Research achievement transformation method and system based on child birth defect research large model

Abstract

This invention provides a method and system for the transformation of scientific research results based on a large-scale research model of birth defects in children. It relates to the field of computer application technology within bioinformatics. The method involves performing correlation mapping on input research data on birth defects, generating a cross-data source data correlation network using the large-scale research model, and performing research discovery mining processing accordingly. This generates a package of research discovery mining results, which are then screened for transformation potential. A transformation potential screening report is generated by combining relevant biological knowledge bases and clinical translation rules. Based on the transformation goals that meet the screening criteria, a scientific research result transformation plan is generated. Finally, a scientific research result transformation implementation link is established based on the transformation plan, generating a transformation implementation plan and matching transformation tasks with clinical research institutions and scientific validation platforms. This ensures the smooth implementation of scientific research result transformation and improves transformation efficiency and success rate.

Description

Technical Field

[0001] This invention relates to the field of computer application technology in bioinformatics, and more specifically, to a method and system for transforming research results based on a large-scale research model of birth defects in children. Background Technology

[0002] In the field of birth defect research, the effective translation of research findings is of paramount importance for improving children's health and enhancing medical standards. Currently, birth defect research involves multiple disciplines and various data types, such as genomics data, proteomics data, clinical phenotypic data, and scientific literature data. These data sources are wide-ranging and diverse in format, containing a wealth of crucial information about birth defects.

[0003] However, existing methods for translating scientific research findings into practical applications suffer from numerous problems. On the one hand, different types of data are often isolated, lacking effective integration and correlation analysis methods, making it difficult to fully uncover the hidden patterns and research discoveries behind the data. For example, the lack of effective mapping and correlation between genomic data and clinical phenotype data makes it difficult for researchers to explain the mechanisms of clinical phenotype occurrence at the genetic level. On the other hand, the lack of scientific and systematic standards and methods for assessing the translational potential of research findings results in many promising research findings failing to be translated into practical clinical applications or products in a timely and accurate manner. Furthermore, the lack of clear translation plans and implementation schemes in the process of translating research findings results in a lack of organization and operability, leading to low translation efficiency. Summary of the Invention

[0004] In view of the aforementioned problems, and in conjunction with the first aspect of the present invention, embodiments of the present invention provide a method for transforming research results based on a large-scale research model of birth defects in children, the method comprising:

[0005] The input research data on birth defects in children is processed by association mapping. The research data on birth defects in children is then imported into the association processing unit of the large-scale research model on birth defects in children, generating a cross-data source research data association network. The research data association network contains the association relationships and association basis between different data types. The research data on birth defects in children includes genomics data, proteomics data, clinical phenotype data, and scientific literature data.

[0006] Based on the aforementioned research data association network, a large-scale research model for child birth defects is executed to mine research findings and generate research findings mining results that include a set of potential pathogenic genes, a network of key signaling pathways, and a list of candidate biomarkers.

[0007] The research findings are processed for translational potential screening. Combined with the biological knowledge base related to birth defects in children and clinical translation rules, a translational potential screening report of the research findings is generated. The translational potential screening report includes the translational screening results and screening basis for each research discovery target.

[0008] Based on the transformation potential screening report, a transformation plan for scientific research results on birth defects in children is generated. The transformation plan includes the target validation path, clinical transformation direction and phased transformation tasks of the transformation target.

[0009] Based on the aforementioned research results transformation plan for children's birth defects, a research results transformation and implementation link is established. The phased transformation tasks in the research results transformation plan for children's birth defects are matched and associated with the corresponding clinical research institutions and scientific research verification platforms to generate a transformation and implementation plan.

[0010] Furthermore, embodiments of the present invention also provide a research results transformation system based on a large-scale research model of birth defects in children, characterized by comprising:

[0011] A processor; a machine-readable storage medium for storing machine-executable instructions of the processor; wherein the processor is configured to execute the aforementioned method for transforming research results based on a large-scale research model of birth defects in children by executing the machine-executable instructions.

[0012] In another aspect, embodiments of the present invention also provide a computer program product, the computer program product including machine-executable instructions, the machine-executable instructions being stored in a computer-readable storage medium, the processor of the research achievement transformation system based on the large-scale research model of child birth defects reading the machine-executable instructions from the computer-readable storage medium, the processor executing the machine-executable instructions, causing the research achievement transformation system based on the large-scale research model of child birth defects to execute the above-mentioned research achievement transformation method based on the large-scale research model of child birth defects.

[0013] Based on the above, by performing association mapping on the input research data on childhood birth defects, a cross-data source research data association network is generated using the association processing unit of the large-scale research model on childhood birth defects. Based on this research data association network, research discovery mining is performed, generating research discovery mining results containing a set of potential pathogenic genes, key signaling pathway networks, and a list of candidate biomarkers, revealing the pathogenesis and potential targets of childhood birth defects. The research discovery mining results undergo translational potential screening, and a translational potential screening report is generated by combining relevant biological knowledge bases and clinical translation rules. Based on the translational targets that meet the screening criteria, a research achievement translation plan is generated, clarifying the target validation path, clinical translation direction, and phased translation tasks. Finally, based on the translation plan, a research achievement translation implementation link is built, and a translation implementation plan is generated, matching and associating translation tasks with clinical research institutions and research validation platforms. This ensures the smooth implementation of research achievement translation, improves translation efficiency and success rate, and promotes the translation process of research results on childhood birth defects from the laboratory to clinical application. Attached Figure Description

[0014] Figure 1 This is a schematic diagram of the execution flow of the scientific research results transformation method based on the large-scale scientific research model of child birth defects provided in the embodiments of the present invention.

[0015] Figure 2 This is a schematic diagram of exemplary hardware and software components of a research achievement transformation system based on a large-scale research model of birth defects in children, provided in an embodiment of the present invention. Detailed Implementation

[0016] The present invention will now be described in detail with reference to the accompanying drawings. Figure 1 This is a flowchart illustrating a method for transforming research results based on a large-scale research model of child birth defects, provided by an embodiment of the present invention. The following is a detailed description of this method for transforming research results based on a large-scale research model of child birth defects.

[0017] Step S110: Perform association mapping processing on the input research data on birth defects in children, import the research data on birth defects in children into the association processing unit of the large research model on birth defects in children, and generate a cross-data source research data association network. The research data association network contains the association relationships and association basis between different data types. The research data on birth defects in children includes genomics data, proteomics data, clinical phenotype data and scientific literature data.

[0018] In this embodiment, congenital heart disease in children is used as the application scenario throughout the text. The aforementioned research data on birth defects specifically includes genomic data of patients with congenital heart disease, such as chromosomal copy number variation data and gene point mutation data; proteomics data, such as protein expression levels and protein modification data in heart tissue samples; clinical phenotype data, such as patients' cardiac structure ultrasound reports, electrocardiogram data, and symptom records; and scientific literature data, such as published gene research papers and protein function analysis literature related to congenital heart disease. In performing this step, the aforementioned multi-source data is first uniformly imported into the association processing unit of the large-scale research model for birth defects in children. This association processing unit has the ability to parse data of different formats and can identify gene names and mutation locations in genomic data, protein IDs and expression levels in proteomics data, symptom terms and examination indicators in clinical phenotype data, and keywords and research conclusions in scientific literature data.

[0019] The collection of research data on birth defects in children strictly adheres to relevant laws and regulations. When collecting genomics, proteomics, and clinical phenotypic data, the data provider (such as medical or research institutions) and the data user (the entity responsible for developing and applying the large-scale research model on birth defects) first sign a data usage authorization agreement, clearly defining the scope, duration, and confidentiality obligations of the data. For clinical phenotypic and genomic data involving patient privacy, written informed consent from the data subject (or their guardian, for minors) must be obtained before collection. This consent clearly states the data's purpose as research for the transformation of research results, data storage methods, and privacy protection measures. During data collection, all personally identifiable information is anonymized, removing direct identifiers such as names, ID numbers, and contact information, replacing them with virtual sample IDs. Indirect identifiers (such as date of birth and address) are also obfuscated to ensure that individuals cannot be identified through the anonymized data. For scientific literature data, only publicly published content without copyright disputes is collected, obtained through legitimate channels from academic databases, and the source and citation information are clearly indicated when using the data to ensure compliance with relevant intellectual property laws. In the data authorization and licensing process, a multi-level review mechanism is established. The ethics committee reviews the data collection plan and the informed consent form template. Data collection can only be started after the review is approved. At the same time, electronic archives of all authorization documents and ethics review opinions are kept for verification by regulatory authorities.

[0020] In this embodiment, the construction of the large-scale research model for birth defects in children includes multiple core modules, which work collaboratively through data interfaces. The association processing unit, as the core module for data preprocessing, includes a terminology standardization submodule, an association extraction submodule, and a network construction submodule. The terminology standardization submodule incorporates a terminology recognition model optimized based on a BERT pre-trained model. The input is the original core descriptive information of various data types, and the output is standardized core descriptive information. This submodule first performs word segmentation on the input text, preliminarily identifies potential terms through terminology dictionary matching, then models the semantic context of the terms using a bidirectional LSTM network, and achieves accurate term boundary positioning using a CRF layer. Finally, it completes the standardization transformation based on the built-in terminology mapping relationship. The association extraction submodule adopts a graph attention network (GAT) architecture. The input is the key association fields in the standardized core descriptive information. The key fields are converted into low-dimensional vector representations through a node embedding layer, and then the association weights between different types of nodes are calculated through a multi-head attention mechanism, outputting association pairs such as gene-protein and protein-clinical phenotype. The network construction submodule takes association pairs as input, uses the Neo4j graph database to store nodes and relationships, and uses the Cypher query language to dynamically update and visualize the association network. The mining and processing unit includes a pathogenic gene mining submodule, a signaling pathway analysis submodule, and a biomarker screening submodule, each built on the Transformer architecture. The pathogenic gene mining submodule takes a research data association network as input, extracts the local network structure features of gene nodes using a graph convolutional network (GCN), combines this with gene expression temporal features (processed using an LSTM network), and inputs it to a fully connected layer for classification, outputting a set of potential pathogenic genes. The signaling pathway analysis submodule uses an attention mechanism and knowledge graph fusion approach, embedding the KEGG signaling pathway knowledge graph into the model training process. Through comparative learning, it optimizes the representation vectors of signaling pathway nodes, thereby identifying key signaling pathways and cross-regulatory relationships. The biomarker screening submodule takes multimodal data (protein expression levels, clinical test indicators, etc.) as input and uses a multilayer perceptron (MLP) to score the clinical value of different types of biomarkers. The number of neurons in the input layer of the MLP is consistent with the feature dimension of the biomarker. The hidden layer is set with two layers (128-dimensional and 64-dimensional neurons, respectively). The output layer is a single score value, which is mapped to the 0-1 interval through the Sigmoid activation function.

[0021] The model training process is divided into two stages: pre-training and fine-tuning. In the pre-training stage, publicly available multi-omics datasets related to birth defects in children (such as a subset of congenital heart disease data from the TCGA database and gene expression data from the GEO database) and scientific literature corpora are used as training data, employing a self-supervised learning approach. For the terminology standardization model of the association processing unit, the semantic context of terms is learned through a masked language model (MLM) task. The training epochs are set to 200, the batch size to 32, and the initial learning rate to 2e-5, using a cosine annealing learning rate scheduling strategy. The association extraction model uses a contrastive loss function, comparing the vector representations of positive samples (true association pairs) and negative samples (randomly generated non-associative pairs) to ensure that the cosine similarity of positive sample pairs is greater than 0.85 and the cosine similarity of negative sample pairs is less than 0.2. During the pre-training of each submodule of the data mining unit, known pathogenic genes, signaling pathways, and biomarkers were used as labeled data, and supervised training was performed using the cross-entropy loss function. The ratio of training set, validation set, and test set for the pathogenic gene mining submodule was 7:2:1. The AUC value was used as the model evaluation metric, and pre-training was stopped when the validation set AUC value did not improve for 10 consecutive rounds. In the fine-tuning phase, actual collected research data on birth defects in children was used as training data, and fine-tuning tasks were constructed for different data types (genomics, proteomics, etc.). For newly added gene mutation data, the model parameters were adjusted using the domain adaptation loss function to improve the model's recognition accuracy for new data types to over 90%. During fine-tuning, gradient accumulation technology (accumulation steps of 4) was used to address the instability problem of training with small sample data. An early stopping mechanism was also introduced, and fine-tuning was completed when the F1 score on the test set reached 0.88 and tended to stabilize. After the model is trained, it is deployed in a server cluster containing GPU accelerator cards. Docker containerization technology is used to achieve rapid deployment and version control of the model, and Kubernetes is used for container orchestration to ensure high availability and load balancing of the model service.

