KR-20260064626-A - Method for providing results of analysis of the relationship between natural products and disease using Artificial Intelligence

KR20260064626AKR 20260064626 AKR20260064626 AKR 20260064626AKR-20260064626-A

Abstract

A method for providing a result of analyzing the correlation between a natural product and a disease using artificial intelligence according to an embodiment of the present invention comprises: a data search step for searching for relevant data regarding natural product information and disease information input by a user; a core target extraction step for extracting a core target based on the relevant data regarding the natural product information and disease information by a target extraction unit of a server; an analysis step for analyzing the correlation between the natural product and the disease using the relevant data regarding the natural product information and disease information by an enrichment analysis unit of a server; and a result provision step for delivering the analysis result regarding the correlation between the natural product and the disease to a customer terminal by an output unit of a server.

Inventors

강재하
신지헌
이래형
정대식
강기성

Assignees

주식회사 인실리콕스

Dates

Publication Date: 20260507
Application Date: 20251030
Priority Date: 20241031

Claims (14)

As a method of providing analysis results of the relationship between natural products and diseases using artificial intelligence, A data retrieval step for searching for relevant data regarding natural product information and disease information entered by the user; A core target extraction step for extracting core targets based on relevant data regarding the natural product information and disease information by the target extraction unit of the server; and An analysis step of analyzing the correlation between natural products and diseases using relevant data regarding the natural product information and disease information by the concentration analysis unit of the server; and A result providing step of transmitting the analysis results regarding the correlation between the natural product and the disease to a customer terminal by the output unit of the server; Characterized by including, A method for providing analysis results on the relationship between natural products and diseases using artificial intelligence.
In paragraph 1, The above data search step is. A step of searching for the components of the input natural product and at least one related target protein for each of the components through a pre-prepared natural product DB; and A step of searching for at least one disease-related gene associated with the input disease through a pre-prepared natural product database; including, A method for providing analysis results on the relationship between natural products and diseases using artificial intelligence.
In paragraph 2, The above core target extraction step is, A step of extracting at least one potential target associated with a component by a potential target extraction unit of a target extraction unit; and A step of extracting at least one core target by filtering the potential targets based on centrality analysis criteria using the core target extraction unit of the target extraction unit; including, A method for providing analysis results on the relationship between natural products and diseases using artificial intelligence.
In paragraph 3, The above core target extraction step is, Prior to the step of extracting the above potential targets, The method further includes a step of extracting common targets commonly contained in the natural product and the disease by a common target extraction unit of a target extraction unit, wherein a protein commonly contained in the set of related target proteins and the set of disease-related genes is extracted as a common target. The step of extracting the potential target is to filter the common targets based on a relevance score entered by the user to extract at least one potential target associated with the component. A method for providing analysis results on the relationship between natural products and diseases using artificial intelligence.
In paragraph 4, After the step of extracting the above common targets, A step of inputting the components of the natural product and the common targets into a drug-protein interaction prediction model; and A step of outputting the result of component-target interaction between a component and a target; Perform more, The above component-target interaction result is the presence or absence of interaction between the component and the target, or an affinity score, and In the step of extracting the potential target, the potential target is extracted based on the component-target interaction result in addition to the relevance score. A method for providing analysis results on the relationship between natural products and diseases using artificial intelligence.
In paragraph 3, The above core target extraction step is, It further includes a step of outputting the results of component-target interactions between the component and the disease-related gene, and The above component-target interaction result is the presence or absence of interaction between the component and the target, or an affinity score, and The step of extracting the above potential target is to extract at least one potential target associated with a component by filtering based on the presence or absence of interaction or affinity score entered by the user. A method for providing analysis results on the relationship between natural products and diseases using artificial intelligence.
In paragraph 6, The step of outputting the above component-target interaction results is to input the above components and the above disease-related genes into a drug-protein interaction prediction model and output the component-target interaction results between the components and the disease-related genes. A method for providing analysis results on the relationship between natural products and diseases using artificial intelligence.
In paragraph 2, The above result provision step includes generating a CTP network, After creating the CTP network, A step of inputting component pairs for the constituent components into a drug-drug interaction prediction model; and A step of outputting the drug interaction type and the probability (p) for each interaction type for each pair of components; Includes, The above drug interaction prediction model is an artificial intelligence model trained to perform calculations on gene expression information based on drug structure information and drug attribute information, and After the step of outputting the probability (p) for each interaction type mentioned above, A step of correcting the probability (p) by calculating a modified probability (p') for each pair of components based on the above drug interaction type and the probability (p) for each interaction type; and A step of reflecting the modified probability (p') for each pair of components into the CTP network; performing more, A method for providing analysis results on the relationship between natural products and diseases using artificial intelligence.
In paragraph 8, The above modified probability (p') is expressed by the following formula, and p' = p * m_s * max(1-λσ, 0) Here, p is the probability for each interaction type, m_s is the severity factor, λ is the uncertainty penalty constant, and σ is the standard deviation (prediction uncertainty), A method for providing analysis results on the relationship between natural products and diseases using artificial intelligence.
In Paragraph 9, The above result provision step includes generating a CTP network, but, The method includes the step of reflecting the modified probability (p') for each of the above-mentioned component pairs into the CTP network. A method for providing analysis results on the relationship between natural products and diseases using artificial intelligence.
In paragraph 3, After the step of extracting the above potential targets, Performing a PPI step of generating a protein interaction network by performing PPI analysis on the potential targets by the network generation unit of the server, A method for providing analysis results on the relationship between natural products and diseases using artificial intelligence.
In Paragraph 11, The above analysis step further includes, after the step of extracting the above core targets, a step of performing a first enrichment analysis and a second enrichment analysis for each core target by the enrichment analysis unit of the server, and The method comprises performing the first enrichment analysis to predict the function of each core target and generating a gene ontology, and performing the second enrichment analysis to generate a biological pathway of each core target. A method for providing analysis results on the relationship between natural products and diseases using artificial intelligence.
In Paragraph 12, After the step of performing the second concentration analysis mentioned above, The method further includes a step of generating a CTP network that expresses the relationships between the components of the natural product, key targets, and biological pathways as a CTP network. A method for providing analysis results on the relationship between natural products and diseases using artificial intelligence.
A computer program stored on a computer-readable recording medium to execute a method of providing results of an analysis of the correlation between a natural product and a disease using artificial intelligence according to any one of claims 1 to 13, combined with hardware.

