Fusions CSV

File name: {Sample_ID}_Fusions.csv

This file contains all candidate fusions identified by the RNA analysis pipeline. Candidate fusions from the RNA splice variant caller are listed in this output with relevant supporting information. Two key output columns in the file describe the candidate fusions: Filter and KeepFusion.

The following table describes the output found in the Filter columns. The output is either a confidence filter or information only, as indicated. If none of these filters are triggered, the Filter column displays PASS.

Filter

Description

Imprecise

A low-resolution candidate, not an assembled fusion call. (Confidence filter)

RepeatOverlap

The fusion is tagged as overlapping with a repeat region. Only used as a confidence filter for nonuniquely mapping fusion candidates, otherwise information only. (Confidence filter)

WeakBreakend

The read/alignment evidence on one side of the fusion is weak. Usually this filter indicates that the reads only overlap the fusion by a few base pairs. Alternatively, it can indicate too much homology (no unique sequence). (Confidence filter)

Homology

The fusion contig is a substring of another fusion contig. (Information only)

DuplicateContig

The two contigs of the fusion are the same contig. (Confidence filter)

ContigIntragenic

The realignment of half-contigs produces alignments that map to the same gene on both sides (or within 1 kb if unannotated). (Confidence filter)

LowQ

Fusion supporting reads (unique) < 5 (+ 1 for every 10 million reads over 16 million reads). (Confidence filter)

LowDupReads

Fusion supporting reads (duplicate) < 5. (Information only)

NonExonic

Fusion breakpoint does not fall within an exon. (Information only)

LocalContigAlign

Contig realignment found a nonfusion alignment for this contig. (Information only)

LowFusionRatio

Few strong evidence reads compared to wild type reads. (Information only)

NoReferenceReads

No reads on either side of the presumed breakpoint are marked as reference (structurally normal) reads.

(Information only)

The KeepFusion column of the output has a value of True when the RNAFusionFilter score is ≥ 0.45 and none of the confidence filters are triggered.

The directionality of a fusion can be determined from the FusionFilter.csv file. If either the Strand1 or Strand2 column of the output has a value of + or −, then the gene in the Gene1 column is in the 5ʹ position, and the gene in the Gene2 column is in the 3ʹ position.

See the headers in the output for more information about each column.

NOTE

When using Microsoft Excel to view this report, genes that are convertible to dates (such as MARCH1) automatically convert to dd-mm format (1-Mar) by Excel.