[0022] In the context of translating research findings on congenital heart disease in children, the model input data is closely linked to specific research objectives. When mining potential pathogenic genes, the input data includes gene mutation site information from genomics data (stored in VCF format), protein expression matrices from proteomics data (rows represent protein IDs, columns represent sample IDs, and values ​​are standardized expression levels), structured symptom codes from clinical phenotype data (represented using a combination of ICD-10 coding and HPO terminology), and gene-disease association triples from scientific literature data (the subject is the gene name, the predicate is the association type, and the object is the disease name). The above input data is converted into a format acceptable to the model through a data preprocessing pipeline, such as converting gene mutation site information into one-hot encoded vectors and performing Z-score standardization on the protein expression matrix. The model's output data is structured based on different parts of the research findings. The potential pathogenic gene set is output in JSON format, including gene ID, gene name, screening criteria (such as association strength score, number of supporting publications), and functional annotation information. The key signaling pathway network is output in GraphML format, including node attributes (node ​​ID, type, name) and edge attributes (association type, regulatory direction, source of evidence). The candidate biomarker list is output in tabular form, including biomarker ID, type, clinical value score, and suggested detection methods. During model application, a validation mechanism for input and output data is established. Schema validation ensures the integrity of input data fields (e.g., genomic data must include chromosome number, start position, and variant type fields), and output data undergoes logical consistency checks (e.g., upstream and downstream relationships in signaling pathways must conform to the regulatory direction in the biological knowledge base). If data anomalies are detected, an alarm mechanism is triggered, prompting the user to correct the data.

[0023] The integration of the model with clinical translation rules is achieved through a rule engine, which incorporates a clinical translation decision tree model. The root node of this decision tree represents the primary indicators for translational potential screening (functional importance, translational feasibility, etc.). Each primary indicator contains multiple secondary indicators as child nodes, and the leaf nodes represent the translation screening results (meeting criteria, pending evaluation, or not yet included). The decision tree is constructed based on historical case data annotated by experts in the clinical translation field. The information gain of each indicator is calculated using the ID3 algorithm to determine the node splitting order; for example, the functional importance indicator, with the highest information gain, is selected as the first splitting node. When research findings are input, the rule engine maps the analysis results of each target to the corresponding nodes in the decision tree, traversing the path to obtain the translation screening results. For example, if a potential pathogenic gene meets the screening criteria for functional importance (information gain greater than 0.6), the target drugability score in the translational feasibility indicator is greater than 0.75, and the disease incidence rate in the clinical need indicator is higher than 5 / 10000, then the decision tree path terminates at the "meets translation screening criteria" leaf node. The model's output translation potential screening report includes a visual representation of the decision tree traversal path, annotating the judgment criteria and indicator value ranges for each node, enabling clinical researchers to clearly understand the logic behind the screening results. During the research outcome translation plan generation stage, the model automatically calls the corresponding template library based on the type of translation target. For example, for potential pathogenic genes, it calls the drug development pathway template, which includes standardized process nodes such as target validation, compound screening, and preclinical trials. Based on the specific scoring results in the translation screening report (such as clinical need matching degree and translation risk level), the model adjusts the task parameters (such as sample size and experimental cycle length) of each process node, ultimately generating a personalized translation plan.

[0024] Step S111: Collect genomic data, proteomics data, clinical phenotype data and scientific literature data related to birth defects in children to form an initial research data set. Transmit the initial research data set to the association processing unit of the large-scale research model for birth defects in children, and extract the core descriptive information of each type of data in the initial research data set. The core descriptive information includes the characteristics of the research object corresponding to the data, the data collection conditions and the core observation indicators of the data.

[0025] In the application scenario of congenital heart disease, the collected genomic data came from peripheral blood samples of children with congenital heart disease and healthy controls in a certain region, obtained through gene sequencing technology. The characteristics of the research subjects included sample ID, gender, age, and disease diagnosis type; data acquisition conditions included sampling time, clinical status at the time of sampling, sequencing platform model, and sequencing depth; core observation indicators included gene name, variant type, and variant site coordinates. Proteomics data were collected from surgically removed heart tissue samples from children with congenital heart disease and normal heart tissue samples. The characteristics of the research subjects also included sample ID and corresponding disease subtype; data acquisition conditions included sample preservation method, protein extraction method, and mass spectrometry detection parameters; core observation indicators included protein name, peptide sequence, and relative expression level. Clinical phenotypic data came from the electronic medical record systems of children with congenital heart disease in multiple hospitals. The characteristics of the research subjects included patient ID, admission time, and diagnosis result; data acquisition conditions included examination equipment model, examination time, and operator qualifications; core observation indicators included atrial and ventricular size, valvular function status, and cardiac function classification in echocardiography. Scientific literature data is obtained through academic database searches. The characteristics of the research objects include document titles, authors, and publishing journals. Data collection conditions include search keywords, search time range, and document selection criteria. Core observation indicators include descriptions of the relationships between genes, proteins, signaling pathways, and clinical phenotypes mentioned in the literature. After integrating the above data to form an initial research dataset, it is transmitted to a correlation processing unit. This unit automatically extracts the aforementioned core descriptive information using natural language processing technology and structured data parsing algorithms.

[0026] Step S112: Based on the terminology unification rules built into the large-scale scientific research model for birth defects in children, perform terminology standardization processing on the core descriptive information of different types of data to generate standardized core descriptive information, and extract the associated key fields in the standardized core descriptive information. The associated key fields include gene name, protein name, clinical symptom name and disease type name.

[0027] The large-scale research model for childhood birth defects incorporates standardized terminology rules from multiple authoritative databases. For example, gene names adopt the HGNC (Human Genome Nomenclature Committee) naming standards, protein names use the standard names from the UniProt database, clinical symptom names use HPO (Human Phenotypic Ontology) terminology, and disease type names use the corresponding disease names from ICD-10 (International Classification of Diseases, 10th Revision). In the congenital heart disease scenario, for gene names in the core descriptive information of genomics data, such as the "GATA4" gene, if different expressions exist in the original data, such as "GATA-bindingfactor4," they are uniformly converted to the HGNC-approved standard name "GATA4." Protein names in proteomics data, such as "cardiactroponin I," are uniformly converted to the standard accession number and corresponding name in UniProt. The clinical symptom "ventricular septal defect" in clinical phenotype data corresponds to a specific identifier and standard name in HPO terminology and is associated with the disease code in ICD-10. Through this standardization process, the terminology describing the same entity is kept consistent across different types of data, generating standardized core descriptive information. Subsequently, key related fields were extracted from the standardized core description information. For example, gene names such as "GATA4" and "NKX2-5" were extracted from the standardized core description information of genomics data; protein names such as "Troponin I" and "Myosin" were extracted from proteomics data; clinical symptom names such as "ventricular septal defect" and "pulmonary hypertension" were extracted from clinical phenotype data; and disease type names such as "congenital heart disease" and "tetralogy of Fallot" were extracted from scientific literature data.

[0028] Step S1121: Analyze the terminology mapping relationship, terminology naming convention, and terminology classification standard in the terminology unification rules.

[0029] In the context of congenital heart disease applications, the terminology mapping relationship in the unified terminology rules specifically manifests as the correspondence between different names for the same biological entity in different databases or literature and the standard terminology. For example, regarding gene names, the terminology mapping relationship lists possible aliases for the "GATA4" gene in different research literature, such as "GATA-4" and "GATAbindingprotein4," and maps these aliases to the standard name "GATA4" as defined by HGNC. The terminology naming convention specifies the naming format for various terms. For gene names, uppercase letters are required, with no spaces between gene symbols; for protein names, the protein family name and a specific functional description must be included, such as "Troponin Itype3 (cardiac)." The terminology classification standard categorizes terms according to their biological category, such as classifying gene terms into transcription factor genes, structural protein genes, etc., and clinical symptom terms into symptoms of cardiac structural abnormalities, symptoms of cardiac functional abnormalities, etc.

[0030] Step S1122: Extract terms from the core descriptive information of different types of data to form a set of terms to be standardized, which includes genomics terms, proteomics terms, clinical phenotype terms and literature terms.

[0031] For various types of data on congenital heart disease, genomic terms such as "chromosome 22q11 deletion" and "GATA4 gene mutation" were extracted from the core descriptive information of genomic data; proteomics terms such as "cardiac myosin heavy chain" and "phosphorylated ERK protein" were extracted from proteomics data; clinical phenotype terms such as "atrial septal defect" and "patent ductus arteriosus" were extracted from clinical phenotype data; and literature terms such as "congenital heart malformation" and "genetic susceptibility factors" were extracted from scientific literature data. The terms extracted above together constitute a set of terms to be standardized. There may be different expressions of the same term, which need to be standardized in subsequent processes.

[0032] Step S1123: For each genomic term in the set of terms to be standardized, according to the term mapping relationship in the terminology unification rules, map the non-standard genomic terms to standard genomic terms to generate standardized genomic terms.

[0033] In the set of terms to be standardized, if the genomic term "deletion of the long arm of chromosome 22" appears, according to the term mapping relationship, "long arm of chromosome 22" corresponds to the standard "22q", and "deletion" corresponds to "deletion". Therefore, it is standardized as the standard genomic term "22qdeletion". For some abbreviated or commonly used gene names, such as "NKX25", its standard name is found to be "NKX2-5" through the term mapping relationship, and the corresponding conversion is performed to ensure that all genomic terms conform to the HGNC standard nomenclature.

[0034] Step S1124: For each proteomics term in the set of terms to be standardized, according to the term mapping relationship in the terminology unification rules, map the non-standard proteomics term to a standard proteomics term, and generate standardized proteomics terms.

[0035] If the term "cardiac troponin I" exists among the proteomics terms to be standardized, its corresponding UniProt standard name is "Troponin I, cardiac cmuscle" according to the terminology mapping relationship. Simultaneously, its standard accession number, such as "P19429," should be obtained, and this information should be used together as the standardized proteomics terminology. For protein modification terms, such as "p-ERK," they should be standardized as "Phospho-ERK1 / 2(Thr202 / Tyr204)," clearly specifying the modification site and type to conform to the standard expression for proteomics research.

[0036] Step S1125: For each clinical phenotype term in the set of terms to be standardized, the naming method of non-standard clinical phenotype terms is modified according to the terminology naming specifications in the unified terminology rules, and standardized clinical phenotype terms are generated.

[0037] In clinical phenotypic terminology, the term "cardiac septal defect" is not specific enough. According to terminology naming standards, the specific location of the defect needs to be clarified, such as "atrial septal defect" or "ventricular septal defect." This should be revised to a specific name conforming to HPO terminology standards and associated with the corresponding HPOID. Regarding the aforementioned symptom description of "heart murmur," its characteristics should be supplemented according to naming standards, such as "systolic heart murmur" or "diastolic heart murmur," to make clinical phenotypic terminology more precise and standardized.

[0038] Step S1126: For each document term in the set of terms to be standardized, classify the document term into the corresponding standard term category according to the term classification standard in the terminology unification rules, and generate standardized document terms.

[0039] The term "genetic factors of congenital heart disease" extracted from scientific literature data, according to the terminology classification standard, belongs to the "genetic etiology terminology" category under "disease etiology terminology," and is therefore classified into this standard category. The term "mechanism of abnormal cardiac development" is classified into the "disease pathogenesis mechanism terminology" category, ensuring that literature terminology is organized according to a unified classification standard, facilitating subsequent correlation analysis and knowledge integration.

[0040] Step S1127: Replace the non-standard terms in the original core description information with the generated standardized genomics terms, standardized proteomics terms, standardized clinical phenotype terms, and standardized literature terms.

[0041] After standardizing various terms, the original non-standard terms in the core descriptive information of various congenital heart disease data were replaced with the corresponding standardized terms. For example, in the core descriptive information of genomics data, "GATA4 gene variant" was replaced with the standardized genomics term "GATA4 gene variant"; in proteomics data, "cardiac myosin" was replaced with the standardized protein name and accession number; and in clinical phenotype data, "foramen in the heart" was replaced with standardized clinical phenotype terms such as "ventricular septal defect," so that the terms in all core descriptive information met a unified standard.

[0042] Step S1128: Perform sentence fluency processing on the core description information after term replacement, generate a standardized core description information set containing all standardized core description information, and complete the term standardization process.

