Bioinformatics analysis of the prognostic role of alternative splicing data in lung adenocarcinoma.

Background: As a post-transcriptional regulatory mechanism, alternative splicing (AS) is engaged in a variety of pathophysiological processes, and it has been widely reported in connection with the occurrence, progression, metastasis, and drug resistance of cancer. However, the research on AS in lung adenocarcinoma (LUAD) is very limited. In addition, the prognostic effect of AS event (ASE) on LUAD and its related mechanism are not clear. This study aimed to explore the role and potential prognostic value of ASE in LUAD.

Methods: Relevant data and ASE datasets of the sample were acquired from The Cancer Genome Atlas (TCGA) and TCGASpliceSeq databases. We constructed a new prognostic criterion based on ASEs. Then, Cox regression and least absolute shrinkage and selection operator (LASSO) regression analysis were used to construct the model. Based on this model, the risk score of each ASE was calculated, and the reliability of this model was evaluated by Kaplan-Meier survival and receiver operating characteristic (ROC) curve analyses. Finally, these results were verified on different network platforms.

Results: We identified seven types of ASEs related to survival. The prognostic risk model for ASEs was established. The Kaplan-Meier curve showed that compared to the low-risk group, the overall survival (OS) rate of LUAD patients in the high-risk group was lower. ROC curve analysis showed that the prognostic risk model of LUAD patients was well predicted, and the area under the curve (AUC) also confirmed this.

Conclusions: This study screened the ASE related to the prognosis of LUAD patients, and provided a theoretical basis for further study of the correlation between ASE and the prognosis of LUAD patients. It has provided new ideas for developing new biomarkers and therapeutic targets for LUAD patients.