Machine Learning Supports Long Noncoding RNAs as Expression Markers for Endometrial Carcinoma

Area of Science:

• Oncology
• Molecular Biology
• Bioinformatics

Background:

• Uterine corpus endometrial carcinoma (UCEC) is a prevalent gynecological malignancy.
• Non-coding RNAs (ncRNAs) are emerging as crucial biomarkers for cancer diagnosis and prognosis.
• Long noncoding RNAs (lncRNAs) have demonstrated potential in various cancer types.

Purpose of the Study:

• To investigate the potential of lncRNAs as accurate biomarkers for differentiating tumor and normal tissues in UCEC.
• To identify specific lncRNAs that can serve as diagnostic markers for UCEC.

Main Methods:

• In silico differential gene expression analysis of UCEC tumor (TP) versus normal adjacent (NT) tissues using TCGA data.
• Identification of highly differentially expressed lncRNAs.
• Receiver Operating Characteristic (ROC) analysis and machine learning (supervised and unsupervised) for biomarker validation.
• Coexpression network and target enrichment analysis for functional assessment.

Main Results:

• A set of 14 differentially expressed lncRNAs were identified as potential biomarkers for UCEC.
• These lncRNAs demonstrated high accuracy in discriminating between TP and NT tissues.
• Functional analysis suggested potential roles and pathways associated with these lncRNAs in UCEC.

Conclusions:

• The identified 14 lncRNAs hold significant potential as accurate diagnostic biomarkers for UCEC.
• These lncRNAs could aid in distinguishing tumor from normal tissues, facilitating improved diagnosis and potentially guiding treatment strategies.
• Further validation is warranted to establish their clinical utility in UCEC management.