Scholarlink

The tumor microenvironment (TME) performs a crucial function in the happening and advancement of malignancy. The general framework of investigation of potential prognostic TME-related biomarkers of a special cancer is explained in this chapter. Initially, extraction of the RNA-sequencing profiles and resembling clinical parameters from various databases such as GEO, TCGA, ICGC, and TARGET has been done, based on which the immune and stromal scores are estimated via the ESTIMATE algorithm. Overlaying differentially expressed genes between these two above score groups are examined by the LASSO and Random Forest algorithms and validated in real cases from the external cohort. Finally, a prognostic multigene signature is created by employing Cox regression. CIBERSORT algorithms are a tool for evaluating the infiltration of hematopoietic cell phenotypes. Characteristics of individuals and the predominant subtype determine the association of the stromal and immune scores. The correlation and direction of the stromal and immune score with prognosis state are detected. Furthermore, multi-TME-related prognostic genes are identified by machine learning algorithms such as LASSO regression and Radom Forest. The multigene signature status indicated a general survival condition. Subsequently, the frequency of subtypes of the immune cells in each of the signature classes is analyzed. The arrangement of the tumor microenvironment particularly influenced the prognosis of cancerous cases. By this method, a novel approach is developed for the answering curiosity exploration of prognostic TME-related biomarkers and immunotherapy.

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

An Immune-Related Gene Signature for Determining Tumor Prognosis Based on Machine Learning

  • 工具书词条

  • 发表日期：

    2025年

查看全文
  • Fereshteh Yazdanpanah,
  • Nima Rezaei

摘要

The tumor microenvironment (TME) performs a crucial function in the happening and advancement of malignancy. The general framework of investigation of potential prognostic TME-related biomarkers of a special cancer is explained in this chapter. Initially, extraction of the RNA-sequencing profiles and resembling clinical parameters from various databases such as GEO, TCGA, ICGC, and TARGET has been done, based on which the immune and stromal scores are estimated via the ESTIMATE algorithm. Overlaying differentially expressed genes between these two above score groups are examined by the LASSO and Random Forest algorithms and validated in real cases from the external cohort. Finally, a prognostic multigene signature is created by employing Cox regression. CIBERSORT algorithms are a tool for evaluating the infiltration of hematopoietic cell phenotypes. Characteristics of individuals and the predominant subtype determine the association of the stromal and immune scores. The correlation and direction of the stromal and immune score with prognosis state are detected. Furthermore, multi-TME-related prognostic genes are identified by machine learning algorithms such as LASSO regression and Radom Forest. The multigene signature status indicated a general survival condition. Subsequently, the frequency of subtypes of the immune cells in each of the signature classes is analyzed. The arrangement of the tumor microenvironment particularly influenced the prognosis of cancerous cases. By this method, a novel approach is developed for the answering curiosity exploration of prognostic TME-related biomarkers and immunotherapy.