[0043] After replacing the terms, check the fluency of the core descriptive information. For example, for the description "This patient has a 22q defect, presenting as a ventricular septal defect," ensure its grammatical correctness and semantic clarity. If there are grammatical inconsistencies due to term replacement, such as "high expression of Troponin I, cardiac septum was detected," adjust it to "elevated expression levels of Troponin I, cardiac septum were detected" to conform to natural language expression habits. After the above processing, a standardized set of core descriptive information is formed.

[0044] Step S113: Based on the aforementioned key association fields, establish an association mapping relationship between genomics data and proteomics data, and generate gene-protein association pairs. The gene-protein association pairs contain information on the associated genes and proteins and a description of the association strength.

[0045] In the context of congenital heart disease, extracted key correlation fields, such as gene and protein names, are used for matching through a gene-protein relationship database built into the large-scale research model for childhood birth defects. This database contains known gene-encoded protein correspondences; for example, the gene "GATA4" encodes the protein "GATA4 transcription factor," and the gene "NKX2-5" encodes the protein "NKX2-5 homeobox protein." For "GATA4" gene mutations detected in genomic data, the proteomics data is searched for changes in the expression level or modification status of the corresponding "GATA4 transcription factor." By analyzing the co-occurrence and expression trend correlation of the two in samples, the strength of the association is determined, such as "strong association," "moderate association," or "weak association." For example, if a specific mutation occurs in the "GATA4" gene in a sample, and the expression level of its corresponding "GATA4 transcription factor" is significantly reduced, and this phenomenon is repeated in multiple samples, then the association strength of this gene-protein pair is described as "strong association."

[0046] Step S114: Based on the aforementioned key association fields, establish an association mapping relationship between proteomics data and clinical phenotype data, and generate protein-clinical phenotype association pairs. The protein-clinical phenotype association pairs include information on the associated proteins and clinical phenotypes, as well as a description of the association strength.

[0047] In the application scenario of congenital heart disease, protein names such as "Troponin I" and clinical symptom names such as "myocardial injury" in key association fields can be used to establish a correlation. By analyzing the expression level of "Troponin I" in proteomics data and the occurrence of "myocardial injury" symptoms in clinical phenotype data, if the frequency of "myocardial injury" symptoms is significantly higher in samples with elevated "Troponin I" expression than in samples with normal expression, then a protein-clinical phenotype correlation pair of "Troponin I-myocardial injury" is established. The strength of the correlation is determined by statistical analysis of the degree of correlation between the two; for example, when the correlation coefficient reaches a certain threshold, the correlation strength is considered "strong."

[0048] Step S115: Based on the aforementioned key fields, establish a mapping relationship between genomics data and scientific literature data, and generate gene-literature association pairs. The gene-literature association pairs contain the associated gene and literature information and a summary of the basis for the association.

[0049] For the "NKX2-5" gene in genomics data, literature containing the gene name is retrieved from scientific literature data. If a paper studies the role of the "NKX2-5" gene in the development of congenital heart disease and concludes that mutations in this gene are associated with abnormal heart structure, then a gene-document association pair of "NKX2-5-document title" is established. The association is based on the abstract, which extracts key research findings from the literature regarding the association between this gene and congenital heart disease, such as "This study found that missense mutations in the NKX2-5 gene can lead to abnormal development of the cardiac outflow tract."

[0050] Step S116: Integrate the gene-protein association pairs, protein-clinical phenotype association pairs, and gene-literature association pairs to construct an initial scientific research data association network, which includes nodes of various association pairs and the connections between nodes.

[0051] The gene-protein association pairs, protein-clinical phenotype association pairs, and gene-document association pairs generated above are integrated. In the initial scientific research data association network, genes, proteins, clinical phenotypes, and documents are used as nodes, and the association relationships in each type of association pair are the connections between nodes. For example, the "GATA4" gene node is connected to the "GATA4 transcription factor" protein node through a "strong association" connection, the "GATA4 transcription factor" protein node is connected to the "ventricular septal defect" clinical phenotype node through a "moderate association" connection, and the "GATA4" gene node is connected to relevant document nodes through association connections, forming an initial network structure with multiple interconnected nodes.

[0052] Step S117: Supplement the initial scientific research data association network with supporting evidence information for various associations, including citations of experimental verification results, citations of clinical observation data, and citations of literature conclusions.

[0053] For the association between "GATA4 gene and GATA4 transcription factor", supplementary supporting evidence should be provided, such as experimental data citing the decrease in GATA4 transcription factor expression after knocking out the GATA4 gene in a certain cell experiment; for the association between "GATA4 transcription factor and ventricular septal defect", supplementary clinical observation data should be provided, namely, statistical data on the correlation between GATA4 transcription factor abnormalities and the incidence of ventricular septal defects in patients with congenital heart disease; for the association between "GATA4 gene and literature", supplementary literature conclusions should be provided, namely, specific research conclusions in the literature regarding the association between the GATA4 gene and congenital heart disease.

[0054] Step S118: Optimize the structure of the initial scientific research data association network after supplementing supporting evidence information, extract the indirect association paths between nodes, and generate a cross-data source scientific research data association network that includes direct and indirect association relationships.

[0055] Based on the initial network supplemented with additional evidence, structural optimization was performed. For example, it was discovered that the NKX2-5 gene is directly associated with the clinical phenotype of cardiac conduction block through the NKX2-5 protein. Furthermore, the NKX2-5 gene can influence the expression of the GATA4 gene, thereby affecting the GATA4 transcription factor, ultimately forming an indirect association pathway with the clinical phenotype of ventricular septal defect. These indirect association pathways were extracted and added to the network, ensuring that the research data association network not only includes direct associations but also rich indirect associations, more comprehensively reflecting the complex connections between multi-source data.

[0056] Step S120: Based on the research data association network, perform research discovery mining processing of the large-scale research model of birth defects in children to generate research discovery mining results containing a set of potential pathogenic genes, key signaling pathway networks and a list of candidate biomarkers.

[0057] The constructed research data association network for congenital heart disease is input into the mining and processing unit of the large-scale research model for birth defects in children. This unit first comprehensively analyzes the nodes and relationships within the network, identifying different types of nodes, such as genes, proteins, signaling pathways, and clinical phenotypes, as well as the direct and indirect relationships between them. Then, using built-in mining algorithms, such as graph-based community detection algorithms and association rule mining algorithms, it extracts potential pathogenic genes, key signaling pathways, and candidate biomarkers related to the occurrence and development of congenital heart disease from the network. For example, by analyzing the strength of associations and the number of pathways between gene nodes and clinical phenotype nodes, potential pathogenic genes are screened; by tracing gene-protein-signaling pathway associations, a key signaling pathway network is constructed; and by analyzing the closeness between protein or metabolite nodes and clinical phenotypes, candidate biomarkers are identified.

[0058] Step S121: Input the research data association network into the mining and processing unit of the large-scale research model of birth defects in children, and analyze the node types and the types of association relationships between nodes in the research data association network.

[0059] The data mining and processing unit analyzes the network of connections in scientific research data, identifying node types including gene nodes (such as GATA4 and NKX2-5), protein nodes (such as GATA4 transcription factor and Troponin I), clinical phenotype nodes (such as ventricular septal defect and cardiac conduction block), signaling pathway nodes (such as Wnt signaling pathway and Notch signaling pathway), and literature nodes. The types of relationships between nodes include gene-protein expression regulation relationships, protein-protein interaction relationships, causal relationships between proteins and clinical phenotypes, and gene-signaling pathway involvement relationships. For example, the analysis reveals an expression regulation relationship between the "GATA4 gene" and "GATA4 transcription factor" nodes, and an interaction relationship between the "GATA4 transcription factor" and "NKX2-5 protein" nodes.

[0060] Step S122: Based on the pathogenic association mining rules built into the large-scale scientific research model for children's birth defects, filter gene nodes in the scientific research data association network that are directly associated with clinical phenotype nodes of children's birth defects, and generate a set of candidate pathogenic genes.

[0061] The pathogenic association mining rules built into the large-scale research model for childhood birth defects stipulate that gene nodes with a "strong association" with clinical phenotype nodes and supported by multiple independent pieces of evidence can be screened as candidate pathogenic genes. In the context of congenital heart disease, clinical phenotype nodes such as "ventricular septal defect" and "tetralogy of Fallot" are screened to identify gene nodes with a direct "strong association" with these nodes, such as GATA4, NKX2-5, and TBX5, and these gene nodes are combined into a set of candidate pathogenic genes.

[0062] Step S123: For each gene node in the candidate pathogenic gene set, extract its associated protein nodes, associated literature nodes, and associated clinical phenotype nodes in the scientific research data association network to generate a gene association information package.

[0063] Taking the "TBX5" gene in the candidate pathogenic gene set as an example, we extract its associated protein nodes in the scientific research data association network, such as "TBX5 protein"; associated literature nodes, such as multiple papers studying the relationship between the TBX5 gene and congenital heart disease; and associated clinical phenotype nodes, such as "atrial septal defect" and "finger deformity" (because TBX5 gene mutations may cause both heart and limb deformities). We integrate and package the above information to form the "TBX5 gene association information package".

[0064] Step S124: Based on the gene association information package, analyze the expression regulation pattern of each candidate pathogenic gene, combine it with the biological knowledge base related to birth defects in children, screen out genes that meet the expression characteristics of pathogenic genes, and generate a set of potential pathogenic genes.

[0065] For the "TBX5 gene association information package," its expression regulation pattern is analyzed, such as the expression site and level changes of the TBX5 gene during embryonic heart development, and the regulatory effect of its expression product, TBX5 protein, on downstream target genes. This is combined with the expression characteristic criteria for pathogenic genes of congenital heart disease in the biological knowledge base, such as high expression during critical periods of heart development, transcriptional regulatory function, and mutation leading to abnormal heart structure. If the expression regulation pattern of the TBX5 gene meets these criteria, it is screened as a gene that meets the expression characteristic criteria for pathogenic genes and added to the potential pathogenic gene set.

[0066] Step S1241: Extract the associated protein node information from the gene association information package of each candidate pathogenic gene, analyze the interaction pattern between the candidate pathogenic gene and the associated protein, and determine the regulatory mode of gene expression products.

[0067] Information on associated protein nodes, such as "GATA4 transcription factor" and "NKX2-5 protein," was extracted from the "GATA4 gene association information package." Analysis of the interaction pattern between the GATA4 gene expression product, GATA4 transcription factor, and the NKX2-5 protein revealed that they can form heterodimers and jointly regulate the expression of downstream target genes. This interaction pattern belongs to the protein-protein interaction regulation mode, thus confirming that one of the regulatory mechanisms of the GATA4 gene expression product is through interaction with other proteins to regulate gene expression.

[0068] Step S1242: Extract the associated literature node information from the gene association information package, summarize the research conclusions on the expression characteristics of candidate pathogenic genes in the literature, and generate a gene expression literature summary.

[0069] The research conclusions of the related literature nodes in the "GATA4 gene association information package" are summarized. For example, literature A indicates that the GATA4 gene is highly expressed during the 3rd to 8th week of embryonic heart formation; literature B shows that the GATA4 gene plays an important role in the development of the atria and ventricles; literature C finds that the expression of the GATA4 gene is positively regulated by certain transcription factors. The above conclusions are integrated to generate a summary of literature on GATA4 gene expression.

[0070] Step S1243: Extract the associated clinical phenotype node information from the gene association information package and analyze the expression differences of candidate pathogenic genes in different clinical phenotypes of birth defects in children.

[0071] The clinical phenotype nodes associated with the GATA4 gene, such as "ventricular septal defect," "atrial septal defect," and "patent ductus arteriosus," were analyzed in the "GATA4 gene association information package." The expression levels of the GATA4 gene in patients with these different clinical phenotypes were compared. It was found that the expression level of the GATA4 gene in patients with ventricular septal defects was significantly lower than that in patients with atrial septal defects, thus revealing the differential expression characteristics of the GATA4 gene across different clinical phenotypes.

[0072] Step S1244: Retrieve the biological knowledge base related to birth defects from the knowledge storage unit of the large-scale scientific research model of birth defects, extract the description of the expression characteristics of pathogenic genes, and generate reference standards for the expression characteristics of pathogenic genes.

[0073] The biological knowledge base was retrieved from the knowledge storage unit, where descriptions of the expression characteristics of pathogenic genes for congenital heart disease included: specific expression during critical stages of heart development; expression products located in the cell nucleus with DNA-binding capacity; gene mutations leading to loss or gain of protein function; and significant differences in expression levels between patient samples and normal controls. These descriptions were then compiled into a reference standard for pathogenic gene expression characteristics.

