Fusion with strand orientation

Question

[mangfu100]

Hi

I have been reading "FACTERA : a practical method for the discovery of genomic rearrangements at breakpoint resolution".

While reading this paper, there is a explanation about its process.

anyway, In this paper, soft-clipping with orientation concept is hard to understand. So I am appreciated if you look at it.

Four orientation of R1 and R2 are possible.(Fig.1D).However, only cases 1a and 2a shown in Figure !d can generate valid fusions, as their reads have soft-clipped sequences facing opposite directions. Thus , before k-mer comparison (Fig.1C), the reverse complement of R1 is taken for cases 1b and 2b, respectively, converting them into cases 1a and 2a.

In this paragraph, I can confirm the 1a 2a has the opposite directions in soft-clipped reads, but why only fusion is to be opposite direction in soft-clipped. I can't understand.

Can anyone help me understand this further more with picture?

Answer 1

[Istvan Albert]

First I think you are right to be confused - the way the rationale is written should be much improved.

I believe that the answer to your question is that the situations in 1a and 2a refer to genes on the same strand getting fused. Whereas cases 1b and 2b refer to a situation where genes on different strand are fused. Thus when mapping the different paired ends reads will have appear to have different orientations.

[mangfu100]

Thank you for your advice.

Anyway How did you know that 1a and 2a go on the same strand while the others are on the different strand?

[mikhail.shugay]

You check their strand on genome. The genes in fusion transcript are in the same strand based RNA-seq reads aligned to transcriptome, yet the initial positioning and orientation of partner genes on genome could vary due to the fact that there is a multitude of mechanisms (large-scale mutations, e.g. deletions, inversions, translocations) that could lead to fusion.

[Istvan Albert]

As mikhail.shugay also stated I suspected this from other types of genome reorganization problems. For example genomic inversions will also show up as reads in the wrong say head-to-head orientation.

[mangfu100]

As I look at it again, I didn't understand.

Why R1 and R2 in case 1a is on the same orientation?

I think that paired end reads is RF orientation and ,therefore, R1 should be reverse strand and R2 also be a forward strand. so they are not on the same orientation.