Abstract
Epithelial-to-mesenchymal transition (EMT) is the underlying mechanism for tumor metastasis and shows the metastatic potential of tumor cells. Although the transcriptional regulation of EMT has been well studied, the role of alternative splicing (AS) regulation in EMT remains largely uncharacterized. The rapid accumulation of RNA-seq datasets has provided the opportunities for developing computational methods to associate mRNA isoform variations with EMT. In this study, we propose regularization models to identify significant AS events during EMT. Our experimental results confirm that the predicted AS events are closely related to apoptosis, focal adhesion-invadopodium shift and tight junction formation that are essential during EMT. Therefore, our study highlights the broad role of posttranscriptional regulation during EMT and identifies key subsets of AS events serving as distinct regulatory nodes.