[0074] Step S1245: Compare and analyze the expression regulation patterns of each candidate pathogenic gene, the summary of gene expression literature, and the fit between expression difference characteristics and the reference standards for pathogenic gene expression characteristics. Mark the candidate pathogenic genes that meet the reference standards for pathogenic gene expression characteristics as suspected pathogenic genes.

[0075] The expression regulation pattern of the GATA4 gene (interacting with the NKX2-5 protein to regulate downstream genes), a summary of gene expression literature (high expression during critical periods of embryonic heart development), and differential expression characteristics (different expression in different clinical phenotypes) were compared with the reference standards for pathogenic gene expression characteristics. If the above characteristics of the GATA4 gene match multiple descriptions in the reference standards, such as specific expression during critical stages of heart development and the DNA-binding ability of the expression product, then the GATA4 gene is marked as a suspected pathogenic gene.

[0076] Step S1246: Further analyze the expression change trend of the suspected pathogenic gene during the critical period of the onset of birth defects in children, and generate gene temporal expression characteristics.

[0077] We analyzed the expression trends of the suspected pathogenic gene GATA4 during critical periods of congenital heart disease development (such as the 4th-6th week of embryonic heart development). By retrieving relevant temporal gene expression data, we found that GATA4 gene expression gradually increased starting from the 4th week of embryonic development, peaked in the 5th week, and gradually decreased after the 6th week. This expression trend is the temporal expression characteristic of the GATA4 gene.

[0078] Step S1247: Combine the description of the pathogenesis of birth defects in children in the biological knowledge base to verify the correlation between the temporal expression characteristics of suspected pathogenic genes and the disease progression.

[0079] In the biological knowledge base describing the pathogenesis of congenital heart disease, the 4th to 6th week of embryonic development is a critical period for the formation of the atrioventricular septum. The high expression sequence of the GATA4 gene during this period closely coincides with the time window for atrioventricular septal formation. Abnormal GATA4 gene expression may lead to incomplete development of the atrioventricular septum, resulting in congenital heart diseases such as ventricular septal defects or atrial septal defects. Through the above matching verification, it is confirmed that the temporal expression characteristics of the GATA4 gene are associated with the disease progression.

[0080] Step S1248: Screen out suspected pathogenic genes whose temporal expression characteristics are definitely associated with the disease pathogenesis. These genes meet the criteria for pathogenic gene expression characteristics. Integrate all genes that meet the criteria for pathogenic gene expression characteristics to generate a set of potential pathogenic genes, and label the screening criteria for each gene.

[0081] Based on the above verification, the temporal expression characteristics of suspected pathogenic genes such as GATA4, NKX2-5, and TBX5 are all definitely associated with the pathogenesis of congenital heart disease. Therefore, these genes were screened as meeting the criteria for pathogenic gene expression characteristics. These genes were integrated to generate a set of potential pathogenic genes, and each gene was labeled with screening criteria. For example, the screening criteria for the GATA4 gene included a strong association with ventricular septal defects, high expression during critical periods of embryonic heart development, and gene mutations leading to abnormal protein function.

[0082] Step S125: Based on the set of potential pathogenic genes, mine the signaling pathway nodes associated with potential pathogenic genes in the scientific research data association network, extract the upstream and downstream regulatory relationships between potential pathogenic genes and signaling pathway nodes, and generate signaling pathway regulatory links.

[0083] Taking the GATA4 gene, a potential pathogenic gene in the dataset, as an example, we explored its associated signaling pathway nodes, such as the "TGF-β signaling pathway" and the "BMP signaling pathway," within the research data association network. Extracting the upstream and downstream regulatory relationships between the GATA4 gene and these signaling pathway nodes revealed that the GATA4 gene is regulated by the Smad protein in the TGF-β signaling pathway. Simultaneously, the GATA4 gene can regulate the expression of the BMP2 gene in the BMP signaling pathway, thus generating the aforementioned signaling pathway regulatory chain: "TGF-β signaling pathway → Smad protein → GATA4 gene → BMP2 gene → BMP signaling pathway."

[0084] Step S126: Integrate the signaling pathway regulatory links to construct a key signaling pathway network, which includes the role positions of potential pathogenic genes in the signaling pathways and the cross-associations between signaling pathways.

[0085] Multiple signaling pathway regulatory links were integrated to construct a key signaling pathway network. Within this network, the roles of potential pathogenic genes in each signaling pathway were identified; for example, the GATA4 gene is located downstream of the TGF-β signaling pathway and upstream of the BMP signaling pathway. Simultaneously, the cross-regulation relationships between signaling pathways were demonstrated, such as the cross-regulation between the TGF-β and BMP signaling pathways through their combined action on the GATA4 gene, forming a complex signaling pathway network structure.

[0086] Step S127: Extract protein nodes, metabolite nodes, and clinical detection indicator nodes associated with the potential pathogenic gene set and key signaling pathway network from the scientific research data association network, and generate a candidate biomarker list.

[0087] From the network of scientific research data, protein nodes associated with potential pathogenic gene sets (such as GATA4, NKX2-5) and key signaling pathway networks (such as TGF-β, BMP signaling pathway) are extracted, such as “Smad4 protein” and “BMPR2 protein”; metabolite nodes, such as “ATP and lactate related to cardiomyocyte energy metabolism”; and clinical test indicator nodes, such as “pulmonary artery pressure value in echocardiography” and “serum troponin level”. The above nodes are summarized to generate a candidate biomarker list.

[0088] Step S128: For each node in the candidate biomarker list, analyze its correlation with the clinical phenotype of child birth defects, combine it with clinical testing feasibility information, score it using the clinical value assessment model built into the child birth defect research model, and screen out nodes with scores higher than a preset threshold.

[0089] For the "serum troponin level" node in the candidate biomarker list, analysis of its association with the clinical phenotype of "myocardial injury" revealed a high correlation. Furthermore, considering its clinical feasibility, serum samples are readily available, and the detection method is mature. This information was input into a clinical value assessment model, which scored the node based on multiple dimensions, including association strength, ease of detection, and stability. If the node's score exceeded a preset threshold, it was selected.

[0090] Step S129: Classify and organize the selected nodes with clinical testing value standards to generate a candidate biomarker list that includes protein biomarkers, metabolite biomarkers and clinical testing indicator biomarkers.

[0091] Nodes with scores exceeding a preset threshold are categorized according to their type. These categories include: protein biomarkers such as "Smad4 protein" and "BMPR2 protein"; metabolite biomarkers such as "ATP" and "lactic acid"; and clinical indicator biomarkers such as "pulmonary artery pressure" and "serum troponin level". The categorized biomarkers are then organized to generate a candidate biomarker list.

[0092] Step S1210: Integrate the potential pathogenic gene set, key signaling pathway network and candidate biomarker list to generate a scientific research discovery mining result that includes the contents of each part and the description of the relationship.

[0093] This study integrates a set of potential pathogenic genes, a network of key signaling pathways, and a list of candidate biomarkers. The research findings explain how potential pathogenic genes influence biomarker expression through key signaling pathways, and the relationship between these biomarkers and clinical phenotypes. For example, the GATA4 gene affects BMPR2 protein expression by regulating the BMP signaling pathway; BMPR2 protein, as a protein biomarker, is associated with the clinical phenotype of ventricular septal defect, forming a complete logical chain of research findings.

[0094] Step S130: The research findings mining results are processed for translational potential screening. Combining the biological knowledge base related to birth defects in children and clinical translation rules, a translational potential screening report of the research findings mining results is generated. The translational potential screening report includes the translational screening results and screening basis of each research discovery target.

[0095] In the context of congenital heart disease, the translational potential of potential pathogenic gene sets, key signaling pathway networks, and candidate biomarkers discovered through scientific research is screened. Each research discovery is evaluated by combining information from a biological knowledge base regarding the function and mechanism of action of these discoveries, as well as the provisions in clinical translation guidelines concerning translational feasibility, clinical needs, and risks. For example, the suitability of a potential pathogenic gene as a drug target and its druggability are assessed; the ease of drug intervention for a key signaling pathway; and the detection cost and clinical application prospects of a candidate biomarker. A translational potential screening report is generated based on the evaluation results, clarifying whether each research discovery meets the translational criteria and the corresponding screening basis.

[0096] Step S131: Extract the potential pathogenic gene set, key signaling pathway network and candidate biomarker list from the scientific research discovery mining results to form a set of scientific research discovery targets to be screened, and transfer the set of scientific research discovery targets to be screened to the screening and processing unit of the large-scale scientific research model for birth defects in children.

[0097] The results of scientific research discovery mining are used to extract potential pathogenic gene sets (such as GATA4, NKX2-5, etc.), key signaling pathway networks (such as TGF-β signaling pathway, BMP signaling pathway, etc.) and candidate biomarker lists (such as serum troponin level, BMPR2 protein, etc.), which are combined to form a set of scientific research discovery targets to be screened, and then transferred to the screening processing unit for subsequent screening processing.

[0098] Step S132: Retrieve the biological knowledge base and clinical translation rules related to birth defects from the knowledge storage unit of the large-scale scientific research model for children's birth defects. The clinical translation rules include translation feasibility judgment criteria, clinical need matching criteria, and translation risk assessment criteria.

[0099] The screening and processing unit retrieves a biological knowledge base from the knowledge storage unit. This knowledge base provides detailed biological functional information on genes, signaling pathways, and biomarkers related to congenital heart disease. Simultaneously, it retrieves clinical translation rules, including criteria for determining translation feasibility such as target drugability and the maturity of detection methods; criteria for matching clinical needs such as disease incidence and limitations of existing treatments; and criteria for assessing translation risks such as safety and ethical risks.

[0100] Step S133: For each potential pathogenic gene in the set of scientific research discovery targets to be screened, analyze its functional importance in the occurrence and development of birth defects in children based on a biological knowledge base, and generate a description of gene functional importance.

[0101] The function of the GATA4 gene, a potential pathogenic gene, was analyzed using a biological knowledge base. The knowledge base describes the transcription factor encoded by the GATA4 gene as regulating the expression of multiple downstream target genes during heart development, participating in key biological processes such as heart morphogenesis and cardiomyocyte differentiation. Abnormal GATA4 gene function directly affects normal heart development, leading to congenital heart disease. Based on this information, a description of gene functional importance was generated, such as "The GATA4 gene plays a core regulatory role in multiple key stages of heart development, and its abnormal function is one of the important causes of congenital heart disease."

[0102] Step S134: Combining the translation feasibility assessment criteria in the clinical translation rules, analyze the feasibility of each potential pathogenic gene as a drug target or diagnostic target, and generate gene translation feasibility analysis results.

[0103] Based on the feasibility assessment criteria in clinical translation rules, the feasibility of using the GATA4 gene as a drug target was analyzed. Factors considered included whether the structure of the GATA4 protein was suitable for small molecule drug binding, whether there were known compounds regulating GATA4 activity, and whether drugs targeting GATA4 might affect other normal tissues. If the analysis showed that the active pocket structure of the GATA4 protein was well-defined and a preliminary compound screening model existed, the gene translation feasibility analysis result was "GATA4 gene as a drug target has a certain degree of feasibility, and a preliminary basis for drug screening has been established."

[0104] Step S135: For each key signaling pathway in the set of research findings to be screened, analyze its regulatory role in the pathological process of birth defects in children based on a biological knowledge base, and generate a description of the importance of signaling pathway regulation.

[0105] Analysis of the TGF-β signaling pathway within the key signaling pathway network reveals that this pathway is involved in pathological processes such as heart valve formation and myocardial fibrosis, according to a biological knowledge base. In congenital heart disease, abnormal activation of the TGF-β signaling pathway can lead to excessive proliferation of cardiac septal tissue, thereby causing septal defects. Based on this information, a description of the importance of signaling pathway regulation is generated, such as "The TGF-β signaling pathway plays a crucial regulatory role in the pathological process of abnormal cardiac septal development in congenital heart disease, and its abnormal activation is a key factor leading to septal defects."

[0106] Step S136: Combining the clinical demand matching criteria in the clinical translation rules, analyze the clinical demand fit of each key signaling pathway as an intervention target, and generate signaling pathway translation demand analysis results.

[0107] Based on clinical needs matching criteria, the clinical need for the TGF-β signaling pathway as an intervention target was analyzed. In congenital heart disease, the treatment of septal defects currently relies mainly on surgery, lacking effective drug interventions, resulting in significant unmet clinical needs. As a key pathway regulating septal development, the TGF-β signaling pathway offers the potential for developing inhibitors that could provide new non-surgical treatment options. Therefore, the analysis of the signaling pathway translational needs concluded that "the TGF-β signaling pathway as an intervention target is highly matched with the clinical treatment needs of congenital heart disease septal defects, and has significant translational value."

