the link between the supporting read of translocation not shown sometimes

[charliechen912ilovbash]

As title, it happened when I using a BAM that proccessed by whatshap haplotag and whatshap split.
The links looks fine when I use the BAM that has not been processed by whatshap. Is this related to the tag that whatshap add? (like the HP tag)

[cmdcolin]

hmmm...I have not tried using the breakpoint split view (which is what I am presuming you are using) with whatshap haplotag/split

The most important component to the breakpoint split view is that it connects reads with the same QNAME using data in the SA tag and CIGAR strings

If whatshap split a read with the same QNAME into two different files that would be disruptive...it would also be a weird decision for whatshap to make since being the same QNAME would literally mean it came from the same source read and thus should not be split into two files...

If you have any sample data or link to live instance i might be able to check it out

[charliechen912ilovbash]

I found the reason. It seems that whatshap haplotag or whatshap split will only leave one side of a split read. So for translocation supporting read, only the read at the start or the end point is remained. Thus, no link between the two chromosomes is generated on JBrowse.
I will discuss with the whatshap team about this issue. Thanks for your reply.

[charliechen912ilovbash]

Hi again. I noticed another possible bug when showing translocation with the SV inspector.
Sometimes (Not always. May be because the supportings reads are too many?), some links of the translocation cannot be shown normally. Instead, it looks like deletion. As the attached figure shown.

I inspected the read that showed abnormally (when with other reads) individually with JBrowse, and everything seems fine.

[cmdcolin]

Thanks for the detailed report. If you have a BAM file that shows this and the regions I could check it out. if not, I can try to reproduce locally too.

The code does allow reads to be connected to the "same row" in the sv inspector/breakpoint split view and this can sometimes look glitchy. It also can get tricked when the read depth is high, which is a limitation I want to fix (old issue #1561) but not sure if that is what you are seeing here

[charliechen912ilovbash]

Sorry that I could not share the BAM file. But the straight link line will be correct (curve) when zoom-out to some level.
In my BAM file, this situation happened to ~30% of the reads.