So, My question is structure variations detected in DNA sequencing results should also exist in whole transcriptome sequencing results?
The breakpoint? No. The vast majority of SVs occur outside of exonic sequence so the breakpoints will not be detected in RNA-seq data. Gene fusions are almost always intron fusions thus the position of the fusion in the DNA-seq will be at the intronic breakpoint position, but the position of the fusion in the RNA-seq will be between the flanking exons (or sometimes between other upstream and downstream exons if there's alternate spliting and/or the flanking exons are out of frame - the details get really complicated).
That said, SVs that result in gene fusions will result in a fusion gene products in the RNA-seq data if that fusion is expressed. On our cohort, we found:
Overall, LINX predicted 84% of 1,478 in-frame fusions detected by STAR-Fusion. Whilst 86%
of LINX intergenic fusion predictions are not detected in RNA expression, this is unsurprising as
many rearrangements may be passenger events (like most point mutations) and, as the
average expression of 50% of genes is less than 20 transcripts per million, fusions in genes
with low of no expression are undetectable by RNA-seq.
That is, passenger fusions are, in general, not expressed, but driver fusions are.
We also found that 17% of driver fusion are cause by complex rearrangments in which the fusion spanned mulitple breakpoints. These mostly occured in TMPRSS2-ERG fusions (i.e. TMPRSS2-> random other location without any exons -> ERG).
[shameless plug]You should try my fusion prediction software as it's the only DNA-seq fusion predictor that can can handle compound fusions.
TLDR: an EML4-ALK fusion of the flanking exons should be there.