Prediction of biologically important features related to intron retention events based on machine learning algorithms

The retention of introns represents a class of alternative splicing (AS) events frequently associated with diseases. Despite the recent development of many AS identification tools, most of the tools do not consider their relationship to relevant biological features. We developed a package capable of accessing many features to AS events and evaluated association between events. We identified that retained introns and nearby exons have lower GC content then their nonretained counterparts.