A computational strategy to uncover fusion genes in prostate cancer cell lines

Abstract

Fusion genes, chimeric transcripts formed by the merging of two distinct genes due to chromosomal structural changes (e.g., inversions or trans/cis-splicing), are established cancer drivers. Advances in genomic technologies, particularly RNA sequencing and improved fusion gene prediction algorithms, have significantly expanded our understanding of fusion genes in cancer. This chapter explores computational methods for identifying fusion genes through RNA sequencing data, using the TMPRSS2::ERG fusion in prostate cancer cell lines as a case study, and includes analysis of both fusion-positive and fusion-negative cell lines. To achieve high-confidence detection, three open-source fusion prediction tools, STAR-Fusion, FusionCatcher, and JAFFA are investigated. These tools were selected for their accessibility, active maintenance, and strong performance in benchmarking studies. Their sensitivity and accuracy in detecting TMPRSS2::ERG is systematically evaluated and validated, ensuring robust and high-resolution detection of fusion events.