A novel framework for chimeric transcript detection based on accurate gene fusion model

Abstract

Next generation sequencing plays a key role in the detection of structural variations. Chimeric transcripts are relevant examples of such variations, as they are involved in several diseases. In this work, we propose an effective methodology for the detection of fused transcripts in RNA-Seq paired-end data. The proposed methodology is based on an accurate fusion model implemented by a set of filters reducing the impact of artifacts. Moreover, the methodology accounts for transcripts consistently expressing in the sample under study even if they are not annotated. The effectiveness of the proposed solution has been experimentally validated on of Chronic Myelogenous Leukemia (CML) samples, providing both the genes involved in the fusion and the exact chimeric sequence.