Identification of the Real Hub Gene and Construction of a Novel Prognostic Signature for Pancreatic Adenocarcinoma Based on the Weighted Gene Co-expression Network Analysis and Least Absolute Shrinkage and Selection Operator Algorithms

Qihang Yuan; Jie Ren; Zhizhou Wang; Li Ji; Dawei Deng; Dong Shang

doi:10.3389/fgene.2021.692953

Identification of the Real Hub Gene and Construction of a Novel Prognostic Signature for Pancreatic Adenocarcinoma Based on the Weighted Gene Co-expression Network Analysis and Least Absolute Shrinkage and Selection Operator Algorithms

Front Genet. 2021 Aug 20:12:692953. doi: 10.3389/fgene.2021.692953. eCollection 2021.

Authors

Qihang Yuan^{1

2}, Jie Ren³, Zhizhou Wang^{1

2}, Li Ji⁴, Dawei Deng^{1

2}, Dong Shang^{1

2}

Affiliations

¹ Department of General Surgery, First Affiliated Hospital of Dalian Medical University, Dalian, China.
² Clinical Laboratory of Integrative Medicine, First Affiliated Hospital of Dalian Medical University, Dalian, China.
³ Department of Oncology, First Affiliated Hospital of Dalian Medical University, Dalian, China.
⁴ Department of Gastroenterology, Dongzhimen Hospital, Beijing University of Chinese Medicine, Beijing, China.

Abstract

Background: Pancreatic adenocarcinoma (PAAD) has a considerably bad prognosis, and its pathophysiologic mechanism remains unclear. Thus, we aimed to identify real hub genes to better explore the pathophysiology of PAAD and construct a prognostic panel to better predict the prognosis of PAAD using the weighted gene co-expression network analysis (WGCNA) and the least absolute shrinkage and selection operator (LASSO) algorithms. Methods: WGCNA identified the modules most closely related to the PAAD stage and grade based on the Gene Expression Omnibus. The module genes significantly associated with PAAD progression and prognosis were considered as the real hub genes. Eligible genes in the most significant module were selected for construction and validation of a multigene prognostic signature based on the LASSO-Cox regression analysis in The Cancer Genome Atlas and the International Cancer Genome Consortium databases, respectively. Results: The brown module identified by WGCNA was most closely associated with the clinical characteristics of PAAD. Scaffold attachment factor B (SAFB) was significantly associated with PAAD progression and prognosis, and was identified as the real hub gene of PAAD. Moreover, both transcriptional and translational levels of SAFB were significantly lower in PAAD tissues compared with normal pancreatic tissues. In addition, a novel multigene-independent prognostic signature consisting of SAFB, SP1, and SERTAD3 was identified and verified. The predictive accuracy of our signature was superior to that of previous studies, especially for predicting 3- and 5-year survival probabilities. Furthermore, a prognostic nomogram based on independent prognostic variables was developed and validated using calibration curves. The predictive ability of this nomogram was also superior to the well-established AJCC stage and histological grade. The potential mechanisms of different prognoses between the high- and low-risk subgroups were also investigated using functional enrichment analysis, GSEA, ssGSEA, immune checkpoint analysis, and mutation profile analysis. Conclusion: SAFB was identified as the real hub gene of PAAD. A novel multigene-independent prognostic signature was successfully identified and validated to better predict PAAD prognosis. An accurate nomogram was also developed and verified to aid in the accurate treatment of tumors, as well as in early intervention.

Keywords: hub gene; least absolute shrinkage and selection operator; pancreatic adenocarcinoma; prognostic signature; weighted gene co-expression network analysis.