Artificial intelligence-based prediction of molecular and genetic markers for hepatitis C-related hepatocellular

Cemil Colak1, Zeynep Kucukakcali1, Sami Akbulut1,2

  • 1Department of Biostatistics and Medical Informatics.

Insights

This study used machine learning to identify key genes for Hepatitis C virus-related hepatocellular carcinoma (HCC). The XGboost model achieved high accuracy, pinpointing HAO2, TOMM20, GPC3, and PSMB4 as potential biomarkers for HCC.

Area of Science:

  • Oncology
  • Genomics
  • Bioinformatics

Background:

  • Hepatocellular carcinoma (HCC) is a leading global cause of cancer mortality.
  • Hepatitis C virus (HCV) infection is a significant risk factor for HCC development.
  • Distinguishing between HCV-related HCC and chronic HCV without HCC is crucial for patient management.

Purpose of the Study:

  • To classify gene expression data from patients with HCV-related HCC and chronic HCV.
  • To identify key genes associated with the development of HCC in HCV-infected patients.
  • To leverage machine learning for biomarker discovery in HCC.

Main Methods:

  • Retrospective case-control study utilizing public gene expression data.
  • Application of the XGboost machine learning algorithm for classification.
  • 10-fold cross-validation and performance metrics including accuracy, sensitivity, and specificity were employed.

Main Results:

  • The XGboost model demonstrated high performance with an overall accuracy of 98.1%.
  • Key genes identified as potential biomarkers for HCV-related HCC include HAO2, TOMM20, GPC3, and PSMB4.
  • The model achieved excellent sensitivity (100%) and specificity (94.1%).

Conclusions:

  • Machine learning effectively identified potential biomarkers for HCV-related HCC.
  • The identified genes (HAO2, TOMM20, GPC3, PSMB4) warrant further clinical investigation.
  • Future research should focus on validating these biomarkers for therapeutic and diagnostic applications.
Abstract