Research Paper Volume 16, Issue 11 pp 9599—9624
Prognosis and diagnosis of prostate cancer based on hypergraph regularization sparse least partial squares regression algorithm
- 1 Department of Urology, First Affiliated Hospital of Gannan Medical University, Ganzhou, Jiangxi, China
- 2 First Clinical Medical College, Gannan Medical University, Ganzhou, Jiangxi, China
Received: July 31, 2023 Accepted: February 29, 2024 Published: May 31, 2024
https://doi.org/10.18632/aging.205889How to Cite
Copyright: © 2024 Huang et al. This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Abstract
Background: Prostate cancer (PCa) is a malignant tumor of the male reproductive system, and its incidence has increased significantly in recent years. This study aimed to further identify candidate biomarkers with prognostic and diagnostic significance by integrating gene expression and DNA methylation data from PCa patients through association analysis.
Material and methods: To this end, this paper proposes a sparse partial least squares regression algorithm based on hypergraph regularization (HR-SPLS) by integrating and clustering two kinds of data. Next, module 2, with the most significant weight, was selected for further analysis according to the weight of each module related to DNA methylation and mRNAs. Based on the DNA methylation sites in module 2, this paper uses multiple machine learning methods to construct a PCa diagnosis-related model of 10-DNA methylation sites.
Results: The results of Receiver Operating Characteristic (ROC) analysis showed that the DNA methylation-related diagnostic model we constructed could diagnose PCa patients with high accuracy. Subsequently, based on the mRNAs in module 2, we constructed a prognostic model for 7-mRNAs (MYH11, ACTG2, DDR2, CDC42EP3, MARCKSL1, LMOD1, and MYLK) using multivariate Cox regression analysis. The prognostic model could predict the disease free survival of PCa patients with moderate to high accuracy (area under the curve (AUC) =0.761). In addition, Gene Set EnrichmentAnalysis (GSEA) and immune analysis indicated that the prognosis of patients in the risk group might be related to immune cell infiltration.
Conclusions: Our findings may provide new methods and insights for identifying disease-related biomarkers by integrating DNA methylation and gene expression data.