Comprehensive analysis of alternative splicing signatures in pancreatic head cancer

Abstract The correlation between dysregulation of splicing and cancers has been increasingly recognised and confirmed. The identification of valuable alternative splicing (AS) in pancreatic head cancer (PHC) has a great significance. AS profiles in PHC were generated using the data from The Cancer Genome Atlas and TCGASpliceSeq. Then, the NMF clustering method was performed to identify overall survival‐associated AS (OS‐AS) subtypes in PHC patients. Subsequently, we used least absolute shrinkage and selection operator Cox regression analysis to construct an AS‐related risk model. The splicing regulatory network was uncovered by Cytoscape 3.7. A total of 1694 OS‐AS events were obtained. The PHC patients were divided into clusters 1 and 2. Cluster 1 had poorer prognosis and lower infiltration of immune cells. Subsequently, a prognostic signature was established that showed good performance in predicting OS and progression‐free survival. The risk score of this signature was associated with the unique tumour immunity. Moreover, a nomogram incorporating the risk score and clinicopathological parameters was established. Finally, a splicing factor‐AS regulatory network was developed. A comprehensive analysis of the AS events in PHC associated with prognosis and tumour immunity may help provide reliable information to guide individual treatment strategies.