Lung cancer medical big data-based treatment pathway key node information processing method

Abstract

The invention belongs to the technical field of information data processing, and discloses a lung cancer medical big data-based treatment pathway key node information processing method, which comprises the following steps: performing multi-source data fusion on data in an existing biomedical data set, and selecting five biomedical databases from the Internet to obtain lung cancer related information; performing data cleaning, data disambiguation and data compression on the source data; converting the preprocessed data into an RDF format to form a knowledge network; taking entities representinggenes, proteins and the like as subjects and objects according to the definition of the RDF, and taking the relation between the entities as a predicate, so that unified organization and representation of data are realized; calling the formed RDF graph as a lung cancer knowledge network graph, and calculating the importance of nodes by adopting a PageRank algorithm to find key nodes on a lung cancer treatment pathway. The correctness of the method is verified, and a new research direction can be provided for a new lung cancer treatment method and a new lung cancer treatment medicine through in-depth research from diseases and genes.

Description

technical field [0001] The invention belongs to the technical field of information data processing, and in particular relates to a method for processing key node information of a treatment pathway based on lung cancer medical big data. Background technique [0002] At present, the closest existing technology: network-based biological research is an academic frontier field that has received extensive attention from the international academic community in recent years, and has been widely used in disease research and drug prediction. RDF (ResourceDescription Framework) is a framework proposed by W3C to describe Semantic Web resources. Therefore, biomedical RDF data has gradually become an important type of structured data on the Internet of Things. There are many medical classifications, and the same data has different semantics in different classification environments; and the data collected through different means has various forms, usually including graphic images, text in...

AI Technical Summary

Problems solved by technology

[0003] According to consulting and researching relevant literature, many scholars have used some data mining algorithms such as PageRank, SimRank, etc. to study protein functions, disease gene prediction, etc., but they all analyze a single problem, so they only focus on and use A single entity (such as a protein) lacks the fusion of information such as genes and proteins, and there is an association between proteins, disease genes, and drugs. The organization of a single entity ignores the rich semantic associations between entities, and cannot The role and influence of multiple entities are considered in the etiology and therapeutic effect, so it is impossible to use more relevant medical literature, medical experiment results and other research results to comprehensively analyze the problem in breadth; ), along its metabolic pathways and signaling pathways, with the help of the relationship between entities, to search and discover important factors in the treatment of important diseases in depth

[0004] To sum up, the problems existing in the existing technology are: the existing algorithms that use data mining to analyze protein functions and disease gene predictions are aimed at a single problem, lack the fusion of information such as genes and proteins, and ignore rich correlations. Semantic associations, unable to simultaneously advance research in breadth and depth

[0006] The current lung cancer-related data format is not uniform, and there is noise in the data. The cleaning and fusion of massive data has always been a technical problem. Using the existing disambiguation and cleaning methods can realize the sorting of part of the data, but for medical, especially for special diseases Research needs to effectively integrate domain knowledge into data fusion methods. How to achieve data fusion with domain knowledge is a technical problem; at the same time, the network graph formed by a large amount of data is relatively complex, and the traditional pagerank algorithm is extremely inefficient when performing calculations. How to design and implement a distributed and parallel pagerank algorithm and enable the algorithm to obtain correct results based on semantic information is also a technical problem

Images