Event Types Identified

Modules are comprised of 1 or more Events, of which the following are currently identified:


  • If no events are identified in a module, the module is discarded. We

    have tested the voila Categorizer on highly complex human samples, and so far, we are able to identify events for all modules. So, we *think* we’re capturing all possible events a module could have. However, if you think you find a case where a module was discarded, but it shouldn’t have been. Please let us know! There is always a bigger fish, and there are always exceptions in Biology :)

  • Another reason modules are discarded is if you don’t turn on


  • Constitutive Junctions are not reported by default, you must turn on

    –keep-constitutive (and we will automatically also turn on –show-all-modules).

Cassette Exons


Note: C1_C2 and C2_C1 refer to the same junction, but the quantifications reported in the output file are from two different LSVs’ points of view (Reference Exon 4 (C1) or 6 (C2), respectively). Normally, these quantifications are identical, but read coverage for C1_A or C2_A may differ, and different read coverages alter the beta-posterior used for quantifications.

The “cassette.tsv” output file contains information about each cassette splicing event. This information includes coordinates for the exons and junctions as well as quantifications for each junction, from the point of view of usually two LSVs (sometimes there is not enough read coverage in one of the LSVs). Each cassette is represented by four rows. Each of the four rows in the file corresponds to a single junction quantification from the point of an LSV (or from the point of view of the reference exon of where there could have been an LSV, but MAJIQ didn’t see enough coverage for the LSV to be quantifiable). So, the skipping junction is represented by two rows: one row from the point of view of C1 (C1_C2) and the second row from the point of view of C2 (C2_C1). Here are the column names in the file:

  • Exon Spliced With: <C1, A, or C2>

  • Exon Spliced With Coordinate: <coordinate range>

  • Junction Name: <C1_A, C1_C2, C2_A, or C2_C1>

  • Junction Coordinate: <coordinate range>

Tandem Cassette


  • Same definition as a cassette exon, but there are 2+ alternative

    exons joined by junctions in the middle

  • The output file is the same as for cassette exon, but also includes a

    column specifying how many middle exons are in the tandem cassette

Alternative Splice Sites - 5’, 3’, or both - (Alt 5/3)


Putative Alternative Splice Sites (P_Alt 5/3)


TODO: compare to actual A3/A5 and decide if this should be renamed

Mutually Exclusive Exons (MXE)


Alternative First/Last Exons (AFE/ALE)


Putative Alternative First/Last Exons (P_AFE/ALE)


Alternative Intron


Orphan Junction


Orphan junctions are where splice junction reads are observed between two loci, but there is no evidence of complete exons. This is due to a lack of corresponding splice junction reads supporting a 3’ splice site upstream of the 5’ side of the orphan junction, and a lack of splice junction reads supporting a 5’ splice site downstream of the 3’ side of the orphan junction.



Constitutive Junction or Persistent Intron


Persistent introns are an artefact from annotations not supported by your data; persistent introns occur when an input GFF defines a junction between two distinct exons, but these splice sites have zero read evidence in your data. Thus, the “intron” persists … (the name comes from a wonderfully contentious in-lab debate and vote).

Constitutive Junctions and Persistent Introns are defined as follows:

  • Two exons connected by 1 junction/intron

  • From the upstream exon, only 1 junction/intron splices out

  • From the downstream exon, only 1 junction/intron splices in

  • If the 1 junction/intron has fewer than a threshold number of reads,

    then it is not “constitutive/persistent”

Multi Exon Spanning



LSVs that are identified and quantified by MAJIQ sometimes do not make it into the event tsvs. Usually, this is due to a transcript annotation that is inconsistent with your data. For example, in the following cassette exon, the alternative exon is a target LSV because of an annotated splice junction that is never used in your data:


The other.tsv file contains a list of LSVs that are not captured by the standard event tsvs.

Although unlikely, sometimes junction(s) within a Module do not fit *any* of our event definitions. These junctions (and their associated exons) will be listed in two columns in other.tsv, “other_junctions” and “other_exons.” In the above example, these two columns would be empty since dotted junction is decomplexified

Putative Multi-Gene Regions