Scholarlink

Abstract <p>Inflammatory Breast Cancer (IBC) is a rare subtype of locally advanced breast cancer, and caused 7% of breast cancer deaths. Studying the function of gene expression profile is one of the strategies that can help to identify and monitor this malignancy. This study aimed to determine the important genes expression and their function in the classification of the IBC patients. Based on microarray dataset with accession number GSE5847, the most important genes among 22 283 genes associated with IBC (<i>n</i> = 13) and non-IBC (<i>n</i>&#xa0;= 34) patients were selected using the eXtreme Gradient Boosting (XGBoost) and Random Forest (RF) method. Then, the relationship between the expression levels of the selected genes in the two groups, an independent sample <i>t</i>-test and Mann–Whitney <i>U</i> test were used. Finally, using decision tree algorithm, performance evaluation of the developed decision tree (DDT) such as F1-score and AUC indices of 3-fold cross-validation were reported. The rules extracted from the <i>DDT</i> were also interpreted. EDN1, NUCB2, REST, CCL5, UBE3A, ARMCX3, ITGBL1, AR, DZIP3, SLC2A10, FOSB, BLNK, IL6, PSD3, ABAT, LIMCH1, STS, ANGPTL4, and ZIC1 were selected using random forest. Statistical tests revealed differences in the expression levels of all genes except CCL5, DZIP3, FOSB, and ANGPTL4. The mean of the 3-fold cross-validation F1-score, and AUC indices of the <i>DDT</i> were 71.1, and 73.73, respectively. This study demonstrated the effectiveness of the decision tree in classifying IBC. It also highlighted the contributions of the AR, REST, UBE3A, STS, and NUCB2 genes to the classification of IBC and non-IBC patients.</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

Identification of Gene Expression Profile in Inflammatory Breast Cancer Patients Using Machine Learning Method

  • 期刊论文

  • 发表日期：

    2025年7月1日

查看全文
  • N. Rabiei,
  • F. Bahreini,
  • S. Moslehi

摘要

Abstract

Inflammatory Breast Cancer (IBC) is a rare subtype of locally advanced breast cancer, and caused 7% of breast cancer deaths. Studying the function of gene expression profile is one of the strategies that can help to identify and monitor this malignancy. This study aimed to determine the important genes expression and their function in the classification of the IBC patients. Based on microarray dataset with accession number GSE5847, the most important genes among 22 283 genes associated with IBC (n = 13) and non-IBC (n = 34) patients were selected using the eXtreme Gradient Boosting (XGBoost) and Random Forest (RF) method. Then, the relationship between the expression levels of the selected genes in the two groups, an independent sample t-test and Mann–Whitney U test were used. Finally, using decision tree algorithm, performance evaluation of the developed decision tree (DDT) such as F1-score and AUC indices of 3-fold cross-validation were reported. The rules extracted from the DDT were also interpreted. EDN1, NUCB2, REST, CCL5, UBE3A, ARMCX3, ITGBL1, AR, DZIP3, SLC2A10, FOSB, BLNK, IL6, PSD3, ABAT, LIMCH1, STS, ANGPTL4, and ZIC1 were selected using random forest. Statistical tests revealed differences in the expression levels of all genes except CCL5, DZIP3, FOSB, and ANGPTL4. The mean of the 3-fold cross-validation F1-score, and AUC indices of the DDT were 71.1, and 73.73, respectively. This study demonstrated the effectiveness of the decision tree in classifying IBC. It also highlighted the contributions of the AR, REST, UBE3A, STS, and NUCB2 genes to the classification of IBC and non-IBC patients.