Identification and characterization of gene fusions in breast cancer - A non-trivial pursuit

Gene fusions represent a very important class of genomic aberrations playing a significant role in certain types of cancer. In this paper we provide a method for gene fusion prioritization and assessment of their functional impact in cancer. Recent discoveries of recurrent and targetable gene fusions in breast cancer suggest the need to characterize the functional significance of such genomic aberrations within larger cohorts. Beyond existing gene fusion detection algorithms, our method identifies and prioritizes high confidence fusion calls and produces full sequence of the fused genes, annotates and visualizes protein domains included in the chimeric protein. We evaluate the sensitivity of downstream interpretive algorithms such as pathway enrichment to the statistical confidence thresholds that are parameters of the gene fusion calling algorithms in a breast cancer cohort. We show that while some pathways are strongly enriched in the results across multiple confidence cutoffs, the pathway enrichment analysis is indeed sensitive to the statistical cutoffs. This suggests that the gene fusion calling algorithms should not be considered plug-and-play tools and require great care in parameter selection before down-stream analysis and interpretation is performed.