[0108] Step S137: For each candidate biomarker in the set of scientific research discovery targets to be screened, analyze its diagnostic relevance to birth defects in children based on a biological knowledge base, and generate a description of the diagnostic value of the biomarker.

[0109] Regarding "serum troponin level" in the candidate biomarker list, the biological knowledge base describes it as a specific indicator of cardiomyocyte damage. In patients with congenital heart disease, myocardial hypoxia and ischemia can lead to elevated serum troponin levels, and the degree of elevation is correlated with the severity of myocardial damage. Based on this, a description of the diagnostic value of the biomarker is generated, such as "serum troponin level can serve as a diagnostic indicator of the degree of myocardial damage in patients with congenital heart disease, and its level changes can reflect the state of myocardial damage."

[0110] Step S138: Combining the translation risk assessment criteria in the clinical translation rules, analyze the detection technology maturity of each candidate biomarker and the relevant risks in the translation process, and generate biomarker translation risk analysis results.

[0111] For the candidate biomarker "serum troponin level," an analysis was conducted based on translational risk assessment criteria. The detection technology is mature, commercially available test kits are available, and the test results are highly accurate, which reduces the risk of technology translation. However, considering that serum troponin level may also be affected by other heart diseases, there is a certain specificity risk. The comprehensive analysis yielded the following biomarker translational risk analysis result: "The serum troponin level detection technology is mature, and the technology translation risk is low, but attention should be paid to the risk of cross-reactivity in other heart diseases."

[0112] Step S139: Based on the analysis results, a screening method is used to determine the transformation screening results of each research discovery target to be screened. The transformation screening results include transformation targets that meet the transformation screening criteria, transformation targets that need further evaluation, and transformation targets that are not included in the transformation for the time being.

[0113] Based on the combined descriptions of the functional importance of potential pathogenic genes and the results of translational feasibility analysis, the descriptions of the regulatory importance of key signaling pathways and the results of translational demand analysis, and the descriptions of the diagnostic value of candidate biomarkers and the results of translational risk analysis, a pre-defined screening method is used for determination. For example, if a potential pathogenic gene is functionally important and has high translational feasibility, it may be determined as a translational target that meets the translational screening criteria; if a candidate biomarker has high diagnostic value but uncertain translational risk, it may be determined as a translational target that requires further evaluation; if a key signaling pathway has low regulatory importance and poor clinical need matching, it may be determined as a translational target that is not included in the translation at this time.

[0114] Step S1391: Determine the primary screening indicators and secondary screening indicators. The primary screening indicators include functional importance indicators, translational feasibility indicators, clinical need indicators, and translational risk indicators. Each primary screening indicator contains multiple secondary screening indicators.

[0115] Secondary screening indicators are set under the functional importance indicators in the primary screening indicators, such as the irreplaceable role of genes in disease occurrence, the core regulatory role of signaling pathways, and the diagnostic specificity of biomarkers; secondary screening indicators are set under the translational feasibility indicators, such as the druggability of targets, the maturity of detection methods, and the difficulty of technology development; secondary screening indicators are set under the clinical need indicators, such as the degree of disease harm, the effectiveness of existing treatments, and the scale of market demand; and secondary screening indicators are set under the translational risk indicators, such as the level of safety risk, the degree of ethical controversy, and the difficulty of regulatory approval.

[0116] Step S1392: Set screening criteria for each secondary screening indicator, which are determined based on expert consensus data in the field of clinical translation of birth defects in children.

[0117] Based on expert consensus data in the field of clinical translation of birth defects in children, screening criteria were set for the secondary screening indicator of "target drugability", such as "the target protein has a clear active pocket structure and there are reports of at least one small molecule compound binding to it"; screening criteria were set for the secondary screening indicator of "diagnostic specificity", such as "the positive detection rate of biomarkers in patient samples is higher than 90%, and the false positive rate in normal control samples is lower than 5%", etc.

[0118] Step S1393: Compare the analysis results of each research discovery target to be screened with the screening criteria of the corresponding secondary screening indicators to generate the screening results for each secondary screening indicator.

[0119] For the GATA4 gene, a research discovery target to be screened, the "core regulatory role" in its description of gene functional importance is compared with the screening criteria of the secondary screening indicator "irreplaceability". If it meets the criteria, the screening result of the secondary indicator is "passed". The results of its gene transformation feasibility analysis are compared with the screening criteria of the secondary screening indicator "target druggability". If it meets the requirement of "reported binding of small molecule compounds", the screening result of the secondary indicator is also "passed".

[0120] Step S1394: Based on the screening results of each secondary screening indicator under each primary screening indicator, comprehensively determine the screening conclusion of each primary screening indicator. Based on the screening conclusion of the primary screening indicators of each research discovery target to be screened, comprehensively determine the overall screening result.

[0121] For the "GATA4 gene," under the functional importance indicator, if the screening results of multiple secondary screening indicators are all "passed," then the overall screening conclusion for the functional importance indicator is "high." Under the translational feasibility indicator, if the screening results of most relevant secondary indicators are "passed," then the screening conclusion for the translational feasibility indicator is "medium." The overall screening result is determined by combining the screening conclusions of each primary screening indicator, such as "high" functional importance, "medium" translational feasibility, "high" clinical need, and "low" translational risk.

[0122] Step S1395: Set the judgment criteria for conversion screening results. The judgment criteria include conditions that meet the conversion screening criteria, conditions that need further evaluation, and conditions that are not included in the conversion for the time being.

[0123] The criteria for meeting the conversion screening standards are set as follows: all primary screening indicators have a conclusion of "high" or "medium" and no conclusion of "low". The criteria for further evaluation are set as follows: there is a primary screening indicator with a conclusion of "low", but the conclusions of other indicators are relatively good, or the results of some secondary screening indicators are unclear. The criteria for not being included in the conversion are set as follows: two or more primary screening indicators have a conclusion of "low", or there are serious conversion obstacles.

[0124] Step S1396: Compare the overall screening results of each research discovery target to be screened with the set judgment conditions. Research discovery targets to be screened that meet the conditions for transformation screening are determined as transformation targets that meet the transformation screening criteria; research discovery targets to be screened that meet the conditions for further evaluation are determined as transformation targets for further evaluation; research discovery targets to be screened that meet the conditions for not being included in transformation are determined as transformation targets not being included in transformation for the time being.

[0125] The overall screening results of the "GATA4 gene" are compared with the judgment criteria. If it meets the criteria for transformation screening, it is identified as a transformation target that meets the criteria for transformation screening. If the overall screening results of a "metabolite biomarker" meet the criteria for further evaluation, it is identified as a transformation target for further evaluation. If the overall screening results of a "signaling pathway" meet the criteria for not being included in transformation for the time being, it is identified as a transformation target that is not included in transformation for the time being.

[0126] Step S1397: Record the screening process for each research discovery target to be screened, the screening results, and the criteria for judging the screening results.

[0127] The results of each secondary screening indicator, the conclusions of the primary screening indicator, the overall screening results, and the comparison with the judgment criteria are recorded in detail for each research discovery target to be screened. For example, the specific performance of the "GATA4 gene" in the secondary indicator of "target druggability" and the comprehensive judgment reasons for the functional importance indicators are recorded as the basis for judging the screening results.

[0128] Step S1310: Integrate the screening criteria, various analysis results, and transformation screening results for each research discovery target to be screened, and generate a transformation potential screening report of the research discovery mining results.

[0129] The report integrates the screening criteria, gene functional importance descriptions, translational feasibility analysis results, and final translational screening results (e.g., meeting translational screening criteria, requiring further evaluation, or not yet included in translation) for each research discovery target. The translational potential screening report lists relevant information for each research discovery target by category, such as translational screening results for potential pathogenic genes, key signaling pathways, and candidate biomarkers, along with detailed explanations of the screening criteria.

[0130] Step S140: Based on the transformation targets that meet the transformation screening criteria in the transformation potential screening report, generate a transformation plan for scientific research results on birth defects in children. The transformation plan for scientific research results on birth defects in children includes the target validation path, clinical transformation direction and phased transformation tasks of the transformation targets.

[0131] Based on the translational potential screening report, translational targets meeting the screening criteria were identified, such as the GATA4 gene as a drug target, the TGF-β signaling pathway as an intervention target, and serum troponin levels as a diagnostic biomarker. For each of these translational targets, target validation pathways were developed, clarifying which experiments are needed to verify their effectiveness; clinical translation directions were determined, such as drug development and diagnostic reagent development; and the translational process was broken down into multiple phased translational tasks, such as basic research, preclinical research, and clinical trials, ultimately integrating them to generate a translational plan for research findings on birth defects in children.

[0132] Step S141: Analyze the transformation potential screening report, extract the transformation targets that meet the transformation screening criteria in the report, and determine the types of transformation targets that meet the transformation screening criteria. The types of transformation targets that meet the transformation screening criteria include potential pathogenic genes that meet the criteria, key signaling pathways that meet the criteria, and candidate biomarkers that meet the criteria.

[0133] The translation potential screening report was analyzed to extract all translation targets that met the translation screening criteria. Based on the attributes of these targets, they were categorized into three types: potential pathogenic genes that meet the criteria (such as the GATA4 gene), key signaling pathways that meet the criteria (such as the TGF-β signaling pathway), and candidate biomarkers that meet the criteria (such as serum troponin levels), in order to develop targeted translation strategies.

[0134] Step S142: For potential pathogenic genes that meet the standards, construct a target verification path by combining the biological knowledge base related to birth defects in children. The target verification path includes an in vitro cell experiment verification step and an in vivo animal experiment verification step.

[0135] Taking the GATA4 gene, a potential pathogenic gene that meets the criteria, as an example, this study constructs a target validation pathway by combining information on the function of the GATA4 gene from a biological knowledge base. The in vitro cell experiment validation phase plans to knock out or overexpress the GATA4 gene using gene editing technology and observe its effects on cardiomyocyte differentiation and proliferation. The in vivo animal experiment validation phase plans to construct animal models with GATA4 gene mutations and observe whether these models exhibit congenital heart disease-related phenotypes, as well as the improvement in phenotypes after intervention with GATA4 gene expression.

[0136] Step S1421: Retrieve functional descriptions and related pathological mechanism information of potential pathogenic genes that meet the criteria from the biological knowledge base related to birth defects in children, and determine the core target of target validation.

[0137] Functional descriptions of the GATA4 gene, such as its role in regulating cardiomyocyte fate and participating in heart valve formation, were retrieved from a biological knowledge base. Information on related pathological mechanisms, such as the loss of transcriptional regulation due to GATA4 gene mutations, leading to abnormal heart development, was also collected. Based on this information, the core objective of target validation was determined to be: to verify whether abnormal GATA4 gene expression directly leads to the occurrence of congenital heart disease-related phenotypes, and whether intervention in GATA4 gene function can reverse or improve these phenotypes.

[0138] Step S1422: Based on the core objective of the target verification, construct an experimental plan for the in vitro cell experiment verification step. The experimental plan includes cell model selection, experimental group setting, and determination of experimental observation indicators.

[0139] To achieve the core objective of target validation, an in vitro cell experimental protocol was constructed. The selected cell model was a cardiomyocyte model induced from human pluripotent stem cells, which can simulate the human heart development process. Experimental groups included a normal control group (cardiomyocytes without gene editing), a GATA4 gene knockout group (cardiomyocytes with the GATA4 gene knocked out using CRISPR-Cas9 technology), and a GATA4 gene overexpression group (cardiomyocytes transfected with a GATA4 overexpression vector). The experimental monitoring indicators included the expression levels of cardiomyocyte markers (such as cTnT), cell morphological changes, cell beating frequency, and calcium transient characteristics.

[0140] Step S1423: Select a cell model associated with a child’s birth defect, the cell model comprising a normal cell model and a defective cell model, wherein the defective cell model must contain a known mutation or abnormal expression of the potential pathogenic gene that meets the criteria.

[0141] Normal cell models are cardiomyocytes induced to differentiate from pluripotent stem cells derived from healthy individuals; defective cell models are cardiomyocytes induced to differentiate from patient-specific pluripotent stem cells carrying known mutations in the GATA4 gene (such as missense mutations that lead to loss of protein function), or cardiomyocytes induced to differentiate after introducing GATA4 gene mutations into normal pluripotent stem cells through gene editing technology, in order to simulate abnormal GATA4 gene expression.

