9
$\begingroup$

I have a dataset of Oxford Nanopore cDNA reads. Many of my reads are full-length or close to full-length transcripts, and I and am interested in examining alternative splicing. For this, I would like to begin by visualising my reads and comparing variants qualitatively.

I have tried visualising my data in both IGV and SeqMonk, but neither has given a satisfying result.

IGV

IGV Visualisation of NOTCH2

IGV shows some links between aligned sections, shown as fine blue lines and thick black lines. I cannot work out what the difference between these lines is, and even more confusingly, we have numerous alignments where two aligned exons come from the same sequence but are not joined by any line. This means that, to confirm or deny a spliced variant, I need to manually examine the read ID of each exon.

SeqMonk

SeqMonk Visualisation of NOTCH2

SeqMonk gives a much cleaner visualisation, but unfortunately shows no links between aligned exons, and worse, I cannot find the read IDs. This means there is functionally (as far as I can see) no way to tell which exons come from the same read.

EDIT: Sashimi Plot

As recommended in this answer, I have tried using IGV's Sashimi Plot. Unfortunately, I believe the high error rate and relatively low coverage is causing this to create somewhat confusing output. The Sashimi Plot shows mostly nearly every junction to have just a single read supporting it.

IGV Sashimi Plot of NOTCH2

Is there a visualisation tool which allows simple examination of splicing of full-length transcripts?

$\endgroup$
  • $\begingroup$ Are you sure you are mapping the reads correctly? Do they have a lot of soft clips, perhaps suggesting that they aren't fully mapped? $\endgroup$ – GWW Jun 14 '17 at 22:09
  • $\begingroup$ I couldn't say for sure, but it looks okay to me.. They were mapped with GMAP, and with read lengths in the range of 1500-4000bp we have soft clipping sitting around 20-50 for most of those reads. About half of the reads have mapping quality 40 (the others are very low, about a quarter have MAPQ 0.) $\endgroup$ – Scott Gigante Jun 14 '17 at 23:15
5
$\begingroup$

GenVision Pro Sashimi PlotDNASTAR's software is for purchase, but high quality. GenVision Pro does genomic visualization, including Sashimi plots.

Edit: not sure why this answer is being downvoted, unless it's because the software isn't free. OP has tried IGV and SeqMonk, I mentioned an alternative he might not have heard of.

Here is a video demonstrating the use of Sashimi plots in GenVision Pro:

http://www.dnastar.com/t-support-videos.aspx?video=YJvcERoSIsg

$\endgroup$
  • 1
    $\begingroup$ Please downvoters, explain the reason for downvoting. $\endgroup$ – bli Jun 14 '17 at 11:41
  • $\begingroup$ Maybe add example plot outputs? $\endgroup$ – zx8754 Jun 14 '17 at 21:29
3
$\begingroup$

The Sashmi Plot feature built into IGV. It gives a nice summary of the spliced transcripts and the coverage of each exon.

For example:

enter image description here

$\endgroup$
3
$\begingroup$

IGB gives the intended result without much hassle. It doesn't have the bells and whistles of IGV, but presents a clean and intuitive view of the individual reads.

In IGB, each read is presented as a single, contiguous line.

IGB zoomed out view

If you zoom in you can see the read ID for each read.

IGB zoomed in view

For a more holistic view, IGV's Sashimi plot can be modified to exclude spurious junctions supported by few reads. Here is a section of the above gene with default settings:

Sashimi plot with default settings

The minimum junction coverage can be set by right-clicking on the Sashimi plot itself, or application-wide under View -> Preferences -> Alignments -> Splice Junction Track Options. With minimum junction coverage 3, the noisy plot above now looks like this:

Sashimi plot with min junction coverage 3

The numbers are still hard to read due to numerous overlapping almost-identical junctions, but at least the plot is now readable.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.