowledge discovery service module specifically include: Call the stem cell knowledge computing environment module to collect metadata of the literature in the knowledge graph, perform data preprocessing and feature extraction, and then identify research fields, evaluate scientific research influence, and analyze scientific research activity and trends. Integrate the above analysis results to build a comprehensive portrait of scientific researchers, and use charts, network diagrams or radar charts to visualize the scientific research portrait.

Images