Development of a new clustering technique for biological component analyses

2023 Fiscal Year Final Research Report

• Project/Area Number: 18K12112
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Review Section: Basic Section 90130:Medical systems-related
• Research Institution: Juntendo University (2022-2023); International University of Health and Welfare (2018-2021)
• Principal Investigator: Sato Shouichi 順天堂大学, 医療科学部, 教授 (90803255)
• Co-Investigator(Kenkyū-buntansha): 市原 清志 山口大学, 大学院医学系研究科, 学術研究員(寄附金) (10144495) ; 山下 哲平 東海大学, 医学部, 助教 (50617420)
• Project Period (FY): 2018-04-01 – 2024-03-31
• Keywords: クロマトグラフィー / フローサイトメトリー法 / クラスター解析 / 正規分布 / シミュレーションプログラム

Outline of Final Research Achievements

A new cluster analysis method has been developed to solve the problem of separating data in which multiple components overlap in separation analysis technology. This technology is particularly effective for overcoming problems such as overlapping between cell components, the appearance of abnormal cells, and the introduction of noise in the classification of white blood cells. Specifically, by using image processing to accurately estimate the types and distribution of cells, and by mathematically separating the overlapping of multiple components using multiple normal distribution theory and iterative truncation correction, it is possible to obtain clear analytical results. We have further developed this technology and improved it into an algorithm that can be applied to the separation of various mixed distributions, and have reported on it at various academic conferences.

Free Research Field

統計学

Academic Significance and Societal Importance of the Research Achievements

本研究では、分離分析技術で問題となる成分の重なりを解決する手法として、画像処理、自己分配方式、反復切断補正法を組み合わせた新しいクラスター分析法を開発した。特に、白血球分類において、細胞成分の正確な分離とノイズ除去を行うことが可能になり、複雑な臨床データに対応する分析精度が向上することを確認した。この技術は、臨床検査の精度向上に寄与し、様々な分離分析場面に対応できるため、医療現場での診断の信頼性を高める社会的意義があると考えられる。