Genome-wide co-localization of RNA-DNA interactions and fusion RNA pairs

Abstract

Fusion transcripts are used as biomarkers in companion diagnoses. Although more than 15,000 fusion RNAs have been identified from diverse cancer types, few common feature has been reported. Here, we compared 16,410 fusion transcripts detected in cancer (from a published cohort of 9,966 tumor samples of 33 cancer types) with genome-wide RNA-DNA interactions mapped in two normal, non-cancerous cell types (using iMARGI, an enhanced version of the MARGI [Mapping RNA-Genome Interactions assay]). Among the top 10 most significant RNA-DNA interactions in normal cells, 5 co-localized with the gene pairs that formed fusion RNAs in cancer. Furthermore, throughout the genome, the frequency of a gene pair to exhibit RNA-DNA interactions is positively correlated with the probability of this gene pair to present documented fusion transcripts in cancer. To test whether RNA-DNA interactions in normal cells are predictive of fusion RNAs, we analyzed in a validation cohort of 96 lung cancer samples using RNA-seq. We found 34 out of 42 fusion transcripts in the validation cohort exhibited RNA-DNA interactions in normal cells. Finally, combining RNA-seq, single-molecule RNA FISH, and DNA FISH, we detected a cancer sample with Eml4-Alk fusion RNA without forming the Eml4-Alk fusion gene. Collectively, these data suggest a novel RNA-poise model, where spatial proximity of RNA and DNA could poise for creation of fusion transcripts.