[0142] Step S1424: Set up experimental groups, which include a blank control group, a negative control group, a positive control group, and a potential pathogenic gene intervention group that meets the criteria. Each experimental group has multiple parallel experimental samples.

[0143] The blank control group consisted of normal cell models that had not undergone any treatment; the negative control group consisted of normal cell models transformed with an empty vector; the positive control group consisted of cell models transformed with a gene known to cause abnormal phenotypes in cardiomyocytes (such as another gene causing congenital heart disease); the GATA4 gene intervention group included a GATA4 gene knockout subgroup and a GATA4 gene overexpression subgroup. Each experimental group had at least three parallel experimental samples to ensure the reliability of the experimental results.

[0144] Step S1425: Determine the experimental observation indicators for the in vitro cell experiment verification step. The experimental observation indicators include gene expression level indicators, protein expression level indicators, and cell function change indicators.

[0145] Gene expression level indicators were detected by real-time quantitative PCR to measure the mRNA expression levels of GATA4 gene and its downstream target genes (such as Nkx2-5 and Tbx5); protein expression level indicators were detected by Western blot to measure the expression levels of GATA4 protein and related myocardial-specific proteins; cell function change indicators were obtained by observing cell morphology and pulsation under a microscope, detecting calcium transient characteristics using calcium imaging technology, and detecting cell proliferation capacity by cell counting.

[0146] Step S1426: Construct the operation procedure for in vitro cell experiments, and determine the sequence of experimental steps, operating conditions, and operation time nodes.

[0147] The in vitro cell experiment procedure includes: cell resuscitation and culture, performed under specific cell culture medium, temperature, and CO2 concentration conditions; gene editing or vector transfection, performed according to the transfection reagent instructions, and setting the culture time after transfection; cell differentiation induction, replacing the culture medium with cardiomyocyte differentiation induction medium, and controlling the induction time and differentiation stage; and detection of experimental observation indicators, including detection of gene expression, protein expression, and cell function indicators at different time points after differentiation (such as day 7, day 14, and day 21 of differentiation).

[0148] Step S1427: Based on the construction results of the in vitro cell experiment verification step, construct the experimental plan for the in vivo animal experiment verification step. The experimental plan includes the selection of animal models, the setting of experimental groups, and the determination of experimental observation indicators.

[0149] Based on the results of in vitro cell experiments, if GATA4 gene knockout leads to abnormal cardiomyocyte function, further in vivo animal experimental protocols will be developed. Commonly used model organisms, such as mice, will be selected as animal models to construct GATA4 gene conditional knockout mouse models or transgenic mouse models that specifically overexpress the GATA4 gene in specific tissues. Experimental groups will include a wild-type control group, a GATA4 gene heterozygous knockout group, and a GATA4 gene homozygous knockout group (if viable). Experimental observation indicators will include embryonic cardiac morphological examination, postnatal cardiac function assessment of mice, and cardiac tissue pathological analysis.

[0150] Step S1428: Select a transgenic animal model or gene-edited animal model carrying the potential pathogenic gene mutation that meets the criteria or that can mimic its abnormal expression.

[0151] Transgenic mouse models carrying known pathogenic mutations in the GATA4 gene, consistent with mutations found in human patients with congenital heart disease, were selected; alternatively, gene-edited mouse models with a specific GATA4 gene knock-in mutation were constructed using CRISPR-Cas9 technology to simulate the pathological state caused by abnormal GATA4 gene expression. Simultaneously, mouse strains with well-defined backgrounds and good genetic stability were selected for experiments.

[0152] Step S1429: Set up experimental groups for in vivo animal experiments, which correspond to the experimental groups in the in vitro cell experiment verification step, and set an appropriate number of animal samples for each group.

[0153] The in vivo animal experiments were grouped in accordance with the in vitro cell experiments, with the following groups: wild-type control group (litter mice without the GATA4 gene mutation), negative control group (transgenic mice carrying the empty vector), GATA4 gene mutation heterozygous group, and GATA4 gene mutation homozygous group (if feasible). The number of animal samples in each group was determined according to statistical requirements to ensure that differences between groups could be detected.

[0154] Step S14210: Determine the experimental observation indicators for the in vivo animal experiment verification step. The experimental observation indicators include histopathological indicators, blood biochemical indicators, and gene expression regulation indicators. Integrate the construction content of the in vitro cell experiment verification step and the in vivo animal experiment verification step to form the target verification path.

[0155] The histopathological indicators in in vivo animal experiments include HE staining of embryonic heart sections to observe cardiac structural abnormalities and immunohistochemical detection of specific protein expression localization; blood biochemical indicators include the detection of serum myocardial enzyme levels (such as CK-MB); and gene expression regulation indicators include the detection of mRNA and protein expression levels of the GATA4 gene and its downstream target genes in cardiac tissue. The experimental protocols, groupings, and observation indicators of both in vitro cell experiments and in vivo animal experiments are integrated to form a complete GATA4 gene target validation pathway.

[0156] Step S143: Based on the target validation path, determine the clinical translation direction of potential pathogenic genes that meet the standards. The clinical translation direction includes drug development and diagnostic reagent development.

[0157] Based on the target validation pathway results of the GATA4 gene, if it is verified that the GATA4 gene is a key pathogenic gene for congenital heart disease and its abnormal expression can be regulated by drugs, then the direction of drug development can be determined, such as developing GATA4 gene expression activators or transcription factor activity regulators. At the same time, if GATA4 gene mutations have high specificity and detection rate, the direction of diagnostic reagent development can be determined, such as developing gene diagnostic kits for detecting GATA4 gene mutations.

[0158] Step S144: Decompose the transformation process of the potential pathogenic genes that meet the criteria, and construct a phased transformation task. The phased transformation task includes gene function verification task, target activity verification task, candidate drug screening task, and preclinical trial preparation task.

[0159] The transformation process of the GATA4 gene is broken down into phased tasks. The gene function validation task corresponds to the in vitro and in vivo experimental validation work in the target validation pathway; the target activity validation task includes screening compounds that can specifically regulate GATA4 gene expression or GATA4 protein activity and validating the effects of the compounds on the target; the candidate drug screening task includes conducting preliminary pharmacodynamic and pharmacokinetic evaluations of the compounds to screen for candidate drugs with development potential; the preclinical trial preparation task includes formulation development of candidate drugs, acute toxicity tests, long-term toxicity tests, etc., to prepare for entry into clinical trials.

[0160] Step S145: For key signaling pathways that meet the standards, and in conjunction with the biological knowledge base related to birth defects in children, construct a verification path for signaling pathway intervention targets. The verification path for signaling pathway intervention targets includes a verification step for key pathway nodes and a verification step for pathway regulation effects.

[0161] Taking the TGF-β signaling pathway, a key signaling pathway that meets the standards, as an example, and combining information from the biological knowledge base regarding its involvement in pathological processes such as cardiac fibrosis and cardiomyocyte proliferation regulation, a validation pathway for intervention targets will be constructed. The validation phase for key pathway nodes plans to verify the roles of key nodes such as Smad protein and TGF-β receptor in congenital heart disease; the validation phase for pathway regulatory effects plans to observe the improvement effect on abnormal cardiac development phenotypes by intervening in these key nodes.

[0162] Step S146: Based on the signaling pathway intervention target validation path, determine the clinical translation direction of key signaling pathways that meet the standards. The clinical translation direction includes the development direction of pathway inhibitors and the development direction of pathway activators.

[0163] Based on the results of the signaling pathway intervention target validation pathway, if it is verified that overactivation of the TGF-β signaling pathway is an important cause of congenital heart disease, then the research and development direction of pathway inhibitors will be determined, and drugs that inhibit the activity of the TGF-β signaling pathway will be developed; if it is verified that insufficient pathway activity leads to the occurrence of disease, then the research and development direction of pathway activators will be determined, and drugs that enhance pathway activity will be developed.

[0164] Step S147: Decompose the transformation process of the key signaling pathway that meets the standard, and construct a phased transformation task. The phased transformation task includes a key pathway node identification task, an intervention drug screening task, a drug efficacy verification task, and a preclinical safety evaluation task.

[0165] The TGF-β signaling pathway transformation process is broken down into phased tasks. The key pathway node identification task corresponds to the work of validating key pathway nodes; the intervention drug screening task includes screening for inhibitors or activators targeting key nodes; the drug efficacy validation task includes validating the drug's regulatory effect on the signaling pathway and its ameliorative effect on disease phenotypes in cell and animal models; and the preclinical safety evaluation task includes assessing the drug's potential toxicity to normal tissues and organs, drug interactions, and other safety issues.

[0166] Step S148: For candidate biomarkers that meet the criteria, construct a biomarker validation path by combining the biological knowledge base related to birth defects in children. The biomarker validation path includes a small-sample clinical validation step and a large-sample clinical validation step.

[0167] For the candidate biomarker "serum troponin level" that meets the criteria, a biomarker validation pathway will be constructed by combining its association with myocardial injury information in the biological knowledge base. The small-sample clinical validation phase plans to collect serum samples from a small number of patients with congenital heart disease and healthy controls to detect serum troponin levels and preliminarily validate its diagnostic value. The large-sample clinical validation phase plans to collect a large number of samples from multiple centers to further validate its sensitivity, specificity, and diagnostic cutoff value.

[0168] Step S149: Based on the biomarker validation pathway, determine the clinical translation direction of candidate biomarkers that meet the standards. The clinical translation direction includes the development of diagnostic reagents and the development of prognostic assessment tools.

[0169] Based on the results of the biomarker validation pathway, if serum troponin levels have good diagnostic sensitivity and specificity in patients with congenital heart disease, then the direction for developing diagnostic reagents will be determined, and a serum troponin level detection kit will be developed for the auxiliary diagnosis of congenital heart disease. At the same time, if changes in its level are related to disease prognosis, then the direction for developing prognostic assessment tools will be determined, and a prognostic assessment model based on serum troponin levels will be constructed.

[0170] Step S1410: Decompose the transformation process of the candidate biomarkers that meet the standards, construct phased transformation tasks, which include detection method optimization tasks, small sample validation tasks, large sample validation tasks, and diagnostic reagent prototype development tasks, and integrate all the contents to generate a transformation plan for scientific research results on birth defects in children.

[0171] The translation process for serum troponin levels is broken down into phased tasks. The detection method optimization task includes optimizing antibody pairing, reaction conditions, and detection time to improve accuracy and efficiency. The small-sample validation task corresponds to the small-sample clinical validation stage in the biomarker validation pathway. The large-sample validation task corresponds to the large-sample clinical validation stage. The diagnostic reagent prototype development task includes designing the reagent kit's composition, packaging, and instructions for use, and developing a diagnostic reagent prototype. Integrating these phased translation tasks targeting potential pathogenic genes, key signaling pathways, and candidate biomarkers generates a complete translation plan for research findings on birth defects in children.

[0172] Step S150: Based on the proposed transformation plan for research results on birth defects in children, establish a transformation and implementation link for research results. Match and associate the phased transformation tasks in the proposed transformation plan for birth defects with the corresponding clinical research institutions and scientific research verification platforms to generate a transformation and implementation execution plan.

[0173] Based on the phased translation tasks in the research results translation plan for birth defects in children, such as gene function verification, candidate drug screening, and preclinical trial preparation, the resources and technical conditions required for each task are analyzed. Then, information on relevant clinical research institutions (such as hospitals with clinical research qualifications for congenital heart disease) and research validation platforms (such as gene editing technology platforms, drug screening platforms, and preclinical safety evaluation platforms) is collected. The phased translation tasks are matched and associated with clinical research institutions and research validation platforms with corresponding capabilities, clarifying which institution or platform will undertake each task, establishing timelines and collaboration methods, and ultimately generating a translation implementation plan to ensure that research results can smoothly transition from the laboratory to clinical application.

[0174] For example, step S151: Analyze the phased transformation tasks in the research results transformation plan for children's birth defects, and determine the task type, task requirements and task completion time limit for each phased transformation task.

[0175] The phased translation tasks in the translation plan are analyzed. For example, the "gene function verification task" is a basic research task. The task requirements include completing the construction of specified cell models and animal models, detecting specific observation indicators, and submitting experimental reports. The task completion deadline is set within six months after the project starts. The "candidate drug screening task" is a drug development task. The task requirements include screening at least three candidate compounds with good activity and completing preliminary pharmacodynamic evaluation. The task completion deadline is set within eight months after the completion of the gene function verification task.

[0176] Step S152: Collect information on clinical research institutions and scientific research verification platforms in the field of birth defects in children to form a set of information on translational partners. The set of information on translational partners includes the institution's name, research direction, technological advantages, and cooperation conditions.

