Machine learning identifies 10 feature miRNAs for lung squamous cell carcinoma.

Area of Science:

• Oncology
• Molecular Biology
• Bioinformatics

Background:

• Lung squamous cell carcinoma (LUSC) is a prevalent malignancy with unclear tumor progression mechanisms.
• Accurate diagnostic biomarkers are crucial for effective LUSC management and treatment.
• MicroRNAs (miRNAs) are increasingly recognized for their roles in cancer development and progression.

Purpose of the Study:

• To utilize machine learning (ML) to identify feature miRNAs as reliable diagnostic biomarkers for LUSC.
• To explore the biological functions and clinical significance of identified miRNAs in LUSC.
• To differentiate between primary tumor tissues and para-carcinoma tissues using miRNA expression profiles.

Main Methods:

• Downloaded miRNA and clinical data from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) databases.
• Employed Support Vector Machine (SVM) and Principal Component Analysis (PCA) for feature miRNA identification.
• Conducted miRNA-mRNA interaction network analysis, Gene Ontology (GO), KEGG pathway, and Kaplan-Meier survival analyses.

Main Results:

• Identified 21 feature miRNAs differentiating LUSC tumor tissues from para-carcinoma tissues.
• A subset of ten miRNAs (including miR-143, miR-100, miR-101-1, miR-101-2, miR-182, miR-183, miR-205, miR-21, miR-30a, miR-30d) demonstrated high accuracy in distinguishing cancer from adjacent tissues.
• These ten miRNAs and their target genes are significantly correlated with cancer pathways and patient survival rates.

Conclusions:

• The identified dysregulated feature miRNAs are potentially involved in LUSC pathology.
• This specific group of ten miRNAs holds promise as potential diagnostic biomarkers for LUSC.
• These miRNAs may also serve as potential therapeutic targets for LUSC, warranting further investigation.