Research Paper|Volume 15, Issue 14|pp 6848—6864

The integration of machine learning and multi-omics analysis provides a powerful approach to screen aging-related genes and predict prognosis and immunotherapy efficacy in hepatocellular carcinoma

Jiahui Shen¹, Han Gao², Bowen Li³, Yan Huang³, Yinfang Shi²

¹Department of Pharmacy, Huzhou Maternity and Child Health Care Hospital, Huzhou, China
²Department of Stomatology, First Affiliated Hospital of Huzhou University, Huzhou, China
³School of Pharmacy, Anhui Medical University, Hefei, China

* Equal contribution

Received: March 31, 2023Accepted: June 15, 2023Published: July 28, 2023

https://doi.org/10.18632/aging.204876

Copyright: © 2023 Shen et al. This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Abstract

Background: Hepatocellular carcinoma (HCC) is a highly malignant tumor with high incidence and mortality rates. Aging-related genes are closely related to the occurrence and development of cancer. Therefore, it is of great significance to evaluate the prognosis of HCC patients by constructing a model based on aging-related genes.

Method: Non-negative matrix factorization (NMF) clustering analysis was used to cluster the samples. The correlation between the risk score and immune cells, immune checkpoints, and Mismatch Repair (MMR) was evaluated through Spearman correlation test. Real Time Quantitative PCR (RT-qPCR) and immunohistochemistry were used to validate the expression levels of key genes in tissue and cells for the constructed model.

Result: By performing NMF clustering, we were able to effectively group the liver cancer samples into two distinct clusters. Considering the potential correlation between aging-related genes and the prognosis of liver cancer patients, we used aging-related genes to construct a prognostic model. Spearman correlation analysis showed that the model risk score was closely related to MMR and immune checkpoint expression. Drug sensitivity analysis also provided guidance for the clinical use of chemotherapy drugs. RT-qPCR showed that TFDP1, NDRG1, and FXR1 were expressed at higher levels in different liver cancer cell lines compared to normal liver cells.

Conclusion: In summary, we have developed an aging-related model to predict the prognosis of hepatocellular carcinoma and guide clinical drug treatment for different patients.