Transcripts are frequently modified by structural variations, which lead to fused transcripts of either multiple genes, known as a fusion gene, or a gene and a previously non-transcribed sequence. Detecting these modifications, called transcriptomic structural variations (TSVs), especially in cancer tumor sequencing, is an important and challenging computational problem. | Ma et al. Genome Biology 2018 19 52 https s13059-018-1421-5 M ET HO D Open Access SQUID transcriptomic structural variation detection from RNA-seq Cong Ma Mingfu Shao and Carl Kingsford Abstract Transcripts are frequently modified by structural variations which lead to fused transcripts of either multiple genes known as a fusion gene or a gene and a previously non-transcribed sequence. Detecting these modifications called transcriptomic structural variations TSVs especially in cancer tumor sequencing is an important and challenging computational problem. We introduce SQUID a novel algorithm to predict both fusion-gene and non-fusion-gene TSVs accurately from RNA-seq alignments. SQUID unifies both concordant and discordant read alignments into one model and doubles the precision on simulation data compared to other approaches. Using SQUID we identify novel non-fusion-gene TSVs on TCGA samples. Keywords Transcriptomic structural variation RNA-seq TCGA Background causing disruption to the function of the altered gene. Large-scale transcriptome sequence changes are known to There have been fewer studies on these TSVs between be associated with cancer 1 2 . Those changes are usually transcribed and non-transcribed regions but their ability a consequence of genomic structural variation SV . By to alter downstream RNA and protein structure is likely to pulling different genomic regions together or separating lead to similar results as fusion-gene TSVs. one region into pieces structural variants can potentially Genomic SVs are typically detected from whole-genome cause severe alteration to transcribed or translated prod- sequencing WGS data by identifying reads and read ucts. Transcriptome changes induced by genomic SVs pairs that are incompatible with a reference genome . called transcriptomic structural variants TSVs can have 6 11 . However WGS data are not completely suitable a particularly large impact on disease genesis and pro- for inferring TSVs
