Systematic Analyses of the Differentially Expressed Alternative Splicing Events in Gastric Cancer and Its Clinical Significance

Changwei Lin; Bowen Yu; Mao Zhang; Yifei Chen; Liang Li; Deze Zhao

doi:10.3389/fgene.2020.522831

Systematic Analyses of the Differentially Expressed Alternative Splicing Events in Gastric Cancer and Its Clinical Significance

Front Genet. 2020 Nov 17:11:522831. doi: 10.3389/fgene.2020.522831. eCollection 2020.

Authors

Changwei Lin^{1

2}, Bowen Yu², Mao Zhang³, Yifei Chen⁴, Liang Li⁵, Deze Zhao⁶

Affiliations

¹ School of Life Sciences, Central South University, Changsha, China.
² Department of Gastrointestinal Surgery, The Third Xiangya Hospital of Central South University, Changsha, China.
³ Department of Hepatobiliary and Pancreatic Surgery, Affiliated Hospital of Qingdao University, Qingdao, China.
⁴ Department of Otolaryngology-Head Neck Surgery, The Fourth Hospital of Changsha (The Changsha Affiliated Hospital of Hunan Normal University), Hunan Normal University, Changsha, China.
⁵ Class 25 Grade 2016, The Five-Year Program in Clinical Medicine, School of Medicine, University of South China, Hengyang, China.
⁶ Department of Thoracic Surgery, Xiangya Hospital, Central South University, Changsha, China.

Abstract

Accumulation of evidence has indicated a close relationship between alternative splicing (AS) and gastric cancer (GC), whereas systematic analyses of the differentially expressed AS events (DEAS) between GC and normal tissues are lacking. RNA-Seq data and the corresponding clinical information were downloaded from TCGA GC cohort. The percent spliced-in (PSI) value calculated in the GC tissues and normal tissues was employed to quantify the DEAS. Further, survival-associated DEAS and DEAS signatures were identified by univariate and multivariate cox regression analyses. To evaluate the association between DEAS and patients' clinical features, Kaplan-Meier analysis, receiver operator characteristic (ROC) curve, Cox proportional regression and nomograms incorporating the DEAS signatures were performed. DEAS and their splicing networks were finally analyzed by bioinformatics methods. In addition, we use the method of random grouping to divide the samples into the training group and the test group. The final results of the two groups are consistent. After strict filtering, a total of 44,935 AS events were identified, among which 11,141 DEAS were preliminarily screened from 5032 genes. A total of 454 DEAS was associated with OS, and 872 DEAS were associated with DFS. The final prognostic signatures were constructed from the survival-associated DEAS with an area under the receiver operating characteristic (ROC) curve (AUC) greater than 0.6. Only ES in ABI1 was simultaneously associated with OS and DFS. Finally, we identified the splicing correlation network between the prognostic splicing factors (SF) and DEAS in GC. Our study provided a systematic portrait of survival-associated DEAS in GC and uncovered splicing networks that are valuable in deciphering the underlying mechanisms of AS in GC.

Keywords: alternative splicing; differentially expressed AS events; gastric cancer; prognostic signature; splicing factor.