Open Access

ARTICLE

WGCNA and LASSO algorithm constructed an immune infiltration-related 5-gene signature and nomogram to improve prognosis prediction of hepatocellular carcinoma

by MENG FANG¹, JING GUO¹, HAIPING WANG¹, ZICHANG YANG², HAN ZHAO^1,*, QINGJIA CHI²

1 School of Medicine, Jianghan University, Wuhan, 430056, China
2 Department of Mechanics and Engineering Structure, Wuhan University of Technology, Wuhan, 430070, China

* Corresponding Author: HAN ZHAO. Email:

(This article belongs to the Special Issue: Noncoding RNAs & Associated Human Diseases)

BIOCELL 2022, 46(2), 401-415. https://doi.org/10.32604/biocell.2022.016989

Received 17 April 2021; Accepted 14 May 2021; Issue published 20 October 2021

Abstract

Hepatocellular carcinoma (HCC) is a common immunogenic malignant tumor. Although the new strategies of immunotherapy and targeted therapy have made considerable progress in the treatment of HCC, the 5-year survival rate of patients is still very low. The identification of new prognostic signatures and the exploration of the immune microenvironment are crucial to the optimization and improvement of molecular therapy strategies. We studied the potential clinical benefits of the inflammation regulator miR-93-3p and mined its target genes. Weighted gene co-expression network analysis (WGCNA), univariate and multivariate COX regression and the LASSO COX algorithm are employed to identify prognostic-related genes and construct multi-gene signature-based risk model and nomogram for survival prediction. Support vector machine (SVM) based Cibersort’s deconvolution algorithm and gene set enrichment analysis (GSEA) is used to evaluate the changes in tumor immune microenvironment and pathway differences. The study found the favorable prognostic performance of miR-93-3p and identified 389 prognostic-related target genes. The risk model based on a novel 5-gene signature (cct5, cdk4, cenpa, dtnbp1 and flvcr1) was developed and has prominent prognostic significance in the training cohort (P < 0.0001) and validation cohort (P = 0.0016). The nomogram constructed by combining the gene signature and the AJCC stage further improves the survival prediction ability of the gene signature. The infiltration level of multiple immune cells (especially T cells, B cells and macrophages) were positively correlated with the expression of prognostic signature. In addition, we found that gene markers of T cells and B cells is monitored and regulated by prognostic signature. Meanwhile, several GSEA pathways related to the immune system are enriched in the high-risk group. In general, we integrated the WGCNA, LASSO COX and SVM algorithms to develop and verify 5-gene signatures and nomograms related to immune infiltration to improve the survival prediction of patients.

---

Keywords

---

Citations