Splice Junctions

IGV supplements each alignment track with (1) a coverage track and (2) if selected in the Alignment Preferences panel, a default splice junctions track. This page describes the default junctions track as well as independently loaded junctions data in the standard .bed format. See Sashimi Plot for how to derive and manipulate interactive junction visualizations within IGV. 

Before loading data, check Show junction track in the Alignment Preferences panel. The panel's settings must be adjusted for your data as default settings are for genomic reads for which splicing is irrelevant.

When enabled, IGV dynamically computes the junctions track from alignment data. The junctions track calls a splicing event when at least a single read splits across two exons in the alignment track.

Splice Junctions Track

The junctions track calls a splicing event when at least a single read splits across two exons in the alignment track.

  • IGV defines exons by your sample's read alignments.
  • IGV does not use the mate inner distance of paired reads for junctions.
  • You must have aligned reads using a splice-aware alignment algorithm, e.g. TopHat, for the library to retain split reads. Furthermore, the aligner must assign reads XS tag values. For example, XS:A:+ for GT-AG intron boundaries and XS:A:– for CT-AC intron boundaries.

Without XS tag strand information, IGV renders junctions using read strand, and therefore IGV's display of the strandedness is inferred.

  • Displaying strandedness in IGV that reflects biology with certainty, minimally, but not necessarily sufficiently, requires strand preserving sample preparation, indicating to the alignment algorithm before it is run a stranded library type, e.g. for TopHat firststrand or secondstrand, and an alignment algorithm that assigns XS tag values.
  • Some alignment algorithm modes may assign XS tag values to reads based on a transcriptome index against which reads were aligned and in which the strandedness of transcripts is presupposed.

Each splice junction is represented by an arc from the beginning to the end of the junction. 

  • Junctions from the + strand are colored red and extend above the center line. 
  • Junctions from the – strand are blue and extend below the center line.
  • The height of the arc, and its thickness, are proportional to the depth of read coverage up to 50 reads (first image).
    • Display a more proportionate representation by selecting Autoscale from the right-click menu (second image).

Hovering the mouse over or clicking on a junction will display coverage information. The first screenshot shows multiple coverage detail panels for each three components of two splice junctions on opposite strands.

  • Read depth for each end of the junction is displayed. For the red junction below, starting flank depth is 109 reads and ending flank depth is at 6606 reads.
  • Other details for a given junction's three hover elements are the same.

 

Right-click pop-up menu options for Junction tracks

Menu options are as detailed for the Feature tracks menu with the following additions or differences.

Command Description

Collapsed
Expanded
Squished

Tracks are collapsed by default. The expanded mode breaks up the junctions track to multiple junctions tracks to minimize visual overlap. IGV does not interpret isoform information.
Autoscale The height of the arc, and its thickness, are proportional to the depth of read coverage.
  • By default, all junctions with more than 50 reads have the same thickness.
  • Select Autoscale to display a more proportionate representation.
Sashimi Plot Displays junctions information for regions within the current IGV view in a new panel with additional options. See Sashimi Plot for details.
Export Features Download junctions track from IGV as a .bed file.

Example showing differential splicing

  • Start IGV and make sure Show junction track is checked in the Alignment Preferences panel and the Visibility range threshold is set to 500.
  • Load the Human hg19 genome.
  • Select File > Load from Server. Select Body Map 2.0 > Alignments > Merged 50bp and 75bp > Heart and Liver.
  • Enter SLC25A3 in the search bar to see an instance where the third exon is differentially spliced for the two tissues (Screenshot 2015.4.15).
    • Here we have colored alignments by XS tag. The library was unstranded, and XS tag values were assigned to reads crossing junctions (in pink) using a predefined transcriptome index.

Enable junctions view for .bed files

The splice junction view displays an alternative representation of .bed files encoding splice junctions, such as the "junctions.bed" file produced by the TopHat program. Display details are as described in the section above.

  • This view is enabled by including a track line that specifies either name=junctions or graphType=junctions.
  • TopHat's "junctions.bed" file includes a track line specifying name=junctions by default, so no action is required for these files.

The track can also be computed dynamically from an Alignment track by enabling the Show splice junctions track option in the alignment preferences as described above.

Junction files should be in the standard .bed format.  The score field is used to indicate depth of coverage.