Description

Method for providing results of analysis of the relationship between natural products and disease using Artificial Intelligence The present invention relates to a method for providing results of an analysis of the correlation between natural products and diseases using artificial intelligence. Network pharmacology is an analytical method based on systems biology that explores drug efficacy by connecting the active ingredient of a drug, target genes (or proteins), and diseases (pathologies) associated with the target genes to form a network. Traditionally, new drug development has proceeded based on the single-component, single-target theory—that is, identifying components that interact with a single target. However, while this methodology offers advantages in terms of safety, it was not suitable for proving the efficacy of natural products, which are multi-component, multi-target. In this context, the concept of multi-compound, multi-target is gaining new attention. In particular, since herbal medicines, a type of natural product, are composed of diverse components, this is a new methodology attracting significant interest in the field of herbal medicine development. Specifically, it involves networking the components contained in herbal ingredients and understanding them in conjunction with biological and disease networks. This approach not only facilitates a comprehensive understanding of herbal ingredients but also enables the analysis of the associations between herbal ingredients, their corresponding target proteins, and disease genes. To conduct research based on such network pharmacology, it is fundamentally necessary to secure a large amount of data at each stage of the study. However, natural products, such as herbal medicines, are extremely diverse, and each product consists of multiple constituent components, resulting in the existence of multiple sets of protein data associated with those components. Furthermore, numerous diseases exist, each with multiple related genes. Moreover, as this vast amount of data is currently dispersed across various institutions, acquiring the necessary data remains a difficult reality. Furthermore, due to the nature of network pharmacology, researchers must perform a relatively large number of activities when conducting drug-based studies on the relationship between natural products and diseases. The primary goal of Korean traditional medicine research is to elucidate the mechanisms of how herbal medicines interact within the human body, and the core of the research is to identify the associations between drugs, genes, and diseases. If a researcher has identified a specific herbal medicine of interest and a related disease, they must undergo numerous trials and errors to determine the connection between the two, given the vast amount of associated data. In addition, various programs must be used to visualize and understand the interactions between data. FIG. 1 is a conceptual diagram of a system that performs a method of providing results of an analysis of the relationship between natural products and diseases using artificial intelligence according to one embodiment of the present invention. FIG. 2 is a configuration diagram of a server that performs a method of providing results of an analysis of the correlation between natural products and diseases using artificial intelligence according to an embodiment of the present invention. Figure 3 is a reference diagram to explain the concept of the components of natural products and related target proteins. Figure 4 is a reference diagram to explain the concept of disease-related genes. FIG. 5 is a flowchart of a method for providing results of an analysis of the correlation between natural products and diseases using artificial intelligence according to an embodiment of the present invention. Figure 6 is a flowchart regarding the data retrieval step of Figure 5. FIG. 7 is a flowchart relating to the core target extraction step in one embodiment of the present invention. FIG. 8a is a flowchart for one embodiment of extracting potential targets based on relevance scores in the core target extraction step of FIG. 7. FIG. 8b is a flowchart relating to an embodiment in which the component-target interaction result is additionally applied in the step of extracting potential targets of FIG. 8a. FIG. 9 is a flowchart relating to another embodiment of extracting potential targets based on relevance scores and affinity scores in the core target extraction step of FIG. 5. Figure 10 is a flowchart regarding the correlation analysis step. Figure 11 is a flowchart regarding the correction of the probability (p) by interaction type of component pairs in the correlation analysis step of Figure 10. Figure 12 is a diagram exemplarily showing the constituent components of a natural product and related target protein information. Figure 13 is a diagram visually illustrating a method for extracting potential targets from a common target ac