Scholarlink

<p>Scaled Principal Component Analysis (sPCA), a further extension of PCA, focuses on its dimensional reduction ability and forecasting application in financial markets. Moreover, the need for dimensional reduction in biomedical data analysis, specifically disease screening, is highlighted since too many features lead to computational and storage challenges. The inadequacy of the PCA method in reducing dimensions effectively prompts the development of sPCA. Yet, directly adapting sPCA to biomedical data is inappropriate due to the differences in application purposes, mathematical formula structure, and more. In this study, we introduce the generalized linear model (GLM) to modify sPCA so that the GLM-sPCA is applicable in disease screening. A real-world data example of using routine hematological tests to screen whether an individual carries the genotype of thalassemia shows the utility of the GLM-sPCA in the biomedical domain for binary traits, and it performs better than PCA. Moreover, an additional simulation study is also conducted, which further proves the robustness of our GLM-sPCA method. Finally, the GLM-sPCA and its high performance can greatly benefit carrier screening and clinical diagnosis.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

GLM-sPCA: An Approach in Carrier Screening for Binary Traits with an Application in Thalassemia Among Adult Males

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年9月9日

查看全文
  • Qiwen He,
  • Hui Liang,
  • Xu Chen,
  • Lu Zhou,
  • Likuan Xiong,
  • Guangxing Mai

摘要

Scaled Principal Component Analysis (sPCA), a further extension of PCA, focuses on its dimensional reduction ability and forecasting application in financial markets. Moreover, the need for dimensional reduction in biomedical data analysis, specifically disease screening, is highlighted since too many features lead to computational and storage challenges. The inadequacy of the PCA method in reducing dimensions effectively prompts the development of sPCA. Yet, directly adapting sPCA to biomedical data is inappropriate due to the differences in application purposes, mathematical formula structure, and more. In this study, we introduce the generalized linear model (GLM) to modify sPCA so that the GLM-sPCA is applicable in disease screening. A real-world data example of using routine hematological tests to screen whether an individual carries the genotype of thalassemia shows the utility of the GLM-sPCA in the biomedical domain for binary traits, and it performs better than PCA. Moreover, an additional simulation study is also conducted, which further proves the robustness of our GLM-sPCA method. Finally, the GLM-sPCA and its high performance can greatly benefit carrier screening and clinical diagnosis.