[0177] Information on potential partners for commercialization was collected through research. Clinical research institutions, such as the "Cardiovascular Disease Research Institute of a University-Affiliated Children's Hospital," focus on clinical and basic research on congenital heart disease. Their technological advantage lies in their abundant patient resources and clinical research experience. Cooperation conditions include funding support and sharing of research results. Research validation platforms, such as the "XX Compound Sample Library Drug Screening Center," focus on drug discovery and screening. Their technological advantage lies in their high-throughput screening equipment and compound library resources. Cooperation conditions include service-based pricing and negotiation of intellectual property ownership. This information was compiled into a collection of potential commercialization partners.

[0178] Step S153: Analyze the task type and requirements of each phase of the transformation task to determine the technical capabilities and resource conditions required to complete the phase of the transformation task.

[0179] The "Drug Screening Task" is classified as a drug development task, requiring the screening of compounds that regulate GATA4 protein activity. The technical capabilities required to complete this task include compound screening experimental design, high-throughput screening equipment operation, and data analysis skills; resources include compound libraries, screening models (such as reporter gene cell lines stably expressing GATA4 protein), and relevant detection reagents.

[0180] Step S154: Based on the technical capabilities and resource conditions required to complete the phased transformation task, select clinical research institutions or scientific research verification platforms with corresponding technical capabilities and resource conditions from the transformation partner information set, and generate a candidate partner list.

[0181] Based on the technical capabilities and resource conditions required for the "drug screening task," a selection process was conducted from the information set of potential partners. If the "XX Compound Sample Library Drug Screening Center" possesses high-throughput screening equipment and compound library resources, and its technical advantages meet the task requirements, it will be included in the list of candidate partners; if other scientific research validation platforms also meet the requirements, they will also be added to the list.

[0182] Step S155: For each partner in the candidate partner list, evaluate them based on the matching degree between their research direction, technological advantages and the technical capabilities required for the phased transformation task, the conformity of their existing equipment and resources with the task requirements, and the compatibility of their project cycle with the task completion deadline, and generate a partner suitability evaluation result.

[0183] The "XX Compound Sample Library Drug Screening Center" in the candidate partner list was evaluated. Regarding research direction matching, its drug discovery and screening focus highly aligns with the "Drug Screening Task"; regarding technological advantages, its high-throughput screening equipment and compound library resources meet the technical capabilities required for the task; regarding equipment resources, the models of screening equipment and the size of its compound library meet the task requirements; and regarding project timeline compatibility, its committed screening cycle is within the task completion timeframe. A partner suitability assessment result, such as "High Suitability," was generated based on these factors.

[0184] Step S156: Based on the partner suitability assessment results, select the partner with the best assessment results for each phase of conversion task, and establish the association between the phase of conversion task and the selected partner.

[0185] Based on the partner suitability assessment results, the "XX Compound Sample Library Drug Screening Center" with the highest suitability was selected as the partner for the "Drug Screening Task". The relationship between the task and the center was established, and the responsibilities and obligations of both parties were clarified.

[0186] Step S157: Extract the relationships between all phased transformation tasks and build a transformation and implementation chain for scientific research results. The transformation and implementation chain includes the task flow sequence, the connection method with partners, and the information transmission path.

[0187] Extract the relationships between all phased translation tasks and partners, and build a translation implementation pipeline according to the task sequence (e.g., gene function verification task → target activity verification task → candidate drug screening task → preclinical trial preparation task). The task flow sequence clarifies how to initiate the next task after the previous one is completed; the partner coordination method specifies the communication and coordination mechanisms between different partners, such as holding regular project progress meetings; the information transmission path determines the methods and confidentiality requirements for the transmission of experimental data, research reports, and other information among partners.

[0188] Step S158: Based on the conversion implementation chain, construct the specific execution steps for each stage of the conversion task, and determine the person responsible for each step, the completion standards for each step, and the connection requirements between steps.

[0189] Taking the "candidate drug screening task" as an example, specific execution steps are constructed based on the translation and implementation chain, such as: Step 1: Both parties sign a technical service contract; Step 2: Provide screening models and compound screening criteria; Step 3: The screening center conducts initial screening of compounds; Step 4: Rescreen the positive compounds from the initial screening; Step 5: Submit the screening report. The responsible person for each step is identified; for example, Step 3 is the responsibility of the technical personnel at the screening center, and Step 4 is the joint responsibility of the project team's R&D personnel and the screening center personnel. The completion criteria for each step are as follows: For example, the initial screening requires testing a specified number of compounds, and the positive compounds in the rescreening must reach a preset activity threshold. The coordination requirements between steps are as follows: For example, after Step 3 is completed, the initial screening results must be fed back to the project team within five working days, and the project team must determine the list of compounds for rescreening within three working days of receiving the results.

[0190] Step S159: Combine the task completion time limit of each stage of conversion task, plan the overall conversion progress, and generate a conversion progress schedule.

[0191] Based on the completion timeframes for each phase of the translation task, such as six months for gene function validation, four months for target activity validation, six months for candidate drug screening, and twelve months for preclinical trial preparation, plan the overall translation schedule. Mark the start and end dates of each task, as well as key milestones (such as candidate drug identification and preclinical trial initiation dates), in the translation schedule to ensure that each task progresses according to plan.

[0192] Step S1510: Integrate the conversion implementation chain, the execution steps of phased conversion tasks, the relationship of partners and the conversion progress schedule, and generate a conversion implementation plan.

[0193] This plan integrates the entire process of technology transfer, the specific execution steps for each stage of the transfer task, the relationships between tasks and partners, and the transfer timeline. The transfer implementation plan clearly outlines the execution flow, responsible organization, timeline, and coordination methods for each task, providing a detailed action guide for the practical implementation of scientific research results.

[0194] Throughout the entire process described above, when collecting privacy-sensitive data (such as genomic data and clinical phenotype data of patients with congenital heart disease), the privacy protection and leak prevention technologies employed include: data anonymization, removing patients' personal identification information (such as name, ID number, contact information, etc.), retaining only the sample ID and medical characteristic data required for the research; encrypted data transmission, using encryption algorithms to encrypt data during data collection and transmission to the large-scale research model of birth defects in children to prevent theft during transmission; access control, setting strict access management for sensitive data such as research data association networks and research findings mining results, authorizing only researchers to access specific ranges of data according to their research needs; and data storage security, using data storage servers that comply with XX information security standards, and regularly backing up and auditing data to prevent leakage and loss during data storage.

[0195] In one exemplary embodiment, a research achievement transformation system based on a large-scale research model of child birth defects is provided. This system can be a terminal, server, etc., and its internal structure diagram can be as follows: Figure 2As shown, this research achievement transformation system based on a large-scale research model of child birth defects includes a processor, memory, input / output interface, communication interface, display unit, and input device. The processor, memory, and input / output interface are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interface. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The input / output interface is used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, near-field communication, or other technologies. When the computer program is executed by the processor, it implements a method for transforming research achievements based on a large-scale research model of child birth defects. The display unit is used to form a visually visible image and can be a display screen, projection device, or virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device can be a touch layer covering the display screen, or a button, trackball, or touchpad set on the shell of the scientific research achievement transformation system based on the large-scale scientific research model of children's birth defects, or an external keyboard, touchpad, or mouse, etc.

[0196] It should be noted that, in order to simplify the description of the present invention and thus help to understand one or more embodiments of the invention, multiple features may sometimes be grouped into one embodiment, drawing or description thereof in the foregoing description of the embodiments of the present invention.

