Research And Application Of MeSH-based Literature Mining Method For Exploring Associations Between Genes And Clinical Terms Of Colorectal Cancer

With the development of medical science and technology, the large-scale biological database have been established and the capabilities to gene sequencing have been improved, resulting in a massive biomedical data. While the accumulation of biological experiment data, large amounts of unstructured literature data has been accumulated. The extraction and meta-analysis of biomedical literature are helpful and useful for understanding the mechanisms and processes of disease development, promoting the diagnosis and treatment of disease. Traditional method of manual analysis and annotation has been difficult to adapt to the rapid growth of literature. Therefore, more and more researchers began to focus on various literature mining methods. However, the existing methods are mostly based on the "co-occurrence" principle, combined with natural language processing including semantic analysis and grammar processing, which limit the potential associations mining and the number of analyzing articles. Besides, many of those research focus on the level of disease. There is no authoritative researches focus on a particular disease, mining the potential associations between disease related clinical terms and gene data.Therefore, this paper proposed a biomedical literature mining method for clinic-genomic associations, and choose colorectal cancer as a research disease case, aiming at promote precision prevention for cancer. Finally, this paper combined biomedical knowledge to analyze and explore the results. The detail works are as follows:Proposing a MeSH-based biomedical literature mining method. Used the information provided by PubMed, this method take advantage of the thought of vector space model, and then use arithmetical operation to achieve quantitative analysis by using MeSH to represent each biomedical entity as a vector.Choosing Colorectal Cancer as the disease object, and then apply the MeSH-based method on the research. Analyzing and explaining the mining result by integrating the biomedical knowledge and applying tool of gene function analysis, such as g:Profiler and KEGG Pathway.Constructing a platform for Colorectal Cancer clinic-genomic associations knowledge sharing, providing index and download of the data, and reuse of the mining method.The results showed that the MeSH-based method provides a new strategy for mining clinic-genomic associations, which can be used for research on other diseases.