interpreting inversion event in long reads

[prasundutta87]

Hi,

Please find an attached inversion event breakpoint (one end of the breakpoint) in the attached image. I am working with ONT data. I tried to follow an example present here: https://www.pacb.com/blog/igv-3-improves-support-pacbio-long-reads/ as they are both long reads, but I am unable to interpret my end of the IGV plot. The inversion breakpoint is present in the red box. What I am failing to understand is how do I interpret this inversion event here? What exactly is "link supplemetary alignment" doing here and how do I understand that its an inversion visually? I would be grateful if I could please be helped by this.

A Zoomed out image:

Regards, Prasun

[jrobinso]

Hi, I don't see any visual evidence of an inversion, has that been called by an external program? There is no evidence against an inversion either. Link supplementary alignments will draw lines between all alignments from the same read, in other words split alignments, but I don't see any in these screenshots.

The only evidence in long reads for an inversion would be in the supplementary (split) alignments, but again I don't see any here. If you are able to share a slice of this bam file around a putative inversion we will look into it. As it happens we are presently working on improving visualization of SVs for long reads.

BTW do you have "show soft clips" turned on?

[prasundutta87]

Hi @jrobinso. Thanks for the quick reply. Yes, the SVs were called by sniffles. I haven't loaded the VCF file, but the breakpoint is from there. The problem is that I can't share any genomic data as I am working with confidential data.

I have kept "show soft clips" on because previous to this, I was trying to visualise a translocation and showing soft clipped base in one location of the chromosome and showing its supplementary location somewhere (BLAT and breakpoint information from VCF) seemed to be a good strategy. Do you think switching it off for this case may make any difference? There are a lot of soft-clipped bases around the inversion breakpoint/s.

Regards, Prasun

[jrobinso]

OK, soft clips are good for all sorts of SVs, I was just confirming. There's probably no specific IGV view now that is going to cause an inversion to jump out at you visually. I can tell from the soft clips that there is a breakpoint there. If you blat the sequence for those soft clips it should tell you where they align. Also, looking at the popup data for the alignment should tell you where the other parts of the read aligned, if the aligner does "split reads" and there is an SA tag.

We have an ongoing project to improve viz for SVs, its complex as you can imagine. But keep this ticket open and we will update it when we have something to try.

[prasundutta87]

Thanks for your reply, @jrobinso! I used minimap2, and it does denote split reads as supplementary alignments. I did the BLAT and it does take me to the other end of the inversion breakpoint. I believe it should a reverese complement of the split read from the other breakpoint, but please correct me if I am wrong. I wonder what's the best way to show this on IGV.

[jrobinso]

Well I'm not sure what you are looking for, but at the moment there is not a great way to visualize these in IGV. @lbergelson is currently working on improvements in this area.

If your split reads are relatively close together I suggest zooming out until they are all in view, then "link by supplementary" to see the connections. If they are very far apart I would suggest trying the Circular View, developed by the jbrowse team and integrated with IGV as described here: https://github.com/igvteam/igv/wiki/Circular-View.

[prasundutta87]

Thanks @jrobinso ..I will look into it..I am also looking into jbrowse2 simultaneously, but will check out the circular view as well in igv..thanks a lot for your help. Much appreciated!

[jrobinso]

JBrowse might have other options, SVs has been a focus of their's recently.

[prasundutta87]

Yes, they seem to have that. Will explore it. Thanks @jrobinso !

[jrobinso]

@lbergelson for your consideration