Claims

1. A scientific research achievement transformation method based on a child birth defect scientific research large model, characterized in that, The method includes: The input research data on birth defects in children is processed by association mapping. The research data on birth defects in children is then imported into the association processing unit of the large-scale research model on birth defects in children, generating a cross-data source research data association network. The research data association network contains the association relationships and association basis between different data types. The research data on birth defects in children includes genomics data, proteomics data, clinical phenotype data, and scientific literature data. Based on the aforementioned research data association network, a large-scale research model for child birth defects is executed to mine research findings and generate research findings mining results that include a set of potential pathogenic genes, a network of key signaling pathways, and a list of candidate biomarkers. The research findings are processed for translational potential screening. Combined with the biological knowledge base related to birth defects in children and clinical translation rules, a translational potential screening report of the research findings is generated. The translational potential screening report includes the translational screening results and screening basis for each research discovery target. Based on the transformation potential screening report, a transformation plan for scientific research results on birth defects in children is generated. The transformation plan includes the target validation path, clinical transformation direction and phased transformation tasks of the transformation target. Based on the aforementioned research results transformation plan for children's birth defects, a research results transformation and implementation link is established. The phased transformation tasks in the research results transformation plan for children's birth defects are matched and associated with the corresponding clinical research institutions and scientific research verification platforms to generate a transformation and implementation plan. 2.The method of transforming research findings based on a child birth defect research big model according to claim 1, wherein, The process of performing association mapping on the input research data on birth defects, importing the research data on birth defects into the association processing unit of the large-scale research model on birth defects, and generating a cross-data source research data association network includes: Genomic data, proteomics data, clinical phenotype data, and scientific literature data related to birth defects in children are collected to form an initial research data set. The initial research data set is then transmitted to the association processing unit of the large-scale research model for birth defects in children. The core descriptive information of each type of data in the initial research data set is extracted. The core descriptive information includes the characteristics of the research object corresponding to the data, the data collection conditions, and the core observation indicators of the data. Based on the terminology standardization rules built into the research model for birth defects in children, the core descriptive information of different types of data is standardized to generate standardized core descriptive information, and the associated key fields in the standardized core descriptive information are extracted. The associated key fields include gene name, protein name, clinical symptom name and disease type name. Based on the aforementioned key association fields, an association mapping relationship is established between genomics data and proteomics data to generate gene-protein association pairs. The gene-protein association pairs contain information on the associated genes and proteins and a description of the association strength. Based on the aforementioned key association fields, an association mapping relationship is established between proteomics data and clinical phenotype data, generating protein-clinical phenotype association pairs. The protein-clinical phenotype association pairs include information on the associated proteins and clinical phenotypes, as well as a description of the association strength. Based on the aforementioned key fields, an association mapping relationship is established between genomics data and scientific literature data, generating gene-literature association pairs. Each gene-literature association pair contains information on the associated genes and literature, as well as a summary of the basis for the association. By integrating the gene-protein association pairs, protein-clinical phenotype association pairs, and gene-literature association pairs, an initial scientific research data association network is constructed, which includes nodes of various association pairs and the association lines between nodes. Supplement the initial scientific research data association network with supporting evidence information for various associations, including citations of experimental verification results, citations of clinical observation data, and citations of literature conclusions; The structure of the initial scientific research data association network after supplementing supporting evidence information is optimized, the indirect association paths between nodes are extracted, and a cross-data source scientific research data association network containing direct and indirect association relationships is generated. 3.The method of transforming research findings based on a child birth defect research big model according to claim 1, wherein, The research discovery mining process based on the research data association network, which performs a large-scale research model on birth defects in children, generates research discovery mining results containing a set of potential pathogenic genes, key signaling pathway networks, and a list of candidate biomarkers, including: The research data association network is input into the mining and processing unit of the large-scale research model of birth defects in children, and the node types and the types of association relationships between nodes in the research data association network are analyzed. Based on the pathogenic association mining rules built into the large-scale scientific research model of child birth defects, gene nodes that are directly associated with clinical phenotype nodes of child birth defects in the scientific research data association network are screened to generate a set of candidate pathogenic genes. For each gene node in the candidate pathogenic gene set, extract its associated protein nodes, associated literature nodes, and associated clinical phenotype nodes in the scientific research data association network to generate a gene association information package. Based on the gene association information package, the rule engine built into the large-scale research model for child birth defects is used for logical judgment to screen out genes that meet the criteria for pathogenic gene expression characteristics. Based on the set of potential pathogenic genes, signaling pathway nodes associated with potential pathogenic genes in the scientific research data association network are mined, and the upstream and downstream regulatory relationships between potential pathogenic genes and signaling pathway nodes are extracted to generate signaling pathway regulatory links. By integrating the aforementioned signaling pathway regulatory links, a key signaling pathway network is constructed, which includes the role positions of potential pathogenic genes in the signaling pathways and the cross-associations between signaling pathways. Protein nodes, metabolite nodes, and clinical detection indicator nodes associated with potential pathogenic gene sets and key signaling pathway networks are extracted from the scientific research data association network to generate a candidate biomarker list. For each node in the candidate biomarker list, analyze its correlation with the clinical phenotype of child birth defects. Combined with clinical testing feasibility information, score it using the clinical value assessment model built into the child birth defect research model, and screen out nodes with scores higher than a preset threshold. The nodes selected with clinical testing value standards are classified and organized to generate a candidate biomarker list that includes protein biomarkers, metabolite biomarkers and clinical testing indicator biomarkers. By integrating the potential pathogenic gene set, key signaling pathway network, and candidate biomarker list, a research discovery mining result containing explanations of each part and their relationships is generated. 4.The method of transforming research findings based on a child birth defect research big model according to claim 1, wherein, The process of screening the translational potential of the research findings, combined with a biological knowledge base related to birth defects in children and clinical translation rules, generates a translational potential screening report for the research findings, including: Extract the potential pathogenic gene set, key signaling pathway network and candidate biomarker list from the scientific research discovery mining results to form a set of scientific research discovery targets to be screened, and transfer the set of scientific research discovery targets to be screened to the screening and processing unit of the large-scale scientific research model for birth defects in children. The biological knowledge base and clinical translation rules related to birth defects are retrieved from the knowledge storage unit of the large-scale scientific research model of birth defects. The clinical translation rules include translation feasibility judgment criteria, clinical need matching criteria, and translation risk assessment criteria. For each potential pathogenic gene in the set of research discovery targets to be screened, its functional importance in the occurrence and development of birth defects in children is analyzed based on a biological knowledge base, and a description of gene functional importance is generated. By combining the translational feasibility assessment criteria in the clinical translation rules, we analyze the feasibility of each potential pathogenic gene as a drug target or diagnostic target, and generate gene translation feasibility analysis results. For each key signaling pathway in the set of research findings to be screened, its regulatory role in the pathological process of birth defects in children is analyzed based on a biological knowledge base, and a description of the importance of signaling pathway regulation is generated. By combining the clinical demand matching criteria in the clinical translation rules, we analyze the clinical demand fit of each key signaling pathway as an intervention target and generate signaling pathway translation demand analysis results. For each candidate biomarker in the set of scientific research discovery targets to be screened, its diagnostic relevance to birth defects in children is analyzed based on a biological knowledge base, and a description of the diagnostic value of the biomarker is generated. By combining the translation risk assessment criteria in the clinical translation rules, we analyze the detection technology maturity of each candidate biomarker and the relevant risks in the translation process, and generate biomarker translation risk analysis results. Based on the analysis results, and the comparison results of each research discovery target to be screened with the preset screening criteria, the transformation screening results of each research discovery target to be screened are determined by weighted scoring algorithm or decision tree model. The transformation screening results include transformation targets that meet the transformation screening criteria, transformation targets that need further evaluation, and transformation targets that are not included in the transformation for the time being. By integrating the screening criteria, analysis results, and transformation screening results for each research discovery target to be screened, a report on the transformation potential of the research discovery mining results is generated. 5.The method of transforming research findings based on a child birth defect research big model according to claim 1, wherein, Based on the transformation potential screening report and the transformation targets that meet the transformation screening criteria, a transformation plan for scientific research results on birth defects in children is generated, including: The transformation potential screening report is analyzed, and transformation targets that meet the transformation screening criteria are extracted from the report. The types of transformation targets that meet the transformation screening criteria are determined. The types of transformation targets that meet the transformation screening criteria include potential pathogenic genes that meet the criteria, key signaling pathways that meet the criteria, and candidate biomarkers that meet the criteria. For potential pathogenic genes that meet the criteria, a target validation pathway is constructed by combining a biological knowledge base related to birth defects in children. The target validation pathway includes in vitro cell experiment validation and in vivo animal experiment validation. Based on the target validation pathway, clinical translation directions for potential pathogenic genes that meet the standards are identified, including drug development and diagnostic reagent development. The transformation process of the potential pathogenic genes that meet the criteria is decomposed, and a phased transformation task is constructed. The phased transformation task includes gene function verification task, target activity verification task, candidate drug screening task, and preclinical trial preparation task. For key signaling pathways that meet the standards, and in conjunction with a biological knowledge base related to birth defects in children, a verification path for signaling pathway intervention targets is constructed. The verification path for signaling pathway intervention targets includes a verification step for key pathway nodes and a verification step for pathway regulation effects. Based on the aforementioned signaling pathway intervention target validation path, clinical translation directions for key signaling pathways that meet the standards are determined. These clinical translation directions include pathway inhibitor development and pathway activator development. The transformation process of the key signaling pathways that meet the standards is decomposed, and a phased transformation task is constructed. The phased transformation task includes the task of identifying key nodes of the pathway, the task of screening intervention drugs, the task of verifying the drug's effect, and the task of preclinical safety evaluation. For candidate biomarkers that meet the criteria, a biomarker validation pathway is constructed by combining a biological knowledge base related to birth defects in children. The biomarker validation pathway includes a small-sample clinical validation stage and a large-sample clinical validation stage. Based on the biomarker validation pathway, clinical translation directions for candidate biomarkers that meet the criteria are determined, including diagnostic reagent development and prognostic assessment tool development. The transformation process of the candidate biomarkers that meet the standards is decomposed, and a phased transformation task is constructed. The phased transformation task includes the task of optimizing the detection method, the task of small sample validation, the task of large sample validation, and the task of developing the prototype of the diagnostic reagent. All the contents are integrated to generate a transformation plan for scientific research results on birth defects in children. 6.The method of transforming research findings based on a child birth defect research big model according to claim 2, wherein, The terminology standardization rules built into the large-scale research model for birth defects in children standardize the core descriptive information of different types of data, generating standardized core descriptive information, including: The terminology mapping relationships, terminology naming conventions, and terminology classification standards in the terminology unification rules are analyzed. Extract terms from the core descriptive information of different types of data to form a set of terms to be standardized, which includes genomics terms, proteomics terms, clinical phenotype terms and literature terms. For each genomic term in the set of terms to be standardized, non-standard genomic terms are mapped to standard genomic terms according to the term mapping relationship in the terminology unification rules, thereby generating standardized genomic terms; For each proteomics term in the set of terms to be standardized, the non-standard proteomics term is mapped to a standard proteomics term according to the term mapping relationship in the terminology unification rules, thereby generating a standardized proteomics term. For each clinical phenotype term in the set of terms to be standardized, the naming method of non-standard clinical phenotype terms is modified according to the terminology naming specifications in the unified terminology rules to generate standardized clinical phenotype terms. For each document term in the set of terms to be standardized, the document term is classified into the corresponding standard term category according to the term classification standard in the terminology unification rules, thereby generating standardized document terms; Replace the non-standard terms in the original core descriptive information with the generated standardized genomics terminology, standardized proteomics terminology, standardized clinical phenotype terminology, and standardized literature terminology; The core descriptive information after term replacement is processed for sentence fluency, generating a standardized core descriptive information set containing all standardized core descriptive information, thus completing the term standardization process. 7.The method of transforming research findings based on a child birth defect research big model according to claim 3, characterized in that, Based on the gene association information package, the expression regulation pattern of each candidate pathogenic gene is analyzed. Combined with a biological knowledge base related to birth defects in children, genes meeting the expression characteristics criteria for pathogenic genes are screened to generate a set of potential pathogenic genes, including: Extract the associated protein node information from the gene association information package of each candidate pathogenic gene, analyze the interaction pattern between the candidate pathogenic gene and the associated protein, and determine the regulatory mode of gene expression products. Extract the associated literature node information from the gene association information package, summarize the research conclusions on the expression characteristics of candidate pathogenic genes in the literature, and generate a gene expression literature summary. Extract clinical phenotype association node information from gene association information packages and analyze the expression differences of candidate pathogenic genes in different clinical phenotypes of birth defects in children. The biological knowledge base related to birth defects is retrieved from the knowledge storage unit of the large-scale scientific research model of birth defects, and the description of the expression characteristics of pathogenic genes is extracted to generate reference standards for the expression characteristics of pathogenic genes. Comparative analysis was conducted on the expression regulation patterns of each candidate pathogenic gene, the summary of gene expression literature, and the fit between expression difference characteristics and the reference standards for pathogenic gene expression characteristics. Candidate pathogenic genes that met the reference standards for pathogenic gene expression characteristics were identified as suspected pathogenic genes. For the suspected pathogenic genes, further analysis was conducted on their expression trends during the critical period of birth defect onset in children to generate gene temporal expression characteristics; By combining descriptions of the pathogenesis of birth defects in children from biological knowledge bases, the correlation between the temporal expression characteristics of suspected pathogenic genes and the disease progression was verified. Suspected pathogenic genes whose temporal expression characteristics are definitely associated with the disease progression are screened out. These genes meet the criteria for pathogenic gene expression characteristics. All genes that meet the criteria for pathogenic gene expression characteristics are integrated to generate a set of potential pathogenic genes, and the screening criteria for each gene are labeled.

8. The method for transforming research results based on a large-scale research model of birth defects in children, as described in claim 4, is characterized in that... Based on the analysis results, a screening method is used to determine the transformation screening results for each research discovery target to be screened, including: The primary and secondary screening indicators are determined. The primary screening indicators include functional importance indicators, translational feasibility indicators, clinical need indicators, and translational risk indicators. Each primary screening indicator contains multiple secondary screening indicators. Screening criteria were set for each secondary screening indicator, and these criteria were determined based on expert consensus data in the field of clinical translation of birth defects in children. The analysis results of each research discovery target to be screened are compared with the screening criteria of the corresponding secondary screening indicators to generate the screening results for each secondary screening indicator. Based on the screening results of each secondary screening indicator under each primary screening indicator, the screening conclusion of each primary screening indicator is comprehensively determined. Based on the screening conclusion of the primary screening indicators of each research discovery target to be screened, the overall screening result is comprehensively determined. Set the criteria for determining the conversion screening results. The criteria include conditions that meet the conversion screening standards, conditions that require further evaluation, and conditions that are not included in the conversion for the time being. The overall screening results of each research discovery target to be screened are compared with the set judgment criteria. Research discovery targets to be screened that meet the criteria for transformation screening are determined to be transformation targets that meet the criteria for transformation screening; research discovery targets to be screened that meet the criteria for further evaluation are determined to be transformation targets that require further evaluation; research discovery targets to be screened that meet the criteria for not being included in transformation are determined to be transformation targets that are not included in transformation for the time being. Record the screening process for each research discovery target to be screened, the screening results, and the criteria for judging the screening results.

9. The method for transforming research results based on a large-scale research model of birth defects in children, as described in claim 5, is characterized in that... The proposed target validation pathway, based on a biological knowledge base related to birth defects in children and targeting potential pathogenic genes that meet the criteria, includes: Retrieve functional descriptions and related pathological mechanism information of potential pathogenic genes that meet the criteria from the biological knowledge base related to birth defects in children, and determine the core objectives of target validation; Based on the core objective of target validation, an experimental protocol for the in vitro cell experiment validation step is constructed. The experimental protocol includes cell model selection, experimental grouping settings, and determination of experimental observation indicators. Select cell models associated with birth defects in children, including normal cell models and defective cell models, wherein the defective cell models must contain known mutations or abnormal expression of the potential pathogenic genes that meet the criteria; Experimental groups were set up, including a blank control group, a negative control group, a positive control group, and a potential pathogenic gene intervention group that met the criteria. Multiple parallel experimental samples were set up in each experimental group. The experimental observation indicators for the in vitro cell experiment verification process are determined, including gene expression level indicators, protein expression level indicators, and cell function change indicators. Construct an operational procedure for in vitro cell experiments, and determine the sequence of experimental steps, operational conditions, and operational time points; Based on the results of the in vitro cell experiment verification, an experimental protocol for the in vivo animal experiment verification is constructed. The experimental protocol includes the selection of animal models, the setting of experimental groups, and the determination of experimental observation indicators. Select transgenic animal models or gene-edited animal models that carry the aforementioned compliant potential pathogenic gene mutations or can mimic their abnormal expression. The in vivo animal experiments are set up with experimental groups that correspond to the experimental groups in the in vitro cell experiments verification stage, and each group has an appropriate number of animal samples. The experimental observation indicators for the in vivo animal experiment verification step are determined, including histopathological indicators, blood biochemical indicators, and gene expression regulation indicators. The construction content of the in vitro cell experiment verification step and the in vivo animal experiment verification step is integrated to form the target verification path.

10. A research achievement transformation system based on a large-scale research model of birth defects in children, characterized in that, include: processor; A machine-readable storage medium for storing machine-executable instructions of the processor; The processor is configured to execute the research results transformation method based on the large-scale research model of child birth defects as described in any one of claims 1 to 9 by executing the machine-executable instructions.

Images