The cox-filter method identifies respective subtype-specific lncRNA prognostic signatures for two human cancers

Area of Science:

• Genomics
• Molecular Biology
• Cancer Research

Background:

• Esophageal cancer has two main subtypes: squamous cell carcinoma (ESCC) and adenocarcinoma (EAC), with distinct gene expression profiles.
• Head and neck squamous cell carcinoma (HNSCC) subtypes, laryngeal (LSCC) and hypopharyngeal (HSCC), exhibit different prognoses.
• Long non-coding RNAs (lncRNAs) are emerging as potential biomarkers for prognostic signatures in various cancers.

Purpose of the Study:

• To develop subtype-specific prognostic lncRNA expression signatures for esophageal cancer and HNSCC.
• To apply a novel Cox-filter method for identifying unique prognostic biomarkers for each cancer subtype.
• To refine identified lncRNA signatures using biological relevancy information.

Main Methods:

• Utilized The Cancer Genome Atlas (TCGA) RNA-Seq data for esophageal cancer and HNSCC.
• Applied the Cox-filter method, a novel feature selection technique, to identify subtype-specific prognostic signatures.
• Incorporated biological relevancy to refine the selected lncRNA lists.

Main Results:

• Identified subtype-specific prognostic lncRNA expression signatures for esophageal cancer and HNSCC.
• The gene H19 demonstrated perfect prognostic value for esophageal cancer.
• The gene NEAT1 showed perfect prognostic value for HNSCC.

Conclusions:

• The Cox-filter method is effective for identifying subtype-specific prognostic lncRNA signatures.
• The identified lncRNA signatures, like H19 and NEAT1, hold significant prognostic value.
• The Cox-filter method is anticipated to have broader applications in cancer research.