Statistical and machine learning based platform-independent key genes identification for hepatocellular carcinoma

Area of Science:

• Oncology
• Genomics
• Bioinformatics

Background:

• Hepatocellular carcinoma (HCC) is a leading cause of cancer mortality with increasing global incidence.
• Previous studies on HCC genetic datasets focused on common differentially expressed genes (DEGs) across platforms, potentially overlooking crucial genes.
• A need exists for a comprehensive approach to identify robust biomarkers for HCC.

Purpose of the Study:

• To develop a statistical and machine learning system for identifying platform-independent key genes (KGs) in hepatocellular carcinoma (HCC).
• To overcome limitations of previous studies by integrating data from multiple platforms.
• To evaluate the diagnostic and prognostic potential of identified KGs in HCC.

Main Methods:

• Utilized a multi-platform dataset approach combined with statistical and machine learning techniques.
• Identified differentially expressed genes (DEGs) per platform, then combined them to find grand combined DEGs (gcDEGs).
• Employed Support Vector Machine (SVM) to determine differentially expressed discriminative genes (DEDGs), constructed a protein-protein interaction (PPI) network, identified hub genes using Maximal Clique Centrality (MCC), and selected optimal modules using MCODE.

Main Results:

• Identified six key genes (KGs): CDC20, TOP2A, CENPF, DLGAP5, UBE2C, and RACGAP1, through the intersection of overlapping hub genes, meta-hub genes, and hub module genes.
• These KGs were selected based on their discriminative accuracy and network properties.
• Evaluated the discriminative power and prognostic potential of the six KGs using Area Under the Curve (AUC) and survival analysis.

Conclusions:

• The identified six KGs represent robust, platform-independent biomarkers for hepatocellular carcinoma (HCC).
• These genes hold significant potential for improving the diagnosis and prognosis of liver cancer.
• The developed methodology offers a novel strategy for biomarker discovery in complex diseases like HCC.